Neuromethods 142

Tatiana Barichello Editor Blood- Barrier N EUROMETHODS

Series Editor Wolfgang Walz University of Saskatchewan Saskatoon, SK, Canada

For further volumes: http://www.springer.com/series/7657 Blood-Brain Barrier

Edited by Tatiana Barichello

Department of Psychiatry & Behavioral Sciences, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, USA Graduate Program in Health Sciences, University of Southern Santa Catarina (UNESC), Criciúma, SC, Brazil Editor Tatiana Barichello Department of Psychiatry & Behavioral Sciences The University of Texas Health Science Center at Houston (UTHealth) Houston, TX, USA Graduate Program in Health Sciences University of Southern Santa Catarina (UNESC) Criciu´ma, SC, Brazil

ISSN 0893-2336 ISSN 1940-6045 (electronic) Neuromethods ISBN 978-1-4939-8945-4 ISBN 978-1-4939-8946-1 (eBook) https://doi.org/10.1007/978-1-4939-8946-1

Library of Congress Control Number: 2018961703

© Springer Science+Business Media, LLC, part of Springer Nature 2019 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Cover illustration: Image courtesy of Tatiana Barichello and Allan Collodel.

This Humana Press imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A. Preface to the Series

Experimental life sciences have two basic foundations: concepts and tools. The Neuro- methods series focuses on the tools and techniques unique to the investigation of the nervous system and excitable cells. It will not, however, shortchange the concept side of things as care has been taken to integrate these tools within the context of the concepts and questions under investigation. In this way, the series is unique in that it not only collects protocols but also includes theoretical background information and critiques which led to the methods and their development. Thus it gives the reader a better understanding of the origin of the techniques and their potential future development. The Neuromethods publishing program strikes a balance between recent and exciting developments like those concerning new animal models of disease, imaging, in vivo methods, and more established techniques, including, for example, immunocytochemistry and electrophysiological technologies. New trainees in neurosciences still need a sound footing in these older methods in order to apply a critical approach to their results. Under the guidance of its founders, Alan Boulton and Glen Baker, the Neuromethods series has been a success since its first volume published through Humana Press in 1985. The series continues to flourish through many changes over the years. It is now published under the umbrella of Springer Protocols. While methods involving brain research have changed a lot since the series started, the publishing environment and technology have changed even more radically. Neuromethods has the distinct layout and style of the Springer Protocols program, designed specifically for readability and ease of reference in a laboratory setting. The careful application of methods is potentially the most important step in the process of scientific inquiry. In the past, new methodologies led the way in developing new dis- ciplines in the biological and medical sciences. For example, physiology emerged out of anatomy in the nineteenth century by harnessing new methods based on the newly discov- ered phenomenon of electricity. Nowadays, the relationships between disciplines and meth- ods are more complex. Methods are now widely shared between disciplines and research areas. New developments in electronic publishing make it possible for scientists that encounter new methods to quickly find sources of information electronically. The design of individual volumes and chapters in this series takes this new access technology into account. Springer Protocols makes it possible to download single protocols separately. In addition, Springer makes its print-on-demand technology available globally. A print copy can therefore be acquired quickly and for a competitive price anywhere in the world.

Saskatoon, SK, Canada Wolfgang Walz

v Preface

The Blood-Brain Barrier (BBB) Methods and Protocols book is focused on experimental research with relevant models to study physiology, biochemistry, and molecular biology of the BBB. The chapters were written by world-renowned scientists who depict their knowl- edge about BBB and its functional measurement. The BBB book offers straightforward guidance for both young and experienced investigators to perform studies using classical and innovative models permitting translational approaches for BBB investigations. The BBB methods and protocols book is organized into six subjects contemplating (1) an overview about the physiology of the BBB; (2) in vitro cell models to investigate the BBB; (3) in vivo and ex vivo techniques to evaluate BBB including Drosophila melanogaster, zebrafish (Danio rerio), and rodent models; (4) techniques to evaluate permeability, influx and efflux trans- portation, and drug delivery through the BBB; (5) invasive and noninvasive imaging techniques to evaluate the BBB such as intravital microscopy, magnetic resonance imaging (MRI), and positron emission tomography (PET); and (6) molecular biomarkers to evaluate the integrity or dysfunction of the BBB. The BBB methods and protocols book brings together many of the specialized methods for evaluating BBB in 20 important chapters with transparency and technical excellence providing practical solutions in the laboratory. We hope that you enjoy this detailed scientific journey about the BBB and this book becomes a great collaborator to unravel the mysteries of the BBB and apply your findings from basic science to enhancing human health and well-being.

Houston, TX, USA Tatiana Barichello Criciu´ ma, SC, Brazil

vii Contents

Preface to the Series ...... v Preface ...... vii Contributors...... xi 1 An Overview of the Blood-Brain Barrier ...... 1 Tatiana Barichello, Allan Collodel, Rodrigo Hasbun, and Rodrigo Morales 2 Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain Diagnostics and Therapeutics ...... 9 Brian M. Kopec, Kavisha R. Ulapane, Mario E. G. Moral, and Teruna J. Siahaan 3 Culturing of Rodent Brain Microvascular Endothelial Cells for In Vitro Modeling of the Blood-Brain Barrier ...... 45 Malgorzata Burek and Carola Y. Fo¨rster 4 In Vitro BBB Models: Working with Static Platforms and Microfluidic Systems...... 55 Mohammad A. Kaisar, Vinay V. Abhyankar, and Luca Cucullo 5 In Vitro Cell Models of the Human Blood-Brain Barrier: Demonstrating the Beneficial Influence of Shear Stress on Brain Microvascular Endothelial Cell Phenotype ...... 71 Keith D. Rochfort and Philip M. Cummins 6 Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure the Integrity of Blood-Brain Barrier ...... 99 Balaji Srinivasan and Aditya Reddy Kolli 7 Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain Barrier Permeability: Implications for Pathogenesis and Therapeutic Treatment of Neurodegenerative Disease ...... 115 Swapna Bera and Anirban Bhunia 8 Microbial Translocation of the Blood-Brain Barrier ...... 137 Charles T. Spencer and Mireya G. Ramos Muniz 9 Transport Across the Choroid Plexus: How to Culture Choroid Plexus Cells and Establish a Functional Assay System ...... 163 Sen Takeda and Keishi Narita 10 Drosophila as a Model to Study the Blood-Brain Barrier ...... 175 Cameron R. Love and Brigitte Dauwalder 11 Zebrafish (Danio rerio) as a Viable Model to Study the Blood-Brain Barrier...... 187 Tianzhi Yang and Shuhua Bai 12 Evans Blue-Albumin as a Marker to Evaluate Blood-Brain Barrier Integrity in Neonatal and Adult Rodents ...... 197 Fabricia Petronilho, Julia L. Goldman, and Tatiana Barichello

ix x Contents

13 Experimental Tools to Study the Regulation and Function of the Choroid Plexus ...... 205 Isabel Gonc¸alves, Telma Quintela, Ana Catarina Duarte, Peter Hubbard, Grac¸a Baltazar, Christian Schwerk, Andrea Carmine Belin, Joana Toma´s, and Cecı´lia Reis A. Santos 14 Techniques for Evaluating Efflux Transport of Radiolabeled Drugs and Compounds from the Cerebrospinal Fluid Across the Blood-Cerebrospinal Fluid Barrier...... 231 Shin-ichi Akanuma, Yoshiyuki Kubo, and Ken-ichi Hosoya 15 In Vivo Analysis to Study Transport Across the Blood-Retinal Barrier...... 249 Yoshiyuki Kubo, Shin-ichi Akanuma, and Ken-ichi Hosoya 16 Increasing BBB Permeability via Focused Ultrasound: Current Methods in Preclinical Research ...... 267 Dallan McMahon, Charissa Poon, and Kullervo Hynynen 17 Evaluation of Blood–Brain Barrier Permeability and Integrity in Juvenile Rodents: Dynamic Contrast-Enhanced (DCE), Magnetic Resonance Imaging (MRI), and Evans Blue Extravasation ...... 299 Trish Domi, Faraz Honarvar, and Andrea Kassner 18 Recording Leukocyte Rolling and Adhesion on Meningeal Vessels by Intravital Microscopy ...... 315 Aline Silva de Miranda, Thiago Macedo Cordeiro, Milene Alvarenga Rachid, and Antoˆnio Lu´ cio Teixeira 19 Molecular Imaging of Blood–Brain Barrier Permeability in Preclinical Models Using PET and SPECT ...... 329 Vijayasree V. Giridharan, Tatiana Barichello, and Sudhakar Selvaraj 20 Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier Injury and Damage Measurement ...... 343 Pavani Sayana, Jean Pierre Oses, Tatiana Barichello, and Vijayasree V. Giridharan

Index ...... 365 Contributors

VINAY V. ABHYANKAR  Department of Biomedical Engineering, Rochester Institute of Technology, Rochester, NY, USA SHIN-ICHI AKANUMA  Department of Pharmaceutics, Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, Toyama, Japan SHUHUA BAI  Department of Basic Pharmaceutical Sciences, School of Pharmacy, Husson University, Bangor, ME, USA GRAC¸A BALTAZAR  CICS-UBI—Health Sciences Research Centre, University of Beira Interior, Covilha˜, Portugal TATIANA BARICHELLO  Department of Psychiatry and Behavioral Sciences, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, USA; Graduate Program in Health Sciences, University of Southern Santa Catarina (UNESC), Criciu´ ma, SC, Brazil ANDREA CARMINE BELIN  Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden SWAPNA BERA  Department of Biophysics, Bose Institute, Kolkata, India ANIRBAN BHUNIA  Department of Biophysics, Bose Institute, Kolkata, India MALGORZATA BUREK  Department of Anaesthesia and Critical Care, University of Wu¨rzburg, Wu¨rzburg, Germany ALLAN COLLODEL  Translational Psychiatry Program, Department of Psychiatry and Behavioral Sciences, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, USA THIAGO MACEDO CORDEIRO  Laboratorio Interdisciplinar de Investigac¸a˜oMe´dica, Faculdade de Medicina, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil LUCA CUCULLO  Department of Pharmaceutical Sciences, Texas Tech University Health Sciences Center, Amarillo, TX, USA; Center for Blood Brain Barrier Research, Texas Tech University Health Sciences Center, Amarillo, TX, USA; School of Pharmacy, Texas Tech University Health Sciences Center, Amarillo, TX, USA PHILIP M. CUMMINS  School of Biotechnology, National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland BRIGITTE DAUWALDER  Department of Biology and Biochemistry, University of Houston, Houston, TX, USA ALINE SILVA DE MIRANDA  Laboratorio Interdisciplinar de Investigac¸a˜oMe´dica, Faculdade de Medicina, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil; Laboratorio de Neurobiologia, Departamento de Morfologia, Institute de Cieˆncias Biologicas, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil TRISH DOMI  Division of Neurology, Department of Pediatrics, Hospital for Sick Children, Toronto, ON, Canada ANA CATARINA DUARTE  CICS-UBI—Health Sciences Research Centre, University of Beira Interior, Covilha˜, Portugal CAROLA Y. FO¨ RSTER  Department of Anaesthesia and Critical Care, University of Wu¨rzburg, Wu¨rzburg, Germany

xi xii Contributors

VIJAYASREE V. GIRIDHARAN  Translational Psychiatry Program, Department of Psychiatry and Behavioral Sciences, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, USA JULIA L. GOLDMAN  Center for Laboratory Animal Medicine and Care, The University of Texas Health Science Center at Houston, Houston, TX, USA ISABEL GONC¸ ALVES  CICS-UBI—Health Sciences Research Centre, University of Beira Interior, Covilha˜, Portugal RODRIGO HASBUN  Department of Infectious Diseases, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, USA FARAZ HONARVAR  Institute of Medical Sciences, University of Toronto, Toronto, ON, Canada KEN-ICHI HOSOYA  Department of Pharmaceutics, Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, Toyama, Japan PETER HUBBARD  Centre of Marine Sciences-CCMAR, University of Algarve, Faro, Portugal KULLERVO HYNYNEN  Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada; Physical Sciences Platform, Sunnybrook Research Institute, Toronto, ON, Canada; Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, ON, Canada MOHAMMAD A. KAISAR  Department of Pharmaceutical Sciences, Texas Tech University Health Sciences Center, Amarillo, TX, USA ANDREA KASSNER  Division of Translational Medicine, Department of Medical Imaging, Hospital for Sick Children, University of Toronto, Toronto, ON, Canada ADITYA REDDY KOLLI  Philip Morris International Research and Development, Neuchaˆtel, Switzerland BRIAN M. KOPEC  Department of Pharmaceutical Chemistry, Simons Laboratories, The University of Kansas, Lawrence, KS, USA YOSHIYUKI KUBO  Department of Pharmaceutics, Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, Toyama, Japan CAMERON R. LOVE  Department of Biology and Biochemistry, University of Houston, Houston, TX, USA DALLAN MCMAHON  Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada; Physical Sciences Platform, Sunnybrook Research Institute, Toronto, ON, Canada MARIO E. G. MORAL  Department of Pharmaceutical Chemistry, Simons Laboratories, The University of Kansas, Lawrence, KS, USA RODRIGO MORALES  Neuroscience Graduate Program, The University of Texas MD Anderson Cancer Center, UTHealth Graduate School of Biomedical Sciences, Houston, TX, USA; Mitchell Center for Alzheimer’s Disease and Related Brain Disorders, Department of Neurology, McGovern Medical School, The University of Texas Health Science Center at Houston, Houston, TX, USA KEISHI NARITA  Department of Anatomy and Cell Biology, University of Yamanashi Faculty of Medicine, Chuo, Yamanashi, Japan JEAN PIERRE OSES  Postgraduate Program in Health and Behavior, Center for Life Sciences and Health, Catholic University of Pelotas (UCPel), Pelotas, RS, Brazil FABRICIA PETRONILHO  Laboratory of Neurobiology of Inflammatory and Metabolic Processes, University of Southern Santa Catarina, Tubara˜o, SC, Brazil Contributors xiii

CHARISSA POON  Physical Sciences Platform, Sunnybrook Research Institute, Toronto, ON, Canada; Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, ON, Canada TELMA QUINTELA  CICS-UBI—Health Sciences Research Centre, University of Beira Interior, Covilha˜, Portugal MILENE ALVARENGA RACHID  Departamento de Patologia Geral, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil MIREYA G. RAMOS MUNIZ  University of Texas at El Paso, El Paso, TX, USA KEITH D. ROCHFORT  School of Biotechnology, National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland CECI´LIA REIS A. SANTOS  CICS-UBI—Health Sciences Research Centre, University of Beira Interior, Covilha˜, Portugal PAVANI SAYANA  Translational Psychiatry Program, Department of Psychiatry and Behavioral Sciences, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, USA CHRISTIAN SCHWERK  Department of Pediatrics, Pediatric Infectious Diseases, Medical Faculty Mannheim, , Mannheim, Germany SUDHAKAR SELVARAJ  Department of Psychiatry and Behavioral Sciences, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, USA TERUNA J. SIAHAAN  Department of Pharmaceutical Chemistry, Simons Laboratories, The University of Kansas, Lawrence, KS, USA CHARLES T. SPENCER  University of Texas at El Paso, El Paso, TX, USA BALAJI SRINIVASAN  Cornell University, Ithaca, NY, USA SEN TAKEDA  Department of Anatomy and Cell Biology, University of Yamanashi Faculty of Medicine, Chuo, Yamanashi, Japan ANTOˆ NIO LU´ CIO TEIXEIRA  Laboratorio Interdisciplinar de Investigac¸a˜oMe´dica, Faculdade de Medicina, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil; Neuropsychiatry Program, Department of Psychiatry & Behavioral Sciences, McGovern Medical School, University of Texas Health Science Center at Houston, Houston, TX, USA JOANA TOMA´ S  CICS-UBI—Health Sciences Research Centre, University of Beira Interior, Covilha˜, Portugal KAVISHA R. ULAPANE  Department of Pharmaceutical Chemistry, Simons Laboratories, The University of Kansas, Lawrence, KS, USA TIANZHI YANG  Department of Basic Pharmaceutical Sciences, School of Pharmacy, Husson University, Bangor, ME, USA Chapter 1

An Overview of the Blood-Brain Barrier

Tatiana Barichello, Allan Collodel, Rodrigo Hasbun, and Rodrigo Morales

Abstract

The blood-brain barrier (BBB) is a highly specialized structure formed by a tight monolayer of brain endothelial cells, which maintain bloodstream cells, neurotoxic compounds, and microorganims outside of the central nervous system (CNS). This barrier has also the ability to orchestrate the flow of some solutes from in and out of the brain. In addition, the BBB constitutes a key component of the neurovascular unit (NVU). The NVU is a functional unit composed of a complex cellular system formed by neurons, interneurons, astrocytic endfeet, microglia, oligodendrocytes, basal lamina covered with smooth muscular cells and pericytes, endothelial cells and extracellular matrix, and circulating blood components. The NVU unit reacts in response to physiological stimuli facilitating the activity-dependent regulation of vascular permeability, regulating the cerebral blood flow, and activating the neuroimmune response to maintain CNS homeostasis. Thus, the NVU facilitates the cross talk between the CNS and the periphery through the BBB. The BBB is organized to prevent undesirable substances from entering the brain while allowing access to necessary compounds. In consequence of its structure, the BBB is an impediment for drug delivery into the CNS and experimental protocols are being investigated to selectively modulate the BBB for delivery of therapeutic drugs to treat neurological diseases. As a consequence of pathological changes experienced in neurological diseases such as Alzheimer’s disease, meningitis, multiple sclerosis, malaria, Parkinson’s disease, stroke, and septic encephalopathy, the BBB loses some of its properties that contribute to exacer- bate disease progression. Understanding the mechanisms operating at the BBB level may help to decrease the development of neurological diseases and improve the design of new strategies that may facilitate the delivery of therapeutics drugs. The BBB Methods and Protocols book is focused on experimental research with relevant models to study the physiology, biochemistry, and molecular biology of the BBB. This book may present a greater relevance to understand several important questions in the field of neurosciences.

Key words Blood-brain barrier, Neurovascular unit, Endothelial junctions, Neurodegenerative diseases

Abbreviations

Afadin AF-6 protein BBB Blood-brain barrier BCSFB Blood-cerebrospinal fluid barrier BNBs Blood-neural barriers BRB Blood-retinal barrier BSCB Blood-spinal cord barrier CNS Central nervous system

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_1, © Springer Science+Business Media, LLC, part of Springer Nature 2019 1 2 Tatiana Barichello et al.

CSF Cerebrospinal fluid ESAM Endothelial cell-selective adhesion molecule JAMs Junctional adhesion molecules MAGUK Membrane-associated guanylate kinase MUPP-1 Multi-PDZ domain protein-1 NVU Neurovascular unit PECAM Platelet endothelial cellular adhesion molecule-1 TER Transepithelial resistance ZO Zona occluden protein

1 Introduction

In 1885, the German scientist Paul Ehrlich (1854–1915) injected vital dyes using parenteral routes in adult animals. All organs were stained, except the brain and spinal cord [1]. The first conclusion was that the central nervous system (CNS) possessed specific fea- tures that resulted in a lack or low affinity for vital dyes. In 1900, the German neurologist Max Lewandowsky (1876–1916) demon- strated that toxins targeting the CNS were more effective in lower doses when administered into the brain of animals compared to higher concentrations administered parenterally. In addition, he demonstrated that strychnine and sodium ferrocyanide, two toxic compounds, presented low bioavailability concentrations in the brain following subcutaneous administration of large amounts of these compounds compared to subarachnoid injections performed in considerably lower amounts [2, 3]. Lewandowsky described this phenomenon as a specific property of the cerebral capillaries to some chemical elements [2, 3]. In 1913, Edwin Goldman (1862–1913), an associate of Paul Ehrlich, injected the acidic dye trypan blue into the brain ventricular system of dogs and rabbits. Results showed staining of the entire CNS but no other organs. On that occasion, it was understood that the blood vessels themselves were responsible for blocking the connection between the blood and the CNS, since no obvious membrane acting at that level was identified [4]. From 1918 to 1925, the Russian neurophysiologist Lina Stern (1878–1968) and her colleague the Swiss physician Raymond Gautier (1885–1957) performed several experiments using different compounds by injecting them into the subarach- noid space, cerebral ventricles, and blood [5, 6]. After these experi- ments, they concluded that (1) the subarachnoid space was anatomically connected with the cerebral ventricles, (2) the cere- brospinal spinal fluid (CSF) and the cerebral ventricles presented the same chemical characteristics, and (3) substances present in the brain easily reached the blood but conversely substances in the bloodstream reached the brain in low concentrations [7, 8]. In 1921, Stern presented the term “Barrie`re he´matoence´phalique” The Blood-Brain Barrier 3

at the Medical Society of Geneva and she published this new concept about blood-brain barrier (BBB) in a Swiss Journal entitled “Schweizer Archiv fu¨r Neurologie und Psychiatrie” [5, 6, 9].

2 The Specialized Blood-Neural Barriers (BNBs)

Specialized neural barriers consist of an interconnection between the blood and neural tissue. Specifically, blood-neural barriers (BNBs) are endothelial structures arranged in an organized net- work including endothelial cells, astrocytes, and neurons that sepa- rate the blood from neural tissues. These structures exclude the entrance of many toxic substances present in the circulation to enter into the brain. Therefore, BNBs contribute to maintain a precisely regulated microenvironment adequate for the normal neuronal activity. The specialized BNBs include the BBB, blood-CSF barrier (BCSFB), blood-retinal barrier (BRB), blood-spinal cord barrier (BSCB), blood-labyrinth barrier, and blood-nerve barrier [10, 11]. A schematic representation of BNBs is displayed in Fig. 1.

Blood Brain Barrier Blood-CSF Barrier Arachnoid Barrier

Astrocyte Choroid plexus Dura mater Pericyte Artery Ependymal cell Arachnoid CSF Blood

Microglia Subarachnoid Ventricle space Pia mater Basolateral Apical Neuron Endothelium surface surface Brain Basement membrane

Blood-retinal barrier Blood-spinal cord Barrier

Outer BRB Artery Astrocyte Choriocapillary Pericyte Retina pgment epithelial cell

Tight Retina juntion Retinal capillary tight junction Endothelium Basement membrane Inner BRB

Fig. 1 The specialized blood-neural barriers (BNBs). The specialized BNBs include the BBB, blood-CSF barrier, blood-retinal barrier, blood-spinal cord barrier, blood-labyrinth barrier, and blood-nerve barrier 4 Tatiana Barichello et al.

3 The Blood-Brain Barrier (BBB)

The BBB is a highly specialized structure formed by a tight mono- layer of brain endothelial cells, which maintain bloodstream cells, neurotoxic compounds, and microorganisms outside of the CNS. This barrier also has the ability to orchestrate the movement of some low-molecular-weight solutes such as peptides, proteins, car- bohydrates, hormones, and vitamins from one compartment to the other. Moreover, some large molecules such as the cytokine- induced neutrophil chemoattractant-1 (CINC-1, 7800 Da) are able to cross the BBB by transmembrane diffusion [12, 13]. The capillary length in a human brain extends for about 650 kilometers (km). In mice, this distance adds up to approximately 0.6 km. Thus, capillaries provide the largest endothelial surface area for the bidi- rectional transport and exchange of solutes between peripheral circulation and the brain [14, 15].

3.1 The The BBB is not a passive barrier but can change in consonance to Neurovascular the demands of the CNS. The structure responsible for this plastic- Unit (NVU) ity is known as the neurovascular unit (NVU) [16]. The NVU is a complex cellular system formed by neurons, interneurons, astro- cytic endfeet, microglia, oligodendrocytes, basal lamina covered with smooth muscular cells and pericytes, endothelial cells, extra- cellular matrix, and circulating blood components [12, 17]; see Fig. 2. In the NVU, pericytes are located in the center between endothelial cells, astrocytes, and neurons. Thus, pericytes receive signals from adjacent cells that trigger pathways which are essential for CNS functioning such as angiogenesis, BBB formation and maintenance, vascular stability, capillary blood flow control, and clearance of toxic cellular products [18, 19].

3.2 Endothelial Tight The interendothelial space of the cerebral microvasculature is char- Junctions acterized by the presence of a junctional complex that includes adherens junctions, tight junctions, and possibly gap junctions [20]. The tight junction is a transmembrane protein complex com- posed by claudins, occludin, and junctional adhesion molecules (JAMs) connected intercellularly and interacting with the cytoplas- mic proteins zonula occludens (ZO)-1 and -2, and cingulin to the actin cytoskeleton [20, 21]. A brief explanation of proteins playing an important role in tight junctions is explained below and is schematized in Fig. 3. Claudins: There are 27 members of the claudin family and most of claudins present a PDZ domain that connects claudins on ZO (ZO-1, ZO-2, and ZO-3 proteins that are members of the membrane-associated guanylate-kinase (MAGUK) protein family) and on multi-PDZ domain protein-1 (MUPP-1) scaffolding pro- teins, which are peripheral proteins localizing at junctional sites The Blood-Brain Barrier 5

Fig. 2 Neurovascular unit (NVU). The neurovascular unit is a complex cellular system formed by neurons, interneurons, astrocytic endfeet, microglia, oligo- dendrocytes, basal lamina covered with smooth muscular cells and pericytes, endothelial cells and extracellular matrix, and circulating blood components

[22–25]. Claudins form paracellular barriers and pores at the tight junction of endothelial cells. Barrier-forming claudins increase the transepithelial resistance (TER), and present a key role in determin- ing the permeability properties of endothelial cells [23]. Claudin-5 is a major cell adhesion molecule of tight junctions in brain endo- thelial cells. In a previous study, claudin-5 knockout mice presented an impaired BBB against small molecules (<800 Da) [26]. The important role of this protein was reinforced by the fact that claudin-5 knockout mice died within 10 h after being born [26]. Occludin: Occludin, a 60–65 kDa tetraspan integral membrane protein, is capable of linking with ZO-1, ZO-2, and ZO-3 via cingulin dimers to the actin cytoskeletal system of the cell [27]. Occludin’s main function is to regulate tight junctions, con- tributing to their stabilization and optimal barrier function [28]. JAMs: The JAMs are present in brain endothelial cells, and are associated with the development and preservation of tight junc- tions. JAMs interact with PDZ domain-containing proteins afadin (AF-6), ZO-1, and cingulin [29, 30]. 6 Tatiana Barichello et al.

Fig. 3 Endothelial tight junctions. The tight junction is a transmembrane protein complex composed by claudins, occludin, and junctional adhesion molecules (JAMs) associated in intercellular contacts and interactions with cytoplasmic proteins zonula occludens (ZO)-1, -2, and -3 and cingulin to the actin cytoskeleton

4 Considerations

The BBB is the principal interface for communication between the CNS and the periphery [31]. Altered function of the BBB has been observed and may contribute to neuropsychiatric diseases (e.g., Alzheimer’s disease), sepsis-associated encephalopathy, and other pathological conditions [32–34]. In this book, we describe a num- ber of important and updated protocols related to the study of the BBB. The chapters were organized in different subjects listed as in vitro techniques, in vivo techniques (Drosophila sp., Danio rerio (zebrafish), and rodents), transport techniques, invasive and non- invasive techniques, and molecular approaches to study this impor- tant anatomical structure. This book may be a useful tool for researchers to find information on the mechanisms by which thera- peutic agents cross the BBB, as well as current strategies used to optimize permeation. Implications and approaches used for drug discovery, delivery of therapeutic agents, and toxicity mechanisms affecting the BBB are also discussed. The Blood-Brain Barrier 7

References

1. B€aumler E (1984) Paul Ehrlich: scientist for 12. Neuwelt E, Abbott NJ, Abrey L, Banks WA, life. Holmes & Meier, New York Blakley B, Davis T, Engelhardt B, Grammas P, 2. Liddelow SA (2011) Fluids and barriers of the Nedergaard M, Nutt J, Pardridge W, Rosen- CNS: a historical viewpoint. Fluids Barriers berg GA, Smith Q, Drewes LR (2008) Strate- CNS 8(1):2. https://doi.org/10.1186/2045- gies to advance translational research into brain 8118-8-2 barriers. Lancet Neurol 7(1):84–96. https:// 3. Saunders NR, Dreifuss JJ, Dziegielewska KM, doi.org/10.1016/s1474-4422(07)70326-5 Johansson PA, Habgood MD, Mollgard K, 13. Banks WA (2009) Characteristics of com- Bauer HC (2014) The rights and wrongs of pounds that cross the blood-brain barrier. blood-brain barrier permeability studies: a BMC Neurol 9(Suppl 1):S3. https://doi.org/ walk through 100 years of history. Front Neu- 10.1186/1471-2377-9-s1-s3 rosci 8:404. https://doi.org/10.3389/fnins. 14. Zlokovic BV (2008) The blood-brain barrier in 2014.00404 health and chronic neurodegenerative disor- 4. Ribatti D, Nico B, Crivellato E, Artico M ders. Neuron 57(2):178–201. https://doi. (2006) Development of the blood-brain bar- org/10.1016/j.neuron.2008.01.003 rier: a historical point of view. Anat Rec B New 15. Montagne A, Zhao Z, Zlokovic BV (2017) Anat 289(1):3–8. https://doi.org/10.1002/ Alzheimer’s disease: a matter of blood-brain ar.b.20087 barrier dysfunction? J Exp Med 214 5. Sarikcioglu L (2017) Lina Stern (1878–1968): (11):3151–3169. https://doi.org/10.1084/ an outstanding scientist of her time. Childs jem.20171406 Nerv Syst 33(7):1027–1029. https://doi. 16. Banks WA, Erickson MA (2010) The blood- org/10.1007/s00381-017-3392-3 brain barrier and immune function and dys- 6. Vein AA (2008) Science and fate: Lina Stern function. Neurobiol Dis 37(1):26–32. (1878–1968), a neurophysiologist and bio- https://doi.org/10.1016/j.nbd.2009.07.031 chemist. J Hist Neurosci 17(2):195–206. 17. Muoio V, Persson PB, Sendeski MM (2014) https://doi.org/10.1080/ The neurovascular unit—concept review. Acta 09647040601138478 Physiol (Oxford, England) 210(4):790–798. 7. Stern L, Gautier R (1921) Recherches sur le https://doi.org/10.1111/apha.12250 liquide ce´phalo-rachidien. Arch Int Physiol 18. Winkler EA, Bell RD, Zlokovic BV (2011) 17:138–192 Central nervous system pericytes in health and 8. Stern L, Gautier R (1922) Recherches sur le disease. Nat Neurosci 14(11):1398–1405. liquide ce´phalo-rachidien. Les rapports entre le https://doi.org/10.1038/nn.2946 liquide ce´phalo-rachidien et les e´le´ments ner- 19. Sweeney MD, Ayyadurai S, Zlokovic BV veux de l’axe ce´re´bro-spinal. Arch Int Physiol (2016) Pericytes of the neurovascular unit: 17:391–448 key functions and signaling pathways. Nat 9. Stern L (1921) Le liquide ce´falo-rachidien au Neurosci 19(6):771–783. https://doi.org/ point de vue de ses rapports avec la circulation 10.1038/nn.4288 sanguine et avec les e´le´ments nerveux de l’axe 20. Hawkins BT, Davis TP (2005) The blood- ce´re´brospinal. Schweiz Arch Neurol Psychiatr brain barrier/neurovascular unit in health and 8(2):215–232 disease. Pharmacol Rev 57(2):173–185. 10. Edelhauser HF, Rowe-Rendleman CL, Robin- https://doi.org/10.1124/pr.57.2.4 son MR, Dawson DG, Chader GJ, Grossnik- 21. Luissint AC, Artus C, Glacial F, laus HE, Rittenhouse KD, Wilson CG, Weber Ganeshamoorthy K, Couraud PO (2012) DA, Kuppermann BD, Csaky KG, Olsen TW, Tight junctions at the blood brain barrier: Kompella UB, Holers VM, Hageman GS, Gil- physiological architecture and disease- ger BC, Campochiaro PA, Whitcup SM, Wong associated dysregulation. Fluids Barriers CNS WT (2010) Ophthalmic drug delivery systems 9(1):23. https://doi.org/10.1186/2045- for the treatment of retinal diseases: basic 8118-9-23 research to clinical applications. Invest 22. Greene C, Campbell M (2016) Tight junction Ophthalmol Vis Sci 51(11):5403–5420. modulation of the blood brain barrier: CNS https://doi.org/10.1167/iovs.10-5392 delivery of small molecules. Tissue Barriers 4 11. Choi YK, Kim KW (2008) Blood-neural bar- (1):e1138017. https://doi.org/10.1080/ rier: its diversity and coordinated cell-to-cell 21688370.2015.1138017 communication. BMB Rep 41(5):345–352 8 Tatiana Barichello et al.

23. Gunzel D, Yu AS (2013) Claudins and the Junctional adhesion molecule interacts with modulation of tight junction permeability. the PDZ domain-containing proteins AF-6 Physiol Rev 93(2):525–569. https://doi.org/ and ZO-1. J Biol Chem 275 10.1152/physrev.00019.2012 (36):27979–27988. https://doi.org/10. 24. Itoh M, Furuse M, Morita K, Kubota K, 1074/jbc.M002363200 Saitou M, Tsukita S (1999) Direct binding of 30. Bazzoni G, Martinez-Estrada OM, Orsenigo F, three tight junction-associated MAGUKs, Cordenonsi M, Citi S, Dejana E (2000) Inter- ZO-1, ZO-2, and ZO-3, with the COOH ter- action of junctional adhesion molecule with the mini of claudins. J Cell Biol 147 tight junction components ZO-1, cingulin, (6):1351–1363 and occludin. J Biol Chem 275 25. Hamazaki Y, Itoh M, Sasaki H, Furuse M, Tsu- (27):20520–20526. https://doi.org/10. kita S (2002) Multi-PDZ domain protein 1074/jbc.M905251199 1 (MUPP1) is concentrated at tight junctions 31. Logsdon AF, Erickson MA, Rhea EM, Salameh through its possible interaction with claudin-1 TS, Banks WA (2018) Gut reactions: how the and junctional adhesion molecule. J Biol Chem blood-brain barrier connects the microbiome 277(1):455–461. https://doi.org/10.1074/ and the brain. Exp Biol Med (Maywood, NJ) jbc.M109005200 243(2):159–165. https://doi.org/10.1177/ 26. Nitta T, Hata M, Gotoh S, Seo Y, Sasaki H, 1535370217743766 Hashimoto N, Furuse M, Tsukita S (2003) 32. Leonardo ED, Dranovsky A (2017) An open- Size-selective loosening of the blood-brain bar- ing for humor in melancholy. Nat Neurosci 20 rier in claudin-5-deficient mice. J Cell Biol 161 (12):1657–1658. https://doi.org/10.1038/ (3):653–660. https://doi.org/10.1083/jcb. s41593-017-0015-y 200302070 33. Le Page A, Dupuis G, Frost EH, Larbi A, 27. Abbott NJ, Patabendige AA, Dolman DE, Pawelec G, Witkowski JM, Fulop T (2018) Yusof SR, Begley DJ (2010) Structure and Role of the peripheral innate immune system function of the blood-brain barrier. Neurobiol in the development of Alzheimer’s disease. Exp Dis 37(1):13–25. https://doi.org/10.1016/j. Gerontol 107:59–66. https://doi.org/10. nbd.2009.07.030 1016/j.exger.2017.12.019 28. Cummins PM (2012) Occludin: one protein, 34. Kuperberg SJ, Wadgaonkar R (2017) Sepsis- many forms. Mol Cell Biol 32(2):242–250. associated encephalopathy: the blood-brain https://doi.org/10.1128/mcb.06029-11 barrier and the sphingolipid rheostat. Front 29. Ebnet K, Schulz CU, Meyer Zu Brickwedde Immunol 8:597. https://doi.org/10.3389/ MK, Pendl GG, Vestweber D (2000) fimmu.2017.00597 Chapter 2

Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain Diagnostics and Therapeutics

Brian M. Kopec, Kavisha R. Ulapane, Mario E. G. Moral, and Teruna J. Siahaan

Abstract

Brain diseases such as Alzheimer’s, Parkinson’s, multiple sclerosis (MS), and brain tumors are difficult to diagnose and treat. This is partly due to the difficulty in delivering diagnostic and therapeutic molecules across the blood-brain barrier (BBB) and into the brain. This chapter describes these and other challenges and the progress that has been made in transporting molecules into the brain. The pathways that molecules can take to cross the BBB and methods to improve the passage of diagnostics and therapeutic molecules into the brain are described. Several advances in improving detection of molecules in the brain using imaging methods (i.e., MRI, PET, SPECT, NIR) are described as potential diagnostic tools for brain diseases.

Key words Blood-brain barrier (BBB), Brain delivery, Brain therapeutics, Brain diagnostics, Brain diseases

1 Introduction

Every year, hundreds of thousands of people are diagnosed with brain disorders such as Alzheimer’s, Parkinson’s, multiple sclerosis (MS), and brain tumors. However, because the brain is protected by the blood-brain barrier (BBB), there are few efficacious drugs available for treating these conditions. Although many agents have been developed for early diagnosis of brain diseases, these, too, cannot readily be delivered to the brain due to the presence of the BBB. The BBB plays a critical role in protecting the brain from various harmful substances (e.g., toxins) that may be present in the bloodstream. The BBB is very selective, allowing only needed nutrients to cross while excluding many other molecules, including those that could be helpful, from the brain. Many modern biological drugs and diagnostic agents such as peptides, proteins (including antibodies), and oligonucleotides can be used to treat brain diseases; unfortunately, the majority of these

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_2, © Springer Science+Business Media, LLC, part of Springer Nature 2019 9 10 Brian M. Kopec et al.

molecules cannot cross the BBB. Large-molecule biologics (e.g., antibodies, enzymes, cytokines, hormones) can be delivered to the brain only via intracerebral ventricular infusion (ICV) [1]. ICV is accomplished by drilling into the skull and surgically installing a port to deliver drugs directly to the brain; therefore, it is not a stress-free procedure for patients. In addition, this method involves a great deal of risk, as the brain becomes directly exposed to potential infection by pathogens. Thus, there is an urgent need to investigate and develop noninvasive and patient-friendly alternative methods to deliver diagnostic and therapeutic agents to the brain. This chapter is focused on progress and new methods in delivering molecules for diagnosing and treating brain diseases.

2 Anatomy of the Blood-Brain Barrier (BBB)

The BBB is composed of tightly packed vascular endothelial cells that are anchored to a basement membrane and surrounded by astrocytes (Fig. 1a)[2]. These endothelial cells are connected at the intercellular junctions by “zipper” or “Velcro” proteins, which are divided into tight junctions, adherens junctions, and desmo- some regions (Fig. 1b). These proteins form homophilic and/or heterophilic protein-protein interactions at these three regions to glue the endothelial cells forming the BBB microvessels. In this case, the surface proteins from one cell membrane of endothelial cells form a bridge with their respective counterparts on the surface of the opposing cell membrane [3–5]. The vascular endothelial cells are also decorated with various efflux pump molecules (i.e., P-glycoprotein, multidrug-resistant protein) and proteolytic and/or metabolizing enzymes that act as gatekeepers to prevent or degrade molecules from crossing the BBB. At the top of the intercellular junction membranes on the blood side of the BBB reside tight junctions or occludens junctions. They are connected by homophilic interactions of occludins, clau- dins, and junctional adhesion (JAM) proteins (Fig. 1b). These tight interactions prevent the permeation of molecules with hydrody- namic radius larger than 11 A˚ through the intercellular junctions or paracellular pathways [6]. The adherens junction below the tight junction is linked by protein-protein interactions of cadherins, nectins, and platelet endothelial cell adhesion molecules (PECAM). These interactions are the primary adhesion force between two opposing cell membranes of the BBB. Thus, the tight junctions are the secondary seal of the intercellular junctions [2]. Finally, desmosomes are established below the adherens junctions and they produce homophilic desmocollin or desmoglein interac- tions as well as heterophilic desmocollin-desmoglein interactions. The desmocollins and desmogleins are in the cadherin family of Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 11

Fig. 1 (a) A cross-section diagram of the vascular endothelial cells of the BBB surrounded by pericytes, astrocytes, and neurons. Molecules that cross the BBB via transcellular- or paracellular-transport pathway are indicated by two different arrows. The presence of efflux pumps prevents some molecules to diffuse transcellularly across the membranes of the BBB. (b) A representation of the intercellular junction of microvessel endothelial cells of the BBB, which is composed of tight junction, adherens junction, and demosome regions. The protein-protein interactions at the extracellular space act as “Velcro” to connect the membranes between opposing cells. The protein-protein interactions in the intercellular junctions are connected via cross talk of the protein-protein interactions in the cytoplasmic domain

proteins. It should be noted that there is cross talk among all the intercellular junction proteins through their interactions with the cytoskeleton proteins.

3 In Vitro Models to Study Brain Delivery of Therapeutic and Diagnostic Agents

3.1 In Vitro Models Because in vivo delivery of molecules into the brain is challenging, of the Blood-Brain much effort has been invested in developing in vitro models for the Barrier BBB to rapidly identify molecules that can cross the BBB. Trans- port mechanisms of molecules across the BBB can be characterized into several categories. First, molecules can passively diffuse through the cell membranes of the endothelial cells, which is called the transcellular pathway (Fig. 1a). It is well accepted that lipid- soluble nonpolar molecules can enter the brain via passive diffusion across the BBB. Unfortunately, some hydrophobic molecules that can partition to the cell membranes cannot cross the BBB because they are expelled by the efflux pumps (e.g., P-glycoproteins or Pgp) from the membranes back into the bloodstream (Fig. 1a). Second, molecules such as nutrients (i.e., folates, amino acid) and molecular carriers (i.e., transferrin) can permeate the BBB via receptor- mediated transport mechanism. This transport system involves a vesicular endocytosis process. Receptor-mediated systems have so far been accepted as one of the most promising approaches for noninvasive delivery of molecules across the BBB. Vesicular endo- cytosis without receptor involvement can also occur and this is 12 Brian M. Kopec et al.

called fluid-phase pinocytosis. Third, very small hydrophilic mole- cules and ions can penetrate the BBB via the intercellular junction in a process called paracellular pathway transport (Fig. 1). Various in vitro BBB models have been developed and widely used to rapidly evaluate the transport properties of molecules across the BBB during the discovery process. Compared to in vivo models, the in vitro BBB models are less expensive and more conducive to high throughput with reproducible results. Thus, the in vitro mod- els are very useful for screening potential candidates in the initial discovery phase of therapeutic or diagnostic agents for the central nervous systems (CNS). However, the in vitro models have their own disadvantages. For example, some of the models have leaky tight junctions; therefore, they are not good models to study penetration of molecules through the paracellular pathway. It is recommended that in vitro studies should be coupled with in vivo studies using the top candidate molecules after in vitro screening. It should be noted that, with appropriate choice and use of an in vitro model, the results from the in vitro model can be well correlated with the in vivo data [7, 8]. A common in vitro model of the BBB consists of a cell mono- layer isolated from primary brain endothelial microvessels or immortalized brain endothelial microvessels with many adapta- tions. Some in vitro models are composed of co-cultured cells with varying culture conditions to closely mimic the in vivo BBB. To generate the model, cerebral vascular endothelial cells from murine, rat, porcine, bovine, monkey, or human are isolated and cultured on the surface of a microporous membrane filter. The culture is fed with the appropriate nutrients to maintain the cell viability. The BBB co-culture models were also developed using a combination of brain endothelial cells with pericytes, astrocytes, and neurons to mimic the anatomic and physiology complexity of the in vivo BBB. Normally, the BBB co-culture model has tight intercellular junctions with high trans-endothelial electrical resis- tance (TEER) values [9, 10]. The in vitro BBB cell monolayer contains components similar to those of the in vivo microvessel with tight intercellular junctions, functional transporters, surface and intracellular metabolic enzymes, and efflux pumps. The limited porosity of the tight junc- tions in allowing movement of ions is reflected in the TEER value of the cell monolayer [11]. The TEER value determines the electri- cal resistance between the apical (AP) and basolateral (BL) sides of the monolayer. A high TEER value indicates very restricted move- ment of ions across the transcellular and paracellular pathways of the cell monolayer. The intercellular junction proteins act as gate- keepers of the paracellular pathways. Therefore, modulation of the interactions among the intercellular junction proteins of the BBB can result in leakiness of the tight junctions, which is reflected in Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 13

lowering of the TEER value due to the higher movement of the ions between the AP and BL sides. To mimic the BBB vasculature in a three-dimensional (3D) mode, the dynamic in vitro BBB (DIV-BBB) model has been developed to complement the monolayer BBB models. In this model, the endothelial cells are cultured in hollow fibers that resemble the blood vessels [12]. The DIV-BBB model can be used to simulate in vivo blood flow with tunable shear stress. This is a good model to evaluate trans-endothelial trafficking of immune cells across the BBB and to study cerebrovascular physiology in various disease states [12]. Unfortunately, the DIV-BBB model is less suitable for evaluating transport of molecules in a high- throughput manner because it is a difficult system to set up and establish. This model is also difficult to replicate because its genera- tion requires special expertise. Newer microfluidic models called “BBB on a chip” have been developed to model the dynamic flow of the BBB using a smaller number of cells and lower amounts of samples. These models consist of human brain endothelial cells (hCMEC/D3) that are cultured on a porous membrane placed at the interface between two microchannels, allowing the cell culture medium to flow through two electrodes. The electrodes are used to continuously measure TEER values of the cell monolayers. The hCMEC/D3 cell monolayers in the microfluidic device have a TEER value compara- ® ble to that of the well-established standard Transwell [13]. This model can be used to simulate disruption of the BBB in neuroin- flammatory diseases such as Alzheimer’s and Parkinson’s diseases. This study showed that the integrity of this BBB model was influ- enced by shear stress and inflammatory cytokines TNF-α as in neurodegenerative diseases.

3.2 In Silico Models As an alternative to laboratory experimental methods, computa- of the Blood-Brain tional simulation or in silico methods have been developed as tools Barrier to predict favorable properties of molecules for crossing the BBB [14]. These computational methods can be used during the early stages of drug design to include desirable properties in the mole- cule for penetrating the BBB. Compared to in vitro and in vivo experiments, computational methods are fast and low-cost meth- ods for screening molecules that can cross the BBB [14]. The algorithm for a certain method was developed using LogBB and LogPS parameters from in vitro and in vivo experiments of several known compounds [14]. LogBB is the concentration of drug in the brain divided by the concentration in the blood while LogPS is the product of permeability and surface area. The unknown values for novel compounds are then derived from regression models [15]. Molecular dynamics simulations were used to calculate the one-dimensional diffusion coefficient for each molecule. A combi- nation of diffusion coefficient and free energy landscape was used to 14 Brian M. Kopec et al.

determine the effective permeability (Peff) of a compound. The Peff was compared to LogBB and LogPS values to predict the ability of a molecule to cross the BBB [15]. This method has been shown to predict the BBB permeabilities of small molecules, and the predic- tions corelate with the experimental data [15].

4 Methods to Improve Brain Delivery of Drugs and Diagnostic Agents

4.1 Transcellular The traditional method to deliver drugs and diagnostic agents to Pathways the brain is via the transcellular pathway by which the molecule permeates from the bloodstream into the brain via the transcellular pathway of the BBB. Unfortunately, this pathway is only viable for small molecules with appropriate physicochemical properties that follow Lipinski’s rule of five [16–18]. In this case, a drug molecule should have a molecular weight of less than 500 Da with cLogP of less than five. The molecule should have fewer than five hydrogen- bonding donors and no more than ten hydrogen-bonding accep- tors. Therefore, during the discovery process, medicinal chemists utilize these rules to guide them in designing drug candidates that can penetrate the BBB. Alternatively, medicinal chemists have clev- erly created prodrugs that shield the hydrogen-bonding moieties in the molecule to improve their BBB permeation [19, 20]. Although a molecule can satisfy the Lipinski rules, its penetration through the BBB can also be inhibited by the efflux pumps and/or degradation by metabolizing enzymes [21]. Due to their physicochemical properties or violation of the Lipinski’s rules, biologics such as peptides, proteins, and oligonu- cleotides cannot passively diffuse through the cell membranes of the BBB (i.e., paracellular pathway). Therefore, it is very challeng- ing to deliver these types of molecules to the brain. There is a very limited number of peptides and proteins that can cross the BBB. These molecules can do this because they have transport receptors that carry them across the BBB. Therefore, many efforts are under way to find alternative routes and methods to deliver peptides, proteins, and oligonucleotides into the brain.

4.1.1 Prodrug Formation Prodrugs have been developed to temporarily alter the physico- for Passive Diffusion chemical properties of hydrophilic and charged molecules (i.e., Through Cell Membranes drugs or diagnostic agents) to improve their BBB permeation. of the BBB The prodrugs are formed by adding promoieties to the functional group(s) (i.e., alcohol, acid, amine) of the drug to change physico- chemical properties of the prodrug to make it more hydrophobic than the parent drug (Fig. 2). Alternatively, prodrug formation can also be used to improve the solubility of hydrophobic and insoluble drugs in water [22, 23]. Low water solubility prevents drug admin- istration in sufficient doses for oral and BBB bioavailability; thus, charged promoieties (e.g., phosphate ester) have been attached to Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 15

Fig. 2 The formation of prodrug molecules has been done to improve the delivery of drug molecules across the BBB. The formation of a prodrug changes the physicochemical properties to allow efficient partition into cell membranes for passive diffusion across the BBB. Ligand-conjugated prodrugs were also used to target BBB transport receptors to shuttle the prodrug from the blood to the brain. The prodrugs are converted to the parent drug by enzymes such as esterase or by chemical reaction upon pH change

the drug to produce salts of the prodrug to increase water solubility and dose. For charged drugs, the formation of uncharged prodrugs enhances their partition into cell membranes for passive diffusion across the BBB. Drugs with an acid functional group (pKa ¼ 4.0) or a basic functional group (pKa ¼ 8) functional group have a charge at physiological conditions, pH ¼ 7.4; thus, they have low membrane permeation (Fig. 2). In contrast, the formation of an ester prodrug for a drug containing a carboxylic acid removes the charge group; thus, the prodrug is expected to have a better BBB permeation than the parent acid drug. After crossing the BBB, the ester prodrug can be converted to the parent drug in the brain tissue by esterase enzyme(s) (Fig. 2). This conversion traps the charged drug molecules in the brain tissues. There are several factors that prevent prodrug molecules from crossing the BBB. First, the prodrug could be recognized by the efflux pumps; thus, although it could partition effectively into the cell membranes, the efflux pumps would expel the prodrug from the cell membranes. Second, it cannot cross the BBB because of its premature 16 Brian M. Kopec et al.

Fig. 3 Cyclic prodrugs (1 and 3) from RGD peptidomimetic (2) and Aggrastat (4) using (acyloxy)alkoxy and phenylpropionic acid promoieties, respectively. The cyclic prodrugs can be converted by clipping the ester bond by esterase to make the intermediate. The intermediates rapidly degraded chemically to release the parent drugs

conversion to the parent drug by an enzyme (i.e., esterases) in the systemic circulation or inside the endothelial cells. Third, the pro- drug could be converted by metabolizing enzymes (e.g., CYP450) to a more hydrophilic molecule to prevent its exit from the BBB endothelial cells. Due to backbone amide bonds, peptides have a very high number of hydrogen-bonding donors and acceptors as well as a high molecular weight; these are the reasons that most peptides cannot readily cross the BBB [24, 25]. One way to improve the partitioning of peptides or peptidomimetics into cell membranes is by forming cyclic peptide prodrugs (Fig. 3)[24, 25]. Formation of a cyclic peptide prodrug increases the number of intramolecular hydrogen bonds and lowers the hydrogen-bonding potential to surrounding water molecules. In addition, cyclic peptide prodrugs have rigid backbone structures with stable and compact conforma- tions for better partitioning into cell membranes of the BBB. Cyclic prodrugs of RGD-peptidomimetics (e.g., Aggrastat) were designed using “trimethyl lock” phenylpropionic acid and acyl(oxy)alkoxy promoiety (Fig. 3)[26–28]. RGD-peptidomimetics and Aggrastat cannot be delivered orally because they cannot penetrate the biological barriers due to their charges. The formation of a cyclic Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 17

prodrug (1) enhanced its cell membrane penetration 4.4 times compared to the parent RGD-peptidomimetic (2) in an in vitro cell culture model of the intestinal mucosa (i.e., Caco-2 cell mono- layer) [26]. Similar results were found in the formation of cyclic prodrugs of the opioid peptides using “trimethyl lock” phenylpro- pionic acid and acyl(oxy)alkoxy promoieties [29–31]. The advan- tage of cyclic peptide prodrugs is that they can be converted to the parent linear peptides by esterase after crossing the biological bar- riers in vitro and in vivo (Fig. 3). The ester bond is clipped by esterase to form an intermediate followed by fast chemical reaction to release the parent compound (Fig. 3). The disadvantage is that the cyclic peptide prodrug can be prematurely converted to the parent peptide before or during transport across BBB. The metab- olism of cyclic peptide prodrugs by CYP450 while crossing the BBB can also be a hurdle for the development of these compounds.

4.1.2 Drug Conjugates Native transport receptors on the BBB have been exploited to for Receptor-Mediated improve drug delivery to the brain. Receptors of transferrin, insu- Transport Through the BBB lin, IGF-I, IGF-II, angiotensin II, and natriuretic peptide receptor- 7 have been targeted by their respective ligands to deliver drugs to the brain as protein–drug conjugates (Fig. 2). Monoclonal antibo- dies (mAbs) against transferrin receptors (TfRs) on the luminal side of the BBB have been used to carry drug payloads into the brain by forming antibody-drug conjugates (ADC) [32–36]. This method has been referred to as the “Trojan horse” method. The mAb targets an epitope on the receptor that is distinct from the binding site of its endogenous ligand without interfering with the receptor function to shuttle its native ligand into the BBB endothelial cells or across the BBB [37]. The ADC is internalized and transported from the luminal through the vascular endothelial cells to the abluminal side of the BBB to exert the drug activity in the brain. Transferrin receptor (TfR) is an endogenous receptor that binds to holo-transferrin (hTf) carrying iron into the cells. Once the hTf-TfR complex enters the endosome intracellular, it experi- ences a pH change from 7.4 to 4.5–6.5, which triggers the release of hTf from TfR and the release of the iron from hTf [37]. Then, the TfR is recycled back onto the cell surface. Thus, this mechanism has been explored to deliver anti-TfR mAbs such as OX26 and 8D3 mAbs that carry drugs or proteins across the BBB into the brain [37]. The I.V. administration of OX26 mAb resulted in its detec- tion in the brain vasculature at 0.44% of the administered dose after injection. OX26 mAb was co-localized with factor-VIII mAb that marked the branched capillaries of the rat brain [38–40]. OX26 mAb was deposited at the parenchyma region that was associated with neurons and closed to the ventricular system. It was proposed that TfR was responsible for the transcytosis of OX26 mAb into the brain; the explanation given was that plasma clearance of OX26 18 Brian M. Kopec et al.

mAb was faster than the IgG control [41]. Using immunohisto- chemistry and capillary depletion methods, it was found that 90–95% of OX26 was in capillaries in the brain parenchyma, sug- gesting that the majority of the mAb was trapped inside the endo- thelial cells without crossing the BBB. Because the mAb was detected in cerebrospinal fluid (CSF), it was also suggested that the mAb entered the brain via cerebroventricular organs or the blood-CSF barrier. Administration of 8D3 anti-TfR mAb via I.V. injection indi- cated a brain deposition of the mAb at the 1-h time point after injection; unfortunately, the extent of brain deposition was difficult to determine because capillary depletion experiments were not carried out [37, 42, 43]. Pegylated immunoliposomes (PILs) con- jugated to 8D3 mAb and loaded with beta-galactosidase were found in various brain regions (septo-striatal, rostal, and caudal) of adult male Balb/c mice at 1–3 days after I.V. administration [44]. In another study, 1% of the I.V. administered dose per gram of biotinylated radiolabeled 125I-amyloid-beta 1–40 (Bio-125I- Aβ1–40) bound to streptavidin-8D3 (SA-8D3) mAb was deposited Swe in the brain of App /Psen1 double-transgenic Alzheimer’s disease (AD) mice at the 1-h time point [45]. The peptide was found sequestered around the amyloid beta (Aβ) plaques in the brain, suggesting the possibility of using this complex as a brain diagnostic tool for AD [45]. The low efficiency of the 8D3 mAb transport across the BBB into the parenchyma can be attributed to trapping of the mAb inside the BBB endothelial cells and the mAb localiza- tion in the brain parenchyma is still controversial [46]. Studies suggest that 8D3 mAb bound to vasculature endothelial cells. Confocal microscopy studies indicated that 8D3 mAb co-localized with collagen IV as a basal laminal marker, suggesting that the mAb traversed from the luminal (blood) to the abluminal side of the brain endothelial cells without releasing the mAb from the TfR at the abluminal side. The inability to release the mAb was due to its high affinity to TfR [46]. One of the potential solutions to solve the problem of mAb release from TfR at the abluminal side is to design low-affinity mAbs [46]. Alternatively, mAbs with pH-sensitive binding proper- ties have been designed to improve their transcytosis. mAbs with lower binding to TfR at pH 5.5 than at pH 7.4 have higher transcytosis into the brain than mAbs that have higher binding at both pH 5.5 and 7.4 [47]. The majority of mAb molecules with high affinity to TfR were found in the brain capillaries even 24 h after administration; in contrast, the low-affinity mAbs were detected at the brain parenchymal and co-localized with a neuronal marker [46]. The high-affinity mAb-TfR complex degraded mainly in lysosomes compared to the low-affinity complex; this is presum- ably due to the long residence time of the high-affinity complex in Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 19

the lysosomes [48]. This finding was congruent with in vivo obser- vation where the high-affinity mAb-TfR complex caused a low- unbound-fraction TfR on the cell surface. This low-unbound frac- tion was due to the degradation of mAb-TfR complex in lysosomes. This was also one reason why the high-affinity mAb had low brain exposure. To attenuate mAb affinity for TfR, a monovalent TfR mAb was developed for improving brain exposure and suppressing degrada- tion in the lysosomes [49]. In this case, the Fc region of mAb31 that recognizes Aβ plaques was linked to one or two Fab fragment (s) of TfR mAb to produce sFab-mAb31 and dFab-mAb31 [49]. The majority of the monovalent sFab-mAb31 was transcy- tosed to the brain more efficiently than was divalent dFab-mAb31. sFab-mAb31 was distributed throughout the parenchyma and around Aβ plaques in the brains of PS2APP mice. The lysosome degradation of dFab-mAb31 was fast, suggesting one reason for its low transcytosis. dFab-mAb31 had no significant decoration of the plaques compared to control, which confirmed its slow transcytosis properties. In summary, the results suggest that mAb with a weak binding to the carrier receptor is necessary for mAb release in the brain. Although extensive research efforts have been carried out to exploit the use of TfR to transcytose Tf mAb with conjugated drugs into the brain, moving this technology into the clinic has been challenging [37]. One of the reasons is the low efficiency of drug transport from blood to the brain. The degradation of mAb in endosomes as well as its trapping in the endothelial cells contribute to its low brain deposition. It is challenging to find an appropriate balance between binding affinity and release property of the mAb from TfR at the abluminal side of the BBB. The presence of TfRs in various tissues or organs could potentially elicit off-target side effects. Although this method has not yet reached clinical applica- tions, it has contributed insights into mechanisms of actions and stimulated interest in evaluating various transport receptors as car- rier molecules on the BBB. Insulin transporters other than TfR on the BBB have been exploited to deliver drugs, including proteins, into the brain by their conjugation to insulin [33, 50]. Because insulin has a short half-life (t1/2 ¼ 10 min) in the systemic circulation, a high dose of the conjugate has to be administered to reach the needed dose in the brain [33, 50]. One disadvantage of this method is that the insulin conjugate can cause hypoglycemia in the subject after sys- temic delivery [50]. Alternatively, a mAb targeted to an insulin receptor has been studied for delivering brain-derived neurotrophic factor (BDNF) to the brain. BDNF has been shown to induce neuroregeneration in animal models of brain diseases (i.e., MS, Alzheimer’s disease) when administered via ICV. In another 20 Brian M. Kopec et al.

study, the heavy-chain C-terminus of the Tf mAb (HIRMab) was fused to BDNF to produce HIRMab-BDNF. The fusion produced binding affinities similar to Tf mAb to TfR and BDNF to TrkB receptors with affinity similar to that of BDNF. The results suggest that fusion of the two proteins did not influence the binding affinity of each protein to its respective target receptor [33]. Three hours after I.V. administration of 3H-HIRMab-BDNF in rhesus monkeys, a significantly higher brain deposition of 3H-HIRMab- BDNF (24 ng/g brain) than 3H-IgG2a was observed in brain homogenates. It was found that the amount of fusion protein in the brain was tenfold higher than that of endogenous BDNF control, indicating that the mAb enhanced BDNF delivery into the brain [33].

4.1.3 Liposome- The abilities of different ligands to deliver liposomes into the brain Mediated Brain Delivery were investigated by targeting their respective receptors on the BBB. Ligands such as proteins (i.e., transferrin, RI7217 mAb, CRM19) or peptides (i.e., ICOG133, angiopep-2) were used to decorate the surface of liposomes [51–53]. The receptors for these ligands are found on the endothelial cells of the BBB. Both trans- ferrin and RI7217 mAb bind to TfR on the BBB. It is well known that TfR transports iron on transferrin (Tf) into the brain. RI7217 mAb has been shown previously to transport loperamide across the BBB [54]. CRM-197 is a ligand for diphtheria toxin receptor (DTR) and heparin-binding epidermal growth factor (HB-EGF) expressed on the BBB surface, and CRM-197 is a nontoxic mutant of diphtheria toxin (DT) that contains a single-point G52D muta- tion. COG133 peptide is derived from apolipoprotein E (apoE) and COG133 peptide binds to low-density lipoprotein receptor (LDLR). LDLR is responsible for transporting cholesterol and lipids into CNS [55]. Angiopep-2 peptide binds to low-density lipoprotein receptor-related protein 1 (LRP1). The liposomes were decorated with targeting ligands via a maleimide PE linker embedded in the liposomes. The liposomes were composed of EPC, cholesterol, EPG, and MPB-PE with a molar ratio of 6.5:2.6:0.8:0.1 [56]. Mouse brain endothelial cells (bEnd.3) and human (hCMEC/ D3) brain endothelial cell monolayers were used to evaluate the receptor-mediated endocytosis of 3H-labeled liposomes decorated with different ligands by incubating them at 4 or 37 C for 1 h. The results showed that only liposomes decorated with CRM-197 have significant binding compared to control liposomes in both endo- thelial cells. In contrast, liposomes decorated with RI7217 mAb bind only to human (hCMEC/D3) brain endothelial cells; these liposomes were engulfed by hCMEC/D3 cells significantly better than liver and kidney cells, indicating the selectivity of the hCMEC/D3 cells. COG133-labeled liposomes did not bind and Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 21

cross the BBB in the in vitro and in vivo models, suggesting that COG133 peptide does not have the necessary sequence or proper- ties for binding and uptake by the LDLR. Although angiopep-2- paclitaxel conjugate has been shown to cross the BBB, there was no uptake of angiopep-2 liposomes by the endothelial cells; this could be due to the density and/or accessibility of peptides on the surface of liposomes for recognition by LRP1 on the endothelial cells. It has been shown previously that ligand density has an important role in the targeting efficacy of nanoparticles or liposomes [57], and different ligands can have different optimal ligand densities to elicit the same binding outcomes [58]. Because RI7217 liposomes were engulfed by hCMEC/D3 cells, further studies were carried out in vivo. RI7217 liposomes labeled with non-exchangeable [3H]cholesteryl hexadecyl ether (5 Ci/mol total lipid) were administered via I.V. in mice and, after 12 h of circulation time, the mice were sacrificed to collect the blood, brain, and other major organs [51]. The liposomes were cleared from the blood in a 12-h time period. More RI7217 lipo- somes were found in the brain fractions compared to untargeted liposomes after capillary depletion. There was no significant differ- ence in the uptake of RI7217 over untargeted liposomes in any other organs, suggesting selectivity of RI7217 for the brain [51]. The RI7217 liposomes were found in significant amounts in parenchyma, cerebrum, and cerebellum. The amounts of RI7217 liposomes in the brain parenchyma fraction were 4.3 and 2.6 times higher than the untargeted liposomes at 6- and 12-h time points after administration, respectively. The amount of RI7217 lipo- somes was ten times higher than that of untargeted liposomes in the brain capillary fraction, suggesting that RI7217 liposomes bound to TfR on the BBB. Finally, a fraction of the RI7217 lipo- somes crossed the BBB endothelial cells into the brain. In conclu- sion, the amount of liposomes that reached the brain was about 0.18% after 12 h, which was considered a low amount of delivered liposomes [51]. It has been suggested that this delivery method be used only for very potent drugs such as opioids, which need a very low dose in the brain.

4.1.4 Brain Delivery of Nanoparticles have been exploited for delivering molecules to the Nanoparticles brain, and the types of nanoparticles such as shapes, sizes, and physical properties of the particles may play a significant factor in their uptake into the brain. The presence of targeting ligands can help direct the nanoparticles to the BBB and possibly into the brain. Ligands that can enhance the uptake of liposomes into the brain can also be used to deliver different types of nanoparticles (e.g., poly- meric, albumin). Magnetic nanocrystals of Fe3O4 have been used as a magnetic resonance contrast agent for early detection of brain diseases. 22 Brian M. Kopec et al.

Nanoparticle surfaces have been decorated with lactoferrin (Lf) to make Lf-Fe3O4 nanoparticles. Lf is involved in host defense mechanisms against severe inflammation or infection and normally accumulates in the brains of neurodegenerative disease patients [59, 60]. Lactoferrin receptors (LfR) are expressed on the BBB endothelial cells for transcytosis of Lf into the brain [59]. The uptake of Lf by LfR is more efficient than that of transferrin or OX-26 antibody by TfR; thus, Lf is potentially a better ligand to deliver nanoparticles across the BBB [61]. Lf was conjugated to the nanoparticles via one of the carboxylic acid terminals of the poly- ethylene glycol (HOOC-PEG-COOH) that coat the surface of iron nanoparticles. The role of PEG molecules is to improve water solubility and biocompatibility of the particles as well as to reduce protein adsorption, reticuloendothelial system (RES) uptake, and immunogenicity stimulation [60]. At all particle con- centrations, PEG-coated Fe3O4 nanoparticles without Lf (untar- geted nanoparticles) disrupt the tight junctions of the primary porcine brain capillary endothelial cell (PBCEC) monolayers as indicated by decreasing TEER values of cell monolayers. In con- trast, incubation of Lf-Fe3O4 nanoparticles maintained the mono- layer integrity at concentrations of 0.04 and 0.1 mg Fe/mL. Lf-Fe3O4 nanoparticles disrupt the BBB tight junction integrity at a high particle concentration (0.3 Fe mg/mL). The results suggest that Lf conjugation suppresses the BBB tight junc- tion disruption by untargeted nanoparticles. The BBB transport properties of Lf-Fe3O4 (Lf-labeled) nano- particles were evaluated in vitro and in vivo. In vitro incubation of Lf-Fe3O4 nanoparticles (0.04 and 0.1 Fe mg/mL) for 18 h on the apical (AP) side of PBCEC cell monolayers resulted in observation of the transported nanoparticle at the basolateral (BL) side as determined using atomic absorption spectrometer (AAS). About 22% of the nanoparticles crossed the BBB monolayers. The nano- particle transport was inhibited by excess Lf alone (16Â) from 22% to 1% transport, suggesting that the particle was transported via LfR-mediated transcytosis. The Lf-Fe3O4 nanoparticles were eval- uated in vivo using SD rats, and the brain distribution of the nanoparticles was detected using a 7.0 T animal magnetic reso- nance imaging (MRI) instrument. Coronal and axial T2 relaxation data were collected at pre- and 24-h post-tail vein injections to detect nanoparticle deposition in the brain. Axial T2 contrast images showed that both Lf-labeled and unlabeled particles reached the brain, but the Lf-labeled nanoparticles showed greater contrast than the unlabeled particles in the thalamus, brain stem, and frontal cortex. These results suggest that the labeled nanopar- ticles traverse the BBB and access the brain via LfR-mediated uptake. There was a higher localization of Lf-labeled nanoparticles in the brain microvessels compared to unlabeled nanoparticles. Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 23

Because the unlabeled nanoparticles were also found in the brain, their transport across the BBB could be due to the disruption of the BBB intercellular junctions [60].

4.1.5 Exosome-Mediated Exosomes are 40–100 nm nanosized bubble-like particles found in Brain Delivery body fluid that are secreted by various types of cells [62]. They are formed via invagination of multivesicular body followed by fusion of cell plasma membranes, but do not contain lysosomes or mito- chondria. The surfaces of exosomes are decorated with endogenous proteins from the producing cells, and these exosomes can be exploited to target certain cells or tissues. The most exciting char- acteristic of exosomes is that they can travel from one cell to another to release their contents (e.g., proteins, RNA) in the intra- cellular space of destination cells. There are several other unique exosome properties that are beneficial for drug delivery vehicles. First, small-to-large molecules can be loaded into exosomes for delivery. Second, surface proteins on exosomes can be used to direct them to intended target cells for uptake. Third, exosomes have high plasma stability with a long half-life in the systemic circulation and various tissues [62]. Rhodamine-loaded exosomes have been delivered to the brains of zebrafish and fluorescence dye was detected in the brain tissue, excluding the vasculature network [63]. The doxorubicin-loaded exosomes have also been shown to target DiD-labeled U-87 MG cancer cells in zebrafish brain. In addition, 5 days after treatment with doxorubicin-loaded exosomes, the tumor size in the brains of zebrafish was reduced. In another example, siRNA-loaded RVG exosomes were delivered to wild-type mouse brain and BACE1 enzyme expression that is important in Alzheimer’s disease was suppressed [64]. This delivery of this RVG exosomes was relatively selective to target siRNA into the brains and avoided uptake by tissues outside of the brain [62]. In summary, although the exosome delivery system seems bet- ter than other synthetic drug carriers, this technology still has many challenges to overcome. The first is to identify the proper donor cells to generate the desired and appropriate exosomes. Second, it is difficult to find the appropriate method to efficiently load the drug into exosomes. Third, the pharmacokinetic properties of exosomes have been challenging to determine. Fourth, it is still difficult to provide consistent targeting molecules on the surface of exosomes. Finally, the ability to scale up exosome production and to load them for clinical use are still barriers that need to be overcome.

4.1.6 Viral Vector Brain Delivering genes across the BBB faces the same challenges as deliv- Delivery ering conventional therapeutics. Due to the chemical and enzy- matic instability of genes, a high dose of genes is required for their successful therapeutic use; unfortunately, this high dose in the 24 Brian M. Kopec et al.

systemic circulation could increase side effects such as hepatotoxic- ity [65]. Viral vectors have been somewhat successful in delivering genes to various tissues other than the brain; however, this success has been tainted by viral infections in treated patients due to impurity of the viral vectors. The viral vectors exploit their viral infection mechanisms to enter into the brain by crossing the BBB. To enter the brain, viral vectors utilize surface proteins to bind target receptors on the endothelial cells for crossing the BBB. The vectors also use their own surface proteins to bind receptors on the target cells to deliver genes into the cells. Inside the cell, the viral vectors follow the intracellular trafficking process and escape the endosome to enter the nuclear of cells. Finally, the capsid is unpacked and the genome is inserted in the cell nuclear to initiate transcription and translation processes. Nonpathogenic recombinant adeno-associated viruses (rAAVs) with certain capsid serotypes have been shown to cross the BBB and deliver genetic material to neurons and astrocytes in the brain. Thus, AAVs have been developed to carry 4.7 kb single-stranded DNA for gene expression in target-host cells. To design viral vec- tors that can cross the BBB for delivering the genes to the desired location or cells in the brain, the 60 viral proteins (VP) with sequence diversity were used to assemble to make AAVs [66]. AAV1 and AAVrh.10 are viral vectors that share 85% sequence homology; they were discovered through DNA shuffling. AAV1 viral vector did not cross the BBB and can transduce the brain vasculature only; in contrast, AAVrh.10 vector has a different capsid domain to traverse through the BBB [65]. To evaluate whether viral vectors could cross the BBB, a GFP-expressing cassette was packaged into AAV1, AAVrh.10, and six other chimeric variants of viral vectors. Then, they were administered via I.V. injections at a dose of 5 Â 1011 viral genomes (vg)/per mouse into 6–8-week-old mice. Three weeks later, the brains were isolated and subjected to immunostaining. It was found that AAV1 vector was located only in the brain vasculature while AAVrh.10 showed robust GFP trans- duction in neurons, glia, and endothelial cells. An AAV1RX vector was developed by grafting eight residues from AAVrh.10 onto AAV1 to make this viral vector able to cross the BBB. This vector can deliver genes to selectively transduce neurons, indicating that the eight residues have a critical role in allowing the vectors to penetrate the BBB. Unfortunately, AAV1RX viral vectors can also transduce liver and cardiac tissues in a manner similar to that of AAV1, suggesting that other struc- tural domains of the AAVrh.10 capsid play a role in a systemic transduction profile. From the results, it is uncertain whether this eight-residue footprint is sufficient to selectively deliver genes to the brain and facilitate a transduction profile similar to that of AAVrh.10 vector. Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 25

AAV1R6 and AAV1R7 are two viral vectors that were designed from AAV1 with 97% identical amino acid sequence that have an important 22-amino-acid residue derived from AAVrh and AAVrh.10. This amino acid residue is important for crossing the BBB. The design was based on the overall beta-barrel structure of core viral protein 3 (VP3) with VP7 antiparallel beta-strands connected by interlocking loop regions. The highly variable and surface-exposed loop regions on AAV capsids have been modified to control tissue tropism, transduction profile, and antigenicity [66, 67]. Similar to the AAVrh.10 vector, AAV1R6 and AAV1R7 vectors have a robust transduction on cortical neurons and other brain regions; in contrast, they reduced transduction in glia and vasculature when compared to AAVrh.10. This result confirms that AAV1R6 and AAV1R7 vectors can cross the BBB. AAV1RX6 and AAV1RX7 have the potential for developing gene therapies against neurological disorders including spinal muscular atrophy. These vectors do not target peripheral organs (i.e., liver and cardiac tissues); thus, they avoid potential hepatotoxicity side effects. Therefore, future successful clinical applications of these vectors will rely on their efficacy in large animals such as nonhuman pri- mates [65, 67].

4.2 Brain Delivery of One alternative method to deliver molecules through the BBB is via Molecules Through the paracellular pathways by increasing the opening or porosity of Paracellular Pathways the tight intercellular junctions. One of the most successful meth- of the BBB ods in the clinic to deliver molecules through paracellular pathways is osmotic delivery, which has been used to deliver anticancer drugs to treat brain tumor patients [68–76]. The osmotic method involves injecting a hypertonic solution of mannitol into the carotid artery to shrink the BBB endothelial cells and creates a disruption of the tight intercellular junctions. This disruption generates larger pores in the intercellular junction to increase the penetration of anticancer drugs into the brain [73, 75, 77]. It should be noted that the hypertonic mannitol works via increasing the osmotic pressure in the bloodstream and does not disrupt the intercellular junction proteins in a selective manner. Several methods have been developed to increase the porosity of the intercellular junctions by disrupting protein-protein interac- tions in the intercellular junctions. Peptides derived from the sequences of the extracellular domains of intercellular junction proteins have been developed to selectively inhibit protein-protein interactions in an equilibrium fashion to increase permeation of molecules through the BBB. Claudin and occludin peptides have been shown to modulate the intercellular junctions of the biological barriers in vitro and in vivo [78–80]. The C1C2 peptide (Table 1) derived from claudin-1 can deliver opioids and tetrodo- toxin into rat brains [78]. The C1C2 peptide can open the BBB for extended period of time (i.e., 3 days); however, this long period of 26 Brian M. Kopec et al.

Table 1 Peptide names and sequences

Peptide Sequence

C1C2 SSVSQSTGQIQSKVFDSLLNLNSTLQATR-NH2 OCC2 GVNPQAQMSSGYYYSPLLAMC(Acm)SQAYGSTYLN QYIYHYC(Acm)TVDPQE; Acm ¼ Acetamido methyl AT-1002 FCIGRL PN-78 FDFWITP

PN-159 KLALKLALKALKLAALKLA-NH2

HAV6 Ac-SHAVSS-NH2 ADTC5 Cyclo1,7(CDTPPVC) cIBR7 Cyclo1,8(CPRGGSVC) cLABL Cyclo1,12(PenITDGEATDSGC)

opening of the BBB may not be desirable. Thus, further investiga- tion is needed to find peptide derivatives that can modulate the tight junctions of the BBB for a shorter period of time. Several peptides that are related to tight junction proteins were designed to increase the paracellular porosity of the biological barriers. OCC2 peptide (Table 1) derived from loop 2 of chick occludin was active in disrupting the tight junctions and decreasing the TEER values of A6 epithelial cell monolayers. OCC2 peptide also enhanced paracellular permeation of dextran-3000 and 40,000 across the A6 epithelial monolayers. However, OCC2 has not been evaluated to improve the delivery of molecules across the BBB in in vitro and in vivo models. A peptide called AT-1002 (Table 1) that was derived from zonula occludens toxin (ZOT) protein improved the paracellular delivery of low-molecular-weight heparin orally. AT-1002 disrupts the tight junctions of Caco-2 and brain endothe- lial cell monolayers. C-CPE peptide from a microbial toxin frag- ment can modulate claudin-3 and -4 in tight junctions of Caco- 2 but not in brain endothelial cell monolayers. PN-78 and PN-159 peptides (Table 1) were discovered from phage display studies, and both peptides significantly lowered the TEER values of Caco-2 and brain endothelial cell monolayers. PN-159 was very potent in enhancing the delivery of fluorescein and albumin through Caco- 2 and brain endothelial cell monolayers. Peptides (i.e., HAV and ADT peptides) that are derived from the extracellular 1 (EC1) domain of E-cadherin have been shown to modulate in vitro biological barriers (i.e., Caco-2 and MDCK cell monolayers) [81–83]. HAV and ADT peptides can modulate the BBB intercellular junctions and enhance brain delivery of molecules Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 27

in mice and rats [84–89]. The proposed mechanism of action of HAV and ADT peptides is via their binding to the EC1 domain of cadherins. Heteronuclear multiple quantum coherence (HMQC) NMR spectroscopy experiments show that HAV and ADT peptides bind to different binding sites on the EC1 domain of E-cadherin [90]. HAV6 and ADTC5 peptides enhanced the brain delivery of 14C-mannitol as a paracellular marker [86, 87]. In addition, HAV6 peptide enhances the brain delivery of efflux pump substrates such as 3H-daunomycin into rat brains using in situ rat brain experi- ments [84] and IRdye R800 into the mouse brain after I.V. administration [87]. These results suggest that cadherin pep- tides can be used as an alternative way to deliver efflux pump substrates to the brain. To further evaluate the BBB modulatory activity of cadherin peptides, an MRI contrast agent, Gd-DTPA, was delivered via I.V. administration together with HAV6 or ADTC5 peptide in mice; the brain depositions of Gd-DTPA were detected using MRI in living animals [86, 87]. The results showed that the depo- sition of Gd-DTPA was observed in the brains within 3.0 min after administration, and both peptides significantly enhanced brain delivery of Gd-DTPA compared to a control vehicle in mice. The duration of tight junction modulation was evaluated using MRI; in this case, the HAV6 peptide was administered alone, and Gd-DTPA was administered via I.V. route after 1 h [87]. The result showed that there was no significant enhancement of Gd-DTPA compared to control, suggesting that the duration of opening of the BBB by HAV6 peptide was less than 1 h. A similar study showed that ADTC5 peptide could open the BBB between 2 and 4 h, which is longer than when using HAV6 peptide [86]. ADTC5 peptide enhances the brain delivery of 8-mer cIBR and 12-mer cLABL peptides in rats and mice, respectively [89]. The deposition of intact cIBR peptide can be detected in rat brain by LC-MS/MS. HAV6 and ADTC5 peptides have been shown to significantly enhance the delivery 65 kDa albumin conjugated with Gd-DTPA (i.e., galbumin) compared to control peptide or PBS as detected by MRI. The deposition of the protein could be detected in the brain 3.0 min after administration [89]. For a large protein such as galbumin, HAV6 could be used only when both HAV6 and galbumin were delivered simultaneously; however, a 10-min delay of galbumin administration after delivery of HAV6 did not enhance galbumin brain delivery. In contrast, ADTC5 was able to enhance the delivery of galbumin into the brain after a 10-min delay; however, a 40-min delay of galbumin after ADTC5 injection did not show any enhancement of brain delivery. The results suggest that the opening of the BBB by cadherin peptides is short for large molecules compared to the opening for small molecules. It is proposed that cadherin peptides generate large, medium, and small pores in the intercellular junctions of the BBB 28 Brian M. Kopec et al.

and the large pores rapidly collapse to medium and small pores followed by the collapse of medium pores to small pores in a time- dependent manner. In summary, many potential disruptors of the intercellular junctions of the BBB have the potential to enhance the delivery of diagnostic and therapeutic molecules into the brain. The applica- bility and safety of these modulators to deliver therapeutic and diagnostic agents to the brain still need further investigation. The next step is to use these modulators to deliver functional molecules to the brains of animal models of brain diseases.

5 Current Methods to Diagnose Brain Disease

Many imaging methods (e.g., MRI, X-ray, CATscan) have been used to diagnose brain diseases (i.e., Alzheimer’s, Parkinson’s, multiple sclerosis, stroke, and brain tumors). Some of these meth- ods utilize molecular imaging techniques by delivering highly sen- sitive probes into the brain in a noninvasive manner. These probe molecules can be powerful detection tools for detecting changes in the brains during the progression of brain diseases; these changes include the brain cellular activities and morphologies as well as the protein compositions to diagnose brain diseases in clinical and preclinical settings. Early detection of brain diseases is important for halting their progress; therefore, the use of specific and sensitive molecular imaging techniques is necessary. In both preclinical and clinical studies, several molecular imaging techniques such as mag- netic resonance imaging (MRI), positron-emission tomography (PET), single-photon-emission computed tomography (SPECT), and optical imaging techniques including near-infrared (NIR) fluo- rescence imaging have been greatly advanced for diagnosing brain diseases ex vivo and in vivo [91]. Neuroinflammation can be a common sign of brain diseases such as stroke, multiple sclerosis, Alzheimer’s, malignant cancer, and Parkinson’s. The neuroinflammation process activates inflam- matory immune cells that infiltrate the brain along with the upre- gulation of inflammatory cytokines. Therefore, changes in the balance between the activation of inflammatory and regulatory immune cells in the brain can be the target of diagnostic agents for brain disorders. Many current diagnostic molecules for brain diseases are limited to small hydrophobic molecules that can cross the BBB. These small molecules typically have a molecular weight less than 500 Da and have high binding capacity to the target tissue (s) in the brain. In contrast, many peptides, nucleotides, and pro- teins (e.g., enzymes and antibodies) are molecules that have high selectivity to bind cellular components of the brain (e.g., proteins and sugars) for detecting changes in the brain. Unfortunately, these types of molecules cannot be delivered to the brain because they Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 29

cannot cross the BBB. Thus, if these types of molecules could be delivered to the brain, they could be utilized to improve the diag- nostics and treatments of brain diseases. Several methods used in attempts to deliver them to the brain have been described above. The infiltration of leukocytes into the brain during inflamma- tion occurs via extravasation through the intercellular junctions of the BBB [92]. This process involves various cell adhesion molecules on the leukocytes as well as on the endothelial cells of the BBB [93]. The process of extravasation of immune cells is initiated by rolling of these cells of the surface of endothelium mediated by L- and E-selectins. When the immune cells stop rolling, they firmly adhere to the endothelial cells, which are mediated by ICAM-1 and VCAM-1 immunoglobulins on the endothelial cells that bind to LFA-1 (αLβ2) and VLA-4 (α4β1) integrins, respectively, on the surface of leukocytes [92, 93]. Natalizumab, a monoclonal anti- body that binds to α4-subunit of VLA-4 integrin, is being used to treat MS because it can prevent adhesion of immune cells to the BBB. Therefore, it prevents brain infiltration of immune cells to suppress disease exacerbation in MS patients. This indicates upre- gulation of cell adhesion molecules (e.g., ICAM-1) on the endo- thelial cells of the BBB, and the increased expression of cell adhesion molecules on the BBB endothelial cells of MS patients can be explored as diagnostic markers of MS [94].

5.1 Magnetic MRI has been used extensively to diagnose brain diseases. The Resonance Imaging experimental autoimmune encephalomyelitis (EAE) disease in (MRI) mice has been used as an animal model for multiple sclerosis (MS). In animal models and patients with MS, MRI detects T1 black holes in the brain and spinal cord [95, 96]. In EAE animals, T1- and T2-weighted images using high-resolution spin-echo sequences are being used to determine the lesion loads. With the help of MRI contrast agents (e.g., Gd-DTPA, Gd-DOTA), type A or B lesions can be detected where the hypo-intensity in both T1- and T2-weighted MRI is usually related to type A lesions. In contrast, type B lesions usually are correlated with the observation of hyper- intensity in T2-weighted MRI with reduced signal of T1-weighted MRI [97]. A high infiltration of inflammatory cells and myelin loss are indications of type A lesions; moderate inflammatory cell infiltra- tion and myelin loss correspond to type B lesions [97]. The brain deposition of an MRI contrast agent, Gd-DTPA, has been used to assess the development of MS in patients as well as in EAE animal models. Prior to the infiltration of immune cells, the BBB breakdown is manifested in the leakiness of the paracellular pathways, allowing higher Gd-DTPA permeation through the BBB. In addition, ultrasmall particles of iron oxide (USPIO) have been used to track brain infiltration of macrophages and monocytes as a diagnostic for the development of MS disease. In this case, USPIO 7228 (600 μmol iron oxide/kg) was administered via 30 Brian M. Kopec et al.

Fig. 4 Structures of molecules used in brain analysis including (a) Gd-bis-5-HT-DTPA, (b) DMPO, (c) 18F-FDG, (d) 11C-PK11195, (e) 11C-DAA1106, (f) 11C-CIC, (g) 11C-flumazenil, (h) IR780 iodide, and (i) P1 as NIR-II dye

I.V. injection into EAE mice. Because macrophages and monocytes can engulf the particles in the bloodstream, the level of immune cell infiltration of the brain can be quantified by determining the accu- mulation of iron oxide particles in the brain tissue using MRI [98]. T2 maps of MRI using a spin-echo sequence acquired 24 h after injection were used to determine the amount of iron oxide particles in the brain. One of the limitations of using USPIO particles is that it is difficult to differentiate the brain infiltration between pro-inflammatory and anti-inflammatory cells [99]. Iron oxide particles have different sizes that can be divided into ultrasmall superparamagnetic iron oxide (USPIO, 10–50 nm), super- paramagnetic iron oxide (SPIO, 50–100 nm), and micrometer-sized iron oxide (MPIO, >1 μm). These particles are usually coated with a polymer shell containing citrate or dextran, which can be detected as a negative contrast on T2-weighted MRI images due to their large negative magnetic properties [91]. The USPIO image in MRI has also been used to detect lesions that have low T2 signal in EAE. As mentioned previously, the presence of USPIO in the brain due to phagocytosis by immune cells was confirmed using ex vivo brain histological studies. The MRI signal enhancement from USPIO is Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 31

normally compared to Gd-DOTA enhancement because Gd-DOTA brain deposition is due to paracellular BBB breakdown. The USPIO lesion volume at the peak of disease in EAE can be correlated with stages of inflammation, phagocyte infiltration, demyelination, and axonal damage in the central nervous systems [100]. Iron oxide particles have also been used to track C6 cells in the in vivo rat glioma model because C6 cells have been shown to phagocytose nanoparti- cles with a diameter of 20 nm [101]. This method was later used in humans suffering from glial tumors, and the tumors were found to have high iron oxide signal compared to the Gd-DTPA signal [101]. The upregulation of VCAM-1 expression on the BBB endothe- lial cell during neuroinflammation can be determined using MPIO decorated with anti-VCAM-1 mAb (VCAM-1-MPIO) [102]. In this case, the animals were treated with IL-1β cytokine intracerebral injection for inducing neuroinflammation. The accumulation of VCAM-1-MPIO was detected by T2-signal of MRI caused by particles bound to the BBB endothelium. This particle accumula- tion was inhibited when anti-VCAM-1 was delivered prior to the delivery of VCAM-1-MPIO, indicating that the particles bind spe- cifically to VCAM-1 on the surface of endothelium [102]. In EAE mice, VCAM-1-MPIO particles were detected in all visible lesions that were detected using Gd-DTPA. However, VCAM-1-MPIO can also detect additional lesions that correspond to leukocyte infiltration across the BBB [103]. Similar to VCAM-1, ICAM-1 is also upregulated during neu- roinflammation and this upregulation can be detected using Gd-loaded liposomes decorated with anti-ICAM-1 mAb [94]. The upregulation of ICAM-1 in EAE was detected using MRI [94]. In the in vivo stroke animal model, MPIO decorated with anti-ICAM-1 (ICAM-1-MPIO) showed increased T2 signal in brain areas 1 h after induction of transient middle cerebral artery occlusion (MCAO) [104]. Similarly, the same ICAM-1-MPIO detected the upregulation of ICAM-1 in the brain after radiation injury [105]. Neuroinflammation in the brain induces oxidative stress that is mediated by activation of myeloperoxidase (MPO) that is secreted by macrophages and monocytes [106]. MPO activation converts H2O2 into HOCl in the inflamed tissues, which can be used as biomarker for imaging of brain tissue at a molecular level. Because oligomerization of Gd-DOTA was previously shown to enhance MRI signal in tissues, Gd-bis-5-HT-DTPA (Fig. 4a) was designed to oligomerize and react with surrounding proteins in the brain tissue during activation of MPO in neuroinflammation process [106]. The activation of MPO in EAE mice was used for early detection of the disease. In this case, oxidation of 5-HT moiety on the probe induces probe reaction and oligomerization to sur- rounding proteins, which causes an increase in T1 relaxation time observed by MRI. This model can detect smaller active brain lesions 32 Brian M. Kopec et al.

better than Gd-DTPA alone in EAE mice [107]. Further, the enhanced MRI images were located at MPO-expressing cells and demyelinated areas in EAE mice. This method has the potential for early detection in patients with a presymptomatic stage of MS. The oxidative stress process generates free radicals, which can be detected or trapped by 5,5-dimethyl-1-pyrroline N-oxide (DMPO, Fig. 4b). An antibody to DMPO was conjugated to bovine serum albumin (BSA) linked to Gd-DTPA-biotin to make a DMPO-mAb-BSA-Gd-DTPA-biotin conjugate or anti-DMPO probe. After DMPO is delivered, it will react with membrane- bound radicals (MBR) to produce DMPO-MBR in various tissues [108]. The DMOP-MBR can then be detected and localized by an anti-DMPO probe in a certain tissue as a measure of free radicals in the tissue. The localization and concentration of the anti-DMPO probe can be determined by MRI via detection of Gd-DTPA [108]. The localization of anti-DMPO using MRI can also be confirmed by delivering streptavidin-Cy3, which strongly binds to the biotin segment linked to DMPO probe. This method has been successful in detecting free radical formation in the lung, kidney, and liver of the streptozotocin-induced diabetic mouse model. Due to the large size of the DMPO probe, this method can only be used in tissues or organs outside of the brain. The use of this method to detect free radicals in the brain is only possible when restriction of large molecules through the BBB can be overcome.

5.2 PET and SPEC During tumor angiogenesis, αvβ3 integrins are upregulated on the Imaging vasculature endothelial cells. Cyclic arginine-glycine-aspartic acid (RGD) peptides have been developed to selectively bind αvβ3 integ- rins; they have been investigated using diagnostic tools for tumor angiogenesis. Cyclic RGD peptides were conjugated with 18F- galacto for detection of αvβ3 upregulation using PET and SPECT, respectively, during angiogenesis in tumor growth in humans [109]. In human studies, 18F-galacto-RGD as a detector of angio- genesis was compared to 18F-fluorodeoxyglucose (18F-FDG, Fig. 4c) as a substrate to measure metabolism in primary and metastatic tumors. The images generated in tumors by both probes were observed by PET [109]. The results showed that detection with 18F-FDG was more sensitive than with 18F-galacto-RGD, suggesting that glucose metabolism was more pronounced than the increase in expression of αvβ3 integrins in the vascular endothe- lium during angiogenesis. Similarly, a 99mTc-labeled RGD peptide called 99mTc-NC100692 has been used to detect upregulation of αvβ3 integrins in the vascular endothelium during angiogenesis. Although this radiotracer can detect breast cancer, the high uptake and clearance by the liver make the use of this molecule inefficient for liver cancer diagnostic purposes. In addition to issues regarding the high clearance of the 99mTc-labeled RGD peptide, its use has not been explored for brain diseases due to its inability to cross the Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 33

BBB. To improve the binding efficiency of RGD-derived radio- tracers, multimeric cyclic RGD peptides were developed. However, most studies of these multimeric RGD peptides have only been carried out in vitro although, in general, the multimeric cyclic RGD radiotracers are more efficient than the monomeric cyclic RGD radiotracers [110]. A translocator protein (TSPO) is upregulated in micro- and macroglial cells during neuroinflammation; thus, the upregulation of this protein has been used as a diagnostic target of neurological diseases in rodents and humans [111]. The increase in expression of TSPO can be detected using 11C-PK11195 (Fig. 4d) as a ligand of TSPO, and the increased uptake of 11C-PK11195 by glial cells in a stroke rat model compared to normal rats has been observed using PET. Similarly, the increase in deposition of 11C-PK11195 at the entorhinal, temporoparietal, and cingulate cortices of the brains of the transgenic Alzheimer’s disease mouse model were observed due to the upregulation of TSPO. In addition, 11C-PK11195 can detect microglia activation in traumatic brain injury (TBI), stroke, and MS patients [112]. Brain lesions were detected using localiza- tion of 11C-PK11195 in the brains of MS patients, and the detected lesions were similar to those identified using MRI. One of the disadvantages of in vivo detection of 11C-PK11195 is that the observed signal-to-noise ratio is poor due to high levels of nonspe- cific binding of the ligand to non-targeted cells or tissues. In addition, 11C-PK11195 has a short in vivo half-life (i.e., 20 min). As an alternative, a higher selectivity ligand, 1C-DAA1106 (Fig. 4e), has been developed and evaluated in rodent models of Parkinson’s disease and TBI to overcome the poor signal-to-noise ratio of 11C-PK11195 [113]. Leukocyte trafficking into the brain can also be followed using SPECT. In this case, the leukocytes are labeled with 99mtechnetium (99mTc) or indium-111 (111I). However, 111I was found to be more toxic than 99mTc and can damage the leukocytes and their DNA [114]. The location of injected 99mTc-labeled leukocytes can be detected in the region of stroke damage in the brain using SPECT [115]. Similarly, PET/CT imaging with 18F-FDG has also been used to monitor the location of injected cells to detect neuroin- flammation; unfortunately, the success in using 18F-FDG has been modest due to the short half-life of the radiolabel. Furthermore, 18F-FDG can be released from the target cell to cause images to have high background noise. Axon demyelination is a hallmark of MS, and early detection of demyelination in vivo is useful for MS patients to determine the course of treatment. Congo red has been shown to bind myelin, and 11C-labeled Congo red has been used for detecting demyelin- ation using PET imaging [116]. The advantage of Congo red is that it can readily cross the BBB due to its favorable physicochemi- cal to diffuse through the cellular membranes of the vascular 34 Brian M. Kopec et al.

endothelial cells of the BBB. This method is promising because it has been shown to detect the demyelination in the brains of baboons. However, due to the hydrophobic nature of Congo red, it has low water solubility; thus, 11C-CIC (Fig. 4f) was also devel- oped to improve its solubility while it still can penetrate the BBB to selectively bind to myelin rat brains to detect demyelination in EAE mice [117, 118]. The ability to detect neuronal cell death can be very useful for early diagnosis of MS, Alzheimer’s, and Parkinson’s diseases because cell death is irreversible during the progression of these diseases. 11C-flumazenil (Fig. 4g) has been evaluated to detect neuronal lost in the brain of Alzheimer’s disease patients because 11 C-flumazenil binds to GABAA receptors that are downregulated during neuronal damage in the brain [119]. Thus, the downregula- 11 tion of GABAA receptors can be detected by method C-flumaze- nil using PET. In early Alzheimer’s disease in patients, there is a decrease in 11C-flumazenil signals in several different parts of the brain including several cortical regions and posterior perisylvian regions. This method has also been used to determine neuronal damage in early stages of stroke [120]. Neuronal cell death is correlated with the apoptotic process that expresses cell surface phosphatidylserine. Thus, 99mTc-labeled annexin-V can be used to image the presence of phosphatidylserine on the surfaces of dead neuronal cells [121]. SPECT imaging has been used to detect 99mTc labeled during fulminant hepatic cell apoptosis in a mouse model. In addition, 99mTc-annexin-V can detect neuronal damage in ischemic stroke patients. In Alzheimer’s disease patients, the uptake of 99mTc-annexin-V in the cortex was increased; although the BBB normally does not allow such a large protein to penetrate, it is plausible that the brain uptake of annexin is due to the leakiness of the BBB [122–124]. Clinically, 99mTc- labeled annexin-V has been used to detect cell death after organ rejection during transplantation [125].

5.3 Near-IR Near-IR fluorescence (NIRF) dyes have been used to detect mole- Fluorescence cules in the brain that cross the BBB as well as in detect tumor angiogenesis and localization of tumors for surgery. NIRF imaging can be done in vivo in a noninvasive manner because of low scatter- ing, tissue absorption, and autofluorescence of tissues at 700–900 nm. Brain depositions of NIRF dyes and their conjuga- tion to molecules (i.e., peptides, proteins, PEG) can be detected after their delivery using BBB modulator peptides to improve their penetration through the BBB [87, 89]. The quantity of the deliv- ered molecules in the brains can be conveniently determined by integrating the fluorescence intensity from the NIRF image. The NIRF tumor image from IR780 dye (Fig. 4h) has been used to determine the location and size of glioma tumors in the brain. IR780 has low toxicity and high tumor-targeting properties; Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 35

however, it has low solubility that limits its preclinical and clinical applications [126]. To overcome the solubility problem, IR780 dye was formulated in liposome (IR780-liposomes, size 95 nm) and phospholipid micelle (IR780-micelles, size 26 nm) nanoparticles. It was found that the IR780-micelles were more stable than IR780- liposomes. Using confocal microscopy, in vitro incubation of IR780-micelles in a glioma cell culture resulted in uptake and detection of micelles in the intracellular space of U87MG glioma cells [126]. IR780-micelles were also intravenously delivered via tail vein into mice with glioma brain tumors in which U87MG glioma cells were ectopically and orthotopically xenografted into the brains [126]. IR780-micelles were accumulated in orthotopic xenograft tumors and could be detected using NIRF imaging 4 days after delivery. It was proposed that solid tumor accumulation of micelles was due to enhanced permeability and retention (EPR) effect of the micelle nanoparticles in the vasculatures. As a negative control, there was no observable NIRF image of micelles found in healthy mice. The tumor-specific targeting of IR780-micelles has been confirmed by ex vivo NIRF imaging of the brain [126]. NIR dyes have also been used to detect brain tumor xenografts using photoacoustic (PA) imaging method with the goal of localiz- ing tumors for surgery [127]. The PA image is produced by recon- structing collected acoustic waves from the NIR dye-conjugated nanoparticles that are excited by a laser beam penetrating deep into the tissue [127]. In general, PA detection of brain tumors using NIR-I wavelengths (650–980 nm) has a weak light signal due to the signal dampening by the skull. To overcome this problem, NIR- II-conjugate nanoparticles using P1 chromophore (Fig. 4i) were designed with an excitation of 1064 nm to avoid tissue suppression for better signal-to-noise ratio than with NIR-I. In this study, orthotopic xenografts of brain tumors were generated with luciferase-labeled U87 cells, and the presence, location, and size of the tumor were determined using MRI. Then, NIR-II nanopar- ticles were administered to the mice with brain tumor, followed by PA scanning. The results showed an increase in tumor detection with high background 1 h after administration; after 24 h, the image was observed exclusively from the tumor. There was 94-fold higher signal from the tumor than from the background before the administration of NIR-II-NP.

6 Conclusion

The progress in diagnosing and treating brain diseases has been very slow because of the difficulty in delivering molecules noninva- sively to the brain. Early diagnosis of brain diseases has been diffi- cult, but it is important for halting the diseases in the early stage. Many advances have been made using small molecules to detect 36 Brian M. Kopec et al.

changes in the brain and to treat brain diseases. However, these advances are not occurring fast enough to overcome many pro- blems in diagnosis and treatments of brain diseases. Many very selective and potent molecules such as peptides, proteins, carbohy- drates, and oligonucleotides that can be used for treatment of brain diseases have failed to advance in the clinic due to their inability to cross the BBB to exert their biological activities in the brain. Thus, much effort should be devoted to improving the delivery of these molecules to the brain. Limited success has been achieved in deliv- ering large molecules into the brain; however, the efficiencies of the methods used are lower than expected. The use of particles and exosomes for delivering diagnostics or therapeutics to the brain is still in the early stages and, due to the nature of the BBB, the use of particles for brain delivery could encounter an even higher barrier than that for delivering large molecules such as antibodies. Overall, there is a need to increase efforts in the brain delivery area to help in basic and applied sciences to solve brain disease problems as well as to study how the brain works at cellular levels.

Acknowledgments

The authors acknowledge the research support from an R01-NS075374 grant from the National Institute of Neurological Disorders and Stroke (NINDS), National Institutes of Health (NIH). B.M.K. thanks the support from NIH Predoctoral Training Program on Pharmaceutical Aspects of Biotechnology (T32-GM008359). M.E.G.M. thanks the NIH for NIH for an IRACDA postdoctoral fellowship (5K12-GM063651). We would like to thank Nancy Harmony for proofreading this manuscript.

References

1. Schabitz WR, Schwab S, Spranger M, Hacke 4. Zheng K, Trivedi M, Siahaan TJ (2006) W (1997) Intraventricular brain-derived neu- Structure and function of the intercellular rotrophic factor reduces infarct size after focal junctions: barrier of paracellular drug delivery. cerebral ischemia in rats. J Cereb Blood Flow Curr Pharm Des 12:2813–2824 Metab 17(5):500–506. https://doi.org/10. 5. Kiptoo P, Laksitorini MD, Siahaan TJ (2013) 1097/00004647-199705000-00003 Blood-brain peptides: peptide delivery. In: 2. Laksitorini M, Prasasty VD, Kiptoo PK, Sia- Kastin A (ed) Handbook of biologically active haan TJ (2014) Pathways and progress in peptides. Academic Press, Boston, MA, pp improving drug delivery through the intesti- 1702–1710 nal mucosa and blood-brain barriers. Ther 6. Adson A, Raub TJ, Burton PS, Barsuhn CL, Deliv 5(10):1143–1163. https://doi.org/ Hilgers AR, Audus KL, Ho NFH (1994) 10.4155/tde.14.67 Quantitative approaches to delineate paracel- 3. Lutz KL, Siahaan TJ (1997) Molecular struc- lular diffusion in cultured epithelial cell ture of the apical junction complex and its monolayers. J Pharm Sci 83:1529–1536 contribution to the paracellular barrier. J 7. Gumbleton M, Audus KL (2001) Progress Pharm Sci 86:977–984 and limitations in the use of in vitro cell cul- tures to serve as a permeability screen for the Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 37

blood-brain barrier. J Pharm Sci development settings. Adv Drug Deliv Rev 90:1681–1698 46(1–3):3–26 8. Kuntz M, Mysiorek C, Petrault O, 17. Lipinski CA (2004) Lead- and drug-like com- Petrault M, Uzbekov R, Bordet R, Fenart L, pounds: the rule-of-five revolution. Drug Cecchelli R, Berezowski V (2014) Stroke- Discov Today Technol 1(4):337–341. induced brain parenchymal injury drives https://doi.org/10.1016/j.ddtec.2004.11. blood-brain barrier early leakage kinetics: a 007 combined in vivo/in vitro study. J Cereb 18. Lipinski CA (2016) Rule of five in 2015 and Blood Flow Metab 34(1):95–107. https:// beyond: target and ligand structural limita- doi.org/10.1038/jcbfm.2013.169 tions, ligand chemistry structure and drug 9. Lippmann ES, Al-Ahmad A, Azarin SM, Pale- discovery project decisions. Adv Drug Deliv cek SP, Shusta EV (2014) A retinoic acid- Rev 101:34–41. https://doi.org/10.1016/j. enhanced, multicellular human blood-brain addr.2016.04.029 barrier model derived from stem cell sources. 19. Lecomte JM, Costentin J, Vlaiculescu A, Sci Rep 4:4160. https://doi.org/10.1038/ Chaillet P, Marcais-Collado H, Llorens- srep04160 Cortes C, Leboyer M, Schwartz JC (1986) 10. Wilhelm I, Krizbai IA (2014) In vitro models Pharmacological properties of acetorphan, a of the blood-brain barrier for the study of parenterally active “enkephalinase” inhibitor. drug delivery to the brain. Mol Pharm 11 J Pharmacol Exp Ther 237(3):937–944 (7):1949–1963. https://doi.org/10.1021/ 20. Oldendorf WH, Hyman S, Braun L, Olden- mp500046f dorf SZ (1972) Blood-brain barrier: penetra- 11. Chen S, Einspanier R, Schoen J (2015) Trans- tion of morphine, codeine, heroin, and epithelial electrical resistance (TEER): a func- methadone after carotid injection. Science tional parameter to monitor the quality of 178(4064):984–986 oviduct epithelial cells cultured on filter sup- 21. Banks WA (2009) Characteristics of com- ports. Histochem Cell Biol 144(5):509–515. pounds that cross the blood-brain barrier. https://doi.org/10.1007/s00418-015- BMC Neurol 9(Suppl 1):S3. https://doi. 1351-1 org/10.1186/1471-2377-9-S1-S3 12. Cucullo L, Marchi N, Hossain M, Janigro D 22. Stella VJ, Nti-Addae KW (2007) Prodrug (2011) A dynamic in vitro BBB model for the strategies to overcome poor water solubility. study of immune cell trafficking into the cen- Adv Drug Deliv Rev 59(7):677–694. https:// tral nervous system. J Cereb Blood Flow doi.org/10.1016/j.addr.2007.05.013 Metab 31(2):767–777. https://doi.org/10. 23. Dhareshwar SS, Stella VJ (2010) A novel pro- 1038/jcbfm.2010.162 drug strategy for beta-dicarbonyl carbon acids: 13. Griep LM, Wolbers F, de Wagenaar B, ter syntheses and evaluation of the physicochemi- Braak PM, Weksler BB, Romero IA, Couraud cal characteristics of C-phosphoryloxymethyl PO, Vermes I, van der Meer AD, van den Berg (POM) and phosphoryloxymethyloxymethyl A (2013) BBB on chip: microfluidic platform (POMOM) prodrug derivatives. J Pharm Sci to mechanically and biochemically modulate 99(6):2711–2723. https://doi.org/10. blood-brain barrier function. Biomed Micro- 1002/jps.22021 devices 15(1):145–150. https://doi.org/10. 24. Borchardt RT, Jeffrey A, Siahaan TJ, 1007/s10544-012-9699-7 Gangwar S, Pauletti GM (1997) Improve- 14. Shityakov S, Salvador E, Fo¨rster C (2013) In ment of oral peptide bioavailability: peptido- silico, in vitro and in vivo methods to analyse mimetics and prodrug strategies. Adv Drug drug permeation across the blood–brain bar- Deliv Rev 27(2–3):235–256. https://doi. rier: a critical review. OA Anaesthetics 1:1–7 org/10.1016/S0169-409X(97)00045-8. 15. Carpenter TS, Kirshner DA, Lau EY, Wong pii: S0169409X97000458 SE, Nilmeier JP, Lightstone FC (2014) A 25. Gangwar S, Pauletti GM, Siahaan TJ, Stella method to predict blood-brain barrier perme- VJ, Borchardt RT (1999) Synthesis of an ability of drug-like compounds using molecu- esterase-sensitive cyclic prodrug of a model lar dynamics simulations. Biophys J 107 hexapeptide having enhanced membrane per- (3):630–641. https://doi.org/10.1016/j. meability and enzymatic stability using an acy- bpj.2014.06.024 loxyalkoxy promoiety. Methods Mol Med 16. Lipinski CA, Lombardo F, Dominy BW, Fee- 23:37–51. https://doi.org/10.1385/0- ney PJ (2001) Experimental and computa- 89603-517-4:37 tional approaches to estimate solubility and 26. He HT, Xu CR, Song X, Siahaan TJ (2003) permeability in drug discovery and Syntheses of cyclic prodrugs of RGD 38 Brian M. Kopec et al.

peptidomimetics with various macrocyclic function in rats with middle cerebral artery ring sizes: evaluation of physicochemical, occlusion. Brain Res 1111(1):227–229. transport and antithrombic properties. J https://doi.org/10.1016/j.brainres.2006. Pept Res 61(6):331–342 07.005 27. Song X, He HT, Siahaan TJ (2002) Synthesis 37. Paterson J, Webster CI (2016) Exploiting of cyclic prodrugs of Aggrastat and its ana- transferrin receptor for delivering drugs logue with a modified phenylpropionic acid across the blood-brain barrier. Drug Discov linker. Org Lett 4(4):549–552 Today Technol 20:49–52. https://doi.org/ 28. Song X, Xu CR, He HT, Siahaan TJ (2002) 10.1016/j.ddtec.2016.07.009 Synthesis of a novel cyclic prodrug of RGD 38. Friden PM, Walus LR, Musso GF, Taylor MA, peptidomimetic to improve its cell membrane Malfroy B, Starzyk RM (1991) Anti- permeation. Bioorg Chem 30(4):285–301 transferrin receptor antibody and antibody- 29. Wang W, Camenisch G, Sane DC, Zhang H, drug conjugates cross the blood-brain barrier. Hugger E, Wheeler GL, Borchardt RT, Wang Proc Natl Acad Sci U S A 88(11):4771–4775 B (2000) A coumarin-based prodrug strategy 39. Jefferies WA, Brandon MR, Hunt SV, Wil- to improve the oral absorption of RGD pepti- liams AF, Gatter KC, Mason DY (1984) domimetics. J Control Release 65 Transferrin receptor on endothelium of brain (1–2):245–251 capillaries. Nature 312(5990):162–163 30. Pauletti GM, Gangwar S, Okumu FW, Sia- 40. Pardridge WM, Buciak JL, Friden PM (1991) haan TJ, Stella VJ, Borchardt RT (1996) Selective transport of an anti-transferrin Esterase-sensitive cyclic prodrugs of peptides: receptor antibody through the blood-brain evaluation of an acyloxyalkoxy promoiety in a barrier in vivo. J Pharmacol Exp Ther 259 model hexapeptide. Pharm Res (1):66–70 13:1613–1621 41. Moos T, Morgan EH (1998) Evidence for 31. Gudmundsson OS, Pauletti GM, Wang W, low molecular weight, non-transferrin- Shan D, Zhang H, Wang B, Borchardt RT bound iron in rat brain and cerebrospinal (1999) Coumarinic acid-based cyclic pro- fluid. J Neurosci Res 54(4):486–494. drugs of opioid peptides that exhibit meta- https://doi.org/10.1002/(SICI)1097-4547 bolic stability to peptidases and excellent (19981115)54:4<486::AID-JNR6>3.0. cellular permeability. Pharm Res 16(1):7–15 CO;2-I 32. Boado RJ, Zhang Y, Wang Y, Pardridge WM 42. Kharitonova T, Ahmed N, Thoren M, Ward- (2009) Engineering and expression of a chi- law JM, von Kummer R, Glahn J, Wahlgren N meric transferrin receptor monoclonal anti- (2009) Hyperdense middle cerebral artery body for blood-brain barrier delivery in the sign on admission CT scan—prognostic sig- mouse. Biotechnol Bioeng 102 nificance for ischaemic stroke patients treated (4):1251–1258. https://doi.org/10.1002/ with intravenous thrombolysis in the safe bit.22135 implementation of thrombolysis in Stroke 33. Boado RJ, Zhang Y, Zhang Y, Pardridge WM International Stroke Thrombolysis Register. (2007) Genetic engineering, expression, and Cerebrovasc Dis 27(1):51–59. https://doi. activity of a fusion protein of a human neuro- org/10.1159/000172634. pii: 000172634 trophin and a molecular Trojan horse for 43. Lee HJ, Engelhardt B, Lesley J, Bickel U, delivery across the human blood-brain barrier. Pardridge WM (2000) Targeting rat anti- Biotechnol Bioeng 97(6):1376–1386. mouse transferrin receptor monoclonal anti- https://doi.org/10.1002/bit.21369 bodies through blood-brain barrier in mouse. 34. Boado RJ, Zhou QH, Lu JZ, Hui EK, Par- J Pharmacol Exp Ther 292(3):1048–1052 dridge WM (2010) Pharmacokinetics and 44. Shi N, Zhang Y, Zhu C, Boado RJ, Pardridge brain uptake of a genetically engineered WM (2001) Brain-specific expression of an bifunctional fusion antibody targeting the exogenous gene after i.v. administration. mouse transferrin receptor. Mol Pharm 7 Proc Natl Acad Sci U S A 98 (1):237–244. https://doi.org/10.1021/ (22):12754–12759. https://doi.org/10. mp900235k 1073/pnas.221450098 35. Pardridge WM (2002) Blood-brain barrier 45. Lee HJ, Zhang Y, Zhu C, Duff K, Pardridge drug targeting enables neuroprotection in WM (2002) Imaging brain amyloid of Alzhei- brain ischemia following delayed intravenous mer disease in vivo in transgenic mice with an administration of neurotrophins. Adv Exp Abeta peptide radiopharmaceutical. J Cereb Med Biol 513:397–430 Blood Flow Metab 22(2):223–231. https:// 36. Zhang Y, Pardridge WM (2006) Blood-brain doi.org/10.1097/00004647-200202000-00 barrier targeting of BDNF improves motor 010 Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 39

46. Yu YJ, Zhang Y, Kenrick M, Hoyte K, Luk W, Downregulation of microglial activation by Lu Y, Atwal J, Elliott JM, Prabhu S, Watts RJ, apolipoprotein E and apoE-mimetic peptides. Dennis MS (2011) Boosting brain uptake of a Exp Neurol 167(1):74–85. https://doi.org/ therapeutic antibody by reducing its affinity 10.1006/exnr.2001.7541 for a transcytosis target. Sci Transl Med 3 56. van Rooy I, Cakir-Tascioglu S, Hennink WE, (84):84ra44. https://doi.org/10.1126/ Storm G, Schiffelers RM, Mastrobattista E scitranslmed.3002230 (2011) In vivo methods to study uptake of 47. Sade H, Baumgartner C, Hugenmatter A, nanoparticles into the brain. Pharm Res 28 Moessner E, Freskgard PO, Niewoehner J (3):456–471. https://doi.org/10.1007/ (2014) A human blood-brain barrier transcy- s11095-010-0291-7 tosis assay reveals antibody transcytosis influ- 57. Fakhari A, Baoum A, Siahaan TJ, Le KB, enced by pH-dependent receptor binding. Berkland C (2011) Controlling ligand surface PLoS One 9(4):e96340. https://doi.org/ density optimizes nanoparticle binding to 10.1371/journal.pone.0096340 ICAM-1. J Pharm Sci 100(3):1045–1056. 48. Bien-Ly N, Yu YJ, Bumbaca D, Elstrott J, https://doi.org/10.1002/jps.22342 Boswell CA, Zhang Y, Luk W, Lu Y, Dennis 58. Koning GA, Schiffelers RM, Wauben MH, MS, Weimer RM, Chung I, Watts RJ (2014) Kok RJ, Mastrobattista E, Molema G, ten Transferrin receptor (TfR) trafficking deter- Hagen TL, Storm G (2006) Targeting of mines brain uptake of TfR antibody affinity angiogenic endothelial cells at sites of inflam- variants. J Exp Med 211(2):233–244. mation by dexamethasone phosphate- https://doi.org/10.1084/jem.20131660 containing RGD peptide liposomes inhibits 49. Niewoehner J, Bohrmann B, Collin L, experimental arthritis. Arthritis Rheum 54 Urich E, Sade H, Maier P, Rueger P, Stracke (4):1198–1208. https://doi.org/10.1002/ JO, Lau W, Tissot AC, Loetscher H, Ghosh A, art.21719 Freskgard PO (2014) Increased brain pene- 59. Gonzalez-Chavez SA, Arevalo-Gallegos S, tration and potency of a therapeutic antibody Rascon-Cruz Q (2009) Lactoferrin: struc- using a monovalent molecular shuttle. Neu- ture, function and applications. Int J Antimi- ron 81(1):49–60. https://doi.org/10.1016/ crob Agents 33(4):301.e1–301.e8. https:// j.neuron.2013.10.061 doi.org/10.1016/j.ijantimicag.2008.07.020 50. Wang YY, Lui PC, Li JY (2009) Receptor- 60. Qiao R, Jia Q, Huwel S, Xia R, Liu T, Gao F, mediated therapeutic transport across the Galla HJ, Gao M (2012) Receptor-mediated blood-brain barrier. Immunotherapy 1 delivery of magnetic nanoparticles across the (6):983–993. https://doi.org/10.2217/imt. blood-brain barrier. ACS Nano 6 09.75 (4):3304–3310. https://doi.org/10.1021/ 51. van Rooy I, Mastrobattista E, Storm G, Hen- nn300240p nink WE, Schiffelers RM (2011) Comparison 61. Ji B, Maeda J, Higuchi M, Inoue K, Akita H, of five different targeting ligands to enhance Harashima H, Suhara T (2006) Pharmacoki- accumulation of liposomes into the brain. J netics and brain uptake of lactoferrin in rats. Control Release 150(1):30–36. https://doi. Life Sci 78(8):851–855. https://doi.org/10. org/10.1016/j.jconrel.2010.11.014 1016/j.lfs.2005.05.085 52. Gaillard PJ, de Boer AG (2006) A novel 62. Jain KK (2012) Nanobiotechnology-based opportunity for targeted drug delivery to the strategies for crossing the blood-brain barrier. brain. J Control Release 116(2):e60–e62. Nanomedicine (Lond) 7(8):1225–1233. https://doi.org/10.1016/j.jconrel.2006.09. https://doi.org/10.2217/nnm.12.86 050 63. Yang T, Martin P, Fogarty B, Brown A, 53. Gaillard PJ, Visser CC, de Boer AG (2005) Schurman K, Phipps R, Yin VP, Lockman P, Targeted delivery across the blood-brain bar- Bai S (2015) Exosome delivered anticancer rier. Expert Opin Drug Deliv 2(2):299–309. drugs across the blood-brain barrier for https://doi.org/10.1517/17425247.2.2. brain cancer therapy in Danio rerio. Pharm 299 Res 32(6):2003–2014. https://doi.org/10. 54. Ulbrich K, Hekmatara T, Herbert E, Kreuter 1007/s11095-014-1593-y J (2009) Transferrin- and transferrin-recep- 64. Alvarez-Erviti L, Seow Y, Yin H, Betts C, tor-antibody-modified nanoparticles enable Lakhal S, Wood MJ (2011) Delivery of drug delivery across the blood-brain barrier siRNA to the mouse brain by systemic injec- (BBB). Eur J Pharm Biopharm 71 tion of targeted exosomes. Nat Biotechnol 29 (2):251–256. https://doi.org/10.1016/j. (4):341–345. https://doi.org/10.1038/nbt. ejpb.2008.08.021 1807 55. Laskowitz DT, Thekdi AD, Thekdi SD, Han 65. Albright BH, Storey CM, Murlidharan G, SK, Myers JK, Pizzo SV, Bennett ER (2001) Rivera RMC, Berry GE, Madigan VJ, Asokan 40 Brian M. Kopec et al.

A (2018) Mapping the structural determi- (2):594–599. https://doi.org/10.1124/ nants required for AAVrh.10 transport across jpet.103.063347. pii: jpet.103.063347 the blood-brain barrier. Mol Ther 75. Neuwelt EA, Specht HD, Barnett PA, Dahl- 26:510–523. https://doi.org/10.1016/j. borg SA, Miley A, Larson SM, Brown P, Eck- ymthe.2017.10.017 erman KF, Hellstrom KE, Hellstrom I (1987) 66. Agbandje-McKenna M, Kleinschmidt J Increased delivery of tumor-specific monoclo- (2011) AAV capsid structure and cell interac- nal antibodies to brain after osmotic blood- tions. Methods Mol Biol 807:47–92. https:// brain barrier modification in patients with doi.org/10.1007/978-1-61779-370-7_3 melanoma metastatic to the central nervous 67. Madigan VJ, Asokan A (2016) Engineering system. Neurosurgery 20(6):885–895 AAV receptor footprints for gene therapy. 76. Neuwelt EA, Specht HD, Hill SA (1986) Per- Curr Opin Virol 18:89–96. https://doi.org/ meability of human brain tumor to 99mTc- 10.1016/j.coviro.2016.05.001 gluco-heptonate and 99mTc-albumin. Impli- 68. Neuwelt EA (1984) Therapeutic potential for cations for monoclonal antibody therapy. J blood-brain barrier modification in malignant Neurosurg 65(2):194–198. https://doi.org/ brain tumor. Prog Exp Tumor Res 28:51–66 10.3171/jns.1986.65.2.0194 69. Neuwelt EA, Barnett PA, McCormick CI, 77. Muldoon LL, Neuwelt EA (2003) BR96- Remsen LG, Kroll RA, Sexton G (1998) Dif- DOX immunoconjugate targeting of chemo- ferential permeability of a human brain tumor therapy in brain tumor models. J Neuro- xenograft in the nude rat: impact of tumor Oncol 65(1):49–62 size and method of administration on opti- 78. Zwanziger D, Hackel D, Staat C, Bocker A, mizing delivery of biologically diverse agents. Brack A, Beyermann M, Rittner H, Blasig IE Clin Cancer Res 4(6):1549–1555 (2012) A peptidomimetic tight junction 70. Neuwelt EA, Brummett RE, Doolittle ND, modulator to improve regional analgesia. Muldoon LL, Kroll RA, Pagel MA, Dojan R, Mol Pharm 9(6):1785–1794. https://doi. Church V, Remsen LG, Bubalo JS (1998) org/10.1021/mp3000937 First evidence of otoprotection against 79. Tavelin S, Hashimoto K, Malkinson J, carboplatin-induced hearing loss with a Lazorova L, Toth I, Artursson P (2003) A two-compartment system in patients with new principle for tight junction modulation central nervous system malignancy using based on occludin peptides. Mol Pharmacol sodium thiosulfate. J Pharmacol Exp Ther 64:1530–1540 286(1):77–84 80. Yuan X, Lin X, Manorek G, Howell SB (2011) 71. Neuwelt EA, Brummett RE, Remsen LG, Challenges associated with the targeted deliv- Kroll RA, Pagel MA, McCormick CI, ery of gelonin to claudin-expressing cancer Guitjens S, Muldoon LL (1996) In vitro and cells with the use of activatable cell animal studies of sodium thiosulfate as a penetrating peptides to enhance potency. potential chemoprotectant against BMC Cancer 11:61. https://doi.org/10. carboplatin-induced ototoxicity. Cancer Res 1186/1471-2407-11-61 56(4):706–709 81. Makagiansar I, Avery M, Hu Y, Audus KL, 72. Neuwelt EA, Frenkel EP, D’Agostino AN, Siahaan TJ (2001) Improving the selectivity Carney DN, Minna JD, Barnett PA, McCor- of HAV-peptides in modulating E-cadherin- mick CI (1985) Growth of human lung E-cadherin interactions in the intercellular tumor in the brain of the nude rat as a model junction of MDCK cell monolayers. Pharm to evaluate antitumor agent delivery across Res 18:446–553 the blood-brain barrier. Cancer Res 45 82. Sinaga E, Jois SD, Avery M, Makagiansar IT, (6):2827–2833 Tambunan US, Audus KL, Siahaan TJ (2002) 73. Neuwelt EA, Hill SA, Frenkel EP (1984) Increasing paracellular porosity by E-cadherin Osmotic blood-brain barrier modification peptides: discovery of bulge and groove and combination chemotherapy: concurrent regions in the EC1-domain of E-cadherin. tumor regression in areas of barrier opening Pharm Res 19:1170–1179 and progression in brain regions distant to 83. Calcagno AM, Fostel JM, Reyner EL, barrier opening. Neurosurgery 15 Sinaga E, Alston JT, Mattes WB, Siahaan TJ, (3):362–366 Ware JA (2004) Effects of an E-cadherin- 74. Neuwelt EA, Pagel MA, Kraemer DF, Peter- derived peptide on the gene expression of son DR, Muldoon LL (2004) Bone marrow Caco-2 cells. Pharm Res 21:2095–2104 chemoprotection without compromise of 84. Kiptoo P, Sinaga E, Calcagno AM, Zhao H, chemotherapy efficacy in a rat brain tumor Kobayashi N, Tambunan US, Siahaan TJ model. J Pharmacol Exp Ther 309 Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 41

(2011) Enhancement of drug absorption Inhibition of LFA-1/ICAM-1 and VLA-4/ through the blood-brain barrier and inhibi- VCAM-1 as a therapeutic approach to inflam- tion of intercellular tight junction resealing mation and autoimmune diseases. Med Res by E-cadherin peptides. Mol Pharm 8 Rev 22(2):146–167 (1):239–249. https://doi.org/10.1021/ 94. Sipkins DA, Gijbels K, Tropper FD, mp100293m Bednarski M, Li KC, Steinman L (2000) 85. Alaofi A, On N, Kiptoo P, Williams TD, Miller ICAM-1 expression in autoimmune encepha- DW, Siahaan TJ (2016) Comparison of linear litis visualized using magnetic resonance and cyclic His-Ala-Val peptides in modulating imaging. J Neuroimmunol 104(1):1–9 the blood-brain barrier permeability: impact 95. Zivadinov R, Rudick RA, De Masi R, on delivery of molecules to the brain. J Pharm Nasuelli D, Ukmar M, Pozzi-Mucelli RS, Sci 105(2):797–807. https://doi.org/10. Grop A, Cazzato G, Zorzon M (2001) Effects 1016/S0022-3549(15)00188-4 of IV methylprednisolone on brain atrophy in 86. Laksitorini MD, Kiptoo PK, On NH, Thli- relapsing-remitting MS. Neurology 57 veris JA, Miller DW, Siahaan TJ (2015) Mod- (7):1239–1247 ulation of intercellular junctions by cyclic- 96. Pirko I, Johnson A, Gamez J, Macura SI, ADT peptides as a method to reversibly Rodriguez M (2004) Disappearing “T1 increase blood-brain barrier permeability. J black holes” in an animal model of multiple Pharm Sci 104(3):1065–1075. https://doi. sclerosis. Front Biosci 9:1222–1227 org/10.1002/jps.24309 97. Nessler S, Boretius S, Stadelmann C, 87. On NH, Kiptoo P, Siahaan TJ, Miller DW Bittner A, Merkler D, Hartung HP, (2014) Modulation of blood-brain barrier Michaelis T, Bruck W, Frahm J, Sommer N, permeability in mice using synthetic Hemmer B (2007) Early MRI changes in a E-cadherin peptide. Mol Pharm 11 mouse model of multiple sclerosis are predic- (3):974–981. https://doi.org/10.1021/ tive of severe inflammatory tissue damage. mp400624v Brain 130(Pt 8):2186–2198. https://doi. 88. Tabanor K, Lee P, Kiptoo P, Choi IY, Sherry org/10.1093/brain/awm105 EB, Eagle CS, Williams TD, Siahaan TJ 98. Floris S, Blezer EL, Schreibelt G, Dopp E, van (2016) Brain delivery of drug and MRI con- der Pol SM, Schadee-Eestermans IL, trast agent: detection and quantitative deter- Nicolay K, Dijkstra CD, de Vries HE (2004) mination of brain deposition of CPT-Glu Blood-brain barrier permeability and mono- using LC-MS/MS and Gd-DTPA using mag- cyte infiltration in experimental allergic netic resonance imaging. Mol Pharm 13 encephalomyelitis: a quantitative MRI study. (2):379–390. https://doi.org/10.1021/acs. Brain 127(Pt 3):616–627. https://doi.org/ molpharmaceut.5b00607 10.1093/brain/awh068 89. Ulapane KR, On N, Kiptoo P, Williams TD, 99. Boven LA, Van Meurs M, Van Zwam M, Miller DW, Siahaan TJ (2017) Improving Wierenga-Wolf A, Hintzen RQ, Boot RG, brain delivery of biomolecules via BBB mod- Aerts JM, Amor S, Nieuwenhuis EE, Laman ulation in mouse and rat: detection using JD (2006) Myelin-laden macrophages are MRI, NIRF, mass spectrometry. Nanothera- anti-inflammatory, consistent with foam cells nostics 1:217–231 in multiple sclerosis. Brain 129 90. Alaofi A, Farokhi E, Prasasty VD, (Pt 2):517–526. https://doi.org/10.1093/ Anbanandam A, Kuczera K, Siahaan TJ brain/awh707 (2017) Probing the interaction between 100. Brochet B, Deloire MS, Touil T, Anne O, cHAVc3 peptide and the EC1 domain of Caille JM, Dousset V, Petry KG (2006) E-cadherin using NMR and molecular Early macrophage MRI of inflammatory dynamics simulations. J Biomol Struct Dyn lesions predicts lesion severity and disease 35(1):92–104. https://doi.org/10.1080/ development in relapsing EAE. NeuroImage 07391102.2015.1133321 32(1):266–274. https://doi.org/10.1016/j. 91. Pulli B, Chen JW (2014) Imaging neuroin- neuroimage.2006.03.028 flammation—from bench to bedside. J Clin 101. Stroh A, Zimmer C, Werner N, Gertz K, Cell Immunol 5:226. https://doi.org/10. Weir K, Kronenberg G, Steinbrink J, 4172/2155-9899.1000226 Mueller S, Sieland K, Dirnagl U, 92. Steinman L (2005) Blocking adhesion mole- Nickenig G, Endres M (2006) Tracking of cules as therapy for multiple sclerosis: natali- systemically administered mononuclear cells zumab. Nat Rev Drug Discov 4(6):510–518. in the ischemic brain by high-field magnetic https://doi.org/10.1038/nrd1752 resonance imaging. NeuroImage 33 93. Yusuf-Makagiansar H, Anderson ME, Yakov- (3):886–897. https://doi.org/10.1016/j. leva TV, Murray JS, Siahaan TJ (2002) neuroimage.2006.07.009 42 Brian M. Kopec et al.

102. McAteer MA, Sibson NR, von Zur MC, 110. Zhou Y, Chakraborty S, Liu S (2011) Radi- Schneider JE, Lowe AS, Warrick N, Channon olabeled cyclic RGD peptides as radiotracers KM, Anthony DC, Choudhury RP (2007) In for imaging tumors and thrombosis by vivo magnetic resonance imaging of acute SPECT. Theranostics 1:58–82 brain inflammation using microparticles of 111. Benavides J, Capdeville C, Dauphin F, iron oxide. Nat Med 13(10):1253–1258. Dubois A, Duverger D, Fage D, Gotti B, https://doi.org/10.1038/nm1631 MacKenzie ET, Scatton B (1990) The quan- 103. Serres S, Mardiguian S, Campbell SJ, McA- tification of brain lesions with an omega 3 site teer MA, Akhtar A, Krapitchev A, Choudhury ligand: a critical analysis of animal models of RP, Anthony DC, Sibson NR (2011) VCAM- cerebral ischaemia and neurodegeneration. 1-targeted magnetic resonance imaging Brain Res 522(2):275–289 reveals subclinical disease in a mouse model 112. Ramlackhansingh AF, Brooks DJ, Green- of multiple sclerosis. FASEB J 25 wood RJ, Bose SK, Turkheimer FE, Kinnu- (12):4415–4422. https://doi.org/10.1096/ nen KM, Gentleman S, Heckemann RA, fj.11-183772 Gunanayagam K, Gelosa G, Sharp DJ 104. Deddens LH, van Tilborg GA, van der (2011) Inflammation after trauma: microglial Toorn A, van der Marel K, Paulis LE, van activation and traumatic brain injury. Ann Bloois L, Storm G, Strijkers GJ, Mulder WJ, Neurol 70(3):374–383. https://doi.org/10. de Vries HE, Dijkhuizen RM (2013) MRI of 1002/ana.22455 ICAM-1 upregulation after stroke: the impor- 113. Venneti S, Lopresti BJ, Wang G, Slagel SL, tance of choosing the appropriate target- Mason NS, Mathis CA, Fischer ML, Larsen specific particulate contrast agent. Mol Imag- NJ, Mortimer AD, Hastings TG, Smith AD, ing Biol 15(4):411–422. https://doi.org/10. Zigmond MJ, Suhara T, Higuchi M, Wiley 1007/s11307-013-0617-z CA (2007) A comparison of the high-affinity 105. Zhu Y, Ling Y, Zhong J, Liu X, Wei K, Huang peripheral benzodiazepine receptor ligands S (2012) Magnetic resonance imaging of DAA1106 and (R)-PK11195 in rat models radiation-induced brain injury using targeted of neuroinflammation: implications for PET microparticles of iron oxide. Acta Radiol 53 imaging of microglial activation. J Neuro- (7):812–819. https://doi.org/10.1258/ar. chem 102(6):2118–2131. https://doi.org/ 2012.120040 10.1111/j.1471-4159.2007.04690.x 106. Querol M, Chen JW, Bogdanov AA Jr (2006) 114. Thakur ML, Gottschalk A, Zaret BL (1979) A paramagnetic contrast agent with Imaging experimental myocardial infarction myeloperoxidase-sensing properties. Org with indium-111-labeled autologous leuko- Biomol Chem 4(10):1887–1895. https:// cytes: effects of infarct age and residual doi.org/10.1039/b601540a regional myocardial blood flow. Circulation 107. Chen JW, Breckwoldt MO, Aikawa E, 60(2):297–305 Chiang G, Weissleder R (2008) 115. Wang PY, Kao CH, Mui MY, Wang SJ (1993) Myeloperoxidase-targeted imaging of active Leukocyte infiltration in acute hemispheric inflammatory lesions in murine experimental ischemic stroke. Stroke 24(2):236–240 autoimmune encephalomyelitis. Brain 131 116. Stankoff B, Freeman L, Aigrot MS, (Pt 4):1123–1133. https://doi.org/10. Chardain A, Dolle F, Williams A, Galanaud D, 1093/brain/awn004 Armand L, Lehericy S, Lubetzki C, Zalc B, 108. Towner RA, Smith N, Saunders D, Bottlaender M (2011) Imaging central nervous Henderson M, Downum K, Lupu F, Silasi- system myelin by positron emission tomogra- Mansat R, Ramirez DC, Gomez-Mejiba SE, phy in multiple sclerosis using [methyl-(1)(1) Bonini MG, Ehrenshaft M, Mason RP (2012) C]-2-(40-methylaminophenyl)-6-hydroxyben- In vivo imaging of immuno-spin trapped radi- zothiazole. Ann Neurol 69(4):673–680. cals with molecular magnetic resonance imag- https://doi.org/10.1002/ana.22320 ing in a diabetic mouse model. Diabetes 61 117. de Paula FD, de Vries EF, Sijbesma JW, (10):2405–2413. https://doi.org/10.2337/ Dierckx RA, Buchpiguel CA, Copray S db11-1540 (2014) PET imaging of demyelination and 109. Beer AJ, Lorenzen S, Metz S, Herrmann K, remyelination in the cuprizone mouse model Watzlowik P, Wester HJ, Peschel C, for multiple sclerosis: a comparison between Lordick F, Schwaiger M (2008) Comparison [11C]CIC and [11C]MeDAS. NeuroImage of integrin alphaVbeta3 expression and glu- 87:395–402. https://doi.org/10.1016/j. cose metabolism in primary and metastatic neuroimage.2013.10.057 lesions in cancer patients: a PET study using 118. Faria Dde P, Copray S, Sijbesma JW, Willem- 18F-galacto-RGD and 18F-FDG. J Nucl sen AT, Buchpiguel CA, Dierckx RA, de Vries Med 49(1):22–29. https://doi.org/10. EF (2014) PET imaging of focal demyelin- 2967/jnumed.107.045864 ation and remyelination in a rat model of Methods of Delivering Molecules Through the Blood-Brain Barrier for Brain... 43

multiple sclerosis: comparison of [11C] Blood Flow Metab 31(5):1311–1320. MeDAS, [11C]CIC and [11C]PIB. Eur J https://doi.org/10.1038/jcbfm.2010.233 Nucl Med Mol Imaging 41(5):995–1003. 123. Yagle KJ, Eary JF, Tait JF, Grierson JR, Link https://doi.org/10.1007/s00259-013-268 JM, Lewellen B, Gibson DF, Krohn KA 2-6 (2005) Evaluation of 18F-annexin V as a 119. Pascual B, Prieto E, Arbizu J, Marti-Climent PET imaging agent in an animal model of JM, Penuelas I, Quincoces G, Zarauza R, apoptosis. J Nucl Med 46(4):658–666 Pappata S, Masdeu JC (2012) Decreased car- 124. Lampl Y, Lorberboym M, Blankenberg FG, bon-11-flumazenil binding in early Alzhei- Sadeh M, Gilad R (2006) Annexin V SPECT mer’s disease. Brain 135(Pt 9):2817–2825. imaging of phosphatidylserine expression in https://doi.org/10.1093/brain/aws210 patients with dementia. Neurology 66 120. Rojas S, Martin A, Pareto D, Herance JR, (8):1253–1254. https://doi.org/10.1212/ Abad S, Ruiz A, Flotats N, Gispert JD, 01.wnl.0000208436.75615.8c Llop J, Gomez-Vallejo V, Planas AM (2011) 125. Narula J, Acio ER, Narula N, Samuels LE, Positron emission tomography with Fyfe B, Wood D, Fitzpatrick JM, Raghunath 11C-flumazenil in the rat shows preservation PN, Tomaszewski JE, Kelly C, Steinmetz N, of binding sites during the acute phase after Green A, Tait JF, Leppo J, Blankenberg FG, 2 h-transient focal ischemia. Neuroscience Jain D, Strauss HW (2001) Annexin-V imag- 182:208–216. https://doi.org/10.1016/j. ing for noninvasive detection of cardiac allo- neuroscience.2011.03.013 graft rejection. Nat Med 7(12):1347–1352. 121. Koopman G, Reutelingsperger CP, Kuijten https://doi.org/10.1038/nm1201-1347 GA, Keehnen RM, Pals ST, van Oers MH 126. Li S, Johnson J, Peck A, Xie Q (2017) Near (1994) Annexin V for flow cytometric detec- infrared fluorescent imaging of brain tumor tion of phosphatidylserine expression on B with IR780 dye incorporated phospholipid cells undergoing apoptosis. Blood 84 nanoparticles. J Transl Med 15(1):18. https:// (5):1415–1420 doi.org/10.1186/s12967-016-1115-2 122. Bahmani P, Schellenberger E, Klohs J, 127. Guo B, Sheng Z, Kenry HD, Lin X, Xu S, Steinbrink J, Cordell R, Zille M, Muller J, Liu C, Zheng H, Liu B (2017) Biocompatible Harhausen D, Hofstra L, Reutelingsperger C, conjugated polymer nanoparticles for highly Farr TD, Dirnagl U, Wunder A (2011) Visual- efficient photoacoustic imaging of orthotopic ization of cell death in mice with focal cerebral brain tumors in the second near-infrared win- ischemia using fluorescent annexin A5, propi- dow. Mater Horizons 4:1151–1156. https:// dium iodide, and TUNEL staining. J Cereb doi.org/10.1039/c7mh00672a Chapter 3

Culturing of Rodent Brain Microvascular Endothelial Cells for In Vitro Modeling of the Blood-Brain Barrier

Malgorzata Burek and Carola Y. Fo¨rster

Abstract

The blood-brain barrier (BBB) is important in the maintenance of the microenvironment of the brain and proper neuronal function. Apart from the protective function, BBB regulates entry of nutrients into the central nervous system by selective transport and metabolism of blood- and brain-borne substances. Successful modeling of BBB in vitro is established since 1970s and has been used to study mechanisms of transport, cellular interaction, and gene regulation. Rodent in vitro BBB models are widely used and have been proven to retain sufficiently the in vivo properties during culturing. In this chapter we describe methodological aspects of culturing the microvascular endothelial cells. Immortalized endothelial cell lines as well as primary brain microvascular endothelial cells in monoculture, co-culture, and triple-culture are discussed.

Key words Blood-brain barrier, Brain microvascular endothelial cells, In vitro model, TEER

1 Introduction

1.1 Structure The blood-brain barrier (BBB) can be crossed by only a small class and Function of Blood- of drugs, which includes small molecules with high lipid solubility Brain Barrier and a low molecular mass of <400–500 Da meaning that most of the therapeutics cannot cross the BBB [1, 2]. The beginnings of BBB discovery and study trace back to an experiment conducted by the German microbiologist Paul Ehrlich in which he observed that intravenous administration of certain dyes stained all organs except the brain and the spinal cord. Later, his associate Edwin E. Goldmann discovered that the same dye applied to the cerebro- spinal fluid exclusively stained brain tissue leading to the concept of a barrier between the blood and the brain [3]. The brain capillaries are built up by microvascular endothelial cells. Other cell types such as astrocytes and pericytes interact tightly with endothelial cells building a dynamic neurovascular unit [4]. Brain microvascular endothelial cells are tightly sealed by tight junctions (TJs) forming

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_3, © Springer Science+Business Media, LLC, part of Springer Nature 2019 45 46 Malgorzata Burek and Carola Y. Fo¨ rster

a diffusion barrier that selects which blood-borne substances could enter the brain [5]. Endothelial cells of the BBB differ from endothelial cells in the rest of the body due to the lack of fenestrations, presence of more extensive TJs, and sparse pinocytic vesicular transport. The extreme tightness of TJs provides a key feature of the BBB. Paracellular flux of hydrophilic molecules across the BBB is limited by endothelial cell TJs [5–7]. TJs located between endothelial cells enable the strict restriction of paracellular diffusion of ions and other polar solutes between the cells. TJs are formed by transmembrane pro- teins of Marvel family (occludin, tricellulin), claudin family (claudin-1, -3, -5, and -12), and immunoglobulin-like protein family (junctional adhesion molecules; JAM-A, -B, -C) [8]. TJ proteins are connected to the cytoplasmic proteins and to the cytoskeleton through their intracellular domain interactions with zonula occludens proteins ZO-1 and -2 [8]. Adherens junctions are built by VE-cadherin which is linked to the cytoskeleton through catenin-α,-β,-γ, and p120. The presence of TJs results in a high electrical resistance of the BBB in vivo which is approximately 1800 Ω cm2 [9]. Due to this high electrical resistance, the move- À ment of small ions such as Na+ and Cl across the barrier is limited. In addition to this, brain endothelial cells have a more negatively charged membrane compared to other cells [10]. The tightness of the barrier greatly restricts transport of material across the BBB. Nonetheless, gaseous molecules and small lipophilic compounds pass through the barrier via passive transcellular diffusion through the membranes of brain endothelial cells. Many compounds how- ever fail to cross the enzymatic barrier and efflux transporters. The main efflux transporters of ABC transporter family are ABCG2/ BCRP, ABCB1/MDR1/P-glycoprotein, ABCC1, ABCC4, ABCC5, ABCA2, and ABCA8 [11]. Nutrients like glucose pass the barrier through carrier-mediated influx by solute-carrier trans- porters, SLC2A1/GLUT1, SLC7A5/LAT1, SLC16A1/MCT1, SLC1A3/EAAT1, and SLC1A2/EAAT2 [11]. Large endogenous molecules, on the other hand, cross the barrier with the help of receptor- and adsorptive-mediated transcytosis, respectively [12]. Another means by which transport is regulated across the BBB is through the ABC transporters expressed at high levels in the apical surface of brain endothelial cells. These transporters mediate the efflux of lipophilic compounds which could easily diffuse across the barrier. Among these transporters, the P-glycoprotein (P-gp) transporter has been a major target for drug delivery studies [13]. Since glucose is a major nutrient source for the brain, glucose transporters are another group of transporters that aid in the move- ment of glucose toward the CNS. The major glucose transporter is the facilitative glucose transporter GLUT-1. The expression level of GLUT-1 is regulated by changes in the metabolic demand of cerebral glucose utilization. It may be upregulated at the BBB Culturing of Rodent Blood-Brain Barrier 47

due to glutamate excitotoxicity, hypoxia, aglycemia, as well as mitochondrial damage [14, 15]. It has been shown in an in vitro model of the BBB that GLUT-1 is upregulated in endothelial cells deprived of oxygen and glucose [16]. Apart from these transpor- ters, transport across the BBB also occurs through the migration of immune cells to the CNS [17]. Endothelial cells are the ones that regulate BBB permeability. However, the overall microvascular function depends on the para- crine interactions existing between endothelial cells, astrocytes, and pericytes. Astrocytes play an important role in the maintenance of BBB and neuronal homeostasis. Astrocytic endfeet cover the brain capil- laries almost completely and have contact with neurons [18]. More- over, the direct contact between endothelial cells and astrocytes is necessary to generate an optimal BBB [19]. In addition, apart from physically supporting brain endothelial cells, astrocytes also synthe- size molecules that have been shown to influence endothelial cells such as transforming growth factor-β (TGF-β), glial-derived neu- rotrophic factor (GDNF), basic fibroblast growth factor (bFGF), interleukin-6 (IL-6), and Sonic hedgehog [20, 21]. Astrocytes have been shown to impart high transendothelial resistance to human or bovine endothelial cell monolayers by culturing them in astrocyte- conditioned medium implicating a role of astrocytes in the induc- tion of BBB characteristics among endothelial cells [22]. Pericytes are present in capillaries, venules, and arterioles. They wrap around the endothelial cells providing structural support as well as vasodynamic capacity to the microvasculature [5]. Similar to astrocytes, pericytes also secrete substances, which may influence endothelial function. These include TGF-β, angiopoetin-1, and vascular endothelial growth factor (VEGF) [23]. It has been observed that the absence of pericytes leads to endothelial hyper- plasia, abnormal vasculogenesis, as well as increased BBB perme- ability suggesting an important role of pericytes in BBB formation [24, 25]. The physiological role of pericytes is poorly understood. However, it has been found out that pericyte-conditioned medium can induce the expression of the TJ protein occludin in brain capillary endothelial cells suggesting its involvement in regulating occludin expression in the BBB [26].

2 Materials

2.1 Brain Freshly isolated primary mouse brain microvascular endothelial Microvascular cells (BMEC) or immortalized cell lines (such as cEND, cere- Endothelial Cells, bEND, or bEND3) can be cultured in monoculture or in Astrocytes, Pericytes co-culture with astrocytes and pericytes. Primary astrocytes can be isolated from brains of wild-type C57/Bl6 mice at postnatal day 4 and used for 0–2 passages [27]. For isolation of pericytes, adult 48 Malgorzata Burek and Carola Y. Fo¨ rster

mouse brains can be used. Culturing the mixture of isolated brain cells in conditions optimized for BMEC for two passages and then changing to optimized pericyte medium (commercially available) led to a generation of a pure pericyte culture [28].

2.2 Endothelial Cell Brain endothelial cells are cultured in sterile conditions in an atmo-  Culture Medium sphere containing 5% CO2 at 37 C. Various cell culture media have been described in the literature for culturing of endothelial cells. Dulbecco’s modified Eagle medium (DMEM) with 4.5 g/L glu- cose and 10% FCS, 50 U/mL penicillin/streptomycin, and 1% L-glutamine is being used for cerebEND cell line and bEND3 [29, 30]. Primary BMEC were cultured in medium containing 20% FCS, 1Â MEM nonessential amino acids, 1 mM sodium pyruvate, 0.1 mg/mL heparin salt, 1 mg/mL sodium bicarbonate, and endothelial cell growth supplement (ECGS) [31].

2.3 Transwells For examination of endothelial cell permeability, measurements of TEER, as well as transport studies, the cells are plated in transwell inserts. There is a huge variety of inserts supplied by companies. Transwells are built of a microporous, semipermeable membrane made of polycarbonate (PC) or polyethylene terephthalate (PET); they can be translucent or transparent, and have different growth area and pore size (such as 0.4, 1, 3, and 8 μm). The pore size of 0.4 and 1 μm has been most often used for studies with endothelial cells.

3 Methods

3.1 In Vitro Models In vitro BBB models have been used for decades. They are proven of Blood-Brain Barrier to be essential in investigating and understanding the cellular and molecular mechanisms underlying BBB establishment. With the increasing number of in vitro BBB models available to date, one important factor for a successful model is the choice of suitable cells. Most of the cells used for this purpose are derived from bovine, porcine, rat, or mouse brain tissue [32]. The many available systems use cerebral endothelial cells from primary cultures or cell lines in monoculture, co-culture, or triple co-cultures with other components of the neurovascular unit, such as astrocytes and peri- cytes. One important characteristic of a cell-based model is the ability to replicate in vivo conditions. The common requirements include the ability of isolated endothelial cells to express TJs with the proper membrane localization and a high transendothelial elec- trical resistance (TEER) [32]. An in vitro TEER of at least 150–200 Ω cm2 is being considered sufficient for most of the experiments [33–35]. Tracer substances are also used to monitor paracellular permeability of in vitro systems. Some of the substances used as tracers include sodium fluorescein, lucifer yellow, and Culturing of Rodent Blood-Brain Barrier 49

fluorescein-isothiocyanate (FITC)-labeled substances [36–38]. Cell-based in vitro models are cost effective in comparison to animal experiments allowing working in a simplified environment [39]. However, one major drawback of in vitro models is the lack of exposure to physiological conditions which can influence the expression of important biological transporters, ligands, and enzymes [39].

3.2 Immortalized Immortalization of bovine, porcine, rat, murine, and human brain Cell Lines endothelial cells can be achieved through transfection with adeno- virus E1A oncogenes, polyomavirus large T protein, and SV40 large T oncogenes [40–44]. Isolated brain microvascular endothe- lial cells and immortalized cell lines could retain many in vivo features [43, 45]. A number of immortalized endothelial cell lines have been reported to produce TJs, and express cellular receptors and transporters [8]. For instance, the murine brain microvascular endothelial cell lines from cerebral cortex, cEND, and from cere- bellar cortex, cerebEND, were proven to retain many properties found in vivo [46, 47]. Other well-described rodent in vitro BBB models are: bEND3, bEND5, MBEC4, and TM-BBB1-5 from mouse and GP8/3.9, GPNT, RBE4, RBEC1, rBCEC4, and TR-BBB from rat [30, 42, 48–56].

3.3 Adult Primary Isolation of adult primary rodent BMEC is a standard method for Rodent Brain modeling the BBB. The cells should ideally be used without pas- Microvascular saging, as these cells are not capable of retaining BBB for a long Endothelial Cell time in cell culture. Freshly isolated adult BMEC dedifferentiate (BMEC) Culture after several passages and lose BBB properties including the forma- tion of TJs [57, 58]. This leads to the tedious work of repeatedly producing and characterizing primary BMEC. In addition, due to this, it cannot be guaranteed that the endothelial cell cultures are uniform for all experiments since the culture can differ from batch to batch.

3.4 Monoculture Endothelial cells are plated as a monoculture on a porous mem- brane and are grown to confluence. This takes usually between 5 and 7 days. In a transwell system, there are two compartments, the upper (“blood side”) and the lower one (“brain side”). One should take care that the cell culture medium level is equal between upper and lower compartment in order to avoid changes in the hydrostatic pressure. Changes of medium should be done very carefully. The safest way not to damage the endothelial cell mono- layer is to avoid touching the cells during aspiration of liquids from the transwell. For transport and permeability studies, the sub- stances are added to the upper chamber. A time curve of substance permeability can be prepared by collecting samples from the lower 50 Malgorzata Burek and Carola Y. Fo¨ rster

chamber in distinct time intervals. Such a monoculture in transwell system has been successfully used in a row of studies in our labora- tory [29, 46, 47, 59, 60]. For mimicking a traumatic brain injury, BMEC grow on collagen-coated silastic membranes, which can be stretched by air pulse [61].

3.5 Co-culture For co-culture experiments, BMEC are plated on the upper side of the porous membrane. Astrocyte may be seeded either on the bottom side of the porous membrane (contact culture) or on the bottom of the plate (noncontact culture). In case of contact cul- ture, astrocyte should be seeded on poly-L-lysine-coated membrane 24 h prior to BMEC seeding. Co-culture of cerebEND with astro- cyte cell line C6 leads to an increased barrier damage under oxy- gen/glucose deprivation conditions as well as to changes in ABC transporter function and expression [62]. However, co-culture with C6 astrocytes alone only moderately influenced barrier proper- ties of cerebEND cells and TJ-protein expression [62]. C6 astro- cytes led to a significant increase in protein expression and function of Abcb1 and Abcc4 transporters in cerebEND cells [62]. Co-culture with primary mouse astrocytes had barrier-stabilizing effects on bEND.5 and hCMEC/D3 cells. These effects were measured by increased TEER values [31].

3.6 Triple Cultures In triple cultures, three types of cells are cultured together in a of BBB Models transwell setup. Similar to co-culture, the BMEC are seeded on the upper side of the membrane and astrocytes are seeded on the bottom side of the transwell membrane. The third type of cells, pericytes, are seeded on the cell culture plate. Alternatively, the astrocytes are grown on the culture plate and the pericytes on the bottom of the transwell membrane. Co-culture of primary rat BMEC with astrocyte and pericytes resulted in the highest TEER values [63].

3.7 Measurements Cells are plated on top of collagen IV-coated transwell chambers for of Transendothelial 6-well plates (24 mm diameter, membrane material: PET, 0.4 μm Electrical Resistance pores, pore density/cm2 1.6 Â 106) at densities of 2.5 Â 104/cm2 (TEER) cells per well. When they had reached confluence at day 5, the different experimental sets of cells can be transferred to differentia- tion medium containing 1–2% FCS. TEER can be measured using an assembly containing current-passing and voltage-measuring electrodes (EVOM), equipped with a STX2 “chopstick” electrode set. Resistances of blank filters are subtracted from those of filters with cells and the final resistances (in Ω cm2) can be calculated. ® Other devices used for TEER measurements are cellZscope , ECIS 1600R, and xCELLigence. Culturing of Rodent Blood-Brain Barrier 51

4 Conclusions

In vitro BBB models became a good tool for analysis of cellular interactions and molecular mechanisms behind the BBB integrity in health and disease. Monoculture, co-culture, and triple culture models have been successfully used for various studies. Although the triple culture has been proven to possess the best BBB proper- ties, simplified models can be chosen for certain types of analysis. Ideally, the results produced with immortalized cell lines should be validated on primary in vitro BBB models or in an in vivo experiment.

References

1. Pardridge W (2001) Brain drug targeting: the 10. Ribeiro M, Domingues MM, Freire JM, Santos future of brain drug development. Cambridge NC, Castanho MARB (2012) Translocating University Press, Cambridge the blood–brain barrier using electrostatics. 2. Pardridge W (2005) The blood–brain barrier: Front Cell Neurosci 6:1–7 bottleneck in brain drug development. Neu- 11. Wilhelm I, Krizbai IA (2014) In vitro models roRx 2:3–14 of the blood-brain barrier for the study of drug 3. Liebner S, Engelhardt B (2005) Development delivery to the brain. Mol Pharm 11 of the blood-brain barrier. In: Prat A, de Vries (7):1949–1963 E (eds) The blood-brain barrier and its micro- 12. Patel M, Goyal BR, Bhadada SV, Bhatt JS, environment: basic physiology to neurological Amin AF (2009) Getting into the brain: diseases, 1st edn. Taylor & Francis Group, approaches to enhance brain drug delivery. Boca Raton, FL, pp 1–26 CNS Drugs 23:35–58 4. Neuwelt E, Abbott NJ, Abrey L, Banks WA, 13. Lo¨scher W, Potschka H (2005) Role of drug Blakley B, Davis T et al (2008) Strategies to efflux transporters in the brain for drug dispo- advance translational research into brain bar- sition and treatment of brain diseases. Prog riers. Lancet Neurol 7(1):84–96. Epub Neurobiol 76:22–76 21 Dec 2007 14. Xiuli G, Meiyu G, Guanhua D (2005) Glucose 5. Ballabh P, Braun A, Nedergaard M (2004) The transporter 1, distribution in the brain and in blood–brain barrier: an overview. Structure, neural disorders: its relationship with transport regulation, and clinical implications. Neurobiol of neuroactive drugs through the blood-brain Dis 16:1–13 barrier. Biochem Genet 43:175–187 6. Begley D, Brightman MW (2003) Structural 15. Espinoza-Rojo M, Iturralde-Rodriguez KI, and functional aspects of the blood–brain bar- Chanez-Cardenas ME, Ruiz-Tachiquin ME, rier. Progr Drug Res 61:39–78 Aguilera P (2010) Glucose transporters regula- 7. Wolburg H, Noell S, Mack A, Wolburg- tion on ischemic brain: possible role as thera- Buchholz K, Fallier-Becker P (2009) Brain peutic target. Cent Nerv Syst Agents Med endothelial cells and the glio–vascular complex. Chem 10:317–325 Cell Tissue Res 335:75–96 16. Neuhaus W, Burek M, Djuzenova CS, Thal SC, 8. Helms HC, Abbott NJ, Burek M, Cecchelli R, Koepsell H, Roewer N, Fo¨rster CY (2012) Addi- Couraud PO, Deli MA et al (2016) In vitro tion of NMDA-receptor antagonist MK801 dur- models of the blood-brain barrier: an overview ing oxygen/glucose deprivation moderately of commonly used brain endothelial cell cul- attenuates the upregulation of glucose uptake ture models and guidelines for their use. J after subsequent reoxygenation in brain endo- Cereb Blood Flow Metab 36(5):862–890 thelial cells. Neurosci Lett 506(1):44–49 9. Butt A, Jones HC, Abbott NJ (1990) Electrical 17. Abbott N, Patabendige AAK, Dolman DEM, resistance across the blood–brain barrier in Yusof SR, Begley DJ (2010) Structure and anaesthetised rats: a developmental study. J function of the blood–brain barrier. Neurobiol Physiol 429:47–62 Dis 37:13–25 52 Malgorzata Burek and Carola Y. Fo¨ rster

18. Engelhardt S, Patkar S, Ogunshola OO (2014) 29. Burek M, Arias-Loza PA, Roewer N, Forster Cell-specific blood-brain barrier regulation in CY (2010) Claudin-5 as a novel estrogen target health and disease: a focus on hypoxia. Br J in vascular endothelium. Arterioscler Thromb Pharmacol 171(5):1210–1230 Vasc Biol 30(2):298–304 19. Rubin L, Barbu K, Bard F, Cannon C, Hall 30. Brown RC, Morris AP, O’Neil RG (2007) DE, Horner H, Janatpour M, Liaw C, Tight junction protein expression and barrier Manning K, Morales J et al (1991) Differentia- properties of immortalized mouse brain micro- tion of brain endothelial cells in cell culture. vessel endothelial cells. Brain Res 1130 Ann N Y Acad Sci 633:420–425 (1):17–30 20. Alvarez J, Dodelet-Devillers A, Kebir H, 31. Czupalla CJ, Liebner S, Devraj K (2014) In Ifergan I, Fabre PJ, Terouz S, Sabbagh M, vitro models of the blood-brain barrier. Meth- Wosik K, Bourbonniere L, Bernard M, van ods Mol Biol 1135:415–437 Horssen J, de Vries HE, Charron F, Prat A 32. Wilhelm I, Fasakas C, Krizbai IA (2011) In (2011) The Hedgehog pathway promotes vitro models of the blood-brain barrier. Acta blood-brain barrier integrity and CNS immune Neurobiol Exp 71:113–128 quiescence. Science 334:1727–1731 33. Deli M, Abraham CS, Kataoka Y, Niwa M 21. Haseloff RFBI, Bauer HC, Bauer H (2005) In (2005) Permeability studies on in vitro blood–- search of the astrocytic factor(s) modulating brain barrier models: physiology, pathology, blood-brain barrier functions in brain capillary and pharmacology. Cell Mol Neurobiol endothelial cells in vitro. Cell Mol Neurobiol 25:59–127 25:25–39 34. Reichel A, Begley DJ, Abbott NJ (2003) An 22. Siddharthan V, Kim YV, Liu S, Kim KS (2007) overview of in vitro techniques for blood–brain Human astrocytes/astrocyte-conditioned barrier studies. Methods Mol Med medium and shear stress enhance the barrier 89:307–324 properties of human brain microvascular endo- 35. Toth A, Veszelka S, Nakagawa S, Niwa M, Deli thelial cells. Brain Res 1147:39–50 MA (2011) Patented in vitro blood-brain bar- 23. Thanabalasundaram G, Schneidewind J, rier models in CNS drug discovery. Recent Pat Pieper C, Galla HJ (2011) The impact of peri- Drug Discov 6:107–118 cytes on the blood-brain barrier integrity 36. Liu R, Zhang T, Wu C, Lan X, Du G (2011) depends critically on the pericyte differentia- Targeting the neurovascular unit: development tion stage. Int J Biochem Cell Biol of a new model and consideration for novel 43:1284–1293 strategy for Alzheimer’s disease. Brain Res 24. Hellstrom MGH, Kalen M, Li X, Eriksson U, Bull 86:13–21 Wolburg H, Betsholtz C (2001) Lack of peri- 37. Gaillard P, Voorwinden LH, Nielsen JL, cytes leads to endothelial hyperplasia and Ivanov A, Atsumi R, Engman H, Ringbom C, abnormal vascular morphogenesis. J Cell Biol Boer AG, Breimer DD (2001) Establishment 153:543–553 and functional characterization of an in vitro 25. Armulik A, Genove G, Mae M, Nisancioglu model of the blood–brain barrier, comprising a MH, Wallgard E, Niaudet C, He L, Norlin J, co-culture of brain capillary endothelial cells Lindblom P, Strittmatter K, Johansson BR, and astrocytes. Eur J Pharm Sci 12:215–222 Betsholtz C (2010) Pericytes regulate the 38. Hellinger E, Veszelka S, Toth AE, Walter F, blood-brain barrier. Nature 468:557–561 Kittel A, Bakk ML, Tihanyi K, Hada V, 26. Hori S, Ohtsuki S, Hosoya K, Nakashima E, Nakagawa S, Duy TDH, Niwa M, Deli MA, Terasaki T (2004) A pericyte-derived angio- Vastag M (2012) Comparison of brain capillary poietin-1 multimeric complex induces occludin endothelial cell-based and epithelial (MDCK- gene expression in brain capillary endothelial MDR1, Caco-2, and VB-Caco-2) cell-based cells through Tie-2 activation in vitro. J Neu- surrogate blood–brain penetration models. rochem 89(2):503–513 Eur J Pharm Biopharm 82:340–351 27. Uliasz TF, Hamby ME, Jackman NA, Hewett 39. Naik P, Cucullo L (2012) In vitro blood–brain JA, Hewett SJ (2012) Generation of primary barrier models: current and perspective tech- astrocyte cultures devoid of contaminating nologies. J Pharm Sci 101(4):1337–1354 microglia. Methods Mol Biol 814:61–79 40. Burek M, Salvador E, Forster CY (2012) Gen- 28. Tigges U, Welser-Alves JV, Boroujerdi A, Mil- eration of an immortalized murine brain ner R (2012) A novel and simple method for microvascular endothelial cell line as an culturing pericytes from mouse brain. Micro- in vitro blood brain barrier model. J Vis Exp vasc Res 84(1):74–80 66:e4022. Epub 7 Sept 2012 Culturing of Rodent Blood-Brain Barrier 53

41. Abbott N, Couraud PO, Roux E et al (1995) 51. Harkness KA, Adamson P, Sussman JD, Studies on an immortalized brain endothelial Davies-Jones GA, Greenwood J, Woodroofe cell line: characterization, permeability and MN (2000) Dexamethasone regulation of transport. In: Greenwood J, Begley DJ, Segal matrix metalloproteinase expression in CNS MB (eds) New concepts of a blood-brain bar- vascular endothelium. Brain 123 rier. Plenum Press, New York, pp 239–249 (Pt 4):698–709. Epub 29 Mar 2000 42. Greenwood J, Pryce G, Devine L, Male DK, 52. Regina A, Romero IA, Greenwood J, dos Santos WL, Calder VL et al (1996) SV40 Adamson P, Bourre JM, Couraud PO et al large T immortalised cell lines of the rat blood- (1999) Dexamethasone regulation of brain and blood-retinal barriers retain their P-glycoprotein activity in an immortalized rat phenotypic and immunological characteristics. brain endothelial cell line, GPNT. J Neuro- J Neuroimmunol 71(1–2):51–63. Epub 1 Dec chem 73(5):1954–1963. Epub 28 Oct 1999 1996 53. Roux F, Durieu-Trautmann O, Chaverot N, 43. Muruganandam A, Herx LM, Monette R, Claire M, Mailly P, Bourre JM et al (1994) Durkin JP, Stanimirovic DB (1997) Develop- Regulation of gamma-glutamyl transpeptidase ment of immortalized human cerebromicro- and alkaline phosphatase activities in immorta- vascular endothelial cell line as an in vitro lized rat brain microvessel endothelial cells. J model of the human blood-brain barrier. Cell Physiol 159(1):101–113. Epub 1 Apr FASEB J 11(13):1187–1197. Epub 21 Nov 1994 1997 54. Kido Y, Tamai I, Okamoto M, Suzuki F, Tsuji 44. Teifel M, Friedl P (1996) Establishment of the A (2000) Functional clarification of MCT1- permanent microvascular endothelial cell line mediated transport of monocarboxylic acids at PBMEC/C1-2 from porcine brains. Exp Cell the blood-brain barrier using in vitro cultured Res 228(1):50–57. Epub 10 Oct 1996 cells and in vivo BUI studies. Pharm Res 17 45. Lechardeur D, Scherman D, Schwartz B (1):55–62. Epub 14 Mar 2000 (1997) Development and characterization of 55. Blasig IE, Giese H, Schroeter ML, Sporbert A, cellular models of the blood-brain barrier. Utepbergenov DI, Buchwalow IB et al (2001) STP Pharma Sci 7:5–11 *NO and oxyradical metabolism in new cell 46. Forster C, Silwedel C, Golenhofen N, lines of rat brain capillary endothelial cells Burek M, Kietz S, Mankertz J et al (2005) forming the blood-brain barrier. Microvasc Occludin as direct target for glucocorticoid- Res 62(2):114–127. Epub 23 Aug 2001 induced improvement of blood-brain barrier 56. Terasaki T, Hosoya K (2001) Conditionally properties in a murine in vitro system. J Physiol immortalized cell lines as a new in vitro model 565(Pt 2):475–486. Epub 26 Mar 2005 for the study of barrier functions. Biol Pharm 47. Silwedel C, Forster C (2006) Differential sus- Bull 24(2):111–118. Epub 24 Feb 2001 ceptibility of cerebral and cerebellar murine 57. Lyck R, Ruderisch N, Moll AG, Steiner O, brain microvascular endothelial cells to loss of Cohen CD, Engelhardt B et al (2009) barrier properties in response to inflammatory Culture-induced changes in blood-brain bar- stimuli. J Neuroimmunol 179(1–2):37–45 rier transcriptome: implications for amino- 48. Yang T, Roder KE, Abbruscato TJ (2007) acid transporters in vivo. J Cereb Blood Flow Evaluation of bEnd5 cell line as an in vitro Metab 29(9):1491–1502 model for the blood-brain barrier under nor- 58. Calabria AR, Shusta EV (2008) A genomic mal and hypoxic/aglycemic conditions. J comparison of in vivo and in vitro brain micro- Pharm Sci 96(12):3196–3213. Epub 11 Sept vascular endothelial cells. J Cereb Blood Flow 2007 Metab 28(1):135–148 49. Kochi S, Takanaga H, Matsuo H, Naito M, 59. Blecharz KG, Haghikia A, Stasiolek M, Tsuruo T, Sawada Y (1999) Effect of cyclo- Kruse N, Drenckhahn D, Gold R et al (2010) sporin A or tacrolimus on the function of Glucocorticoid effects on endothelial barrier blood-brain barrier cells. Eur J Pharmacol 372 function in the murine brain endothelial cell (3):287–295. Epub 8 Jul 1999 line cEND incubated with sera from patients 50. Hosoya K, Tetsuka K, Nagase K, Tomi M, with multiple sclerosis. Mult Scler 16 Saeki S, Ohtsuki S et al (2000) Conditionally (3):293–302 immortalized brain capillary endothelial cell 60. Kleinschnitz C, Blecharz K, Kahles T, lines established from a transgenic mouse har- Schwarz T, Kraft P, Gobel K et al (2011) Glu- boring temperature-sensitive simian virus cocorticoid insensitivity at the hypoxic blood- 40 large T-antigen gene. AAPS PharmSci 2 brain barrier can be reversed by inhibition of (3):E27. Epub 14 Dec 2001 the proteasome. Stroke 42(4):1081–1089 54 Malgorzata Burek and Carola Y. Fo¨ rster

61. Salvador E, Burek M, Forster CY (2015) Stretch the blood-brain barrier. Front Cell Neurosci and/or oxygen glucose deprivation (OGD) in 8:352 an in vitro traumatic brain injury (TBI) model 63. Nakagawa S, Deli MA, Nakao S, Honda M, induces calcium alteration and inflammatory Hayashi K, Nakaoke R et al (2007) Pericytes cascade. Front Cell Neurosci 9:323 from brain microvessels strengthen the barrier 62. Neuhaus W, Gaiser F, Mahringer A, Franz J, integrity in primary cultures of rat brain endo- Riethmuller C, Forster C (2014) The pivotal thelial cells. Cell Mol Neurobiol 27 role of astrocytes in an in vitro stroke model of (6):687–694 Chapter 4

In Vitro BBB Models: Working with Static Platforms and Microfluidic Systems

Mohammad A. Kaisar, Vinay V. Abhyankar, and Luca Cucullo

Abstract

To date, many studies aimed at characterizing the multiple pathological paradigms associated with the development and progression of brain diseases are performed in vivo. However, major limitations of this approach, including limited translational significance to human patients, high cost, time, and labor to develop the appropriate model (e.g., transgenic or inbred models), have greatly favored the parallel development of in vitro models. These artificial systems attempt to mimic the physiological characteristics and behavioral responses of the corresponding tissue/organ in vivo. This is of particular interest for studies related to the cerebrovascular system and the blood-brain barrier (BBB). Because the BBB plays a critical role in maintaining the CNS homeostasis and dynamically responds to many events associated with rheological impairments, inflammation, oxidative stress, and exposure of potentially harmful xenobiotics, the study of the BBB in vitro is gaining momentum. As new cell culture techniques and improved in vitro technologies and biomaterials become available, in vitro modeling of the BBB is progressing rapidly. In this book chapter, we provide a detailed view and analysis of currently available static and dynamic in vitro BBB models including microfluidic systems and we also discuss recent and future developments in this expanding field of research.

Key words Microfluidic, Drug development, Permeability, Alternative, Endothelial, Tight junction, Biotechnology, Translational, Neurovascular

1 Introduction

Homeostatic milieu at neurovascular unit (NVU) is critically important for optimum function of the central nervous system (CNS). Highly functional and dynamic interface between the sys- temic circulation and the brain parenchyma—the blood-brain bar- rier (BBB)—plays a vital role in strictly and precisely maintaining the homeostasis by restricting the influx of potentially harmful substances (blood borne and xenobiotics) while facilitating trans- port of essential nutrients for neuronal activity. The innermost luminal component of BBB consists of endothelial cell (EC) on a support of basal lamina surrounded by an additional continuous envelop of pericytes and astroglial foot processes that separate

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_4, © Springer Science+Business Media, LLC, part of Springer Nature 2019 55 56 Mohammad A. Kaisar et al.

Fig. 1 Schematic illustration of the BBB anatomy. A cross section of brain microcapillary depicting the innermost luminal compartment of endothelial cell (EC) on a support of basal lamina surrounded by an additional continuous envelop of pericytes and astroglial foot processes that separate blood vessels from brain tissue. Tight junctions (TJs) present between the cerebral endothelial cells selectively block paracellular trafficking of substances from entering into brain

blood vessels from brain tissue (see Fig. 1). The basal lamina is approximately 40–50 nm thick and composed of three-dimensional networks of type IV collagen or laminin which acts as a scaffold for other intercalated matrix protein such as fibronectin and proteo- glycans. It is not only self-associated extracellular matrix proteins providing mechanical support or maintaining cytoskeleton struc- ture, interactions between matrix protein and cell surface receptors facilitate numerous biological processes, e.g., adhesion, migration, differentiation, and proliferation [1]. This lamina also regulates the cell-to-cell communication and embodies an additional trafficking barrier to the macromolecules between the vascular system and the brain [1, 2]. Brain microvascular ECs are the structural basis of BBB characterized by inter-endothelial tight junctions (TJ) consisting of three main integral protein types [claudins, occludins, and junctional adhesion molecules (JAM)] which selec- tively hinders paracellular diffusion of polar molecules across BBB In Vitro BBB Models: Working with Static Platforms and Microfluidic Systems 57

while efflux transporters deport substances through transcellular pathways. Numerous cytoplasmic adjunct proteins such as zonulae occludentes (ZO) and cingulin tether these transmembrane pro- teins to the cytoskeleton (see Fig. 1). Another notable feature of brain microvascular ECs is asymmetrical distribution of transpor- ters between the luminal and the basolateral membranes rendering them highly polarized to maintain functional integrity and homeo- static brain microenvironment [3–5]. The BBB phenotype develops under the influence of endothelial neighboring cells. Among them the role of astrocytes is prominent in induction and maintenance of the tight junctions and BBB properties. Astrocytes interact with brain microvascular ECs and modulate their differentiation, protein expression, and permeability [6, 7]. The highly controlled traffick- ing across BBB renders the brain microenvironment a privileged site but creates obstacles to CNS drug delivery. Thus to further our understanding of BBB physiology and its dynamic response to stimuli and assess/improve drug delivery into the CNS, interest in developing in vitro platforms closely mimicking the behavior and physiology of the BBB in situ has become of primary interest across basic, translational, and pharmaceutical research. Featuring a number of desirable advantages encompassing basic, translational, pharmaceutical studies including versatility, high-throughput (HTS) capability, relative simplicity, and flexibil- ity, in vitro BBB platforms have rapidly progressed during the last decade evolving from sort of companion “omni tools” to highly sophisticated research devices with specific sets of characteristics, adapted to fill particular research niches and requirements [8]. From a translational and pharmacological standpoint, typically an in vitro BBB model is platform aimed at reproducing the in situ brain microvascular environment through the development of a realistic neurovascular system demonstrating physiological and functional properties identical to the BBB in vivo. At the same time, this artificial BBB is also expected to provide a cost-effective, user-friendly, and highly controllable working platform capable of carrying out the desired experiments with high accuracy and repro- ducibility [9, 10]. Several in vitro BBB models have already been developed. The most widely used and commercially available platform is the trans- well chamber for assessing BBB integrity, barrier property, perme- ability, activation, and migration of inflammatory cells (see also Note 1). Nevertheless, recent advancements in the biotechnology field and materials engineering and understanding of BBB biology have enabled the development of innovative and highly integrated “quasi-physiological” (e.g., capable of reproducing a number of environmental complexities and biological features of the BBB including vascular hemodynamics and physiological/pathological responses to stimuli) in vitro BBB platform which are discussed herein. 58 Mohammad A. Kaisar et al.

2 Methods

2.1 Static Transwell inserts are a widely used in vitro BBB platform for its Culture-Based BBB simplicity and feasibility to assess permeability across BBB and its Model in Transwell integrity, metabolic function, and pathophysiological changes. In this easy-to-use tool, typically brain microvascular endothelial cells are cultured on a microporous semipermeable support (transwell insert) about a millimeter off the bottom of the well in which the insert is hanging from top (see Fig. 2). Transwell inserts are avail- able in a range of diameters, membrane types, and pore sizes to satisfy the purpose of the study. Standard membrane diameters for transwells are 6.5 mm (24-well plate), 12 mm (12-well plate), 24 mm (6-well plate), and 75 mm (100 mm dish) having insert membrane growth area of 0.33, 1.12, 4.67, and 44 cm2.

Fig. 2 In vitro BBB models on transwell platform using co-culture of astrocytes and endothelial cells. Endothelial cells are cultured on top of microporous polyester inserts while astrocytes are seeded at the bottom of the insert In Vitro BBB Models: Working with Static Platforms and Microfluidic Systems 59

Commonly used membrane materials include polycarbonate, poly- ester (polyethylene terephthalate, PET or PETE), and polytetra- fluoroethylene (PTFE). Polycarbonate transwell inserts are thin, translucent while polyester membranes are transparent and provide better cell visibility under phase-contrast microscopy which allows assessment of cell viability and monolayer formation. Selecting the correct pore size according to the experimental needs and expected outcomes is critically important. For drug transport studies across BBB membrane with smaller pore size (0.4–3.0 μm) are primarily used. Transmigration of immune cells on inflammatory stimulus is usually studied in transwell membrane with 3.0 μm or larger pore sizes because cells cannot cross through pores smaller than 3.0 μm. The simplest way of using transwell apparatus as in vitro BBB model is to grow monoculture of brain microvascular endothelial cells on the insert which allows the passage of solutes (including cell-derived factors) from and to the growth medium between the apical (luminal) and basolateral (abluminal) compartments sepa- rated by the insert. The cerebrovascular ECs obtained from differ- ent origins such as mouse, rat, bovine, porcine, monkey, or human have been utilized to establish BBB in transwell. From a practical perspective this model is user friendly, cost effective, and less trou- blesome but this arrangement lacks the barriergenic modulatory stimuli imparted by neighboring cells (astrocytes and pericytes) and mechanical stimuli (e.g., shear stress). Consequently, more sophis- ticated models including co- and triple-culture systems have been developed. Owing to the close spatial relationship, co-cultures of cerebrovascular endothelial cells with astrocytes are widely used since astrocytes play a crucial role in the development of the para- cellular tightness of the BBB. Different experimental systems were developed to mimic the astrocytic influence on the BBB endothe- lium. One of the most commonly used configurations include endothelial cells seeded on the apical surface (luminal) of a micro- porous membrane and juxtaposed astrocytes loaded on its basolat- eral surface (abluminal). Although this arrangement allows for the direct contact between endothelial cells and astrocytes, the relative higher thickness of the artificial membrane compared to the basal lamina in vivo limits cell-cell interaction (see Fig. 2). An alternative approach is to culture astrocytes at the bottom of the wells and allow for the diffusible released factors to reach the BBB endothe- lium on the other side of the membrane. Although pericytes are in close contact with ECs they are relatively less characterized. However, their role in modulating endothelial functions has been well established in recent years and accordingly their use in BBB modeling has also been explored. Pericytes from different origins have been incorporated in a range of co- and triple-cultures BBB setups in transwell platforms using various configurations. A common triple-culture setup consists of EC monolayers laid on the top of a microporous insert with 60 Mohammad A. Kaisar et al.

juxtaposed pericytes on the basal side of the membrane and astrocytes seeded at the bottom of the well. Astrocyte and pericyte culture mixes at the bottom of the well have also been tested with in few occasions [11, 12]. Triple-cell-culture models using BMECs, neurons, and astrocytes/pericytes have also been reported. Neuron-based triple cultures recreate the basic structure, function, and cell-cell interaction of the NVU in vivo. Therefore, NVU models incorporating neurons can also be used to assess (to some extent) the effect of the drug treatment on the CNS. However, the complexity of the culture environment increases significantly, dras- tically reducing the manageability of these systems. Since brain cancer- and neurodegenerative and neuroinflamma- tory disorder-related preclinical research is predominantly per- formed in mice, using mouse-derived primary brain microvascular endothelial cells is reasonably best suited to provide anticipated and reproducible results during in vitro to in vivo transition phase. A functional parameter-trans-endothelial electrical resistance (TEER, Ohms (Ω)cm2) measures the electrical impedance posed by the barrier. Low TEER indicates disruption/loss of BBB integrity and paracellular tightness. BBB models have been developed from freshly prepared mouse BMEC monocultures or in combination with pericytes. The reported TEER values were in the range of ~50 and ~150 Ω cm2, respectively [13], and rising up to ~200 Ω cm2 when astrocytes are incorporated into the culture [14, 15]. Although primary BMECs closely mimic the in vivo BBB phenotype, presence of non-endothelial cell contaminants (e.g., pericytes and fibroblasts) is one of the major issues researchers have to deal with during the extraction process. These cells can disrupt the intactness and uniformity of the endothelial culture by forming holes or void spaces in the monolayer. This can impact the reliability and accuracy of permeability testing. Primary rat BMECs exhibit higher TEER values (ranging from 300 to more than 600 Ω cm2) than mouse-derived ECs either in mono, co-, or triple cultures with astrocytes, pericytes, or both [16–18]. However, pri- mary bovine and porcine BMECs exhibit even higher electrical impedance (TEER >1000 Ω cm2) and expression levels of TJ proteins ZO-1 and claudin-5 [19–21].

2.1.1 Cell Lines or A recurrent dilemma about establishing an effective BBB model Primary Cultures? in vitro is whether cell lines or primary cultures (human or animal derived) are better suited for the scope. Human brain microvascular endothelial cells (HBMEC) and astrocytes can be isolated from fresh tissue (including fetal human brain specimens, autopsy, or tissue resections from brain surgeries) to generate primary cultures. One of the advantages of using primary cells lies in the possibility (although limited) to obtain disease-specific cells which can be helpful to dissect out basic BBB pathogenic mechanisms and rele- vant pathological traits of the brain microvascular system (see also Note 2). In Vitro BBB Models: Working with Static Platforms and Microfluidic Systems 61

An alternative to primary human cells includes animal-derived primary cultures such as rodents (usually rats), bovine, and porcine primary BBB ECs as well as non-brain vascular endothelium such as umbilical vein endothelial cells (HUVEC) and human intestinal epithelial cells such as Caco-2. Although it is important to mention that while Caco-2 cells express tight junctions (TJs) and are consid- ered good models for studies of passive diffusion across the BBB, they do not recapitulate the central role of transporters in the regulation of BBB selective permeability. More recently, a viable solution to the difficulty to access human brain tissue to generate primary cell cultures (also one of the main limiting factors hinder- ing the widespread use of humanized in vitro BBB models) is to generate human BBB cells from human pluripotent stem cells (hPSCs) by forcing them to acquire the desired BBB cellular phe- notype through the use of proper physiological cues including exposure to neuronal cells. Although this is an interesting solution maintaining the cells in the desired phenotype for a prolonged period of time is still problematic. An additional alternative to the use of primary cells (either human or animal derived) in BBB modeling (also to reduce cost and labor associated with the procurement of primary cells) is to opt for immortalized BBB cell lines. HMEC-1, HCMEC/D3, and TY08 are among the immortalized human brain endothelial cell lines that have been established and used in BBB modeling. Immor- talized rat brain endothelial cells (RBE4) are among the most extensively animal-derived cell line used for BBB modeling in static co-culture systems. RBE4 cells express a variety of BBB transporters including multidrug resistance-associated protein (Mrp1) and P-glycoprotein (Pgp) and exhibit drug-metabolizing activities although they fall short of forming TJ complexes resulting in high paracellular permeability [22]. bEnd.3 and bEnd.5 are instead widely used mouse-derived immortalized cell lines. Of these two bEnd.3 cells grow rapidly and seem to retain the endothelial phe- notype over multiple passages. Their ability to express higher levels of TJ proteins (claudin-5, occludin, and ZO-1) allows them to develop a more stringent and physiological responsive in vitro BBB [23, 24].

2.1.2 Transwell Insert All the steps should be performed under sterile laminar airflow in a Coating biosafety cabinet, (BSC) class II. We recommend exposing the hood to UV light for at least an hour prior to the start of a cell culture. The transwell package needs to be sprayed with 70% alco- hol thoroughly before transferring into the hood. The inserts and the wells must be opened only in the hood and sterile tweezers (dipped in 70% alcohol for at least 5 min in hood and air-dried before use) should be used to handle the inserts. For convenience, either a sterile Petri dish (bottom of the inserts) or another tissue 62 Mohammad A. Kaisar et al.

culture plate with equal number of wells (top of the inserts) can be used to keep the inserts intermittently during coating process. We recommend using a vacuum pump with sterile glass Pasteur pip- ettes (5.75 in.) to aspirate the liquid from top of the insert and the wells cautiously to avoid accidentally touching the insert membrane or disruption of the monolayer. Careless, rough, and fast handling of Pasteur pipette (sharp glass tip) may damage the insert or make a hole. We recommend holding the inserts/wells in slightly tilted position during aspiration of liquid to avoid damage. Here we describe step-by-step coating process of 6.5 mm insert with 0.4 μm pores and 0.33 cm2 area made of polyester membrane (24-well plate, Costar, Corning Inc.) for a static co-culture of human BBB endothelial cells on the apical (luminal, on top of insert) side and human astrocytes (purchased from commercially available source, ScienCell Research Laboratory) in juxtaposition to ECs on the basal (abluminal, on bottom side of the insert) compartment. As per supplier’s recommendation poly-L-lysine (PLL) should be used for coating to facilitate effective attachment of the astro- cytes. Thaw PLL (stored at À20 C) stock solution in a water bath at 37 C. Once thawed disinfect the external surface of the tube with 70% alcohol before transferring to the hood. Add 2 μL stock solution (10 mg/mL) to 2 mL sterile water and vortex under hood to prepare 10 μg/mL PLL solution. Open the lid of the 24-well plate, carefully transfer one insert to a sterile Petri dish (sufficiently large enough to keep 24 inserts well spaced) at a time with upside- down position. Add 75 μL of PLL solution on the bottom side of the insert placed on the Petri dish at inverted position. Once addition of PLL solution to 24 inserts is complete, close the lid of the Petri dish to prevent evaporation and leave in the incubator at 37 C for at least an hour (longer incubation up to 3 h may result in better coating). Meanwhile, prepare recommended coating solution for the EC to be seeded on top of the inserts. Take out collagen type I (5 g/ mL) bottle from refrigerator (4–8 C) and leave under the hood for few minutes to reduce the consistency for uninterrupted and smooth pipetting. Add 4 μL of collagen type I stock solution to 4 mL of 0.2 M acetic acid to yield a final concentration of 50 μg/ mL. Filter the solution using 0.2 μ syringe filter and 10 mL syringe. Filter slowly but forcefully as collagen solution creates too much back pressure to the plunger. Add 500 μL of sterile PBS to each of the 24 well before transferring the inserts from Petri dish. Aspirate PLL carefully by 200 μL pipette, flip the insert to original position, and put back into the well to which PBS was added previously. Apply 150 μLof50μg/mL collagen solution to each well, close the lid to prevent evaporation, and transfer to the incubator at 37 C for an hour. In Vitro BBB Models: Working with Static Platforms and Microfluidic Systems 63

2.1.3 Astrocytes and EC When the culture reaches 90% confluency or above the astrocyte is Cell Seeding ready to seed. Warm PBS, complete growth medium, and trypsin/ EDTA solution to room temperature (using water bath is not recommended) prior to initiation of seeding. Rinse the cells with warm PBS twice. We recommend using 2.5 mL PBS if the cells are cultured in T-25 flask. Add 600 μL of trypsin/EDTA solution; make sure that the entire surface of the flask has been covered by gently rocking the flask. Incubate the flask at 37 C for 1–2 min or until cells completely round up and detach. The change in cell morphology and complete detachment must be monitored under microscope. Add 3.5 mL of complete astrocyte growth medium to neutralize the trypsin. Centrifuge at 1000 rpm for 5 min in a 15 mL tube. Resuspend the cells in 2 mL of culture medium and count them in a hemocytometer. We recommend preparing a suspension of 50,000 cells in 2 mL of complete growth medium (dilution might be required to achieve this cell density) sufficient enough to seed 24 inserts. Aspirate PBS from the wells and collagen solution from the insert following 1-h incubation, and transfer the inserts to the Petri dish again facing upside down (inverted). Apply 75 μL cell suspen- sion on the bottom side of the insert (mix well by using pipette before seeding), cover the Petri dish with lid to prevent evapora- tion, and incubate at 37 C for an hour to let the astrocytes to attach. Once the cells are being attached transfer them immediately to the wells in original orientation to which 700 μL of warm complete growth medium has already been added. Carefully add 200 μL of astrocyte media to the upper compartment and incubate  overnight in a 37 C, 5% CO2, incubator. It is difficult to visualize the proper seeding density and attachment under microscope. The morphology of the cells is circular at the beginning but they stretch and start to form process as the time goes by. EC cells (passage 1–4) are seeded on collagen-coated insert top in EBM-2 basal medium supplemented with 5% FBS, chemically defined lipid concentrate, growth factors, antibiotic/antimycotic (1:1), and HEPES (10 mM). Wash cells with prewarm sterile PBS. Add 2 mL of trypsin/EDTA solution to a T-25 flask and incubate at 37 C for 4–5 min. Once the cells are detached and spherical in shape neutralize the trypsin/EDTA with 8 mL of complete growth medium. Spin at 1700 rpm for 10 min. Aspirate the media and resuspend the cell pellet in 5 mL of complete growth medium. Count the cells and dilute it to 40,000 cells per mL of media. Aspirate the astrocyte media from apical chamber and add 200 μL of the EC cell suspension. Transfer to the 37 C, 5%, incubator and monitor the seeding density and cell morphology the following day. A confluent monolayer is expected to be formed in less than a week. 64 Mohammad A. Kaisar et al.

2.2 Flow-Based BBB Microfluidic based microphysiological systems (MPS) are becom- Model: Assessing the ing increasingly attractive for developing advanced in vitro models Option of Microfluidic of barrier tissues such as the gut, lung, or BBB [25]. These systems Devices combine microengineering techniques with physiologically organized living cell populations and aim to replicate the unique biochemical and biophysical microenvironments found in vivo. This work is underpinned by the hypothesis that recapitulation of key in vivo elements within scalable in vitro models can provide clinically relevant insights into drug mechanism of action, molecu- lar transport, and host-pathogen interaction [26]. Stepping beyond conventional barrier models, the integration of sensors (e.g., TEER, temperature, pH) allows real-time sensing and monitoring to assess barrier function [27, 28] while multiple MPS modules can be linked together to explore multi-tissue responses [29]. In a typical MPS, top and bottom microfluidic channels are separated with a porous cell culture substrate and the barrier tissue is established by seeding appropriate cell populations on either side of the porous substrate [30]. Cells are maintained under flow to support cell proliferation and mimic the tissue-specific shear stres- ses that induce physiological junction formation and establish polarized tissues. Given the importance of the BBB in biomedical research, several microfluidic models incorporating closely apposed blood and brain compartments have been established [31–35]. In this section we discuss emerging techniques to create BBB models using 3D gels and other biomaterials that have been published over the past year. As shown in Fig. 3, Xu and colleagues [36] presented an elegant approach to create a high-throughput microfluidic system to create and validate a BBB model composed of astrocytes (brain compartment), microvascular endothelial cells (blood compart- ment), and a collagen gel to explore metastatic brain tumors and therapeutic responses to chemotherapy drugs. The authors reported interesting findings that relate specific interactions between cancer cells and astrocytes play a role in the effectiveness of a malignant tumor crossing the BBB. One of the key challenges in developing microfluidic approaches is the inherent complexity both in fabrication and oper- ation of the devices. This complexity poses a challenge for nonengineering-focused laboratories to integrate microfluidic tools into their experiments. Thus, there is a growing trend in the microfluidics community to develop techniques that can be readily translated from engineering-focused labs to life science labora- tories. As shown in Fig. 4, a possible alternative to these issues has been proposed and worked on by the group of Dr. Abhyankar and colleagues who came out with an easy-to-use modular approach (scalable easy-access modular—SEAM), adopting a magnetic cou- pling that simplifies the operational workflow from cell seeding to cell isolation for downstream analysis [37]. A differentiating feature In Vitro BBB Models: Working with Static Platforms and Microfluidic Systems 65

Fig. 3 (a) Cellular components of the BBB and a schematic view of the surface transporters and functional proteins validated in the study (b). (c) Highlights of (i) overall device design; (ii) view of functional units and expanded schematics of the highlighting cellular components and flow directions (panels iii, iv). (d) Workflow describing procedures to establish BBB including (i) gel introduction, (ii) astrocyte seeding, (iii) endothelial seeding, (iv) static culture, and (v) co-culture under defined microfluidic flow on the endothelial side

of this platform is that culture membrane is independent from the microfluidic architecture and can be transferred from specialized cell seeding modules and culture modules, and then removed for high-resolution imaging or analysis using commercial workflows including nucleic acid extraction [37]. This approach also allows tailored biomaterials with tissue-specific thickness and properties to be easily integrated and then seeded with cells. As shown in Fig. 4b, the proposed platform integrates a hyaluronic-acid-based culture membrane with thickness and Young’s modulus that mimics the in vivo separation between the brain and blood compartments. 66 Mohammad A. Kaisar et al.

Fig. 4 (a) Schematic of magnetic latching platform that simplifies the experimental workflow for microfluidic model development. (b) The culture membrane and microfluidic architecture are independent from one another and tailored biomaterials can be easily integrated. A self-assembled hyaluronic acid membrane is shown with Young’s modulus (E) and thickness that mimic the separation between the brain and blood compartments. (c) MAP2 staining of astrocytes cultured on the hyaluronic acid membrane

Then astrocytes are seeded on the surface of the biomimetic mem- branes. Results demonstrated that the use of biomimetic mem- branes provide a more in vivo-like culture support than that provided by more rigid hyaluronic acid-coated surfaces, thus high- lighting the importance of selecting culture materials that recapitu- late in vivo characteristics to help develop more clinically meaningful models (see also Note 3).

2.3 Workflow and The modular workflow used in SEAM is shown in Fig. 5 (adapted Experimental Protocol from Abhyankar et al. [37]). The culture membrane is independent for the SEAM Platform of the seeding and culture modules and can therefore be transferred from one specialized module to another. In conventional micro- fluidic systems, uniform cell seeding and localization are challeng- ing because the cells are introduced to the culture channel under flow. Here, the culture membrane is sealed between two housing layers with direct access to the surface of the membrane (Fig. 5a). Cells can be added directly to the membrane surface at a defined seeding density, incubated, and allowed to attach. Cells are then inspected visually to ensure that desired surface density is achieved and can be reseeded as needed. Co-cultures are established by disassembling the magnetically sealed housings, flipping over the membrane, and then repeating the seeding process (Fig. 5b, steps 1–4). Once the desired cell populations are in place, the seeding module is disassembled and the membrane (with attached cell populations) is removed and transferred to the microfluidic culture platform. Figure 5c shows co-cultured primary human alveolar epithelium and microvascular endothelium ready to transfer to nVtoBBMdl:Wrigwt ttcPafrsadMcoudcSses67 Systems Microfluidic and Platforms Static with Working Models: BBB Vitro In

Fig. 5 (a) Magnetically latched seeding module used to efficiently introduce cells onto a suspended cell culture membrane. (b) Magnetic latching allows the membrane to be reversible sealed within the system to simplify the process required to establish tissue interfaces (steps 1–4). (c) After cell attachment, the membrane is transferred and magnetically sealed within a microfluidic culture architecture. (d) Schematics of the culture module including microfluidic (left panel) and overview of the fully assembled device (right panel). (e) Endothelial LPS exposure with imaging and gene expression readouts were simplified by removable culture membrane. Adapted from Abhyankar et al. [37] 68 Mohammad A. Kaisar et al.

the culture module. In Fig. 5d, the membrane is magnetically sealed between two microfluidic channels that provide access to the apical and basolateral surfaces of the tissue interface. If bubbles form in the channels, they can be easily removed by disassembling the magnetically sealed housings. This is a key advantage from conventional microfluidic systems where trapped bubbles are noto- riously difficult to remove from a system. In conventional microfluidic systems, fluid flow is required to lyse and collect lysate for downstream analysis. The removable culture insert provides a simple way to access cells within the system without requiring fluidic manipulation steps that can cause sample degradation or loss. Figure 5e shows results from an exposure experiment with gene expression and surface protein imaging after a 4-h apical stimulation with the bacterial endotoxin lipopo- lysaccharide (LPS) under flow. The membrane was removed from the culture platform and directly immersed in lysis buffer to extract RNA for gene expression or fixed and stained for fluorescence imaging. As shown in the workflow, the modular approach provides experimental flexibility with simplified steps for seeding, culture, and analysis. Future Directions: As microfluidic models continue to mature and widespread accessibility concerns are addressed these tools could help significantly advance BBB and other barrier tissue research.

3 Notes

1. Transwell platforms, while providing scalability, cost- effectiveness, and ease of use, are still considered the gold standard in the field of BBB modeling; they lack the ability to effectively mimic the NVU milieu and many physiological fea- tures expressed by the BBB in vivo including endothelial expo- sure to physiological shear stress [38]; the edge effect compromises barrier effectiveness and enables paracellular pas- sages along the edges of the transwell membrane; and they provide significantly lower barrier selectivity and higher perme- ability to polar molecules [39, 40]. 2. Problematic issues that need to be taken into consideration when working with primary cells are as follows: (1) primary cells ex situ tend to dedifferentiate quickly, thus limiting the number of passages; prior to use these cells can withstand before losing their defining characteristics; (2) purity of the original primary culture where the presence of cellular bystan- ders (e.g., fibroblasts, pericytes) left behind during the purifi- cation process can affect the quality, behavior, and responsiveness of the derived BBB model; and (3) isolation In Vitro BBB Models: Working with Static Platforms and Microfluidic Systems 69

procedures are generally limited by the ability to obtain the desired brain tissue and yield of the isolation procedures. In some instances, it is possible to exploit the intrinsic biological features of the BBB endothelium such as high-expression-level P-glycoprotein to purify the cell cultures using toxic levels of puromycin (a P-glycoprotein substrate) to kill the contami- nants without affecting EC viability [16]. 3. Unfortunately, there are other factors that must be factored in when selecting the “microfluidic option.” The first one is plat- form availability. Given the “experimental” nature of these platforms, availability is generally very limited. In other words, these are not systems that can be purchased from a vendor’s catalog. Being “experimental” also means that tech- nical support and user guides are also in short supply as well. Direct collaboration with the developer/s is almost 100% a must since transferring/lending a complex technology to other laboratories that may lack bioengineering expertize could be problematic. The second one is sampling; microfluidic systems (as the definition of the word implies) work with a minimal amount of culture medium in the order of μL; there- fore, sampling volumes dictated by the experimental needs and sampling frequency need to be well evaluated before moving on with the use of these platforms. Microfluidic pumps and micro-samplers will also likely be needed.

References

1. Tilling T et al (2002) Expression and adhesive 8. Palmiotti CA et al (2014) In vitro cerebrovas- properties of basement membrane proteins in cular modeling in the 21st century: current and cerebral capillary endothelial cell cultures. Cell prospective technologies. Pharm Res 31 Tissue Res 310(1):19–29 (12):3229–3250 2. Tilling T et al (1998) Basement membrane 9. Wang YI, Abaci HE, Shuler ML (2017) Micro- proteins influence brain capillary endothelial fluidic blood-brain barrier model provides barrier function in vitro. J Neurochem 71 in vivo-like barrier properties for drug perme- (3):1151–1157 ability screening. Biotechnol Bioeng 114 3. Correale J, Villa A (2009) Cellular elements of (1):184–194 the blood-brain barrier. Neurochem Res 34 10. Alavijeh MS et al (2005) Drug metabolism and (12):2067–2077 pharmacokinetics, the blood-brain barrier, and 4. Huber JD, Egleton RD, Davis TP (2001) central nervous system drug discovery. Neu- Molecular physiology and pathophysiology of roRx 2(4):554–571 tight junctions in the blood-brain barrier. 11. Lippmann ES et al (2011) Blood-brain barrier Trends Neurosci 24(12):719–725 modeling with co-cultured neural progenitor 5. Tietz S, Engelhardt B (2015) Brain barriers: cell-derived astrocytes and neurons. J Neuro- crosstalk between complex tight junctions and chem 119(3):507–520 adherens junctions. J Cell Biol 209(4):493–506 12. Xue Q et al (2013) A novel brain neurovascular 6. Abbott NJ (2002) Astrocyte-endothelial inter- unit model with neurons, astrocytes and micro- actions and blood-brain barrier permeability. J vascular endothelial cells of rat. Int J Biol Sci 9 Anat 200(6):629–638 (2):174–189 7. Haseloff RF et al (2005) In search of the astro- 13. Daneman R et al (2010) Pericytes are cytic factor(s) modulating blood-brain barrier required for blood-brain barrier integrity dur- functions in brain capillary endothelial cells ing embryogenesis. Nature 468(7323): in vitro. Cell Mol Neurobiol 25(1):25–39 562–566 70 Mohammad A. Kaisar et al.

14. Shayan G et al (2011) Murine in vitro model of 27. Ferrell N et al (2010) A microfluidic bioreactor the blood-brain barrier for evaluating drug with integrated transepithelial electrical resis- transport. Eur J Pharm Sci 42(1–2):148–155 tance (TEER) measurement electrodes for 15. Stamatovic SM et al (2003) Potential role of evaluation of renal epithelial cells. Biotechnol MCP-1 in endothelial cell tight junction ‘open- Bioeng 107(4):707–716 ing’: signaling via rho and rho kinase. J Cell Sci 28. Young EW et al (2010) Technique for real-time 116(Pt 22):4615–4628 measurements of endothelial permeability in a 16. Perriere N et al (2005) Puromycin-based puri- microfluidic membrane chip using laser- fication of rat brain capillary endothelial cell induced fluorescence detection. Anal Chem cultures. Effect on the expression of blood- 82(3):808–816 brain barrier-specific properties. J Neurochem 29. Xiao S et al (2017) A microfluidic culture 93(2):279–289 model of the human reproductive tract and 17. Abbott NJ et al (2012) An improved in vitro 28-day menstrual cycle. Nat Commun blood-brain barrier model: rat brain endothe- 8:14584 lial cells co-cultured with astrocytes. Methods 30. Huh D et al (2012) A human disease model of Mol Biol 814:415–430 drug toxicity-induced pulmonary edema in a 18. Nakagawa S et al (2009) A new blood-brain lung-on-a-chip microdevice. Sci Transl Med 4 barrier model using primary rat brain endothe- (159):159ra147 lial cells, pericytes and astrocytes. Neurochem 31. Booth R, Kim H (2012) Characterization of a Int 54(3–4):253–263 microfluidic in vitro model of the blood-brain 19. Helms HC, Brodin B (2014) Generation of barrier (muBBB). Lab Chip 12 primary cultures of bovine brain endothelial (10):1784–1792 cells and setup of cocultures with rat astrocytes. 32. Griep LM et al (2013) BBB on chip: microflui- Methods Mol Biol 1135:365–382 dic platform to mechanically and biochemically 20. Patabendige A et al (2013) A detailed method modulate blood-brain barrier function. for preparation of a functional and flexible Biomed Microdevices 15(1):145–150 blood-brain barrier model using porcine brain 33. Prabhakarpandian B et al (2013) SyM-BBB: a endothelial cells. Brain Res 1521:16–30 microfluidic blood brain barrier model. Lab 21. Thomsen LB, Burkhart A, Moos T (2015) A Chip 13(6):1093–1101 triple culture model of the blood-brain barrier 34. Booth R, Kim H (2014) Permeability analysis using porcine brain endothelial cells, astrocytes of neuroactive drugs through a dynamic micro- and pericytes. PLoS One 10(8):e0134765 fluidic in vitro blood-brain barrier model. Ann 22. Roux F, Couraud PO (2005) Rat brain endo- Biomed Eng 42(12):2379–2391 thelial cell lines for the study of blood-brain 35. Srinivasan B et al (2015) TEER measurement barrier permeability and transport functions. techniques for in vitro barrier model systems. J Cell Mol Neurobiol 25(1):41–58 Lab Autom 20(2):107–126 23. Watanabe T et al (2013) Paracellular barrier 36. Xu H et al (2016) A dynamic in vivo-like orga- and tight junction protein expression in the notypic blood-brain barrier model to probe immortalized brain endothelial cell lines metastatic brain tumors. Sci Rep 6:36670 bEND.3, bEND.5 and mouse brain endothe- 37. Abhyankar VV et al (2016) A reversibly sealed, lial cell 4. Biol Pharm Bull 36(3):492–495 easy access, modular (SEAM) microfluidic 24. Brown RC, Morris AP, O’Neil RG (2007) architecture to establish in vitro tissue inter- Tight junction protein expression and barrier faces. PLoS One 11(5):e0156341 properties of immortalized mouse brain micro- 38. Cucullo L et al (2011) The role of shear stress vessel endothelial cells. Brain Res 1130 in blood-brain barrier endothelial physiology. (1):17–30 BMC Neurosci 12:40 25. Esch EW, Bahinski A, Huh D (2015) Organs- 39. Santaguida S et al (2006) Side by side compar- on-chips at the frontiers of drug discovery. Nat ison between dynamic versus static models of Rev Drug Discov 14(4):248–260 blood-brain barrier in vitro: a permeability 26. Herland A et al (2016) Distinct contributions study. Brain Res 1109(1):1–13 of astrocytes and pericytes to neuroinflamma- 40. Cucullo L et al (2005) Drug delivery and tion identified in a 3D human blood-brain bar- in vitro models of the blood-brain barrier. rier on a chip. PLoS One 11(3):e0150360 Curr Opin Drug Discov Devel 8(1):89–99 Chapter 5

In Vitro Cell Models of the Human Blood-Brain Barrier: Demonstrating the Beneficial Influence of Shear Stress on Brain Microvascular Endothelial Cell Phenotype

Keith D. Rochfort and Philip M. Cummins

Abstract

The field of translational cerebrovascular research routinely employs blood-brain barrier (BBB) cell models. Using in vitro culture models to accurately mimic the physiological complexity of the in vivo BBB continues to be a challenge, however. To meet this challenge, in vitro BBB models have evolved significantly over the last three decades, from static monocultures to dynamic multicellular flow-based systems. In this chapter, we initially focus on three key aspects that have helped to drive the evolution of in vitro BBB models, namely (1) the availability of suitable BBB cell lines; (2) a fuller understanding of the neurovascular unit (NVU); and (3) an appreciation of the relevance of blood flow shear stress to BBB physiology. We then put this knowledge into a more practical context by experimentally demonstrating two alternate means of applying physiological shear stress to primary-derived human brain microvascular endothelial cells (HBMvECs), and showing how BBB phenotype is improved in response to flow by using immunocytochemical localization of tight-junction zonula occludens-1 (ZO-1) as a reporter index. The ability of applied shear to attenuate the pro-oxidant effects of inflammatory TNF-α in HBMvECs will also be demonstrated using flow cytometry, further highlighting the relevance of introducing shear into BBB models.

Key words Blood-brain barrier, Endothelial, Shear stress, Neurovascular unit, Permeability, ROS, ZO-1

1 Introduction

A steady increase in the aging population (+65 years), coupled with an increased age-related incidence of debilitating central nervous system (CNS) disorders [1–4], has led to the field of neurother- apeutics becoming a multibillion-dollar industry [5, 6]. With neu- rological healthcare costs for Europe alone estimated at $800 billion in 2010 [7], and with a projected 85% increase in costs by 2030 [8], the challenge to develop medicines that can effectively access the brain is therefore a high priority. A key obstacle however is the blood-brain barrier (BBB), a formidable microvascular struc- ture that functions as a metabolic barrier, dynamically controlling

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_5, © Springer Science+Business Media, LLC, part of Springer Nature 2019 71 72 Keith D. Rochfort and Philip M. Cummins

and restricting the exchange of endogenous/exogenous macromo- lecules and fluids between the circulating blood and the brain parenchyma, thereby maintaining neural homeostasis by actively partitioning the latter from the former [9–12]. It follows therefore that loss of BBB integrity leading to an increase in barrier perme- ability is a central feature of common neurological disorders that are instigated by environmental factors such as proinflammatory sti- muli [13–15], brain trauma [16, 17], and rheological impairment [18]. For these reasons, the BBB commands huge clinical impor- tance; a natural barrier that protects and regulates the brain micro- environment also renders drug delivery to the brain very difficult [19–26]. As a result, drug discovery programs now integrate quan- titative BBB penetration studies, a reflection of the growing empha- sis on design and delivery of novel CNS-directed therapeutics. BBB complexity however has extended the average timeline for success- ful CNS drug development from 10 to 12 years [27]. To improve upon this, the development of better translational research models is widely viewed as a priority [28], as this could significantly shorten the drug-development timeline [29–31]. This concise introduction will examine the evolution of BBB modeling and the key attempts to recapitulate its sophisticated behavior and architecture in order to tackle this issue. The BBB is a complex multicellular interface responsive to both physiological and pathophysiological events [32]. It collectively comprises the microvascular endothelial cells (or endothelium) of the cerebrovascular capillary network, which are unique in various ways to peripheral endothelial cells. BBB endothelia lack fenestra- tions, display reduced pinocytosis, exhibit distinct differences in cell morphology and expression of specific transporters, and in particu- lar exhibit considerably elevated tight junction formation and bar- rier properties [12]. Highly expressed inter-endothelial tight junctions and adherens junctions, comprised of transmembrane proteins such as claudins [33–35], occludin [36, 37], junctional adhesion molecules [38], and vascular endothelial cadherin [39], working in conjunction with cytosolic scaffolding proteins such as zonula occludens [40–42], ensure that the paracellular pathway of the cerebrovascular endothelium is up to 100-fold “tighter” than that seen for endothelia in non-CNS vascular beds [43]. Indeed, while small lipid-soluble molecules (<500 Da) and blood gases necessary for brain metabolism can passively diffuse across the BBB endothelium, larger and potentially harmful macromolecules and pathogens within the circulation are typically denied entry [12, 44] (although it should be pointed out that large biologically important molecules such as glucose and amino acids may cross the BBB endothelium transcellularly by means of carrier-mediated transport systems [12, 24]). By virtue of its unique location and properties, the BBB endothelium is therefore a crucial participant in both normal brain homeostatic regulation and, in the event of its In vitro BBB Cell Models 73 dysfunction, neurodegenerative diseases [2, 45–51], stroke [52–54], and brain injury/edema [55], which typically manifest a reduction of inter-endothelial tight junction expression [56–58]. As such, the endothelium is central to any attempt to examine the BBB by in vitro means. In vitro BBB models are important tools that allow us to study fundamental processes within this unique and complex microvas- cular environment, frequently preceding and often complementing both animal and clinical studies [59–61]. Over the years, numerous experimental BBB endothelial models of varying degrees of com- plexity and applicability have been implemented [62, 63]. In this respect, the steady evolution of BBB models is of central thematic importance to this chapter as all of our knowledge of the BBB is based on correlation of in vitro, ex vivo, and in vivo models with human clinical studies. While in vivo modeling (e.g., mouse, rat, pig, bovine) of BBB physiology collectively captures all of the multicellular and hemodynamic aspects of the BBB, it is not with- out its limitations. For example, in vivo studies can often be time consuming and expensive, with several animals required to yield statistically valid data. They can also lack translational relevance to a human setting [64], while mounting pressure from ethical bodies demanding a reduction in animal experiments further highlights the growing need for more robust and relevant in vitro human cell screening tools. Progress with in vitro BBB models is therefore the main focus for this chapter. In vitro cellular models should model conditions that closely mirror those in the proceeding animal model [65]. While in vitro cell modeling has improved considerably in recent years [66–68], there is still arguably an unmet need for a more physiologically relevant and predictive in vitro human BBB model [69, 70]. The importance of using well-designed in vitro models means that pharmacologically promising drugs that lack effective BBB permeation could be either rapidly excluded from in vivo/clinical trials or further modified to improve their intrinsic permeability, thus reducing the timeline for successful CNS drug development [71, 72]. In the last three decades, this necessity has burgeoned the field of cerebrovascular research, although mimicking in vitro the physiology and functional complexity of the in vivo BBB still remains a challenging task. In addressing this, an ever-increasing understanding and appreciation of important factors (e.g., species [64], the neurovascular unit (NVU) [12], blood flow [18], and predictive in silico modeling [73, 74]) that influence the BBB have resulted in an impressive increase in the variety and translational value of in vitro BBB models. To shed more light on this specific topic, we focus on three key aspects that have helped to drive the evolution of in vitro BBB models: (1) the availability of suitable BBB endothelial cells; (2) a fuller understanding of the neurovas- cular unit (NVU); and (3) an appreciation of the relevance of blood flow shear to BBB physiology. 74 Keith D. Rochfort and Philip M. Cummins

1.1 Endothelial Cell The most common and widely used BBB model constitutes the Models BBB at its most fundamental level—the brain microvascular endo- thelial cells that form the inner surface of cerebral capillaries. High- resolution electron microscopy has clearly identified the cerebro- vascular endothelium as the site for traffic between the blood and brain tissue [75], a fact supported by experiments demonstrating a fully functioning BBB at just the endothelial level, without the assistance of other cell types [76]. Early BBB models employed epithelial cell lines such as Madin-Darby canine kidney (MDCK) cells [77, 78] and Caco-2 cells [69, 79, 80] that could act as pseudo-BBB models. While being of non-cerebral origin is an obvious limitation [81, 82], epithelial cells offer a robust model in their applicability to monolayer permeability studies owing largely to their highly upregulated tight junctions [80]. Subse- quently, advancement in cell isolation techniques since the 1990s (see Fig. 1) led to BBB models being established with primary- derived endothelial cells isolated from several mammalian species including, but not limited to, murine [83], rat [84], bovine [85], porcine [86], and eventually human. Primary culture of brain microvascular endothelial cells (e.g., from rodents) has the advantage of providing the closest pheno- typic approximation to the in vivo environment [87–91]. A major limitation of isolating cells from small animals is cell number (~1–- 2 million cells/brain), limiting the ability to perform meaningful numbers of experiments [92]. In contrast, bovine [85, 93–96] and porcine [97–99] models provide substantially more cells (>50 million cells/brain) while also mimicking many of the

Fig. 1 Evolution of in vitro BBB models. Note: Timelines are approximate In vitro BBB Cell Models 75 morphological and biochemical properties of the human BBB (although it is worth noting that unlike rodents, access to larger animals may be more restricted) [92]. As is the problem with many in vitro cell models, primary endothelial cultures obtained from animal tissue undergo morphological and functional changes ex vivo, with extended culturing and lack of proper stimuli leading to accelerated endothelial dedifferentiation and transient diminish- ment of BBB characteristics [69, 100]. Furthermore, factors such as user handling and age/condition of starting material have been shown to cause intra/inter-batch variation, issues that have con- tributed to variability in properties (e.g., transendothelial electrical resistance, TEER) of primary-derived BBB models across the liter- ature [101–103]. Technologies to immortalize BBB endothelial cell lines offer a solution to several of the aforementioned problems. Well- characterized immortal cerebral endothelial cell lines derived from rat (RBE4 [104, 105], GP8 [106], and GPNT [107]) and mouse (bEND.3 [108, 109], bEND.5 [110], cEND [111]) are now extensively used today. While the advantages of the immortalization process are evident (e.g., extended passage limits, improved batch- to-batch homogeneity, well-characterized barrier properties), doubts remain concerning the impact on endothelial barrier func- tion of the genetic manipulation process (e.g., retroviruses, poly- oma T-cell antigen [112]). It can also be noted that most models employing immortalized cells do not reach TEER values >500 (i.e., significantly less than in vivo values [76, 113]), although media formulation strategies may improve on this [98, 111, 114–118]). While nonhuman endothelial BBB models have clearly played a critical role in progressing our understanding of the BBB, the identification of species-specific differences has raised many ques- tions as to their suitability for preclinical investigations [61, 64, 119–121]. As a consequence, human brain microvascular endothe- lial cells (HBMvECs) are now widely recognized as being more translationally relevant for modeling BBB behavior. Early chal- lenges to HBMvEC availability included ethical barriers, as well as technical constraints associated with harvesting human cerebral biopsy tissue for primary cultures (e.g., cost/labor-intensive prepa- ration, significant batch-to-batch variability, and difficulties with obtaining sufficiently pure endothelial cultures that would retain critical barrier properties—namely, high TEER, low paracellular permeability, and functional polarization of key endothelial cell markers [122]). With improved understanding of primary culture behavior, and the refinement of culturing techniques, many of these obstacles were overcome, culminating in the commercial availabil- ity of primary-derived HBMvECs. It should be noted that there are distinct limitations with primary-derived HBMvECs in the absence of physiologically relevant humoral and hemodynamic stimuli 76 Keith D. Rochfort and Philip M. Cummins

(discussed below), with cells losing phenotypic definition after a short number of passages. Moreover, immortalized human cell lines [123–125] such as hCMEC/D3 (HBMvECs immortalized with hTERT and SV40 large T antigen [126]) have also been widely adapted by the scientific and industrial communities as a human BBB endothelial model due to their phenotypic stability and rigorous characterization [127–132], although transcriptomic profiling of hCMEC/D3 against primary brain endothelia has highlighted significant differences [112, 133, 134]. Efforts are therefore ongoing to develop more stable cell lines to improve upon these current limitations. Irrespective of the BBB cell type used, the simplest in vitro model for BBB permeability studies is cerebral endothelial mono- cultures seeded onto microporous transwell membranes, thus cre- ating distinct luminal and subliminal compartments analogous to the in vivo microvessel architecture [85, 135]. The next section will highlight the overall importance of the NVU concept to BBB physiology and examine how it has been practically incorporated into transwell models to improve their translational value.

1.2 The The NVU is comprised of five cell types that collectively form the Neurovascular BBB; endothelial cells, pericytes, astrocytes, neurons, and microglia Unit (NVU) [12, 136]. In order to capture the collective input of the NVU into BBB models, thereby ensuring more translationally accurate func- tional readouts (e.g., drug permeability, toxicity, behavior, and/or pharmacodynamics), in vivo (and ex vivo) BBB models from vari- ous species were initially favored. Aside from various nonhuman mammalian models (e.g., rodents), more recent approaches have implemented Drosophila [137] (although phylogenetic differences to mammals have limited translational impact). The relatively recent adoption of the zebrafish model surmounts these shortcom- ings however, demonstrating a superior sequence similarity for pertinent BBB transporters and receptors, as well as a high degree of anatomical and structural similarity to human BBB physiology [138–142], while also offering ease of genetic manipulation and cost-effectiveness [143]. Notwithstanding these approaches, in vivo models (even at the rodent level) are becoming increasingly more expensive, limiting their applicability for high-throughput screening. Moreover, a considerable proportion of results gener- ated from animal models still cannot be directly translated into human responses due to species-specific responses [64, 120, 121, 144, 145]. To overcome these issues, much research has focused on human BBB cell models that recapitulate the NVU environment in vitro—a major evolution in BBB model development. Evidence from animal studies has improved our understanding of the BBB. Several studies demonstrate that the cell types compris- ing the NVU contribute via paracrine cross talk to the unique phenotype of the BBB endothelium [146]. As such, identification In vitro BBB Cell Models 77 of these paracrine factors has become a key objective for improving in vitro BBB models. In this respect, incorporating NVU cell types into BBB transwell models has yielded valuable information. Indeed, proteomics has identified differences in proteins involved in cell structure, motility, and metabolism in BBB endothelial cells when cultured with glial cells [147, 148], underlining the impor- tance of cellular co-interactions within the NVU. Within brain capillaries, endothelial cells sit on the basal lamina composed of cell-derived matrices such as collagen IV, fibronectin, laminin, and proteoglycans [149], matrix proteins synthesized in vivo by glial cells, and which are often incorporated into in vitro models to coat growth surfaces, with demonstrable improvements to BBB phenotype [150]. Architecturally, this basal lamina shares a split basement membrane with pericytes, which have been shown to synthesize a large number of the aforemen- tioned matrices. Covering approximately 20–30% of the endothe- lium via their long processes [151], the location of pericytes along the exterior vessel surface has been shown to correlate with the “tightness” of endothelial junctions in their immediate vicinity [60, 152–154]. Aside from their role in matrix deposition, in recent years pericytes have been shown to play an increasingly integral role in BBB phenotype, influencing angiogenesis/vasculogenesis, trans- cellular transport, inflammatory responses, and BBB permeability [153–160]. Unlike other NVU cell types that influence the BBB through noncontact means, direct contact is critical for pericytes to exert their influence [153, 161]. In addition to pericytes, the external surface of capillaries is also encapsulated (>99%) by astrocytic foot processes [162], which, through a combination of glial signaling and physical obstruction, can restrict the passage of drugs and other agents across the BBB endothelium [163]. Astrocytes function as NVU scaffolds, guiding cells within their vicinity into place and ultimately dictating the development of their immediate vascular network [164–168]. Astrocytes may also modulate the BBB through the release of regulatory factors [169]. An early landmark study by DeBault et al. for example demonstrated how aspects of BBB phenotype (γ-glutamyl transpeptidase) could be enhanced by cul- turing endothelial cells in the presence of astrocytes [170]. Since then, there has been an increasing recognition of the influence of astrocytes on transcellular transport at the BBB, specifically with respect to efflux transporters [43, 171, 172]. Moreover, transwell co-culture models that facilitate “direct physical contact” between brain astrocytes and endothelial cells have demonstrated up to ninefold improvement in endothelial TEER values [173–175]in comparison to just twofold in “noncontact” co-cultures [176]. As such, models involving co-culture of BBB endothelials with glial cells and/or glial conditioned media are in widespread use today. 78 Keith D. Rochfort and Philip M. Cummins

Of note, the remaining NVU cell types, neurons, and microglia are not considered to be directly involved in the gross structural organization of the BBB, although they have been shown to influ- ence NVU function via regulation of a number of BBB functions, such as induced expression of enzymes [177] and enacting immu- nological responses, respectively [178–180]. While the presence of both these cells is indeed prominent in the NVU, and several studies acknowledge their behavior having consequence for barrier effectiveness [181, 182], their respective degree of influence toward BBB phenotype is significantly less characterized by com- parison to astrocytes and pericytes. Several approaches have been employed to recapitulate the complex environment of the NVU in vitro. Use of transwell inserts is by far the most popular approach, presenting a versatile and efficient means of co-culturing within a single well two or more NVU cell types deriving from either the same [183–185] or differ- ent species [186–188]. Typically, BBB endothelials are grown within the transwell insert, with a second NVU cell type cultured either within the base of the underlying well (noncontact co-culture) [161, 189] or on the immediate underside of the transwell insert (pseudo-contact co-culture) [161, 190]. Inclusion of cell-conditioned media into the lower well beneath the insert (i.e., in lieu of cells) has also been routinely employed with this transwell model [191, 192]. Using this approach, co-culture of BBB endothelial cells with astrocytes for example has been found to improve barrier properties of the former in part through astrocyte-dependent modulation of the expression of endothelial transporters, enzymes, and specialized transport and efflux systems [193]. Depending on the application, the complexity of the co-culture environment can be enhanced as required. Relative to monocultures for example, tri-cultures have demonstrated signifi- cant improvements with respect to barrier integrity [154, 194–196], but may be unnecessarily costly to set up and maintain for high-throughput drug transport studies, while condition-specific environments (e.g., ischemia) may command a quad-culture at minimum for useful modeling [197, 198]. Impor- tantly, transwell co-culture models are not without their limita- tions. The thickness (greater than the basal lamina) and pore density of the transwell inserts can restrict cell-cell contact [199]. - Co-culture models are also highly labor and cost intensive to estab- lish and maintain for meaningful, translatable BBB screening studies. Despite these limitations, the transwell approach remains the most practical and preferred approach to co-culture modeling, though 3D modeling is now becoming increasingly more popular (see below). Notwithstanding the innovative contributions that our knowl- edge of the NVU has made to the development of in vitro BBB models, a disadvantage of many current models has been their In vitro BBB Cell Models 79

“static” nature. An oft-overlooked aspect of BBB physiology when implementing an in vitro cell model is the influences of hemody- namic force (e.g., shear flow) on BBB phenotype.

1.3 Hemo- Most BBB models fail to incorporate the blood flow-associated dynamic Flow shear environment that the macrovascular (and microvascular) endothelium is constantly subjected to. Hemodynamic flow is a well-documented differentiating influence on endothelia [200–202] and has a corresponding critical modulatory role in BBB phenotype [18, 203–206]. Laminar shear stress, the frictional force generated by exposing the luminal surface of the endothelium to flow, activates several apical mechanosensors (e.g., integrins, caveolae, G-proteins, ion channels), which in turn transduce mechanical stimuli into cell signals [207–210]. A wealth of evi- dence exists on the mechanotransduction pathways that trigger phenotypic changes in endothelial cells in response to flow [209, 210]. With respect to morphology, BBB endothelial cells exposed to physiological shear become flatter and larger, and mani- fest increases in endocytotic vesicles, microfilaments, and clathrin- coated pits, thereby more closely resembling that of BBB ECs in vivo [202, 211]. BBB models incorporating shear stress report the lowest permeability to mannitol and sucrose tracers [212], comparable to the levels demonstrated in co-culture models [18, 213], clearly highlighting the critical role that laminar shear plays in promoting stable BBB phenotype. Early attempts to recreate the flow environment in cell culture models employed cone-and-plate viscometers [214, 215], a leading method in the initial phase of devices designed to model rheological stimuli on cultured endothelial cells. Customization of these instru- ments allowed researchers to examine the effects of different hemo- dynamic flow patterns on BBB endothelial cells, although they tend to be limited to monoculture models. More sophisticated approaches to incorporating fluid flow into a BBB co-culture envi- ronment have since been developed. Several purpose-built flow channel-orientated models, such as the parallel-plate flow chamber ® or the ibidi microfluidic slide (ibidi GmbH, Martinsried, Ger- many), are now routinely employed in studies spanning basic research to advanced drug discovery [216, 217]. More advanced 3D hollow fiber co-culture systems such as the DIV-BBB system (FloCell Inc., OH, USA) employ porous capillary bundles to model the three-dimensional architecture of brain capillaries. This system successfully merges the co-culture of endothelial cells with other NVU cell types under prolonged laminar flow (up to 5 weeks), and has been successfully employed in studies utilizing primary cultures and cell lines of both human and animal origin, with reported increases in TEER of up to tenfold as compared to static cultures [211, 218]. While the elaborate nature of these models offers obvious advantages over conventional static BBB systems, it should 80 Keith D. Rochfort and Philip M. Cummins

be noted that the multifactorial nature of these approaches, techni- cal skills, added costs, and time required to establish and maintain this perfused capillary system are significantly higher in comparison to simpler static models, thereby limiting assay throughput. An additional drawback specific to capillary systems includes the inabil- ity to easily view the intracapillary cell compartment to monitor cultures. Microfluidics perhaps represents the next key advancement in BBB modeling. BBB microfluidics has surged in recent years with similar aims to the DIV-BBB system, but reducing the amount of cellular materials, costs, and time involved. A number of these designs [175, 204, 219–221] have even been able to maintain the co-culture aspects and have improved upon the limitations of capil- lary systems by reducing the amount of porous dividing material between the different cultured cell types, increasing cell contacts, and more closely mimicking the neurovascular microenvironment. In conclusion, in vitro BBB models have evolved significantly over the last three decades, from static monocultures to dynamic multicellular systems. While it is still beyond our ability to recapit- ulate all of the complex functions of the BBB in one in vitro culture model, attempts to merge cell co-culture and biomechanical prin- ciples in order to reflect the complex contributions to neurovascu- lar homeostasis of the NVU and fluid shear stress, respectively, constitute major advances. Sects. 2 and 3 attempt to put this information into context by experimentally demonstrating two alternate means of applying shear stress to human brain microvas- cular endothelial cells (HBMvECs) in order to show how BBB phenotype is improved in response to flow.

2 Materials

2.1 Cell Culture Primary human brain microvascular endothelial cells (Cat# ACBRI-376, Cell Systems Corporation, WA, USA). 2.1.1 Cell Lines

2.1.2 Cell Culture Cell culture media in our experiments is Endo-GRO™-MV Com- Reagents plete Media (Cat# SCME004, Merck Millipore) comprising of 475 mL EndoGRO™ Basal Medium, 25 mL fetal bovine serum, 1 mL of EndoGRO-LS Supplement, 5 ng/mL recombinant human epidermal growth factor, 10 mM L-glutamine, 1 μg/mL hydrocortisone hemisuccinate, 0.75 U/mL heparin sulfate, and 50 μg/mL ascorbic acid. Other reagents include Attachment Fac- tor (Cat# S-006-100, Thermo Fisher); phosphate-buffered saline (Cat# D8537, Sigma Aldrich); trypsin-EDTA solution (Cat# T4174, Sigma Aldrich); recombinant human TNF-α (Cat# In vitro BBB Cell Models 81

GF023, Merck Millipore), to be reconstituted at 0.1 mg/mL in culture-grade H2O, allowed to sit at room temperature for 2 h, aliquoted, and stored at À20 C.

2.2 Immuno- 3.7% (w/v) Paraformaldehyde (Cat# P6148, Sigma Aldrich) dis- fluorescence solved in PBS: Heat the mixture to 60 C and adjust to pH 10 until the solution goes clear. Cool to room temperature and readjust to 2.2.1 Microscopy pH 7.0. Filter using a 0.2 μm membrane. Solution is stable for up Reagents to 1 month at room temperature, or for several months at À20 C. Other reagents include the following: Perma-block solution is 0.1% (w/v) saponin, 0.25% (v/v) fish gelatin, and 0.02% (w/v) sodium azide made up in PBS and stored at 4 C for up to 6 months; mouse anti-ZO-1 monoclonal antibody (Cat# 33-9100, Thermo Fisher Scientific), aliquoted and stored at À20 C; Donkey anti-Mouse IgG Highly Cross-Adsorbed Secondary Antibody with Alexa ® Fluor 488 (Cat# A-21202, Thermo Fisher Scientific), aliquoted ® and stored in the dark at 4 C; Alexa Fluor 546 Phalloidin (Cat# A-22283, Thermo Fisher Scientific), stored in the dark at À20 C; Fluorescent Mounting Media (Cat# S302380-2, Agilent Technologies).

2.2.2 Materials 0.13–0.16 mm Coverslips (RA 1.5) (Cat# 474030-9000, ZEISS); and Equipment Adhesion Slides 25 Â 75 Â 1 mm (Cat# 10219280, Thermo Fisher Scientific); Olaf Humidifying Chamber (Cat# 80008, ibidi); μ-slide I0.6 Leur (Cat# 80186, ibidi); Orbital Shaker (Cat# SSM1, Stuart Scientific); Zeiss 710 Confocal Microscope (or similar fluorescent microscope).

2.3 Flow Cytometry Dihydroethidium (Cat# 37291, Sigma Aldrich) and 20,70-dichlorofluorescin diacetate (Cat# D6883, Sigma Aldrich) 2.3.1 Flow Cytometry are prepared as needed by dissolving in high-molecular-grade Reagents DMSO and stored in the dark for up to 3 months at À20 C; flow cytometry buffer is comprised of 2% (v/v) fetal bovine serum, 0.05% (w/v) sodium azide, and 0.01% EDTA in PBS. The solution is 0.2 μM filtered and is stable for 6 months when stored at 4 C.

3 Methods

The aim of this section is to experimentally demonstrate two alter- nate means of applying physiological shear stress to cultured HBMvECs and to show how BBB phenotype is improved in response to flow using tight-junction zonula occludens-1 (ZO-1) immunolocalization as a reporter index. The ability of physiological shear to attenuate the ability of pro-inflammatory cytokines (e.g., TNF-α) to induce reactive oxygen species (ROS) in HBMvECs will also be demonstrated using flow cytometry. For clarity, an overview of the work scheme for this section is provided in Fig. 2. 82 Keith D. Rochfort and Philip M. Cummins

Fig. 2 Method section workflow. Overview of the methodology we employ to culture HBMvECs, expose them to shear stress, and subsequently monitor improvements in BBB phenotype using immunocytochemistry (to monitor locali- zation of tight-junction proteins to the cell-cell border) and flow cytometry (to demonstrate how fluid shear can attenuate TNF-α-induced ROS induction and oxidative stress)

3.1 Cell Culture (1) Pipette 2 mL of Attachment Factor™ solution to each P100 plate. Using a sterile cell scraper or spreader, gently coat the plate 3.1.1 Culture surface with Attachment Factor™, incubating for 30 min at 37 C. of HBMvECs Remove any excess Attachment Factor™ solution from the culture dishes before adding 6 mL of complete cell culture medium. (2) *Seed HBMvECs at 7000 cells/cm2 and grow to ~80% con- fluency. (3) Trypsinize the cells using trypsin:EDTA, minimizing the time of trypsin exposure. (4) Once a cell suspension is obtained and a cell count has been performed, adjust the volume of pre-warmed complete media containing the cells to obtain a cell suspension of 1 Â 106 cells/mL. Subculture the cells at an optimal seeding density to propagate further cultures, or resuspend in a cryo-preservation reagent for storage in liquid nitrogen. *Note: Seeding densities should be optimized for each cell system depend- ing on cell-doubling time, matrices, culture dish area, adherence properties, etc. In vitro BBB Cell Models 83

3.2 Application (1) In advance, 0.13–0.16 mm thick coverslips (refractive of Shear Stress index ¼ 1.5) are placed in 100% ethanol and sterilized by passing them, with the aid of a forceps, through a Bunsen flame. The 3.2.1 Application coverslips are placed into individual wells of a 6-well dish and of Shear Stress by Orbital exposed to ultraviolet light for 30 min. (2) Add 200 μL of Attach- Rotation ment Factor™ to each coverslip and gently coat the surface of the coverslip and entire culture well using a sterile cell scraper. Place the dish in the incubator for 30 min. (3) Trypsinize and prepare the cell culture suspension at 1 Â 106 cells/mL. Add 1 mL of cell suspen- sion to each well of the 6-well plate containing coverslips. Add a further 1 mL of complete media before returning the plate to the incubator. Leave overnight to allow cells to adhere. Once the cultures are ~100% confluent, proceed with the shear protocol. (4) Remove the media and replace with 4 mL of fresh complete cell culture medium. Affix the culture dish to an orbital rotator within a dedicated incubator. Set rotator to 160 rpm (equivalent to 8 dyn/cm2 of shear). (5) After 24 h, the slides in each well can be visualized by light microscopy to confirm cell morphological realignment in the direction of flow (see Fig. 3 left). *Note: The degree of shear stress applied in this model depends on a number of factors (see Note 1; see equation in Fig. 3, left).

® ® 3.2.2 Application (1) ibidi 0.6 Luer ibiTreat slides (Cat# 80186, ibidi , Martinsried, of Shear Stress by ibidi® Germany) were opened in the laminar and placed on a slide rack (or clean flat surface). (2) A mixture of Attachment Factor™ and complete cell culture media (2:1) is made. Gently pipette 150 μLof this mixture to the “upstream” reservoir, taking care not to intro- duce air bubbles to the channel. Allow the slide to equilibrate such that the channel fills with suspension and both reservoirs have an equal level. Place the reservoir caps back on the Luer openings and place the coated slides in the Olaf humidity chamber (or similar) before placing it in an incubator. (3) Trypsinize and prepare the cell culture suspension at 1 Â 106 cells/mL. (4) Collect the humidity chamber from the incubator and place the slides on a slide rack (or a flat, clean surface). (5) Add 150 μL of cell suspension to the “upstream” reservoir, taking care not to introduce air bubbles to the channel. Allow the slide to equilibrate such that the channel fills with cell suspension and both reservoirs have an equal level. (6) Draw 200 μL of cell suspension from the “downstream” reser- voir to perfuse the cell suspension through the channel and remove excess Attachment Factor. Add a further 200 μL of cell suspension to the “upstream” reservoir as previously. Allow the volumes in each reservoir to equilibrate before drawing 200 μL of cell suspen- sion from the other reservoir. (7) Repeat the previous step a further two times to ensure that the entire cell channel has a homogeneous cell distribution. (8) Place the caps back on the Luer openings before placing the slides into the humidity chamber. Place the chamber into the incubator and leave overnight to allow the cells 84 Keith D. Rochfort and Philip M. Cummins

Fig. 3 Applying shear stress to HBMvECs. In our laboratory, we sheared cultured endothelial cells using either (left) orbital rotation or (right) the ibidi® microfluidic slide system (ibidi GmbH, Martinsried, Germany). Equations for calculating shear rate are shown for either method. The shear-dependent improvement in localization of ZO-1 (green) to the cell-cell border with either method is also shown using immunocytochemi- cal staining (representative images on bottom left and right)

to adhere. (9) Cells in the slide channel should be adhered, display a healthy morphology, and be ~100% confluent. *The ibidi system was set up as per the manufacturer’s protocol, with the perfusion towers holding 12 mL of media each, before being housed in a dedicated incubator. (10) Pump is set at 10-s cycles with a ramp-up in flow rate of 0.5mL every 30 min until 13.3 mL/min was achieved (8 dyn/cm2). After 24-h flow, cells were visualized to confirm cell morphological realignment in the direction of flow In vitro BBB Cell Models 85

(see Fig. 3 right). *Note: In this particular setup, 15 cm of connec- tive tubing (ID 1.6 mm) was employed. Based on media viscosity, channel height, and tubing diameter, flow rate is automatically calculated by the software based on the requested shear rate (see Note 1; see equation in Fig. 3, right).

3.3 Fluorescent (1) Post-shear, culture media is aspirated and the coverslips in each Detection of ZO-1 well are gently washed twice in PBS (see Note 2). 1 mL of ice-cold 3.7% paraformaldehyde is added to each well and the dish placed on 3.3.1 Fluorescent ice for 10 min. (2) Post-fixation, coverslips are gently washed twice Labeling of Fixed Cell in PBS. 1 mL of a 50 mM ammonium chloride solution is then Cultures (Orbital Rotation) added to each well for 10 min to quench unreacted formaldehyde groups (background fluorescence source). (3) 1 mL of Perma- block solution is added to each well for 30 min to gently permea- bilize the cells and concurrently reduce the unspecific binding of antibodies in the latter steps. (4) In this example, ZO-1 indirect staining will be described with an F-actin direct counterstain also. To reduce reagent consumption, immunofluorescent labeling* is carried out by pipetting 100 μL of primary antibody (10 μg/mL) into the center of each well of a fresh 6-well dish and inverting the harvested coverslips (see Note 3), cell-side downward, onto the antisera. (5) To prevent the coverslips from drying out, dH2O was placed in any empty wells and/or gaps in the 6-well dish to maintain humidity. The dish was then incubated at 4 C overnight, following which coverslips are removed from their individual wells and placed cell-side upwards in a fresh well of a 6-well dish for rinsing in 1 mL of Perma-block solution. (6) FITC-conjugated secondary antibody (4 μg/mL) is made up in fresh Perma-block and 100 μL pipetted into the center of a 6-well dish well. Once again the coverslips are carefully removed from 6-well dishes using a forceps, and placed into the new 6-well dish containing secondary antibody, followed by incubation in the dark for 2 h. Post- incubation, coverslips are again removed from their individual wells and placed cell-side upwards in a fresh well of a 6-well dish for rinsing in 1 mL of Perma-block solution. (7) F-actin staining is often carried out as a direct counterstain to validate shear-induced changes in cytoskeletal realignment. F-actin phalloidin (0.13 μM) is prepared in Perma-block solution and applied to the coverslip as before for 30 min in the dark. (8) Prior to the end of incubation, 10 μL of DAKO mounting medium is added to a clean microscopy slide. After two more Perma-block washes, the coverslips are inverted cells-down onto the DAKO mounting medium to form a liquid-tight seal and slides set aside for 1 h in the dark to allow the mounting media to set. (9) The edges of the coverslip are sealed using a small amount of clear nail polish before they are placed in a slide box at 4 C until required (Sect. 3.3.3). *Note: Primary antibody dilutions are based on in-lab optimizations. 86 Keith D. Rochfort and Philip M. Cummins

® 3.3.2 Fluorescent (1) Due to the narrow channel of the ibidi slides, it is critical that Labeling of Fixed Cell the channel contains liquid to reduce the chances of having to Cultures (ibidi® Apparatus) remove air bubbles and stripping the slide clean of cells. Using the clips, clamp off the lines attached to the slide. Gently remove the lines and place the slides on a firm flat surface. (2) With the reservoirs full, remove 200 μL from the reservoir at the outflow end of the slide. Wash the slide five times by adding 200 μL of PBS to the upstream reservoir and removing 200 μL from the downstream reservoir. (3) 5 Â 200 μL of ice-cold 3.7% paraformaldehyde is added to the upstream reservoir, with subsequent removal of 200 μL from the downstream reservoir, and slides placed on a pre-chilled slide rack for 10 min. (4) 5 Â 200 μLof50mM ammonium chloride solution is added to the upstream reservoir, with subsequent removal of 200 μL from the downstream reservoir to quench unreacted formaldehyde groups. This reaction is left for 10 min at room temperature. (5) 5 Â 200 μL of Perma-block solution is added to the upstream reservoir, with subsequent removal of 200 μL from the downstream reservoir. Slides are left at room temperature for 30 min. (6) In this example, ZO-1 indirect staining will be detailed with an F-actin direct counterstain. (7) ZO-1 primary antibody* is made up in fresh Perma-block (10 μg/mL) and 5 Â 100 μL is pipetted into the upstream reser- voir, with subsequent removal of 100 μL from the downstream reservoir. To prevent the slides from drying out, the slides are relocated to the humidity chamber at 4 C overnight. (8) 5 Â 200 μL of Perma-block solution is added to the upstream reservoir, with subsequent removal of 200 μL from the downstream reservoir to rinse the channel and remove traces of unbound anti- body. (9) FITC-conjugated secondary antibody (4 μg/mL) was made up in fresh Perma-block and 5 Â 200 μL is pipetted into the upstream reservoir, with subsequent removal of 200 μL from the downstream reservoir. Slides are then placed in the dark and left to incubate for 2 h at room temperature. (10) Repeat step 8 to remove traces of unbound antibody. (11) F-actin phalloidin (0.13 μM) is diluted in Perma-block solution and applied to the coverslip as before for 30 min in the dark. (12) Repeat step 8 to remove traces of unbound phalloidin. (13) 3 Â 100 μL of DAKO mounting medium is added to the upstream reservoir and slides set aside for ~1 h in the dark to allow medium to set. The open reservoirs are then sealed with parafilm to prevent any potential liquid loss, and slides stored at 4 C(see Sects. 3.3.3). *Note: Primary antibody dilutions are based on in-house optimizations.

3.3.3 Fluorescent (1) Allow slides to equilibrate to room temperature. (2) Control Detection of ZO-1 by samples for immunofluorescent staining may include primary anti- Immunofluorescent body omission control sample is incubated with antibody diluent Microscopy not containing the primary antibody of interest. Control slide staining should be negligible and allow identification of In vitro BBB Cell Models 87

background fluorescence in the samples, isotype control—using immunoglobulin of the same isotype of your monoclonal primary antibody of interest—and tissue-type control—using a sample that definitively expresses the epitope of interest. (3) Using the appro- priate controls, the apparatus can be set up for detection of labeled ZO-1 and F-actin in endothelial cultures using excitation of 495 and 560 nm and emission filters at 520 and 570 nm, respec- tively. The beneficial effects of shear using either method on ZO-1 immunolocalization are shown in Fig. 3 (bottom left and right).

3.4 Measurement Refer to Sect. 3.2.1 on exposing culture to shear stress by orbital of ROS by Flow rotation. In applications such as immunofluorescence (above), Cytometry seeding cells onto coverslips facilitates relocating intact sheared cells ex situ for practical purposes. In applications such as flow cytometry (below), seeding cells onto coverslips is not necessary, thereby allowing the cells to be cultured and treated/sheared in situ in 6-well dishes. In this next section, we demonstrate how shear preconditioning of HBMvECs can reduce the oxidative stress induced by pro-inflammatory cytokines such as TNF-α (see Fig. 4). (1) Cells are sheared (8 dyn/cm2, 24 h) as per Sect. 3.2.1 to induce the shear phenotype, with parallel static controls employed. (2) TNF-α (100 ng/mL, 18 h) is added to both static and shear cultures, with vehicle controls also employed. (3) Proceed with either Sects. 3.4.1 or 3.4.2 to detect oxidative stress.

3.4.1 ROS Staining One of the two different dyes can be used for ROS staining (see Note 4). (1) Dihydroethidium (DHE) is a superoxide specific indicator. It colors the cytoplasm of living cells blue until it becomes oxidized in the presence of ROS to produce ethidium, which is free to intercalate with double-stranded DNA. (a) Add DHE at a final concentration of 3 μM into each culture well 30 min prior to completion of the 18-h TNF-α incubation time. (b) Proceed to Sect. 3.4.2. (2) 2070-Dichlorofluorescein diacetate (DCFDA) is a fluorogenic dye that measures ROS activity in a cell. Once taken up by the cell, cellular esterases cause the DCFDA to become deacetylated and subsequently oxidized upon interaction with ROS to create the highly fluorescent 20,70-dichlorofluorescein. (a) Add DCFDA at a final concentration of 5 μM to each culture well at the same time the TNF-α is added. The cultures are left to condition for 18 h. (b) Proceed to Sect. 3.4.2.

3.4.2 Preparing Cells (1) The culture media is removed and placed to one side in centri- for Flow Cytometry fuge tubes. Cells are trypsinized and added to their respective Analysis harvested culture media to analyze all events. (2) Cells are pelleted and washed in room-temperature PBS before being pelleted again. (3) Cell pellets are resuspended in 500 μL of flow buffer and transferred to sterile flow tubes. The samples are kept on ice and protected from light until read at the earliest possible moment. 88 Keith D. Rochfort and Philip M. Cummins

Fig. 4 Effect of shear preconditioning on cytokine-induced ROS generation in HBMvECs. Confluent cells were initially “preconditioned” for 24 h under static or shear conditions (0 or 8 dyn/cm2) prior to treatment of cells with TNF-α (100 ng/mL, 18 h). Static and shear conditions were also maintained during cytokine treatments. ROS production was subsequently monitored by flow cytometry using DCFDA (upper) and DHE (lower). Histograms (left hand side) represent the fold change in fluorescent signal normalized to untreated control. Representative flow cytometry scans (right-hand side) are also shown for cytokine treatments in the absence and presence of shear. *P  0.05 versus untreated controls. φP  0.05 versus cytokine without shear

(4) An unstained control culture is often employed to establish the appropriate gates for endothelial cell populations based on size and granularity. A positive control for ROS generation may also be employed by culturing a sample in the presence of 0.1 mM H2O2 for comparison. Using the appropriate controls the instrument can be set up for DHE- and DCFDA-stained HBMvEC cultures using excitation of 562 and 488 nm and emission filters at 610 and 520 nm, respectively. In vitro BBB Cell Models 89

4 Notes

1. In Sect. 3.2, it is critical to know the exact experimental para- meters in order to calculate the maximal exerted shear stress. While the majority of these parameters are selected/inherent of the device being employed, the viscosity of the solution used to apply shear stress should be determined and incorporated into the equation. Companies that sell complete culture media for- mulations often have this information and will provide it upon request. If you intend to use an in-lab formulation or are interested in fluid mechanic studies, whereby viscosity-altering compounds are being implemented, the resultant fluid viscosity can be determined using simple devices (e.g., viscometers) if available; otherwise the fluid should be sent for analysis. 2. For proper analyses of tight-junction proteins such as ZO-1, it is imperative that cell-cell junctions are maintained throughout the experimental procedure and sample processing. Experi- mental and processing conditions such as pH, temperature, and humidity can affect cellular monolayers to reduce cell-cell contact and create inter-endothelial gaps. This loss of cell-cell contact will affect subsequent staining for tight junction pro- teins and will ultimately confound results. 3. In Sect. 3.3.1, the experimental cultures on coverslips are routinely kept within the confines of a 6-well dish for the duration of the staining protocol. The use of this system creates individual compartments for each experimental sample, an easy means of transporting samples (if required), and an easy means of maintaining a humid environment. Alternately, the staining protocol has also been shown to work in other common sys- tems (e.g., the coverslips are placed on parafilm containing small volumes of immunofluorescence buffers). While these systems offer their own unique benefits such as reducing the scale of buffers used, they present their own unique limitations (e.g., increased physical contact with sample slide increasing the risk of damaging the sample during processing). 4. Both DHE and DCFDA can be utilized at the same concentra- tions for fluorescent microscopy analysis of ROS. Depending on the experimental conditions (e.g., application of shear stress/TNF-α treatment), and the platform utilized for micro- scopic evaluation (e.g., coverslip/ibidi™ slide), the staining protocol for each compound can be easily applied as detailed in Sect. 3.4.1, before preparing the slides for subsequent anal- ysis by microscopy. When using DHE, unreacted compound will stain the cytosol blue while the oxidized product (ethi- dium) intercalates with DNA promoting a correlative event to ROS in that the nuclei increasingly turn red. DCFDA in 90 Keith D. Rochfort and Philip M. Cummins

contrast is nonfluorescent until acetate groups are cleaved from the compound during oxidation, generating a highly fluores- cent and correlative product to ROS, 2070-dichlorofluorescein.

Acknowledgments

The authors are grateful for financial support from the National Development Plan/Higher Education Authority of Ireland Programme for Research in Third Level Institutes (HEA/PRTLI Cycle 4—T3 Targeted Therapeutics and Theranostics) and from Science Foundation Ireland (US-Ireland R&D Partnership Programme, Grant No. 14/US/B3116).

References

1. Bake S, Friedman JA, Sohrabji F (2009) 12. Abbott NJ, Patabendige AA, Dolman DE et al Reproductive age-related changes in the (2010) Structure and function of the blood- blood brain barrier: expression of IgG and brain barrier. Neurobiol Dis 37:13–25 tight junction proteins. Microvasc Res 13. Greenwood J, Heasman SJ, Alvarez JI et al 78:413–424 (2011) Leucocyte-endothelial cell crosstalk at 2. Farrall AJ, Wardlaw JM (2009) Blood-brain the blood-brain barrier: a prerequisite for suc- barrier: ageing and microvascular disease- cessful immune cell entry to the brain. Neu- systematic review and meta-analysis. Neuro- ropathol Appl Neurobiol 37:24–39 biol Aging 30:337–352 14. Obermeier B, Daneman R, Ransohoff RM 3. De Reuck JL (2012) Histopathological stain- (2013) Development, maintenance and dis- ings and definitions of vascular disruptions in ruption of the blood-brain barrier. Nat Med the elderly brain. Exp Gerontol 47:834–837 19:1584–1596 4. Montagne A, Barnes SR, Sweeney MD et al 15. Rochfort KD, Cummins PM (2015) The (2015) Blood-brain barrier breakdown in the blood-brain barrier endothelium: a target for aging human hippocampus. Neuron pro-inflammatory cytokines. Biochem Soc 85:296–302 Trans 43:702–706 5. Kola I, Landis J (2004) Can the pharmaceuti- 16. Shlosberg D, Benifla M, Kaufer D et al (2010) cal industry reduce attrition rates? Nat Rev Blood-brain barrier breakdown as a therapeu- Drug Discov 3:711–715 tic target in traumatic brain injury. Nat Rev 6. Alavijeh MS, Chishty M, Qaiser M et al Neurol 6:393–403 (2005) Drug metabolism and pharmacokinet- 17. Nag S, Kapadia A, Stewart DJ (2011) Review: ics, the blood-brain barrier, and central ner- molecular pathogenesis of blood-brain barrier vous system drug discovery. NeuroRx breakdown in acute brain injury. Neuropathol 2:554–571 Appl Neurobiol 37:3–23 7. Olesen J, Gustavsson A, Svensson M et al 18. Cucullo L, Hossain M, Puvenna V et al (2012) The economic cost of brain disorders (2011) The role of shear stress in blood- in Europe. Eur J Neurol 19:155–162 brain barrier endothelial physiology. BMC 8. Wimo A, Jonsson L, Bond J et al (2013) The Neurosci 12:1–15 worldwide economic impact of dementia 19. Oldendorf WH (1971) Blood brain barrier 2010. Alzheimers Dement 9:1–11 permeability to lactate. Eur Neurol 6:49–55 9. Ehrlich P (1885) Das Sauerstoff-Bedurfnis 20. Oldendorf WH (1971) Brain uptake of radi- des Organismus: eine farbenanalytische Stu- olabeled amino acids, amines, and hexoses die. Hirschward, after arterial injection. Am J Phys 10. Lewandowsky M (1900) Zur lehre der cere- 221:1629–1639 brospinal flussigkeit. Z Klin Med 40:480–494 21. Oldendorf WH, Hyman S, Braun L et al 11. Goldman EE (1913) Vitalf €arbung am zen- (1972) Blood-brain barrier: penetration of tralnervensystem. Abh Preuss Akad Wiss morphine, codeine, heroin, and methadone Phys Math Kl 1:1–60 after carotid injection. Science 178:984–986 In vitro BBB Cell Models 91

22. Schinkel AH (1999) P-glycoprotein, a gate- constitutes tight junction strands in endothe- keeper in the blood-brain barrier. Adv Drug lial cells. J Cell Biol 147:185–194 Deliv Rev 36:179–194 35. Ohtsuki S, Yamaguchi H, Katsukura Y et al 23. O’Kane RL, Hawkins RA (2003) Na+- (2008) mRNA expression levels of tight junc- dependent transport of large neutral amino tion protein genes in mouse brain capillary acids occurs at the abluminal membrane of endothelial cells highly purified by magnetic the blood-brain barrier. Am J Physiol Endo- cell sorting. J Neurochem 104:147–154 crinol Metab 285:E1167–E1173 36. Furuse M, Hirase T, Itoh M et al (1993) 24. Roberts LM, Black DS, Raman C et al (2008) Occludin: a novel integral membrane protein Subcellular localization of transporters along localizing at tight junctions. J Cell Biol the rat blood-brain barrier and blood- 123:1777–1788 cerebral-spinal fluid barrier by in vivo biotiny- 37. Balda MS, Flores-Maldonado C, Cereijido M lation. Neuroscience 155:423–438 et al (2000) Multiple domains of occludin are 25. Dutheil F, Jacob A, Dauchy S et al (2010) involved in the regulation of paracellular per- ABC transporters and cytochromes P450 in meability. J Cell Biochem 78:85–96 the human central nervous system: influence 38. Martin-Padura I, Lostaglio S, Schneemann M on brain pharmacokinetics and contribution et al (1998) Junctional adhesion molecule, a to neurodegenerative disorders. Expert Opin novel member of the immunoglobulin super- Drug Metab Toxicol 6:1161–1174 family that distributes at intercellular junc- 26. Ghosh C, Puvenna V, Gonzalez-Martinez J tions and modulates monocyte et al (2011) Blood-brain barrier P450 transmigration. J Cell Biol 142:117–127 enzymes and multidrug transporters in drug 39. Breier G, Breviario F, Caveda L et al (1996) resistance: a synergistic role in neurological Molecular cloning and expression of murine diseases. Curr Drug Metab 12:742–749 vascular endothelial-cadherin in early stage 27. Karande P, Trasatti JP, Chandra D (2015) development of cardiovascular system. Blood Chapter 4—Novel approaches for the delivery 87:630–641 of biologics to the central nervous system. In: 40. Huber JD, Egleton RD, Davis TP (2001) Singh M, Salnikova M (eds) Novel approaches Molecular physiology and pathophysiology and strategies for biologics, vaccines and can- of tight junctions in the blood-brain barrier. cer therapies. Academic Press, San Diego, CA, Trends Neurosci 24:719–725 pp 59–88 41. Vorbrodt AW, Dobrogowska DH (2003) 28. Berezowski V, Landry C, Lundquist S et al Molecular anatomy of intercellular junctions (2004) Transport screening of drug cocktails in brain endothelial and epithelial barriers: through an in vitro blood-brain barrier: is it a electron microscopist’s view. Brain Res Brain good strategy for increasing the throughput Res Rev 42:221–242 of the discovery pipeline? Pharm Res 42. Rochfort KD, Cummins PM (2015) 21:756–760 Cytokine-mediated dysregulation of zonula 29. Abbott NJ (2004) Prediction of blood-brain occludens-1 properties in human brain micro- barrier permeation in drug discovery from vascular endothelium. Microvasc Res in vivo, in vitro and in silico models. Drug 100:48–53 Discov Today Technol 1:407–416 43. Abbott NJ (2002) Astrocyte-endothelial 30. Talevi A, Bellera CL, Di Ianni M et al (2012) interactions and blood-brain barrier perme- CNS drug development—lost in translation? ability. J Anat 200:629–638 Mini Rev Med Chem 12:959–970 44. Pardridge WM (2003) Blood-brain barrier 31. Arrowsmith J, Miller P (2013) Trial watch: drug targeting: the future of brain drug devel- phase II and phase III attrition rates opment. Mol Interv 3:90–105 2011–2012. Nat Rev Drug Discov 12:569 45. Poduslo JF, Curran GL, Wengenack TM et al 32. Keaney J, Campbell M (2015) The dynamic (2001) Permeability of proteins at the blood- blood-brain barrier. FEBS J 282:4067–4079 brain barrier in the normal adult mouse and 33. Furuse M, Fujita K, Hiiragi T et al (1998) double transgenic mouse model of Alzhei- Claudin-1 and -2: novel integral membrane mer’s disease. Neurobiol Dis 8:555–567 proteins localizing at tight junctions with no 46. Ujiie M, Dickstein DL, Carlow DA et al sequence similarity to occludin. J Cell Biol (2003) Blood-brain barrier permeability pre- 141:1539–1550 cedes senile plaque formation in an Alzhei- 34. Morita K, Sasaki H, Furuse M et al (1999) mer disease model. Microcirculation 10: Endothelial claudin: claudin-5/TMVCF 463–470 92 Keith D. Rochfort and Philip M. Cummins

47. Friedman A, Kaufer D, Heinemann U (2009) brain microvascular endothelial cell model. Blood-brain barrier breakdown-inducing Cell Microbiol 19:e12664 astrocytic transformation: novel targets for 59. Summerfield SG, Lucas AJ, Porter RA et al the prevention of epilepsy. Epilepsy Res (2008) Toward an improved prediction of 85:142–149 human in vivo brain penetration. Xenobiotica 48. Bennett J, Basivireddy J, Kollar A et al (2010) 38:1518–1535 Blood-brain barrier disruption and enhanced 60. Nakagawa S, Deli MA, Kawaguchi H et al vascular permeability in the multiple sclerosis (2009) A new blood-brain barrier model model EAE. J Neuroimmunol 229:180–191 using primary rat brain endothelial cells, peri- 49. Nishitsuji K, Hosono T, Nakamura T et al cytes and astrocytes. Neurochem Int (2011) Apolipoprotein E regulates the integ- 54:253–263 rity of tight junctions in an isoform- 61. Lacombe O, Videau O, Chevillon D et al dependent manner in an in vitro blood-brain (2011) In vitro primary human and animal barrier model. J Biol Chem cell-based blood-brain barrier models as a 286:17536–17542 screening tool in drug discovery. Mol Pharm 50. Kim DW, Moon Y, Gee Noh H et al (2011) 8:651–663 Blood-brain barrier disruption is involved in 62. Ogunshola OO (2011) In vitro modeling of seizure and hemianopsia in nonketotic hyper- the blood-brain barrier: simplicity versus glycemia. Neurologist 17:164–166 complexity. Curr Pharm Des 17:2755–2761 51. Pfeiffer F, Schafer J, Lyck R et al (2011) 63. Naik P, Cucullo L (2012) In vitro blood-brain Claudin-1 induced sealing of blood-brain bar- barrier models: current and perspective tech- rier tight junctions ameliorates chronic exper- nologies. J Pharm Sci 101:1337–1354 imental autoimmune encephalomyelitis. Acta 64. Perel P, Roberts I, Sena E et al (2007) Com- Neuropathol 122:601–614 parison of treatment effects between animal 52. Huang J, Upadhyay UM, Tamargo RJ (2006) experiments and clinical trials: systematic Inflammation in stroke and focal cerebral review. BMJ 334:197–202 ischemia. Surg Neurol 66:232–245 65. Reichel A (2006) The role of blood-brain 53. Sandoval KE, Witt KA (2008) Blood-brain barrier studies in the pharmaceutical industry. barrier tight junction permeability and ische- Curr Drug Metab 7:183–203 mic stroke. Neurobiol Dis 32:200–219 66. Cecchelli R, Dehouck B, Descamps L et al 54. Knowland D, Arac A, Sekiguchi KJ et al (1999) In vitro model for evaluating drug (2014) Stepwise recruitment of transcellular transport across the blood-brain barrier. Adv and paracellular pathways underlies blood- Drug Deliv Rev 36:165–178 brain barrier breakdown in stroke. Neuron 67. Mabondzo A, Bottlaender M, Guyot AC et al 82:603–617 (2010) Validation of in vitro cell-based 55. Stamatovic SM, Dimitrijevic OB, Keep RF human blood-brain barrier model using clini- et al (2006) Inflammation and brain edema: cal positron emission tomography radioli- new insights into the role of chemokines and gands to predict in vivo human brain their receptors. Acta Neurochir Suppl penetration. Mol Pharm 7:1805–1815 96:444–450 68. Cecchelli R, Aday S, Sevin E et al (2014) A 56. Rochfort KD, Collins LE, Murphy RP et al stable and reproducible human blood-brain (2014) Downregulation of blood-brain bar- barrier model derived from hematopoietic rier phenotype by proinflammatory cytokines stem cells. PLoS One 9:e99733 involves NADPH oxidase-dependent ROS 69. Garberg P, Ball M, Borg N et al (2005) In generation: consequences for interendothelial vitro models for the blood-brain barrier. Tox- adherens and tight junctions. PLoS One 9: icol In Vitro 19:299–334 e101815 70. Avdeef A (2011) How well can in vitro brain 57. Rochfort KD, Collins LE, McLoughlin A et al microcapillary endothelial cell models predict (2016) Tumour necrosis factor-alpha- rodent in vivo blood-brain barrier permeabil- mediated disruption of cerebrovascular endo- ity? Eur J Pharm Sci 43:109–124 thelial barrier integrity in vitro involves the production of proinflammatory interleukin- 71. Summerfield SG, Read K, Begley DJ et al 6. J Neurochem 136:564–572 (2007) Central nervous system drug disposi- tion: the relationship between in situ brain 58. McLoughlin A, Rochfort KD, McDonnell CJ permeability and brain free fraction. J Phar- et al (2017) Staphylococcus aureus-mediated macol Exp Ther 322:205–213 blood-brain barrier injury: an in vitro human In vitro BBB Cell Models 93

72. Potschka H (2010) Targeting regulation of 85. Bowman PD, Ennis SR, Rarey KE et al (1983) ABC efflux transporters in brain diseases: a Brain microvessel endothelial cells in tissue novel therapeutic approach. Pharmacol Ther culture: a model for study of blood-brain bar- 125:118–127 rier permeability. Ann Neurol 14:396–402 73. Rebitzer S, Annibali D, Kopp S et al (2003) In 86. Patabendige A, Skinner RA, Abbott NJ silico screening with benzofurane- and (2013) Establishment of a simplified in vitro benzopyrane-type MDR-modulators. Farm- porcine blood-brain barrier model with high aco 58:185–191 transendothelial electrical resistance. Brain 74. Ecker GF, Noe CR (2004) In silico prediction Res 1521:1–15 models for blood-brain barrier permeation. 87. Joo F, Karnushina I (1973) A procedure for Curr Med Chem 11:1617–1628 the isolation of capillaries from rat brain. 75. Reese TS, Karnovsky MJ (1967) Fine struc- Cytobios 8:41–48 tural localization of a blood-brain barrier to 88. Panula P, Joo F, Rechardt L (1978) Evidence exogenous peroxidase. J Cell Biol for the presence of viable endothelial cells in 34:207–217 cultures derived from dissociated rat brain. 76. Crone C, Olesen SP (1982) Electrical resis- Experientia 34:95–97 tance of brain microvascular endothelium. 89. DeBault LE, Kahn LE, Frommes SP et al Brain Res 241:49–55 (1979) Cerebral microvessels and derived 77. Wang Q, Rager JD, Weinstein K et al (2005) cells in tissue culture: isolation and prelimi- Evaluation of the MDR-MDCK cell line as a nary characterization. In Vitro 15:473–487 permeability screen for the blood-brain bar- 90. Markovac J, Goldstein GW (1988) Trans- rier. Int J Pharm 288:349–359 forming growth factor beta activates protein 78. Nazer B, Hong S, Selkoe DJ (2008) LRP kinase C in microvessels isolated from imma- promotes endocytosis and degradation, but ture rat brain. Biochem Biophys Res Com- not transcytosis, of the amyloid-beta peptide mun 150:575–582 in a blood-brain barrier in vitro model. Neu- 91. Silbergeld DL, Ali-Osman F (1991) Isolation robiol Dis 30:94–102 and characterization of microvessels from nor- 79. Zhu ZB, Makhija SK, Lu B et al (2004) Trans- mal brain and brain tumors. J Neuro-Oncol port across a polarized monolayer of Caco- 11:49–55 2 cells by transferrin receptor-mediated ade- 92. Deli MA, Abraham CS, Kataoka Y et al (2005) novirus transcytosis. Virology 325:116–128 Permeability studies on in vitro blood-brain 80. Hellinger E, Veszelka S, Toth AE et al (2012) barrier models: physiology, pathology, and Comparison of brain capillary endothelial cell- pharmacology. Cell Mol Neurobiol based and epithelial (MDCK-MDR1, Caco-2, 25:59–127 and VB-Caco-2) cell-based surrogate blood- 93. Audus KL, Borchardt RT (1986) Characteri- brain barrier penetration models. Eur J Pharm zation of an in vitro blood-brain barrier Biopharm 82:340–351 model system for studying drug transport 81. Lohmann C, Huwel S, Galla HJ (2002) Pre- and metabolism. Pharm Res 3:81–87 dicting blood-brain barrier permeability of 94. Audus KL, Borchardt RT (1987) Bovine drugs: evaluation of different in vitro assays. brain microvessel endothelial cell monolayers J Drug Target 10:263–276 as a model system for the blood-brain barrier. 82. Lundquist S, Renftel M, Brillault J et al Ann N Y Acad Sci 507:9–18 (2002) Prediction of drug transport through 95. Audus KL, Ng L, Wang W, Borchardt RT the blood-brain barrier in vivo: a comparison (1996) Brain microvessel endothelial cell cul- between two in vitro cell models. Pharm Res ture systems. Pharm Biotechnol 8:239–258 19:976–981 96. Audus KL, Rose JM, Wang W et al (1998) 83. Coisne C, Dehouck L, Faveeuw C et al (2005) Brain microvessel endothelial cell culture sys- Mouse syngenic in vitro blood-brain barrier tems. In: Pardridge WM (ed) An introduction model: a new tool to examine inflammatory to the blood-brain barrier: methodology and events in cerebral endothelium. Lab Investig biology. Cambridge University Press, 85:734–746 New York, pp 86–93 84. Watson PM, Paterson JC, Thom G et al 97. Tewes B, Franke H, Hellwig S et al (1997) (2013) Modelling the endothelial blood- Preparation of endothelial cells in primary CNS barriers: a method for the production cultures obtained from 6 month old pigs. In: of robust in vitro models of the rat blood- de Boer AG, Sutanto W (eds) Transport brain barrier and blood-spinal cord barrier. across the blood brain barrier: in vitro and BMC Neurosci 14:59 94 Keith D. Rochfort and Philip M. Cummins

in vivo techniques. Harwood Academic, 109. Williams RL, Risau W, Zerwes HG et al Amsterdam, pp 91–97 (1989) Endothelioma cells expressing the 98. Franke H, Galla HJ, Beuckmann CT (1999) polyoma middle T oncogene induce heman- An improved low-permeability in vitro-model giomas by host cell recruitment. Cell of the blood-brain barrier: transport studies 57:1053–1063 on retinoids, sucrose, haloperidol, caffeine 110. Wagner EF, Risau W (1994) Oncogenes in and mannitol. Brain Res 818:65–71 the study of endothelial cell growth and dif- 99. Franke H, Galla H, Beuckmann CT (2000) ferentiation. Semin Cancer Biol 5:137–145 Primary cultures of brain microvessel endo- 111. Burek M, Salvador E, Forster CY (2012) thelial cells: a valid and flexible model to Generation of an immortalized murine brain study drug transport through the blood- microvascular endothelial cell line as an brain barrier in vitro. Brain Res Brain Res in vitro blood brain barrier model. J Vis Exp Protoc 5:248–256 66:e4022 100. Wolburg H, Neuhaus J, Kniesel U et al 112. Urich E, Lazic SE, Molnos J et al (2012) (1994) Modulation of tight junction struc- Transcriptional profiling of human brain ture in blood-brain barrier endothelial cells. endothelial cells reveals key properties crucial Effects of tissue culture, second messengers for predictive in vitro blood-brain barrier and cocultured astrocytes. J Cell Sci 107 models. PLoS One 7:e38149 (Pt 5):1347–1357 113. Butt AM, Jones HC (1992) Effect of hista- 101. Claudio L, Raine CS, Brosnan CF (1995) mine and antagonists on electrical resistance Evidence of persistent blood-brain barrier across the blood-brain barrier in rat brain- abnormalities in chronic-progressive multiple surface microvessels. Brain Res 569:100–105 sclerosis. Acta Neuropathol 90:228–238 114. Hoheisel D, Nitz T, Franke H et al (1998) 102. Claudio L (1996) Ultrastructural features of Hydrocortisone reinforces the blood-brain the blood-brain barrier in biopsy tissue from barrier properties in a serum free cell culture Alzheimer’s disease patients. Acta Neuro- system. Biochem Biophys Res Commun pathol 91:6–14 244:312–316 103. Najjar S, Pearlman DM, Devinsky O et al 115. Weidenfeller C, Schrot S, Zozulya A et al (2013) Neurovascular unit dysfunction with (2005) Murine brain capillary endothelial blood-brain barrier hyperpermeability contri- cells exhibit improved barrier properties butes to major depressive disorder: a review of under the influence of hydrocortisone. Brain clinical and experimental evidence. J Neuroin- Res 1053:162–174 flammation 10:906 116. Kroll S, El-Gindi J, Thanabalasundaram G 104. Begley DJ, Lechardeur D, Chen ZD et al et al (2009) Control of the blood-brain bar- (1996) Functional expression of rier by glucocorticoids and the cells of the P-glycoprotein in an immortalised cell line of neurovascular unit. Ann N Y Acad Sci rat brain endothelial cells, RBE4. J Neuro- 1165:228–239 chem 67:988–995 117. Paolinelli R, Corada M, Ferrarini L et al 105. Roux F, Durieu-Trautmann O, Chaverot N (2013) Wnt activation of immortalized brain et al (1994) Regulation of gamma-glutamyl endothelial cells as a tool for generating a transpeptidase and alkaline phosphatase activ- standardized model of the blood brain barrier ities in immortalized rat brain microvessel in vitro. PLoS One 8:e70233 endothelial cells. J Cell Physiol 159:101–113 118. Mizee MR, Wooldrik D, Lakeman KA et al 106. Greenwood J, Pryce G, Devine L et al (1996) (2013) Retinoic acid induces blood-brain SV40 large T immortalised cell lines of the rat barrier development. J Neurosci blood-brain and blood-retinal barriers retain 33:1660–1671 their phenotypic and immunological charac- 119. Warren MS, Zerangue N, Woodford K et al teristics. J Neuroimmunol 71:51–63 (2009) Comparative gene expression profiles 107. Regina A, Romero IA, Greenwood J et al of ABC transporters in brain microvessel (1999) Dexamethasone regulation of endothelial cells and brain in five species P-glycoprotein activity in an immortalized including human. Pharmacol Res rat brain endothelial cell line, GPNT. J Neu- 59:404–413 rochem 73:1954–1963 120. Uchida Y, Tachikawa M, Obuchi W et al 108. Williams RL, Courtneidge SA, Wagner EF (2013) A study protocol for quantitative tar- (1988) Embryonic lethalities and endothelial geted absolute proteomics (QTAP) by tumors in chimeric mice expressing polyoma LC-MS/MS: application for inter-strain dif- virus middle T oncogene. Cell 52:121–131 ferences in protein expression levels of In vitro BBB Cell Models 95

transporters, receptors, claudin-5, and marker model for drug transport studies. J Neuro- proteins at the blood-brain barrier in ddY, chem 107:1358–1368 FVB, and C57BL/6J mice. Fluids Barriers 132. Fischer S, Nishio M, Peters SC et al (2009) CNS 10:21 Signaling mechanism of extracellular RNA in 121. Hoshi Y, Uchida Y, Tachikawa M et al (2013) endothelial cells. FASEB J 23:2100–2109 Quantitative atlas of blood-brain barrier 133. Ohtsuki S, Ikeda C, Uchida Y et al (2013) transporters, receptors, and tight junction Quantitative targeted absolute proteomic proteins in rats and common marmoset. J analysis of transporters, receptors and junc- Pharm Sci 102:3343–3355 tion proteins for validation of human cerebral 122. Avdeef A, Deli MA, Neuhaus W (2015) In microvascular endothelial cell line hCMEC/ vitro assays for assessing BBB permeability: D3 as a human blood-brain barrier model. artificial membrane and cell culture models. Mol Pharm 10:289–296 In: Di L, Kerns EH (eds) Blood-brain barrier 134. Eigenmann DE, Xue G, Kim KS et al (2013) in drug discovery: optimising brain exposure Comparative study of four immortalized of CNS drugs and minimizing brain side human brain capillary endothelial cell lines, effects for peripheral drugs. Wiley, hCMEC/D3, hBMEC, TY10, and BB19, New York, pp 188–223 and optimization of culture conditions, for 123. Durieu-Trautmann O, Foignant-Chaverot N, an in vitro blood-brain barrier model for Perdomo J et al (1991) Immortalization of drug permeability studies. Fluids Barriers brain capillary endothelial cells with mainte- CNS 10:33 nance of structural characteristics of the 135. Brendel K, Meezan E, Carlson EC (1974) blood-brain barrier endothelium. In Vitro Isolated brain microvessels: a purified, meta- Cell Dev Biol 27A:771–778 bolically active preparation from bovine cere- 124. Muruganandam A, Herx LM, Monette R et al bral cortex. Science 185:953–955 (1997) Development of immortalized human 136. Hawkins BT, Davis TP (2005) The blood- cerebromicrovascular endothelial cell line as brain barrier/neurovascular unit in health an in vitro model of the human blood-brain and disease. Pharmacol Rev 57:173–185 barrier. FASEB J 11:1187–1197 137. DeSalvo MK, Mayer N, Mayer F et al (2011) 125. Stins MF, Badger J, Sik Kim K (2001) Bacte- Physiologic and anatomic characterization of rial invasion and transcytosis in transfected the brain surface glia barrier of Drosophila. human brain microvascular endothelial cells. Glia 59:1322–1340 Microb Pathog 30:19–28 138. Abdelilah-Seyfried S (2010) Claudin-5a in 126. Weksler BB, Subileau EA, Perriere N et al developing zebrafish brain barriers: another (2005) Blood-brain barrier-specific proper- brick in the wall. BioEssays 32:768–776 ties of a human adult brain endothelial cell 139. Eliceiri BP, Gonzalez AM, Baird A (2011) line. FASEB J 19:1872–1874 Zebrafish model of the blood-brain barrier: 127. Schreibelt G, Kooij G, Reijerkerk A et al morphological and permeability studies. (2007) Reactive oxygen species alter brain Methods Mol Biol 686:371–378 endothelial tight junction dynamics via 140. Umans RA, Taylor MR (2012) Zebrafish as a RhoA, PI3 kinase, and PKB signaling. model to study drug transporters at the FASEB J 21:3666–3676 blood-brain barrier. Clin Pharmacol Ther 128. Wilhelm I, Farkas AE, Nagyoszi P et al (2007) 92:567–570 Regulation of cerebral endothelial cell mor- 141. Watanabe K, Nishimura Y, Nomoto T et al phology by extracellular calcium. Phys Med (2012) In vivo assessment of the permeability Biol 52:6261–6274 of the blood-brain barrier and blood-retinal 129. Lim JC, Kania KD, Wijesuriya H et al (2008) barrier to fluorescent indoline derivatives in Activation of beta-catenin signalling by zebrafish. BMC Neurosci 13:101 GSK-3 inhibition increases p-glycoprotein 142. Fleming A, Diekmann H, Goldsmith P expression in brain endothelial cells. J Neuro- (2013) Functional characterisation of the chem 106:1855–1865 maturation of the blood-brain barrier in larval 130. Wilhelm I, Nagyoszi P, Farkas AE et al (2008) zebrafish. PLoS One 8:e77548 Hyperosmotic stress induces Axl activation 143. Stewart AM, Braubach O, Spitsbergen J et al and cleavage in cerebral endothelial cells. J (2014) Zebrafish models for translational Neurochem 107:116–126 neuroscience research: from tank to bedside. 131. Poller B, Gutmann H, Krahenbuhl S et al Trends Neurosci 37:264–278 (2008) The human brain endothelial cell line 144. Syvanen S, Lindhe O, Palner M et al (2009) hCMEC/D3 as a human blood-brain barrier Species differences in blood-brain barrier 96 Keith D. Rochfort and Philip M. Cummins

transport of three positron emission tomog- associated proteins in an in vitro model of raphy radioligands with emphasis on the BBB. Brain Res 1018:1–9 P-glycoprotein transport. Drug Metab Dis- 156. Lai CH, Kuo KH (2005) The critical compo- pos 37:635–643 nent to establish in vitro BBB model: Peri- 145. Ito K, Uchida Y, Ohtsuki S et al (2011) cyte. Brain Res Brain Res Rev 50:258–265 Quantitative membrane protein expression 157. Dore-Duffy P (2008) Pericytes: pluripotent at the blood-brain barrier of adult and youn- cells of the blood brain barrier. Curr Pharm ger cynomolgus monkeys. J Pharm Sci Des 14:1581–1593 100:3939–3950 158. Daneman R, Zhou L, Kebede AA et al (2010) 146. Engelhardt B, Liebner S (2014) Novel Pericytes are required for BBB integrity dur- insights into the development and mainte- ing embryogenesis. Nature 468:562–566 nance of the blood-brain barrier. Cell Tissue 159. Hellstrom M, Gerhardt H, Kalen M et al Res 355:687–699 (2001) Lack of pericytes leads to endothelial 147. Pottiez G, Duban-Deweer S, Deracinois B hyperplasia and abnormal vascular morpho- et al (2011) A differential proteomic approach genesis. J Cell Biol 153:543–553 identifies structural and functional compo- 160. Armulik A, Genove G, Mae M et al (2010) nents that contribute to the differentiation Pericytes regulate the blood-brain barrier. of brain capillary endothelial cells. J Proteome Nature 468:557–561 75:628–641 161. Hayashi K, Nakao S, Nakaoke R et al (2004) 148. Deracinois B, Pottiez G, Chafey P et al (2013) Effects of hypoxia on endothelial/pericytic Glial-cell-mediated re-induction of the co-culture model of the blood-brain barrier. blood-brain barrier phenotype in brain capil- Regul Pept 123:77–83 lary endothelial cells: a differential gel electro- phoresis study. Proteomics 13:1185–1199 162. Kacem K, Lacombe P, Seylaz J et al (1998) Structural organization of the perivascular 149. Yousif LF, Di Russo J, Sorokin L (2013) astrocyte endfeet and their relationship with Laminin isoforms in endothelial and perivas- the endothelial glucose transporter: a confo- cular basement membranes. Cell Adhes Migr cal microscopy study. Glia 23:1–10 7:101–110 163. Hamm S, Dehouck B, Kraus J et al (2004) 150. Hartmann C, Zozulya A, Wegener J, Galla HJ Astrocyte mediated modulation of blood- (2007) The impact of glia-derived extracellu- brain barrier permeability does not correlate lar matrices on the barrier function of cerebral in vitro with a loss of tight junction proteins from the endothelial cells: an study. Exp Cell cellular contacts. Cell Tissue Res Res 313:1318–1325 315:157–166 151. Kim JA, Tran ND, Li Z et al (2006) Brain 164. Tao-Cheng JH, Nagy Z, Brightman MW endothelial hemostasis regulation by peri- (1987) Tight junctions of brain endothelium cytes. J Cereb Blood Flow Metab in vitro are enhanced by astroglia. J Neurosci 26:209–217 7:3293–3299 152. Hori S, Ohtsuki S, Hosoya K et al (2004) A 165. Laterra J, Guerin C, Goldstein GW (1990) pericyte-derived angiopoietin-1 multimeric Astrocytes induce neural microvascular endo- complex induces occludin gene expression in thelial cells to form capillary-like structures brain capillary endothelial cells through Tie-2 in vitro. J Cell Physiol 144:204–215 activation in vitro. J Neurochem 89:503–513 166. Webersinke G, Bauer H, Amberger A et al 153. Dohgu S, Takata F, Yamauchi A et al (2005) (1992) Comparison of gene expression of Brain pericytes contribute to the induction extracellular matrix molecules in brain micro- and up-regulation of blood-brain barrier vascular endothelial cells and astrocytes. Bio- functions through transforming growth chem Biophys Res Commun 189:877–884 factor-beta production. Brain Res 1038:208–215 167. Abbott NJ, Ronnback L, Hansson E (2006) Astrocyte-endothelial interactions at the 154. Nakagawa S, Deli MA, Nakao S et al (2007) blood-brain barrier. Nat Rev Neurosci Pericytes from brain microvessels strengthen 7:41–53 the barrier integrity in primary cultures of rat brain endothelial cells. Cell Mol Neurobiol 168. Taber KH, Hurley RA (2008) Astroglia: not 27:687–694 just glue. J Neuropsychiatry Clin Neurosci 20:124–129 155. Berezowski V, Landry C, Dehouck MP et al (2004) Contribution of glial cells and peri- 169. Liberto CM, Albrecht PJ, Herx LM et al cytes to the mRNA profiles of (2004) Pro-regenerative properties of P-glycoprotein and multidrug resistance- cytokine-activated astrocytes. J Neurochem 89:1092–1100 In vitro BBB Cell Models 97

170. DeBault LE, Cancilla PA (1980) Gamma- 1 regulation of blood-brain barrier perme- glutamyl transpeptidase in isolated brain ability. J Cereb Blood Flow Metab 25: endothelial cells: induction by glial cells 593–606 in vitro. Science 207:653–655 184. Veszelka S, Pasztoi M, Farkas AE et al (2007) 171. Davson H, Oldendorf WH (1967) Sympo- Pentosan polysulfate protects brain endothe- sium on membrane transport. Transport in lial cells against bacterial lipopolysaccharide- the central nervous system. Proc R Soc Med induced damages. Neurochem Int 60:326–329 50:219–228 172. Hori S, Ohtsuki S, Tachikawa M et al (2004) 185. Hutamekalin P, Farkas AE, Orbok A et al Functional expression of rat ABCG2 on the (2008) Effect of nicotine and polyaromtic luminal side of brain capillaries and its hydrocarbons on cerebral endothelial cells. enhancement by astrocyte-derived soluble Cell Biol Int 32:198–209 factor(s). J Neurochem 90:526–536 186. Raub TJ (1996) Signal transduction and glial 173. Gaillard PJ, Voorwinden LH, Nielsen JL et al cell modulation of cultured brain microvessel (2001) Establishment and functional charac- endothelial cell tight junctions. Am J Phys terization of an in vitro model of the blood–- 271:C495–C503 brain barrier, comprising a co-culture of brain 187. Zysk G, Schneider-Wald BK, Hwang JH et al capillary endothelial cells and astrocytes. Eur J (2001) Pneumolysin is the main inducer of Pharm Sci 12:215–222 cytotoxicity to brain microvascular endothe- 174. Jeliazkova-Mecheva VV, Bobilya DJ (2003) A lial cells caused by Streptococcus pneumoniae. porcine astrocyte/endothelial cell co-culture Infect Immun 69:845–852 model of the blood-brain barrier. Brain Res 188. Smith M, Omidi Y, Gumbleton M (2007) Brain Res Protoc 12:91–98 Primary porcine brain microvascular endothe- 175. Booth R, Kim H (2012) Characterization of a lial cells: biochemical and functional charac- microfluidic in vitro model of the blood-brain terisation as a model for drug transport and barrier (muBBB). Lab Chip 12:1784–1792 targeting. J Drug Target 15:253–268 176. Colgan OC, Collins NT, Ferguson G et al 189. Cohen-Kashi Malina K, Cooper I, Teichberg (2008) Influence of basolateral condition on VI (2009) Closing the gap between the the regulation of brain microvascular endo- in-vivo and in-vitro blood-brain barrier tight- thelial tight junction properties and barrier ness. Brain Res 1284:12–21 function. Brain Res 1193:84–92 190. Dehouck MP, Meresse S, Delorme P et al 177. Tontsch U, Bauer HC (1991) Glial cells and (1990) An easier, reproducible, and mass- neurons induce blood-brain barrier related production method to study the blood-brain enzymes in cultured cerebral endothelial barrier in vitro. J Neurochem 54:1798–1801 cells. Brain Res 539:247–253 191. Culot M, Lundquist S, Vanuxeem D et al 178. Ginhoux F, Greter M, Leboeuf M (2010) (2008) An in vitro blood-brain barrier Fate mapping analysis reveals that adult model for high throughput (HTS) toxicolog- microglia derive from primitive macrophages. ical screening. Toxicol In Vitro 22:799–811 Science 330:841–845 192. Vandenhaute E, Sevin E, Hallier-Vanuxeem D 179. Aguzzi A, Barres BA, Bennett ML (2013) et al (2012) Case study: adapting in vitro Microglia: scapegoat, saboteur, or something blood-brain barrier models for use in early- else? Science 339:156–161 stage drug discovery. Drug Discov Today 180. da Fonseca AC, Matias D, Garcia C et al 17:285–290 (2014) The impact of microglial activation 193. Rubin LL, Hall DE, Porter S et al (1991) A on blood-brain barrier in brain diseases. cell culture model of the blood-brain barrier. J Front Cell Neurosci 8:362 Cell Biol 115:1725–1735 181. Cestelli A, Catania C, D’Agostino S et al 194. Schiera G, Sala S, Gallo A et al (2005) Perme- (2001) Functional feature of a novel model ability properties of a three-cell type in vitro of blood brain barrier: studies on permeation model of blood-brain barrier. J Cell Mol Med of test compounds. J Control Release 9:373–379 76:139–147 195. Vandenhaute E, Dehouck L, Boucau MC et al 182. Weidenfeller C, Svendsen CN, Shusta EV (2011) Modelling the neurovascular unit and (2007) Differentiating embryonic neural pro- the blood-brain barrier with the unique func- genitor cells induce blood-brain barrier prop- tion of pericytes. Curr Neurovasc Res erties. J Neurochem 101:555–565 8:258–269 183. Stamatovic SM, Shakui P, Keep RF et al 196. Xue Q, Liu Y, Qi H et al (2013) A novel brain (2005) Monocyte chemoattractant protein- neurovascular unit model with neurons, 98 Keith D. Rochfort and Philip M. Cummins

astrocytes and microvascular endothelial cells 209. Ballermann BJ, Dardik A, Eng E et al (1998) of rat. Int J Biol Sci 9:174–189 Shear stress and the endothelium. Kidney Int 197. Al Ahmad A, Gassmann M, Ogunshola OO Suppl 67:S100–S108 (2009) Maintaining blood-brain barrier 210. Ando J, Yamamoto K (2009) Vascular integrity: pericytes perform better than astro- mechanobiology: endothelial cell responses cytes during prolonged oxygen deprivation. J to fluid shear stress. Circ J 73:1983–1992 Cell Physiol 218:612–622 211. Cucullo L, Couraud PO, Weksler B et al 198. Appelt-Menzel A, Cubukova A, Gunther K (2008) Immortalized human brain endothe- et al (2017) Establishment of a human lial cells and flow-based vascular modeling: a blood-brain barrier co-culture model mimick- marriage of convenience for rational neuro- ing the neurovascular unit using induced plu- vascular studies. J Cereb Blood Flow Metab ripotent and multipotent stem cells. Stem Cell 28:312–328 Rep 8:894–906 212. Santaguida S, Janigro D, Hossain M et al 199. Wilhelm I, Krizbai IA (2014) In vitro models (2006) Side by side comparison between of the blood-brain barrier for the study of dynamic versus static models of blood-brain drug delivery to the brain. Mol Pharm barrier in vitro: a permeability study. Brain 11:1949–1963 Res 1109:1–13 200. Ziegler T, Nerem RM (1994) Effect of flow 213. Stanness KA, Westrum LE, Fornaciari E et al on the process of endothelial cell division. (1997) Morphological and functional charac- Arterioscler Thromb 14:636–643 terization of an in vitro blood-brain barrier 201. Cucullo L, McAllister MS, Kight K et al model. Brain Res 771:329–342 (2002) A new dynamic in vitro model for 214. Dewey CF Jr, Bussolari SR et al (1981) The the multidimensional study of astrocyte- dynamic response of vascular endothelial cells endothelial cell interactions at the blood- to fluid shear stress. J Biomech Eng brain barrier. Brain Res 951:243–254 103:177–185 202. Tarbell JM (2010) Shear stress and the endo- 215. Bussolari SR, Dewey CF Jr, Gimbrone MA Jr thelial transport barrier. Cardiovasc Res (1982) Apparatus for subjecting living cells to 87:320–330 fluid shear stress. Rev Sci Instrum 203. Siddharthan V, Kim YV, Liu S et al (2007) 53:1851–1854 Human astrocytes/astrocyte-conditioned 216. Walsh TG, Murphy RP, Fitzpatrick P et al medium and shear stress enhance the barrier (2011) Stabilization of brain microvascular properties of human brain microvascular endothelial barrier function by shear stress endothelial cells. Brain Res 1147:39–50 involves VE-cadherin signaling leading to 204. Griep LM, Wolbers F, de Wagenaar B et al modulation of pTyr-occludin levels. J Cell (2013) BBB on chip: microfluidic platform to Physiol 226:3053–3063 mechanically and biochemically modulate 217. Takeshita Y, Obermeier B, Cotleur A et al blood-brain barrier function. Biomed Micro- (2014) An in vitro blood-brain barrier devices 15:145–150 model combining shear stress and endothelial 205. Rochfort KD, Cummins PM (2015) Throm- cell/astrocyte co-culture. J Neurosci Meth- bomodulin regulation in human brain micro- ods 232:165–172 vascular endothelial cells in vitro: role of 218. Cucullo L, Marchi N, Hossain M et al (2011) cytokines and shear stress. Microvasc Res A dynamic in vitro BBB model for the study 97:1–5 of immune cell trafficking into the central 206. Rochfort KD, Collins LE, McLoughlin A et al nervous system. J Cereb Blood Flow Metab (2015) Shear-dependent attenuation of cellu- 31:767–777 lar ROS levels can suppress proinflammatory 219. Yeon JH, Na D, Choi K et al (2012) Reliable cytokine injury to human brain microvascular permeability assay system in a microfluidic endothelial barrier properties. J Cereb Blood device mimicking cerebral vasculatures. Flow Metab 35:1648–1656 Biomed Microdevices 14:1141–1148 207. Ballermann BJ, Ott MJ (1995) Adhesion and 220. Prabhakarpandian B, Shen MC, Nichols JB differentiation of endothelial cells by expo- et al (2013) SyM-BBB: a microfluidic blood sure to chronic shear stress: a vascular graft brain barrier model. Lab Chip 13: model. Blood Purif 13:125–134 1093–1101 208. Traub O, Berk BC (1998) Laminar shear 221. Alcendor DJ, Block FE 3rd, Cliffel DE et al stress: mechanisms by which endothelial cells (2013) Neurovascular unit on a chip: implica- transduce an atheroprotective force. Arterios- tions for translational applications. Stem Cell cler Thromb Vasc Biol 18:677–685 Res Ther 4(Suppl 1):S18 Chapter 6

Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure the Integrity of Blood-Brain Barrier

Balaji Srinivasan and Aditya Reddy Kolli

Abstract

Transepithelial/transendothelial electrical resistance (TEER) is a widely accepted electrical parameter to assess barrier integrity and suitability of in vitro cellular barriers for transport studies. TEER measurement has advantages over traditional permeability measurements as a quick, label-free, and non-invasive method. TEER measurement has an added advantage that it can be performed in real -time if the measurement electrodes are integrated into a microfluidic organ-on-chip device such as BBB-on-chip. As evidenced by literature, TEER measurements for various cell types have been reported with both commercially available equipment and custom-built microfluidic implementations. The design versatility of BBBs-on-chips along with the numerous factors affecting TEER can complicate comparison of TEER results reported from various laboratories. Therefore, to achieve meaningful comparison and consensus between TEER reported from various BBBs-on-chips, it is important to understand various factors that affect TEER. The aim of this chapter is to introduce TEER and its significance, explore the different TEER measurement protocols along with their strengths and weaknesses, and review numerous factors that affect TEER.

Key words TEER, BBB-on-chip, Barrier integrity, Electrical Resistance, Drug transport, Organ-on- chip

1 Introduction

Measurement of transepithelial/transendothelial electrical resis- tance (TEER [1–6]) is a widely accepted non-invasive technique for determination of barrier integrity and the extent of differentia- tion of in vitro cellular barriers. TEER is an excellent indicator of barrier integrity since the electrical impedance across an epithelium or endothelium barrier depends on the formation of robust tight junctions [7, 8] between adjacent cells. TEER is not only widely ® applied to traditional Transwell [9–11] inserts for cell culture but more recently also in the emerging technology of organs-on-chips. Organs-on-chips [12–17] are miniaturized biomimetic systems containing microfluidic [18–20] channels lined by living cells, and are designed to closely mimic key functional units of living organs

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_6, © Springer Science+Business Media, LLC, part of Springer Nature 2019 99 100 Balaji Srinivasan and Aditya Reddy Kolli

to reconstitute integrated organ-level pathophysiology in vitro. Organs-on-chips translate the biological, geometrical, and bio- chemical factors from the in vivo environment to a microfluidic platform and play a vital role in drug absorption studies by precisely controlling significant transport parameters and experimental con- ditions. The microenvironment of organs-on-chips with microflui- dic channels that support cell growth and differentiation hinders access to the cellular layer and thereby makes continuous TEER measurement challenging for monitoring permeability changes. There are technical challenges to be addressed for TEER measure- ments in organs-on-chips since practical and validated approaches have not yet been established. Numerous organ-on-chip models of the blood-brain barrier (BBBs-on-chips [15, 21]) have been reported in recent years. One of the key challenges in the field of BBB-on-chip development is the current lack of standardized quantification of parameters such as barrier permeability and shear stress [15, 22]. It is critical to quantify and confirm barrier integrity through qualitative and quantitative techniques before and while performing experimental work on in vitro barrier models. There are numerous approaches besides TEER measurement that have been reported for confirming barrier integrity. Some of these other approaches include freeze-fracture electron microscopy [23]of transmembrane fibrils and immunostaining [24] for tight-junction proteins [25]. Another widely applied approach relies on the barrier permeability to paracellular tracer compounds of various molecular weights. For example, the application of sucrose (molecular weight: 342 Da) labeled with carbon-14 for flux measurement on brain endothelial monolayer [26]. Additionally, non-radioactive fluores- cence-labeled marker proteins such as fluorescein isothiocyanate (FITC)-labeled dextrans [27] have also been reported. Enzymatic markers such as horseradish peroxidase (HRP) have been reported to study macromolecule diffusion across endothelial monolayers by tracking supernatant HRP activity [28]. In these approaches, endo- thelial permeability coefficient [29, 30] to quantify the permeability is calculated based on the measured flux of the selected tracer across cellular layers and cell-free inserts. Though widely used, the appli- cation of tracer compounds is known to interfere with the transport process under study and can also affect the barrier integrity. More- over, the use of chemical dyes renders the tested cells unusable for further experiments. Therefore, non-invasive techniques are needed and are more suitable for continuous monitoring of barrier integrity of in vitro barrier models. TEER approach, based on the measurement of electrical resistance across a cellular barrier, has proven to be a highly sensitive and reliable method to confirm the integrity and permeability of in vitro barrier models. Even though TEER and transepithelial passage of tracer markers are both indi- cators of integrity of the tight junctions and cellular barrier, they determine different entities [31]. TEER reflects the ionic Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure... 101

conductance of the paracellular pathway in the epithelial mono- layer, whereas the flux of non-electrolyte tracers which is calculated as permeability coefficient indicates the paracellular water flow and pore size of the tight junctions [31]. The wide acceptance of TEER as a standard is due to its non-invasive nature with an advantage of continuous monitoring [32–34] of live cells during their various stages of growth and differentiation. This chapter discusses some of the theoretical aspects of TEER measurement protocols, examples of commercially available TEER measurement systems and their applications, a survey of TEER measurement for in vitro BBB models and microfluidic implementations, and a brief discussion on factors that affect TEER.

2 TEER Measurement Methods

2.1 Ohm’s Law Measurement of electrical resistance of the cellular barrier in ohms Method [35] is a simplified approach to quantify the barrier integrity. The experimental setup for TEER measurement, as shown in Fig. 1, typically consists of a cellular layer cultured on a porous membrane insert that separates upper (apical) and lower (basolateral) compart- ments. Electrodes are placed on either side of the cellular layer, with one electrode each placed in the upper and lower compartment. The ohmic resistance, in theory, can be determined by applying a direct current (DC) voltage to electrodes placed on either sides of the cellular layer and measuring the resulting current. The ohmic

Fig. 1 Transepithelial/transendothelial electrical resistance (TEER) measurement with chopstick electrodes. The total electrical resistance includes the ohmic resistance of the cell layer RTEER, the cell culture medium RM, the porous membrane insert RI, and the electrode-medium interface REMI 102 Balaji Srinivasan and Aditya Reddy Kolli

resistance is calculated as the ratio of the voltage and current as per Ohm’s law. However, DC currents can adversely affect cells and measurement electrodes. To overcome this issue, DC voltage is replaced with an alternating current (AC) voltage signal with a square waveform. Epithelial Voltohmmeter (EVOM™)[36–39], a commercially available TEER measurement system, is based on an AC square wave input at 12.5 Hz that avoids charging effects on the electrodes and cells. The EVOM system has a measurement range of 0–9999 Ω with a 1 Ω resolution and uses a pair of electro- des popularly known as STX2/“chopstick” electrode set. Each electrode is 4 mm wide and 1 mm thick and contains a silver/silver chloride pellet for measuring voltage along with a silver electrode for passing current. The measurement procedure includes measur- ing the blank resistance (Rblank-membrane) of the semipermeable membrane only (without cells) and measuring the resistance across the cell layer on the semipermeable membrane (Rtotal). The cell- specific resistance (Rtissue), in units of Ω, can be calculated as Rtissue (Ω) ¼ Rtotal – Rblank-membrane, where the resistance is inversely proportional [40] to the effective surface area of the semipermeable 2 membrane (Marea), which is reported in units of cm . TEER values 2 are typically reported [41, 42] (TEERreported) in units of Ω.cm and 2 calculated as: TEERreported ¼ Rtissue (Ω) Â Marea (cm ). TEER measurements with EVOM system are highly dependent on the electrode positions. Also, it is critical to ensure that cells under study are not disturbed while the electrodes are introduced into the wells. TEER also depends on the uniformity of the current density generated by the measurement electrodes across the cellular layer. The STX2/chopstick electrode configuration causes a non-uniform current density, particularly when used with a large membrane [41] similar to the one in tissue culture inserts with 24 mm diameter, and in effect would cause an overestimation of TEER. EndOhm chamber [43–45] is one of the alternatives to STX2/chopstick electrodes and allows the cups from culture wells to be inserted. The chamber and the cap in EndOhm chamber contain a pair of concentric electrodes that consist of a voltage- sensing silver/silver chloride pellet in the center plus an annular current electrode. EndOhm chamber generates a more uniform current density across the membrane due to the symmetrical arrangement of circular disk electrodes on either side of the mem- brane. The fixed electrode geometry of EndOhm reduces variation of measurements on a given sample to 1–2 Ω [29] when compared to 10–30 Ω observed with STX2/chopstick electrode setup. Some of the other commercial systems available for TEER measurements include Electric Cell-Substrate Impedance Sensing (ECIS) [46–48] (Applied BioPhysics Inc., Troy, NY), REMS AutoSampler [49, 50] (World Precision Instruments, Sarasota, FL), Millicell-ERS system [31, 51] (Millipore Corp., Bedford, USA), and Ussing Chamber Systems [52, 53] (Warner Instruments, Hamden, CT). Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure... 103

2.2 Impedance Impedance spectroscopy in combination with a fitting algorithm Spectroscopy provides a more accurate representation of TEER when compared to traditional DC or single-frequency AC measurement systems [54]. Impedance spectroscopy [48, 55, 56] is performed by apply- ing a small-amplitude AC excitation signal with a frequency sweep and measuring the amplitude and phase response of the resulting current. Figure 2a shows a schematic of the impedance measure- ment concept. Electrical impedance (Z) is the ratio of the voltage- time function V(t) and the resulting current-time function I(t) where Vo and Io are the peak voltage and current, respectively; f is the frequency; t is the time; Φ is the phase shift between V(t) and I (t); and Y is the complex conductance or admittance. Z is a com- plex function and can be described by the modulus |Z| and the phase shift Φ or by the real part ZR and the imaginary part ZI,as illustrated in Fig. 2b. An in-depth analysis of impedance spectros- copy is available elsewhere [57]. Measurement of impedance spec- trum provides additional information about the capacitance of the cellular layer when compared to a DC or single-frequency AC-TEER measurement. Automated measurement systems such as cellZscope™ (nanoAnalytics GmbH, Germany) are commer- cially available for measuring the transendothelial/epithelial imped- ance of various barrier-forming cells cultured on permeable membranes of standard cell culture inserts. Equivalent circuit anal- ysis [58] of the impedance spectrum is performed to extract elec- trical parameters that can be applied to determine the cellular barrier properties. Figure 3a shows a typical equivalent circuit diagram for analyz- ing the impedance spectrum of cellular systems [35]. The current can flow through the junctions between cells (paracellular route) or through the cell membrane of the cells (transcellular route). The tight-junction proteins in the paracellular route cause an ohmic resistance (RTEER) in the equivalent circuit. Each lipid bilayer in the transcellular route contributes to a parallel circuit [35] consist- ing of ohmic resistance (RC) and an electrical capacitance (CC).

Fig. 2 (a) Transepithelial/transendothelial electrical resistance (TEER) measurement concept based on impedance spectroscopy. (b) Components of impedance 104 Balaji Srinivasan and Aditya Reddy Kolli

Fig. 3 (a) Typical equivalent circuit diagram that can be applied to analyze the impedance spectrum of cellular layer. (b) Simplified equivalent circuit. (c) Typical impedance spectrum with distinct frequency-dependent regions

Besides these elements, the resistance of the cell culture medium (Rmedium) and the capacitance of the measurement electrodes (CE) also affect the impedance. The high values of RC cause the current to mostly flow across the capacitor and allow an approximation in which RC is negligible [35] and the lipid bilayers can be represented with CC alone. Based on this approximation, the equivalent circuit diagram can be further simplified to as shown in Fig. 3b and the impedance spectrum observed will have a nonlinear frequency dependency [35] as shown in Fig. 3c. Typically, there are three distinct frequency regions in the impedance spectrum, with specific circuit elements playing a dominant role in each region. In the low-frequency range, the impedance is dominated by CE. In the mid-frequency range, the impedance signal is dominated by circuit elements related to the cells, namely, RTEER and CC. In the high- frequency range, CC and CE provide a more conductive path and the impedance is dominated by Rmedium. These equivalent circuit parameters can be estimated by fitting the experimental impedance spectrum data to the equivalent circuit model using nonlinear least- squares fitting techniques to obtain the best fit parameters. Table 1 lists a range of TEER values reported for BBB models with various cell types and measurement method. Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure... 105

Table 1 TEER for various BBB models

TEER (Ω.cm2) Equipment Reference In vivo BBB (humans) 5000 Estimated [102] BBB (rat) 5900 Two microelectrodes [103] 8000 Permeability [104] coefficients of radioisotopic ions Brain arterial vessels (rat) 1490 Æ 170 Two microelectrodes [103] Brain venous vessels 918 Æ 127 Two microelectrodes [103] In vitro model Primary HBMECs 100 EVOM/chopstick [105] electrodes hCMEC/D3 100 EVOM/chopstick [105] electrodes hCMEC/D3 and primary human astrocytes 140 EVOM/chopstick [105] electrodes BBMCE and MDCK epithelial cells 2020 EVOM/EndOhm [106] chamber b.End3 endothelial cells and C8-D1A 20 EVOM2/EndOhm [97] astrocytes Endothelial (RBE4) and rat astrocytes 490–510 Millicell-ERS/ [107] EndOhm Endothelial (BMCE) and rat astrocytes 250–300 Millicell-ERS/ [107] EndOhm hCMEC/D3 cells, SC-1800 astrocytes, 1200 EVOM/STX2 [108] and HBVPs electrodes hiPS-ECS + pericytes + hiPS-NSCs 1723 Æ 90 Millicell ERS-2 [4] hiPS-ECs, hiPS-NSCs, astrocytes, and 1757 Æ 320 Millicell ERS-2 [4] pericytes hPSC-derived BMECs + primary human 5160 Æ 320 EVOM/STX2 [109] brain pericytes + human astrocytes + electrodes neurons derived from neural progenitor cells Porcine brain microvessel endothelial cells 300–500 (serum) EVOM/EndOhm [41] 600–800 (serum EVOM/EndOhm [41] free) 1200–1800 Impedance analyzer [86] Microfluidic model hCMEC/D3 28.5 Æ 7.2 Gold planar electrodes [100] with EVOM2/ EndOhm hCMEC/D3 36.9 Æ 0.9 Impedance analyzer/ [21] Pt electrodes b.End3 endothelial cells and C8-D1A 250 EVOM/custom [97] astrocytes electrodes Rat brain microvascular endothelial cells 1298 Æ 86 Resistance meter [110] (BMECs), cerebral astrocytes (continued) 106 Balaji Srinivasan and Aditya Reddy Kolli

Table 1 (continued)

TEER (Ω.cm2) Equipment Reference

hIPSC-derived BMECs and rat primary 2000 Millicell-ERS Volt- [111] astrocyte co-culture Ohm Meter Primary human brain-derived microvascular 1950–2210 Custom device [22] endothelial cells (hBMVEC), pericytes, astrocytes, human cortical glutamatergic neurons from hIPSCs

3 Factors Influencing TEER

3.1 Co-culture Cells TEER has been observed to be influenced by the presence of co-cultured cells. The presence of additional cells can act as an obstacle to ion transport and can cause higher resistance when compared to an endothelium-only case. It has been demonstrated that the addition of astrocytes to brain microvascular endothelial cells (BMECs) leads to a significant increase in TEER and decreases their permeability to various molecules in vitro [59–62]. Pericytes have been observed to increase TEER in the BMEC-astrocyte co-culture BBB model [63, 64]. Similarly, neurons may regulate BBB permeability indirectly by modulating BMECs and astrocytes [65, 66]. Neurons have been shown to decrease sucrose leakage across BBB in vitro [67], possibly by regulating the localization of occludin [68, 69].

3.2 Temperature TEER measurements have been reported [70, 71] to be tempera- ture dependent. TEER measurements would be preferable in an incubator at 37 C, which requires that the electrical measurement setup have access to or be placed within the incubator. It is recom- mended that temperature is equilibrated to room temperature before performing TEER measurements to avoid any temperature fluctuation-induced TEER changes. However, this could be detri- mental to cell physiology and function. Typically, equilibration from 37 C to room temperatures could take around 20 min. To overcome these limitations and to allow TEER measurements where temperature fluctuations are expected, a mathematical method has been developed [72] to correct TEER values for the actual temperature at which they were measured, and is referred to as temperature-corrected TEER (tcTEER). To calculate tcTEER, it is required to record temperature accurately during the TEER measurement experiment. The calculation of tcTEER would per- mit comparison of TEER measured at various temperatures among independent experiments and perhaps even between different laboratories [72]. The tcTEER approach not only eliminates the need to perform temperature equilibration when experiments are Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure... 107

performed outside an incubator, but also saves time and minimizes the temperature fluctuations which may be detrimental to the cell function.

3.3 Cell Passage The effect of passage number of Caco-2 cells on TEER has been Number studied [73]. Brain endothelial cells have been extracted from various species to implement in vitro BBB models based on low-passage primary cultures from bovine [74], porcine [75], rat [76], mouse [77] and human [78]. It is known that with high passage number, astrocytes lose their ability to induce differentia- tion of the endothelial cells in co-culture models [79]. Also, pri- mary endothelial cells should be used at low passage (P1) and reconnected, at least in part, with their environment by co-culture with astrocytes or medium conditioned by astrocytes [80, 81]. The use of endothelial cell lines for BBB modeling helps to avoid the disadvantages of primary cells [4] as they are usable over many passages with a higher reproducibility of results when compared to primary cells.

3.4 Composition The origin of cell lines and the variations between various cell of Cell Culture Media culture protocols among laboratories can influence the spontane- ous differentiation that leads to a phenotype expressing many mor- phological and functional characteristics in mature cells. TEER measurements are sensitive to the ionic composition of the culture medium [82]. Also, various components of the culture medium can influence the formation of a tight barrier and in this regard intra- endothelial cAMP levels have been reported to be significant [83]. Hydrocortisone [84] within physiological concentrations has been shown to improve barrier properties in serum-free culture system in such a way that TEER and sucrose permeability came close to corresponding parameters in vivo. An increase in the buffer capacity of the media during growth [85] has been shown to significantly alter the tightness of the BBB model. In this study, they demonstrated that increased buffer concentration by addition of HEPES, MOPS, or TES to the media during differentiation increased the TEER up to 1638 Æ 256 Ω cm2, independent of the type of buffer. In the case of porcine models, it has been shown that serum prevents differentiation of cultured porcine brain endo- thelial cells [86] and weakens already established tight monolayers mainly from the abluminal side, causing a decrease in TEER. Some of the other factors [87] that have proved to be useful in tightening the barrier are insulin, transferrin, sodium selenite, putrescine, and progesterone.

3.5 Membrane Many porous membranes for cell culture are available in varied Properties materials and pore sizes. A comprehensive screening study to opti- mize membrane configuration, with aims to unveil influential membrane effects on the ability of cerebral endothelial cells to 108 Balaji Srinivasan and Aditya Reddy Kolli

form a tight monolayer has been reported [88]. Material and pore size of the filter membrane of the tissue culture inserts strongly affect the adherence of cells and barrier tightness. Some of these commercially available membranes are coated or made with matrix material and specially coated with collagen. Most of the collagen- made membranes, though effective for brain capillary endothelial cell culture, are not permeable enough for sucrose [89] and cannot be used for drug transport studies. Therefore, membranes must be tested with suitable coating for cell culture before conducting drug transport experiments for a specific drug.

3.6 Cell Culture The post-seeding culture period affects formation of tight junctions Period to achieve a uniformly differentiated culture and reach a steady- state TEER. In a study comparing TEER in co-culture model and single-layer BMEC groups at various time instances between 2-96 h, it was observed that TEER increased with time until it plateaued at 72 h in both groups [11]. In dynamic in vitro BBB models with microporous hollow fibers, time taken to reach steady- state TEER is much longer in the range of 9–12 days [90, 91].

3.7 TEER The position of TEER measurement electrodes such as the STX2 Measurement \chopstick can introduce variability between measurements if the Technique positioning is not consistent. The introduction of electrodes into the culture well under test also requires careful handling to prevent any disturbance to the cells under study. These issues can be over- come by integrating microelectrodes within these systems. The positioning of integrated microelectrodes near the cellular layer provides additional advantage of reducing the electrical resistance contribution from cell culture medium and prevention of electrical noise due to electrode motion. The uniformity of current density affects TEER and can be achieved by ensuring that diameter of the permeable membrane is compatible with the electrode geometry. In custom-made microfluidic implementations, the symmetry of electrode geometry on either side of the membrane provides uniform current densities. In TEER setup with non-conventional electrode design, electrical simulation or modeling can be a useful tool to verify uniformity of current density. In custom-designed TEER measurement electronics for impedance spectroscopy-based TEER measurements, it is important to apply parameter extraction using theoretical analyses and equivalent circuit simulation to iso- late parasitic capacitance [92] effects.

3.8 Shear Stress Shear stress caused by the flow of blood in physiological conditions has a mechanotransductive effect [93] on several endothelial molecular pathways through activation of membrane-bound recep- tors [94]. These pathways are reported to stimulate increased gene and protein expression that causes production of tight-junction proteins such as ZO-1 [61] and also modulate cytoskeletal Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure... 109

structure to alter cell orientation and structure [95]. Therefore, dynamic BBB models [91], which induce physiologically relevant shear stress have been developed. In a dynamic model [36] with bovine aortic endothelial cells and C6 glioma cell line, a TEER of ~600 Ω.cm2 has been achieved. Moreover in a side-by-side com- parative study [96], dynamic model generated ten times higher TEER than the equivalent static Transwell co-culture model, and permeability to sucrose and phenytoin was, respectively, ten times and five times less, due to a tighter barrier function of the dynamic model. In a microfluidic BBB model [97], TEER was significantly improved to 140 Ω.cm2 after 3 days of culture under a shear stress 0.023 dyn/cm2, when compared to TEER of 15 Ω.cm2 in the static model. Also, the stability of this microfluidic BBB model was indicated by an increase in permeability on histamine exposure, followed by recovery.

4 Conclusions

Measurement of TEER has numerous advantages over permeability measurements as it is a quick, label-free and non-invasive way to assess the barrier integrity of cells during their various stages of growth and differentiation. Moreover, TEER can be measured in real-time [35, 97] with suitable electrode designs that are integrated within a microfluidic BBB-on-chip device. The success of the various in vitro barrier models to accurately predict drug absorption depends on how closely they can mimic the complexity of the drug absorption in vivo. TEER measurements based on impedance spectroscopy are more reliable and provide additional characterization of the cell culture when compared to the Ohm’s law method. Even though many commercial TEER measurement equipment are now available, integrating TEER electronics within the microfluidic BBB-on-chip provides real-time measurement without disrupting the cell culture. A wide range of TEER values have been reported in the literature for the same cell type. These discrepancies can arise due to many variables affecting TEER such as the measurement protocols, selection and usage of electrodes, device design, temperature fluctuations during measurement, com- position of the culture medium, presence of other cells, static or dynamic culture conditions, cell culture period, and passage num- ber of cells used in the model. As evidenced by numerous publica- tions on various types of BBB models, current focus is more toward developing physiologically relevant models that closely mimic in vivo conditions. Many of the recently reported BBB models show promising advancements [21, 97–101] by incorporating flow-induced shear stress that results in better barrier functions when compared to earlier static Transwell models. Application of in vitro barrier models that do not closely reproduce TEER under 110 Balaji Srinivasan and Aditya Reddy Kolli

in vivo conditions would lead to incorrect conclusions while eval- uating drug transport experiments. The versatility of BBBs-on- chips along with the numerous variables affecting TEER as dis- cussed in previous sections complicates comparison of TEER results reported from various laboratories. Therefore, to achieve meaningful comparison and consensus between TEER reported from various BBBs-on-chips, it is important to identify and clearly report the above-listed variables along with the measurement pro- tocol. BBB-on-chip models are required to advance BBB-related research for facilitating the development of new drugs for many neurological diseases. TEER will be a useful parameter for research- ers in addressing the standardization challenges ahead of them for developing such BBB-on-chip models.

References

1. Katt ME et al (2018) Functional brain- engineering the liver sinusoid. Biofabrication specific microvessels from iPSC-derived 5(4):045008–045008 human brain microvascular endothelial cells: 10. Sip CG, Bhattacharjee N, Folch A (2014) the role of matrix composition on monolayer Microfluidic transwell inserts for generation formation. Fluids Barriers CNS 15(1):7 of tissue culture-friendly gradients in well 2. Jeong S et al (2018) A three-dimensional plates. Lab Chip 14(2):302–314 arrayed microfluidic blood-brain barrier 11. Wang Y et al (2015) In vitro model of the model with integrated electrical sensor array. blood-brain barrier established by co-culture IEEE Trans Biomed Eng 65(2):431–439 of primary cerebral microvascular endothelial 3. Yang S et al (2017) Identification of two and astrocyte cells. Neural Regen Res 10 immortalized cell lines, ECV304 and bEnd3, (12):2011–2017 for in vitro permeability studies of blood- 12. Huh D et al (2013) Microfabrication of human brain barrier. PLoS One 12(10):e0187017 organs-on-chips. Nat Protoc 8(11):2135–2157 4. Appelt-Menzel A et al (2017) Establishment 13. Chan CY et al (2013) Accelerating drug dis- of a human blood-brain barrier co-culture covery via organs-on-chips. Lab Chip 13 model mimicking the neurovascular unit (24):4697–4710 using induced pluri- and multipotent stem 14. van der Meer AD, van den Berg A (2012) cells. Stem Cell Rep 8(4):894–906 Organs-on-chips: breaking the in vitro 5. Ni Y et al (2017) TNFalpha alters occludin impasse. Integr Biol 4(5):461–470 and cerebral endothelial permeability: role of 15. van der Helm MW et al (2016) Microfluidic p38MAPK. PLoS One 12(2):e0170346 organ-on-chip technology for blood-brain 6. Paradis A, Leblanc D, Dumais N (2016) barrier research. Tissue Barriers 4(1): Optimization of an in vitro human blood–- e1142493 brain barrier model: application to blood 16. Perestrelo AR et al (2015) Microfluidic monocyte transmigration assays. MethodsX organ/body-on-a-chip devices at the conver- 3:25–34 gence of biology and microengineering. Sen- 7. Luissint A-C et al (2012) Tight junctions at sors (Basel, Switzerland) 15 the blood brain barrier: physiological archi- (12):31142–31170 tecture and disease-associated dysregulation. 17. Ronaldson-Bouchard K, Vunjak-Novakovic G Fluids Barriers CNS 9:23–23 (2018) Organs-on-a-chip: a fast track for 8. Bauer H-C et al (2014) “You Shall Not engineered human tissues in drug develop- Pass”—tight junctions of the blood brain bar- ment. Cell Stem Cell 22(3):310–324 rier. Front Neurosci 8:392 18. Sackmann EK, Fulton AL, Beebe DJ (2014) 9. Kang YB et al (2013) Layered long term The present and future role of microfluidics in co-culture of hepatocytes and endothelial biomedical research. Nature 507 cells on a transwell membrane: toward (7491):181–189 Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure... 111

19. Acevedo JP et al (2018) Microtechnology blood-brain barrier model. Biotechnol J 7 applied to stem cells research and develop- (3):409–417 ment. Regen Med 13:233–248 33. Eigenmann DE et al (2013) Comparative 20. Aziz AUR et al (2017) The role of microflui- study of four immortalized human brain cap- dics for organ on chip simulations. Bioengi- illary endothelial cell lines, hCMEC/D3, neering 4(2):39 hBMEC, TY10, and BB19, and optimization 21. Griep LM et al (2013) BBB on chip: micro- of culture conditions, for an in vitro blood–- fluidic platform to mechanically and biochem- brain barrier model for drug permeability ically modulate blood-brain barrier function. studies. Fluids Barriers CNS 10(1):33 Biomed Microdevices 15(1):145–150 34. Alexander FA, Eggert S, Wiest J (2018) Skin- 22. Brown JA et al (2015) Recreating blood-brain on-a-chip: transepithelial electrical resistance barrier physiology and structure on chip: a and extracellular acidification measurements novel neurovascular microfluidic bioreactor. through an automated air-liquid interface. Biomicrofluidics 9(5):054124 Genes 9(2):114 23. Wolburg H et al (1994) Modulation of tight 35. Benson K, Cramer S, Galla HJ (2013) junction structure in blood-brain barrier Impedance-based cell monitoring: barrier endothelial cells. Effects of tissue culture, sec- properties and beyond. Fluids Barriers CNS ond messengers and cocultured astrocytes. J 10(1):5 Cell Sci 107(Pt 5):1347–1357 36. Epithelial voltohmmeter. Available from: 24. Mullier A et al (2010) Differential distribu- http://www.wpi-europe.com/products/cell- tion of tight junction proteins suggests a role and-tissue/teer-measurement/evom2.aspx for tanycytes in blood-hypothalamus barrier 37. Griffin JM et al (2016) Statins inhibit fibrillary regulation in the adult mouse brain. J Comp β-amyloid induced inflammation in a model of Neurol 518(7):943–962 the human blood brain barrier. PLoS One 11 25. Jiao H et al (2011) Specific role of tight junc- (6):e0157483 tion proteins claudin-5, occludin, and ZO-1 38. Stebbins MJ et al (2016) Differentiation of the blood-brain barrier in a focal cerebral and characterization of human pluripotent ischemic insult. J Mol Neurosci 44 stem cell-derived brain microvascular endo- (2):130–139 thelial cells. Methods (San Diego, CA) 26. Bowman PD et al (1983) Brain microvessel 101:93–102 endothelial-cells in tissue-culture—a model 39. Janga H et al (2018) Site-specific and for study of blood-brain-barrier permeability. endothelial-mediated dysfunction of the Ann Neurol 14(4):396–402 alveolar-capillary barrier in response to lipo- 27. Horibe Y et al (1997) Polar solute transport polysaccharides. J Cell Mol Med 22 across the pigmented rabbit conjunctiva: size (2):982–998 dependence and the influence of 8-bromo 40. Haorah J et al (2008) Activation of protein cyclic adenosine monophosphate. Pharm Res tyrosine kinases and matrix metalloproteinases 14(9):1246–1251 causes blood-brain barrier injury: novel mech- 28. Duffy SL, Murphy JT (2001) Colorimetric anism for neurodegeneration associated with assay to quantify macromolecule diffusion alcohol abuse. Glia 56(1):78–88 across endothelial monolayers. BioTechni- 41. Franke H, Galla HJ, Beuckmann CT (1999) ques 31(3):495–496, 498, 500–1 An improved low-permeability in vitro-model 29. Audus K, Borchardt R (1986) Characteriza- of the blood-brain barrier: transport studies tion of an in vitro blood–brain barrier model on retinoids, sucrose, haloperidol, caffeine system for studying drug transport and and mannitol. Brain Res 818(1):65–71 metabolism. Pharm Res 3(2):81–87 42. Watson PMD et al (2013) Modelling the 30. Dehouck MP et al (1992) In vitro reconsti- endothelial blood-CNS barriers: a method tuted blood-brain barrier. J Control Release for the production of robust in vitromodels 21(1–3):81–91 of the rat blood-brain barrier and blood- 31. Zucco F et al (2005) An inter-laboratory spinal cord barrier. BMC Neurosci 14(1):59 study to evaluate the effects of medium com- 43. EndOhm chamber. Available from: http:// position on the differentiation and barrier www.wpi-europe.com/products/cell-and-tis function of Caco-2 cell lines. Altern Lab sue/teer-measurement/endohm-24snap.aspx Anim 33(6):603–618 44. Sheller RA, Cuevas ME, Todd MC (2017) 32. Wuest DM, Lee KH (2012) Optimization of Comparison of transepithelial resistance mea- endothelial cell growth in a murine in vitro surement techniques: Chopsticks vs Endohm. Biol Proced Online 19:4 112 Balaji Srinivasan and Aditya Reddy Kolli

45. Edington CD et al (2018) Interconnected 58. Yoon Y et al (2017) Impedance spectroscopy microphysiological systems for quantitative analysis and equivalent circuit modeling of biology and pharmacology studies. Sci Rep graphene oxide solutions. Nanomaterials 8:4530 (Basel) 7(12):446 46. Bernas MJ et al (2010) Establishment of pri- 59. Neuhaus J, Risau W, Wolburg H (1991) mary cultures of human brain microvascular Induction of blood-brain barrier characteris- endothelial cells to provide an in vitro cellular tics in bovine brain endothelial cells by rat model of the blood-brain barrier. Nat Proto- astroglial cells in transfilter coculture. Ann N cols 5(7):1265–1272 Y Acad Sci 633:578–580 47. Bagnaninchi PO, Drummond N (2011) Real- 60. Tao-Cheng JH, Nagy Z, Brightman MW time label-free monitoring of adipose-derived (1987) Tight junctions of brain endothelium stem cell differentiation with electric cell- in vitro are enhanced by astroglia. J Neurosci substrate impedance sensing. Proc Natl Acad 7(10):3293–3299 Sci U S A 108(16):6462–6467 61. Siddharthan V et al (2007) Human 48. Szulcek R, Bogaard HJ, van Nieuw Ameron- astrocytes/astrocyte-conditioned medium gen GP (2014) Electric cell-substrate imped- and shear stress enhance the barrier properties ance sensing for the quantification of of human brain microvascular endothelial endothelial proliferation, barrier function, cells. Brain Res 1147:39–50 and motility. J Vis Exp 85:51300 62. Cohen-Kashi Malina K, Cooper I, Teichberg 49. Cain MD et al (2017) Virus entry and replica- VI (2009) Closing the gap between the tion in the brain precedes blood-brain barrier in-vivo and in-vitro blood-brain barrier tight- disruption during intranasal alphavirus infec- ness. Brain Res 1284:12–21 tion. J Neuroimmunol 308:118–130 63. Dente CJ et al (2001) Pericytes augment the 50. D’Souza WN et al (2017) Differing roles for capillary barrier in in vitro cocultures. J Surg short chain fatty acids and GPR43 agonism in Res 97(1):85–91 the regulation of intestinal barrier function 64. Nakagawa S et al (2009) A new blood-brain and immune responses. PLoS One 12(7): barrier model using primary rat brain endo- e0180190 thelial cells, pericytes and astrocytes. Neuro- 51. Ferruzza S et al (2013) Serum-reduced and chem Int 54(3–4):253–263 serum-free media for differentiation of Caco- 65. Hawkins BT, Davis TP (2005) The blood- 2 cells. Altern Anim Exp 30(2):159–168 brain barrier/neurovascular unit in health 52. Li H, Sheppard DN, Hug MJ (2004) Transe- and disease. Pharmacol Rev 57(2):173–185 pithelial electrical measurements with the 66. Cohen Z, Molinatti G, Hamel E (1997) Ussing chamber. J Cyst Fibros 3(Suppl Astroglial and vascular interactions of nor- 2):123–126 adrenaline terminals in the rat cerebral cortex. 53. Anabazhagan AN et al (2017) Methods to J Cereb Blood Flow Metab 17(8):894–904 study epithelial transport protein function 67. Schiera G et al (2005) Permeability properties and expression in native intestine and Caco- of a three-cell type in vitro model of blood- 2 cells grown in 3D. J Vis Exp 121. https:// brain barrier. J Cell Mol Med 9(2):373–379 doi.org/10.3791/55304 68. Savettieri G et al (2000) Neurons and ECM 54. Douville NJ et al (2010) Fabrication of regulate occludin localization in brain endo- two-layered channel system with embedded thelial cells. Neuroreport 11(5):1081–1084 electrodes to measure resistance across epithe- 69. Schiera G et al (2003) Synergistic effects of lial and endothelial barriers. Anal Chem 82 neurons and astrocytes on the differentiation (6):2505–2511 of brain capillary endothelial cells in culture. J 55. Kuzmanov I et al (2016) An in vitro model of Cell Mol Med 7(2):165–170 the blood-brain barrier using impedance spec- 70. Matter K, Balda MS (2003) Functional analy- troscopy: a focus on T cell-endothelial cell sis of tight junctions. Methods 30 interaction. J Vis Exp 118:e54592 (3):228–234 56. Rahim S, Uren A (2011) A real-time electrical 71. Torres R et al (2007) GTX 2/3 epimers per- impedance based technique to measure inva- meate the intestine through a paracellular sion of endothelial cell monolayer by cancer pathway. J Toxicol Sci 32(3):241–248 cells. J Vis Exp 50:2792 72. Blume LF et al (2010) Temperature corrected 57. Macdonald JR, Johnson WB (2005) Funda- transepithelial electrical resistance (TEER) mentals of impedance spectroscopy. In: measurement to quantify rapid changes in Impedance spectroscopy. Wiley, pp 1–26 paracellular permeability. Pharmazie 65 (1):19–24 Transepithelial/Transendothelial Electrical Resistance (TEER) to Measure... 113

73. Lu S et al (1996) Transport properties are not 87. Wilhelm I, Krizbai IA (2014) In vitro models altered across Caco-2 cells with heightened of the blood-brain barrier for the study of TEER despite underlying physiological and drug delivery to the brain. Mol Pharm 11 ultrastructural changes. J Pharm Sci 85 (7):1949–1963 (3):270–273 88. Wuest DM, Wing AM, Lee KH (2013) Mem- 74. Dehouck MP et al (1990) An easier, repro- brane configuration optimization for a murine ducible, and mass-production method to in vitro blood-brain barrier model. J Neurosci study the blood-brain barrier in vitro. J Neu- Methods 212(2):211–221 rochem 54(5):1798–1801 89. Cecchelli R et al (1999) In vitro model for 75. Patabendige A, Skinner RA, Abbott NJ evaluating drug transport across the blood–- (2013) Establishment of a simplified in vitro brain barrier. Adv Drug Deliv Rev 36 porcine blood–brain barrier model with high (2):165–178 transendothelial electrical resistance. Brain 90. Cucullo L et al (2011) A dynamic in vitro Res 1521:1–15 BBB model for the study of immune cell traf- 76. Abbott NJ et al (2012) An improved in vitro ficking into the central nervous system. J blood-brain barrier model: rat brain endothe- Cereb Blood Flow Metab 31(2):767–777 lial cells co-cultured with astrocytes. Methods 91. Cucullo L et al (2002) A new dynamic in vitro Mol Biol 814:415–430 model for the multidimensional study of 77. Coisne C et al (2005) Mouse syngenic in vitro astrocyte-endothelial cell interactions at the blood-brain barrier model: a new tool to blood-brain barrier. Brain Res 951 examine inflammatory events in cerebral (2):243–254 endothelium. Lab Investig 85(6):734–746 92. Zhang G, Zhu R (2010) Effect of parasitic 78. Lacombe O et al (2011) In vitro primary capacitance on impedance measurement and human and animal cell-based bloodÀbrain model extraction. Electroanalysis 22 barrier models as a screening tool in drug (3):351–358 discovery. Mol Pharm 8(3):651–663 93. Partyka PP et al (2017) Mechanical stress reg- 79. Molino Y et al (2014) Setting-up an in vitro ulates transport in a compliant 3D model of model of rat blood-brain barrier (BBB): a the blood-brain barrier. Biomaterials focus on BBB impermeability and receptor- 115:30–39 mediated transport. J Vis Exp 88:e51278 94. Booth RH, Kim H (2012) A parallel array 80. Abbott NJ et al (1992) Development and microfluidic blood-brain barrier model for characterisation of a rat brain capillary endo- high-throughput quantitation of shear stress thelial culture: towards an in vitro blood-brain effects. In: Proceedings of the 16th interna- barrier. J Cell Sci 103(Pt 1):23–37 tional conference on miniaturized systems 81. Rubin L, Sanes J (1993) Neuronal and glial for chemistry and life sciences, MicroTAS cell biology. Editorial overview: tools of the 2012 trade. Curr Opin Neurobiol 3(5):673–675 95. Galbraith CG, Skalak R, Chien S (1998) Shear 82. Wong AD et al (2013) The blood-brain bar- stress induces spatial reorganization of the rier: an engineering perspective. Front Neu- endothelial cell cytoskeleton. Cell Motil Cyto- roeng 6:7 skeleton 40(4):317–330 83. Rubin LL et al (1991) A cell culture model of 96. Santaguida S et al (2006) Side by side com- the blood-brain barrier. J Cell Biol 115 parison between dynamic versus static models (6):1725–1735 of blood-brain barrier in vitro: a permeability 84. Hoheisel D et al (1998) Hydrocortisone rein- study. Brain Res 1109(1):1–13 forces the blood-brain barrier properties in a 97. Booth R, Kim H (2012) Characterization of a serum free cell culture system. Biochem Bio- microfluidic in vitro model of the blood-brain phys Res Commun 244(1):312–316 barrier (muBBB). Lab Chip 12 85. Helms HC et al (2010) Paracellular tightness (10):1784–1792 and claudin-5 expression is increased in the 98. Booth R, Kim H (2014) Permeability analy- BCEC/astrocyte blood–brain barrier model sis of neuroactive drugs through a dynamic by increasing media buffer capacity during microfluidic in vitro blood-brain barrier growth. AAPS J 12(4):759–770 model. Ann Biomed Eng 42(12): 86. Nitz T et al (2003) Serum-derived factors 2379–2391 weaken the barrier properties of cultured por- 99. Prabhakarpandian B et al (2013) SyM-BBB: a cine brain capillary endothelial cells in vitro. microfluidic blood brain barrier model. Lab Brain Res 981(1–2):30–40 Chip 13(6):1093–1101 114 Balaji Srinivasan and Aditya Reddy Kolli

100. Walter FR et al (2016) A versatile lab-on-a- In: 13th International conference on biomed- chip tool for modeling biological barriers. ical engineering. Springer Berlin Heidelberg, Sensors Actuators B Chem 222:1209–1219 Berlin Heidelberg 101. Chaitanya GV et al (2011) Gliovascular and 107. Balbuena P et al (2010) Comparison of two cytokine interactions modulate brain endo- blood-brain barrier in vitro systems: cytotox- thelial barrier in vitro. J Neuroinflammation icity and transfer assessments of malathion/ 8:162 oxon and lead acetate. Toxicol Sci 114 102. Lauschke K, Frederiksen L, Hall VJ (2017) (2):260–271 Paving the way toward complex blood-brain 108. Maherally Z et al (2018) Real-time acquisi- barrier models using pluripotent stem cells. tion of transendothelial electrical resistance in Stem Cells Dev 26(12):857–874 an all-human, in vitro, 3-dimensional, blood- 103. Butt AM, Jones HC, Abbott NJ (1990) Elec- brain barrier model exemplifies tight-junction trical resistance across the blood-brain barrier integrity. FASEB J 32(1):168–182 in anaesthetized rats: a developmental study. J 109. Lippmann ES et al (2014) A retinoic acid- Physiol 429:47–62 enhanced, multicellular human blood-brain 104. Smith QR, Rapoport SI (1986) Cerebrovas- barrier model derived from stem cell sources. cular permeability coefficients to sodium, Sci Rep 4:4160 potassium, and chloride. J Neurochem 46 110. Xu H et al (2016) A dynamic in vivo-like (6):1732–1742 organotypic blood-brain barrier model to 105. Daniels BP et al (2013) Immortalized human probe metastatic brain tumors. Sci Rep cerebral microvascular endothelial cells main- 6:36670 tain the properties of primary cells in an 111. Wang YI, Abaci HE, Shuler ML (2017) in vitro model of immune migration across Microfluidic blood–brain barrier model pro- the blood brain barrier. J Neurosci Methods vides in vivo-like barrier properties for drug 212(1):173–179 permeability screening. Biotechnol Bioeng 106. Amin R et al (2009) Permeability of an 114(1):184–194 in vitro model of blood brain barrier (BBB). Chapter 7

Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain Barrier Permeability: Implications for Pathogenesis and Therapeutic Treatment of Neurodegenerative Disease

Swapna Bera and Anirban Bhunia

Abstract

Over the past few decades, the blood-brain barrier (BBB) has been acknowledged as the prime defense mechanism for the brain against exterior deadly substances. This dynamic barrier, which is primarily comprised of precisely arranged, specialized endothelial cells, astrocytes, and pericytes, not only prevents the brain interstitial fluid from mixing with the components of blood but also maintains a delicate balance between central nervous system (CNS) and blood circulatory system by transferring essential substances across the BBB. Thus, any dysfunction in the BBB activity leads to several neuroinflammatory disorders such as Alzheimer’s, Parkinson’s, and prion-related diseases. The risk posed by such neurodegenerative diseases has become the foremost public health concern worldwide because of the lack of BBB-permeable therapeutic agents. This points to the need for studying the characteristic traits of various BBB-permeable cell-penetrating peptides (CPPs), which have emerged as a potential drug delivery tool in CNS therapeu- tics. In this chapter, we would like to review the benefits of employing nuclear magnetic resonance (NMR) techniques toward understanding the structural, functional, and dynamic behavior of CPPs at the molecular level, during their interaction with BBB-model membranes. We conclude that CPPs can be utilized for developing promising therapeutic drugs against CNS diseases.

Key words BBB, CNS, CPPs, NMR

1 Introduction

Inasmuch as breakthroughs in the field of medicine have helped increase the life span of humans, the risks posed by deadly neuro- logical diseases like Alzheimer’s, Parkinson’s, and Huntington’s continue to be enduring threats worldwide. This is exemplified by the fact that as of 2015, the outraging number of people living with dementia has been estimated at 46.8 million worldwide, with over 3.7 million of those people coming from various underprivileged parts of India. This number is anticipated to double within next 20 years, also breaking the age barrier in the process. Despite its

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_7, © Springer Science+Business Media, LLC, part of Springer Nature 2019 115 116 Swapna Bera and Anirban Bhunia

rapid progression toward deteriorating health, a major part of the population does not appear to be concerned about these neuroin- flammatory diseases. This is likely because of the gradually develop- ing and unrecognizable symptoms, which remain for several years before the onset of worse symptoms that result in severe interfer- ence with normal day-to-day activities. Besides gross ignorance and negligence, the scarcity of available treatments makes these deadly diseases more complicated to be dealt with. Attempts to develop drugs for delivery across the blood-brain barrier (BBB) for the treatment of neurodegenerative disorders have not been successful [1, 2]. It is well known that BBB, which is the most complex barrier for providing protection to the central nervous system (CNS), is also involved in the execution of several other critical physiological functions. This highly organized network of tight endothelial cells not only prohibits the passage of blood-borne molecules, thereby defending the CNS, but also accomplishes the synchronization between the CNS and blood circulatory system by adaptating to physiological needs [3]. However, under certain circumstances, BBB can also be directly impacted by a degenerative disease or can predispose the host body for the onset of a CNS disease. For example, the rapid deposition of misfolded Aβ aggregates in the cerebral vasculature, also termed as cerebral amyloid angiopathy (CAA), can directly influence the BBB morphology, permeability, and CNS homeostasis, which can result in pro-inflammatory and cytotoxic events leading up to Alzheimer’s disease (AD) [4]. Collectively, this usefulness and complexity of BBB make them more amenable targets for therapeutic studies aimed at treating CNS disorders [5]. Despite the volume of research work that has gone into addressing this, scientists have been unable to come up with permanent remedies for neurodegenerative diseases. The search for practical approaches to accomplish this has become crucial. Unlike other classical methods employed previously, the therapeutic utilization of CPPs as CNS drug delivery system has emerged as an effective strategy [6]. However, the conventional CPPs will need to be further improvised for expanding the efficacy as CNS therapeutics. The structures and dynamics of membrane- bound CPP molecules have been poorly investigated till date. Atomic-resolution structures of well-folded proteins or complexes can be obtained using X-ray crystallography. However, a large number of proteins or domains of large proteins (e.g., in signaling cascades) and bioactive peptides (e.g., CPPs) appear to be dynamic, thus limiting the application of X-ray-based methods. On the other hand, it is possible to gain insights into such molecular systems at the atomic level using advanced nuclear magnetic resonance (NMR) spectroscopy. In this chapter, we elaborate on a few most powerful NMR methods that are frequently being used to explore the molecular interactions, three-dimensional structures, and dynamics of ligand molecules upon interaction with specific Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 117

receptors. The limitations of these techniques and important aspects of the experimental design are also explicated here. Even though the chapter mainly focuses on techniques pertaining to the treatment of Alzheimer’s disease, the methods described for drug delivery to CNS can also be applied for the treatment of other relevant BBB-directed neurodegenerative diseases.

2 BBB Being the Active Performer of Alzheimer’s Disease Onset: BBB-Permeable CPPS as a Promising Tool for CNS Drug Delivery

While Alzheimer’s disease is majorly referred as a progressive dis- ease of neuronal cell dysfunction, many other severe symptoms like metal ion dyshomeostasis, oxidative stress, and disruption of BBB integrity are also shown to be allied with the dreadful disease [7]. In non-pathological condition, the level of soluble Aβ peptide, which is the cleaved product of membrane-bound signaling receptor pro- tein, amyloid precursor protein (APP) by the neuronal activity of α-secretase and γ-secretase, remains near to the ground. However, under certain physiological strain, the same APP now becomes the substrate of β-secretase and γ-secretase and produces the misfolded version of Aβ peptide which tends to get aggregate and deposits in several regions of cerebral neutrophils and vasculature [8, 9]. Thus, the massive accumulation of insoluble, toxic amyloid plaques (Aβ aggregates along with hyperphosphorylated tau proteins) in the brain leads to the adverse condition of neurodegeneration and Alzheimer’s disease onset [10]. So far, several conflicting theories exist based on the crucial link between BBB and amyloidogenesis process [11]. While the influence of BBB in the early onset of Aβ aggregation has been reported, the loss of its structural integrity as a consequence of amyloid plaque deposition in the cerebral vascu- lature has also come into the limelight. The BBB is mainly composed of tight endothelial cell junctions which strictly separates the brain parenchyma from the endothelial cells of the blood capillaries by forming two distinct functional units, abluminal and luminal sides [12–15]. The abluminal side of BBB, facing the brain parenchyma, forms the “neurovascular unit” (NVU) of CNS, which is again formed by the interaction of a variety of cells including cerebral endothelial cells, basal lamina, pericytes, and astrocytic foot processes [16]. The NVU has a dual role to play in CNS activity. It not only restricts the entry of lipophobic molecules to CNS but also helps in carrying essential molecules across BBB through transporter/receptor-mediated endocytosis method, and thus retains CNS homoeostasis. In addi- tion, BBB is also believed to be directly responsible for the clear- ance of continuously produced Aβ peptide within the brain and thus prohibits its accumulation and aggregation, which further 118 Swapna Bera and Anirban Bhunia

leads to Alzheimer’s disease [17]. Earlier studies which were carried out on the effect of ageing and neurodegenerative diseases on BBB implied an indication toward alterations in NVU permeability over time, thus making the abnormal access of nearly all blood-borne molecules to the CNS much easier. Several other pathological con- ditions like deteriorated BBB integrity, vascular smooth muscle dysfunction, and cerebral microhemorrhage are also shown to be allied with it. On the contrary, current evidence denies neurovas- cular degeneration and BBB collapse due to AD onset [18]. Rather their theories have something important to say about BBB mostly remaining intact throughout the amyloidogenesis or tauopathy. Hence the challenge imposed by the intact integrity of BBB, even under diseased condition, remains in CNS therapeutics due to which scarcity in efficient BBB-permeable drugs can be found till date. Besides, several distinct traits of therapeutic drugs have also shown to bring additional burden in drug accessibility to the brain. For instance, the passage of any therapeutic agents to CNS is broadly directed by its molecular size, surface charge, lipophilicity, functional groups, and specific transporters. Therefore, scientists are now trying hard to bypass the complications of BBB through a number of ways including nanotechnology, liposome-mediated drug delivery, antibody facilitated transfer, microinjections, drugs attached to biodegradable substance, use of proline-rich peptides/ chimeric peptides/radionuclides, DNA biotechnology or viral- based vectors, etc. [5, 19–21]. Few of them are like microbubble- encapsulated drugs which after intranasal (IN)/intravenous (IV) injection were directed to enter CNS by applying ultrasound pulses (FUS) [22]. The fusion of microbubbles and FUS thus becomes more efficient to overcome some previous application- oriented restriction in the intranasal route of CNS drug delivery. Meanwhile, in another eventful approach use of nanowires is also shown to be effective due to its neuroprotective ability. Regardless of the usefulness, most of these methods have now become incom- patible for CNS therapeutics owing to certain limitations. There- fore, as an alternative strategy, popular CPPs are now extensively being used as CNS drug carrier and unexpectedly they are rising as the effective one regarding their high efficacy and low cytotoxicity to combat the neurodegenerative diseases [6, 23].

3 Overview on CPPs: Features, Functions, and Feasible Utility in Numerous Therapeutics

CPPs are in general short membrane-active peptides (sizes of less than 40 amino acids), which are found to be internalized effort- lessly within the cells without showing cell toxicity [24–26]. Earlier, CPPs are believed as similar of membrane-interacting antimicrobial Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 119

peptides (AMPs) due to their structural and functional similarities; however, unlike conventional AMPs, CPPs are less active in mem- branolytic activity. In the era of 90s, two CPPs were discovered, namely human immunodeficiency virus type-1 (HIV-1) derived trans-activating factor TAT [27, 28] and homeodomain of Dro- sophila antennapedia transcription factor, penetratin [29, 30]. These two CPPs were considered to be the first to reveal the translocating ability of CPP through the semi-impermeable cell membrane. After several years of research, a group of versatile CPPs like Transportan/TP10 from galanin-Lys-mastoparan [31], VP22 peptide of herpes simplex virus [32], Pep-1 [33], MAP (amphipathic model peptide) [34], etc. have been identified. Regardless of their conserved sequence similarities what brings the uniqueness is their ability to influence the delivery of a variety of covalently/non-covalently linked cargo molecules across the plasma membrane irrespective of their size and solubility [35–40]. The property of CPP being a cargo delivery agent is merely governed by several distinct features including short size, amphipathicity, highly distributed cationic charges, and most important its secondary structure which facilitates the initial peptide-membrane interaction followed by its cellular uptake [41–44]. In addition to this, the membrane composition and the nature and concentration of cargo molecules also play a significant role in the CPP-cargo translocation mechanisms. Several theories have concurred to date, which reports more than one pathway for the intracellular uptake of CPPs. In general, direct penetration and endocytosis are considered as the most facilitated translocating route of CPP and CPP-cargo complexes inside cells [45]. The translocation by endocytotic pathways is however further classified into caveolae-mediated endocytosis (~60 nm), caveolae- and clathrin-independent endocytosis (~90 nm), clathrin-mediated endocytosis (~120 nm), and macropinocytosis (>500 nm), which activates under specific stimuli [46–48]. Although the uptake mechanisms by CPP remain elusive, the concept of using them as a molecular vehicle to promote drug delivery in therapeutics is emerging with time. To date, they are extensively employed in the treatment of tumor, chronic infection (both viral and bacterial), anti-prion diseases, CNS disorder and neurodegenerative diseases, muscular dystrophy, cardiology, inhibition in NF-κB signaling pathway, etc. [49, 50]. Nevertheless, depending on the need, these existing CPPs are now continuously being replaced with the more advanced one having high efficacy and minimal side effects [51]. Here we review few of these CPP applications.

3.1 CPPs in With the aim of diminishing side effects of chemotherapy in cancer Anticancer Drug treatment, scientists are now showing their concern in using che- Delivery motherapeutic drugs (cyclosporine A, doxorubicin, methotrexate, chlorambucil and paclitaxel, etc.) conjugated to various cell-borne 120 Swapna Bera and Anirban Bhunia

CPPs through labile linkage and more to our surprise they have come out as the effective one [52–56]. Likewise, a new CPP-drug complex, dPasFHV-p53C0, is also reported to be active in glioma- initiating cell death in mice models [57]. Furthermore, in another eventful approach the application of RI-HA2-p53C0 and RI-TAT-p53C0 (transducible D-isomer of retro-inverso CPP connected to p53 C-terminus) in introducing particular tumor- suppressor proteins in cells has also shown prominent cure in bladder cancer, peritoneal lymphoma, and peritoneal carcinomato- sis [58, 59]. The use of CPP however not only facilitates the drug delivery but under certain circumstances CPP itself shows signifi- cant contribution in increased drug activity by influencing cell- signaling pathways. For example, covalently linked doxorubicin to TAT and penetratin (CPP-Dox conjugates) construct exhibits dif- ferent apoptotic pathways resulting in comparatively higher apo- ptotic efficiency than Dox alone [60, 61]. Currently, CPPs are further modified to activatable cell-penetrating peptide (ACPP) to improve tumor-targeted drug delivery [62].

3.2 Use of CPPs in Earlier studies have already demonstrated how CPPs conjugated to Anti-prion and human prion protein-derived peptides (PrP-CPPs) can influence Muscular Dystrophy PrPC protein level, responsible for prion diseases [63]. In addition, therapeutic strategies against muscular dystrophy (DMD) using both oligonucleotides and PMOs-CPP construct (termed as PPMOs or peptide-conjugated PMO) have been thoroughly inspected and apparently; it is also proven to be useful in restoring dystrophin expression in muscles [64]. Here, PMOs are defined as small, uncharged antisense agents, which exhibit the job of pre- venting gene expression, only after their preferential binding to arginine-rich CPP molecules.

3.3 CPP-Based Unlike other diseases in case of bacterial and fungal infection, CPPs Approaches Against like MAP, TAT, pVEC, penetratin, Pep-1, and ε-poly-L-lysines itself Antibacterial/Fungal/ act as antimicrobial peptides depending on peptide concentration Viral Diseases and membrane composition [65]. Very recently, Christian Ru¨tera and his co-workers have published one interesting article where they showed successful delivery of commonly used antibiotic, gen- tamicin using bacterium-derived CPPs [66]. Here, two novel CPPs are derived from the Yersinia enterocolitica YopM effector protein, α1H and α2H, having a similar activity like Tat peptide to transfer antibiotic gentamicin in endothelial and epithelial cells. Further- more, similar to DMD treatment use of PPMOs against viral infec- tions is also reported to be useful in current clinical studies [67, 68].

3.4 Effectivity in Irrespective of the nature of stroke, i.e., ischemic or hemorrhagic, Cardiological CPP-conjugated compounds are also considered as an effective Disorders approach to selective drug delivery. For example, all these Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 121

covalently conjugated peptide inhibitors, protein kinase C (δV1-1) and D-JNKI-1 to TAT and hemagglutinin (HA) to a shorter frag- ment of TAT, are able to defend against cerebral ischemic reperfu- sion impairment [69–71]. In addition, studies have also found that administration of TAT-bound anti-apoptotic and neurotrophic fac- tors, Bcl-XL and GDNF just after the cerebral ischemia, can reduce the effect of severe brain damage [72]. It was earlier thought that CPP doesn’t have any control over the adverse effect of ischemic stroke. However, recently one TAT-mGluR1 fusion construct has been discovered where beside mGluR1, the protein transduction domain (PTD) of TAT itself shows neuroprotective effect through its active involvement in the regulation of membrane channel function [73].

3.5 In Topical Certain other successful therapeutic applications of CPPs could also Delivery of Therapeutic be found in medicinal research. For instance, penetratin [74], TAT Agents [75], polyarginine [76], and meganin [77] are known to be active in the transdermal delivery of drug molecules to dermal layers of skin’s protective barriers. However, considering the challenges imposed by various skin diseases, an arginine-rich CPP, IMT-P8 (derived from human protein), has been discovered and character- ized which shows better internalization into skin cells in compari- son to TAT or other CPPs [78].

3.6 Application of Apart from their use in several other diseases CPPs are also shown CPPs in CNS to have a positive impact in CNS drug delivery. For example, both Therapeutics TAT and penetratin peptide are rapidly being used as the molecular vehicle for BBB-impermeable therapeutic molecules. In another study, the use of dNP2 (another BBB-permeable CPP) in delivering cytoplasmic domain of Tc-lymphocyte antigen in brain tissue cells seems to be effective against autoimmune encephalomyelitis [79]. Similarly, it has also been seen that acetyltransferase-tagged TAT peptides not only experts BBB permeability, but also helps in the recovery of spatial memory in AD mouse. Numerous applications could be found on CPPs in the treat- ment of neurodegenerative diseases [80]. In spite of various appli- cations, limited information is available on CPP molecules in the presence of lipid membrane. Therefore the need for extensive studies about such CPPs in BBB mimicking model membrane environment has become crucial in advance therapeutics. Several low-resolution biophysical techniques (fluorescence/circular dichroism spectroscopy) can provide information about CPP bind- ing affinity, change in secondary conformation, solvent accessibility in membrane models, etc. Likewise, to evaluate the prominence of CPP residues in BBB membrane permeability, in vivo studies could also be accomplished using fluorescent-labeled CPP molecules in mice models. However, besides the low-resolution evidence, one must monitor the binding events between CPP and BBB 122 Swapna Bera and Anirban Bhunia

membrane models at the atomic level and for that nuclear magnetic resonance (NMR) technique is considered as the most useful one. Here, a detailed account of several powerful NMR experiments is provided in the context of their applications, significance, and precautions.

4 Characterization of Atomic Details About CPP-BBB Membrane Interaction Through High-Resolution NMR Techniques

As discussed earlier, the deeper understanding of such specific interaction between BBB membrane and CPP tagged-therapeutic agents at an atomic level needs serious consideration in improvising CNS-directed therapeutics. Undoubtedly, NMR has emerged as one of the powerful methods that led to a broader recognition of the structural, conformational, and dynamic characteristics of the interacting molecules having biomedical significance at atomic res- olution. Unlike other techniques that are applied for structural studies including X-ray crystallography, circular dichroism, and mass spectroscopy, NMR has an overall advantage of accruing conformational and dynamical information on most of the individ- ual nuclear sites in the molecule at the same time and that too for picosecond to hour timescales [81, 82]. These unique features of NMR methods, however, make them approachable and useful for extensive studies of both ligand and receptor molecules in their native form. For example, the information related to the three- dimensional conformation, specific interacting partner, epitope mapping, change in dynamics, relaxation rates, and diffusion coef- ficients could be easily acquired through several one- and two-dimensional based NMR techniques. Several structural studies through NMR have already been performed on various CPP molecules in the presence of model membrane and all these attempts are proven to be successful in their respective context [64, 83–86]. Usually, two major experi- mental approaches are believed to be there in NMR, one is ligand- based and another one is receptor-based methods. While a large number of ligand parameters can be directly assessed through the chemical shift perturbation and line broadening of ligand signals, analysis of the ligand-induced effect on receptor molecules is also shown to be possible through NMR studies. We have, therefore, carefully chosen few useful NMR techniques in the context of their principle, application, and limitations that can serve the purpose of exploring characteristic features of both CPP and membrane really well (Fig. 1). Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 123

Fig. 1 A flowchart representation of several useful NMR techniques that can be significant enough to explore the interaction study of both CPP and membrane at atomic resolution

4.1 In-Depth The discrimination between specific CPP and model BBB lipid Analysis of CPP membrane interactions from the nonspecific interaction can be Parameters in the easily made through either chemical shift perturbations or peak 1 Presence of BBB broadening of H NMR spectra of CPP molecules in the presence Membrane Model at of BBB membrane model. Thus, the understanding of overall Atomic Resolution proton nuclei connectivity (both through bond and space connec- tivity) of CPP molecules in membrane environment has become 4.1.1 Molecular much easier now using various two-dimensional correlation NMR Recognition Between CPP methods, total correlation spectroscopy (TOCSY), and nuclear and BBB Membrane Overhauser effect spectroscopy (NOESY). And, based on the pro- Models Through ton connectivity map, the refined structural calculation can be Solution NMR accomplished [81, 87]. However, the size of CPPs always remains challenging using homonuclear two-dimensional NMR experiment and hence the implementation of two-dimensional 1H–15N hetero- nuclear single-quantum correlation (HSQC) NMR could be useful to identify the backbone amide proton of each amino acid residue [88, 89]. Thus, much more resolved cross peaks for a uniformly labeled 15N sample at very low concentration can be obtained within a short timescale. Recently, this method is further impro- vised to band-selective optimized-flip-angle short transient (SOFAST) heteronuclear multiple quantum correlation (HMQC) 124 Swapna Bera and Anirban Bhunia

for the purpose of faster data acquisition and increased signal-to- noise (S/N) ratio [90]. Recently developed HNN NOESY (also known as HSQC-NOESY-HSQC) is relevant to get the backbone amide assignment of a uniformly 15N-labeled peptide [91]. Therefore, implying all these routine NMR experiments, one can acquire various information about the CPP-BBB membrane interaction. For instance, the membrane-induced changes in chem- ical shift values and relaxation rates of CPP molecules can be directly assessed through the NMR cross-peak signals, provided that assignment for each nuclei is known. However, one should be concerned about certain aspects while performing the experiments: 1. NMR has a size limit of approximately 50 kDa for the protein samples beyond which the signals can’t be acquired because of the fast T2 relaxation of the proton nuclei. 2. To circumvent the signal ambiguities often seen in overlapped spectra of a large peptide it is always preferable to go for 15N- labeled samples. 3. The cross-peak resonances are very sensitive to several para- meters including sample concentration, temperature, pH, etc., which should be optimized carefully. Preferably, the solution NMR experiments for peptide resonance assignment are per- formed at low temperature to avoid the conformational exchange (however, for some peptides, the exchange can be prominent at low temperature). 4. The selection of accurate mixing time for all these two-dimensional experiments also needs to be done cautiously to evade the spurious effect of spin diffusion (applicable while doing the transfer NOESY experiments in the presence of membrane). However, in case of free peptide in solution, the NOESY experiments with low mixing time (<120 ms) aid the assignment process through providing sequential NOEs, whereas long mixing time (>120 ms) to obtain the NOE constraints for structural modeling of free peptide.

4.1.2 Determining Higher Despite the preliminary information about the CPP-BBB mem- Resolution Solution brane model interaction, the membrane-induced high-resolution Structure of CPP on BBB conformation of CPP molecules can be determined using trans- Membrane Model Through ferred NOESY (tr-NOESY) NMR which mainly relies on the Transferred NOESY nuclear Overhauser effect (NOE) of NMR-active nuclei (tr-NOESY) NMR [92, 93]. tr-NOESY is one such powerful technique, which holds Experiment their uniqueness by emphasizing the conformational transition of any ligand molecules upon interaction with macro-sized receptor molecules in NMR timescale [94, 95]. The structural studies through tr-NOESY mainly rely on the appearance of comparatively stronger tr-NOE cross peaks for receptor-bound ligand molecules, Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 125

which are initially absent in their free form. And it happens because of the increased rotational correlation time (τc) of the small ligand molecules, possessing similar motional features as the receptor in their bound state [96, 97], thus unveiling fast NOE buildups as a consequence of the spin diffusion effect. It is believed that the ligand-receptor interactions, which are of fast kinetics, are an ideal system for the tr-NOESY experiment. Considering that, one can easily avail BBB membrane-induced three-dimensional conforma- tion of CPP molecules with the help of tr-NOE peaks appeared in the tr-NOESY spectra. However, one must be careful enough about the relaxation timescale being comparatively slow than the chemical exchange between the bound and free forms of ligand for the successful execution of tr-NOESY. Besides, spurious cross peaks can also be seen from the interference of receptor molecules due to spin diffusion effects.

Precautions 1. The ligand molecules should be small enough. For example, large ligand molecules cannot be studied because of their high rotational correlation time. 2. The experiment deals with a small amount of ligand sample (isotopically unlabeled). However, the appearance of tr-NOEs is highly influenced by the ligand/receptor ratio, mostly being excess ligand concentration over receptor molecules (usually, 5–50 times more, depending on receptor-binding affinity). 3. Ligand-receptor-binding affinity and kinetic parameters need to be ensured properly before performing the experiment. In general, medium- to low-range receptor affinity (within μMto mM range) is preferred. 4. NMR mixing time plays a significant role in tr-NOE buildup. Therefore, the selection of proper mixing time is very crucial here. In general, short NOESY mixing times of ~80–100 ms for the complex are mostly preferred to avoid spin diffusion. 5. Correct assignment of all the tr-NOEs in the spectra is highly requisite for accurate structural calculation of ligand molecules, which sometimes creates difficulty but not so impossible [98–100]. So far, many successful attempts could be seen on determining the solution conformation of membrane-bound ligand molecules. Recently, in our lab, we have solved the three-dimensional structure of penetratin peptide in three different BBB mimicking membrane model systems using tr-NOESY experiment [101]. For our work, penetratin CPP was chosen purposely owing to their intrinsic property to cross the tight endothelial network of BBB, and thus showed the prominence of using them as a vehicle for CNS thera- peutic drug delivery. However, a lack of structural evidence about the penetratin peptide in BBB membrane could be found till date. 126 Swapna Bera and Anirban Bhunia

Hence, we performed a series of biophysical experiments with three different large unilamellar vesicles (LUVs), i.e., ganglioside GM1, POPC/POPG/cholesterol/GM1 (7.65:0.85:1:0.5), and total brain lipid extract (TLBE) along with in vivo studies. All the LUVs were prepared with the consideration that they mimic the BBB membrane environment closely. Here, the preliminary fluo- rescence data provided the glimpse of penetratin peptide’s prefer- ential interaction with GM1 and POPC/POPG/cholesterol/GM1 membrane system compared to TLBE LUVs, which was further emphasized by CD data. Later, the tr-NOESY spectra revealed the secondary conformational information about DK17 in each model membrane. While in ganglioside GM1 LUVs, penetratin adopted a definite α-helical conformation; the α-helical propensity was only seen in the N-ter segment of the peptide (Arg2–Arg12) in the POPC/POPG/cholesterol/GM1 membrane system. Unlike the above two membrane mimics, the CPP mostly remained in their random coil conformation in TLBE except for the central region (Trp7-Arg11). Surprisingly, the central hydrophobic segment (Ile6-Arg11) of penetratin peptide retained their structural integ- rity irrespective of membrane composition, signifying their role in BBB penetration which was further proven to be right in in vivo studies with mice model. The in vivo studies were mainly carried out for a better understanding of the structural-functional events under physiological condition and for that the central Ile6-Arg11 section of penetratin was replaced with alanine residues. In the coarse of the experiment, 5–6-week-old female SJL/J mice were individually injected with Alexa Fluor 680 (an infrared dye)-tagged wild-type and mutant penetratin peptide [102]. As expected, unlike the wild-type peptide, the mut-penetratin had failed to enter the central nervous system, thus emphasizing the contribution of hydrophobic interaction between CPP and BBB, in BBB perme- ability (Fig. 2). Apart from that the structural studies also helped to identify the driving force behind the CPP-membrane interaction and conformational switching. This is somehow a preliminary work, which can be extended further prior to the need. Overall, these data highlighted the structural importance of penetratin pep- tide in its corresponding functional traits that may permit designing novel CPP for successful drug delivery in CNS.

4.1.3 Epitope Mapping The epitope mapping of CPP in BBB membrane model can be CPP-BBB Membrane Model mostly resolved through saturation transfer difference (STD) Interactions in Solution NMR experiments [97, 103]. Similar like tr-NOESY, STD also deals with the change in ligand’s rotation correlation time upon its preferential interaction with counterpart receptor molecules [104]. In principle, the magnetization from the selectively saturated receptor molecules is first transferred to the nearby pro- tons of interacting ligand molecules and then saturates to the rest of Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 127

Fig. 2 The overlapped high-resolution, three-dimensional solution structure of DK17 peptide in three different BBB mimicking model membrane systems, ganglioside GM1 (green), POPC/POPG/cholesterol/GM1 (7.65:0.85:1:0.5) (cyan), and total brain lipid extract (TLBE) (purple), provides significant information about the central segment (Ile6–Arg11) being structurally conserved throughout in all lipid environments despite their considerable conformational variation (a). To prove it further, both AF680-conjugated wt-DK17 (red) and mut-DK17 (blue) peptide (b) were injected to 5–6-week-old female SJL/J mice and more to our surprise AF680 signal was found in the different parts of the brain and spinal cord for the wt-DK17. However, the passage of mut-DK17 seems to be restricted only in the tail region (c). CD spectra of mut-DK17 in both the absence and presence of membrane further support the in vivo results showing majorly random coil conformation of peptide in the POPC/POPG/cholesterol/GM1 LUVs, specifying the reason of its missing translocation ability to CNS (d). Figure is adopted with permission from [102], Copyright ©2016, American Chemical Society

the ligand. This helps in creating an epitope map for the specific molecular recognition site of ligand molecule when bound to its specific receptors [105, 106]. Likewise, identifying membrane- binding site of CPP molecules can easily be achieved through STD NMR, provided that binding has taken place.

Precautions 1. Isotopically unlabeled ligand sample is enough for STD experi- ments [107]. However, for successful execution of the experi- ment, the molar concentration of CPP molecule has to be always much higher (molar ratio of receptor:ligand ¼ 1:100). 128 Swapna Bera and Anirban Bhunia

2. STD experiment is feasible for low-intermediate binding affin- ity systems. 3. Careful selection of the on- and off-resonance saturation fre- quencies and saturation time (tsat) is highly recommended to avoid false results. 4. In many occasions, “spin-lock” filter can be used to suppress the “unwanted” background membrane signals.

4.1.4 Characterization In addition to the three-dimensional structure of CPPs in BBB of Conformational model membrane mimic, the membrane-binding site of CPP mole- Dynamics Using Relaxation cules can also be characterized by tracking the changes in peptide Experiments backbone dynamics through several 1D/2D-based relaxation experiment [108–110]. Here, a series of sequential delays are applied for the measurement of both spin-lattice (T1) and spin- spin (T2) relaxation. Depending on the changes in the longitudinal (R1) and transverse (R2) relaxation rate profile of CPP residues upon membrane binding, one can easily differ the peptide region that is directly associated with the BBB membrane model. How- ever, the accessibility of spin relaxation studies has a limit of review- ing molecular motions on picosecond-nanosecond (ps-ns) timescale only. Therefore, one can additionally perform CPMG (Carr-Purcell-Meiboom-Gill) relaxation dispersion experiment so as to evaluate the chemical exchange between free and membrane- bound CPP molecules that occurs in the range of microsecond to millisecond (μs-ms) timescale [111, 112]. In CMPG experiment, a series of 180 pulses with different magnitude are applied during a fixed relaxation delay [113]. And depending on the significantly different relaxation dispersion profile of both free and bound pep- tide, it would be easier to characterize the exchange rate (Rex) and the population of two distinct conformations (major/minor state).

Precautions 1. The two-dimensional experiments, T1/T2 relaxation and CMPG, are carried out using 1H–15N HSQC method with 15N-labeled sample. 2. For successful execution of relaxation dispersion experiment, transient interaction between CPP and BBB membrane models is a prerequisite. 3. Relaxation experiments need to be run separately for both free and membrane-bound CPP molecules and then only the changes in relaxation parameters can be obtained. 4. While performing the relaxation experiment, the delay values should be chosen carefully. 5. Correct peak assignment and intensity calculation are highly requisite for the determination of relaxation parameters. Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 129

4.1.5 Probing the Solvent Paramagnetic relaxation enhancement (PRE) is another important Accessibility of CPP NMR method to acquire information about the orientation of CPP Molecules in their BBB molecules in their membrane-bound state [110]. Here, the inter- Membrane Model-Bound molecular distances between studying nucleus and paramagnetic Conformation center can easily be monitored using a different paramagnetic probe. The PRE effect mainly arises due to the dipolar interaction between NMR-active nucleus and unpaired electrons of the para- magnetic probe causing perturbations of peak intensity, which enhanced the spin relaxation rate of nearby ligand nuclei [113]. Thus the CPP residues, which are not in direct contact with membrane, will get more affected by the PRE effect compared to the one directly associated with the BBB membrane model. Taken together with the identification of CPP residues responsible for the preliminary BBB membrane-binding event, PRE NMR can be counted as the effective one [114–116].

Precautions 1. 2D 1H–1H TOCSY or 1H–15N HSQC cross-peaks are obtained before (Io) and after (I) addition of paramagnetic molecules like MnCl2 or spin labeled lipids such as 5-doxyl stearic acid (5- DSA) or 16- doxyl stearic acid (16-DSA), keeping all the para- meters constant. The intensity ratio (I/Io) is considered as remaining amplitude and can be plotted against residues. 2. The concentration of both CPP and BBB membrane model should be considerable enough for the successful binding event. 3. Hence, one must be careful enough about the added concen- tration of paramagnetic probe; otherwise signals will be lost due to excessive peak broadening.

4.1.6 Information About Residual dipolar coupling (RDC) NMR experiment has already Molecular Alignment of succeeded in providing information about the alignment of all CPP Molecule on BBB probable bond angles of ligand molecules in their major orientation Membrane Model Using frame [97, 113]. Here, the measurement of RDC parameter is Residual Dipolar Coupling mainly carried out through the molecular alignment of NMR sam- (RDC) ple in both the isotropic and anisotropic environment, under the influence of strong external magnetic field [117–119]. According to principle, a notable difference in J-couplings should be noticed definitely for ligand molecules if placed under an anisotropic atmo- sphere, which further causes a difference in RDC profile [120]. Thus, the purpose of monitoring the direct change in molecular orientation of CPP molecules under anisotropic (in the presence of BBB membrane model) media can be thrived easily using RDC experiment.

Precautions RDC signals are achieved mainly through the molecular average of signals of both free and bound conformation and hence suffer limitations in determining the orientation of large and multi- domain proteins if a remarkable ligand proportion is not bound. 130 Swapna Bera and Anirban Bhunia

Therefore, it is always recommended to implement RDC along with other NMR experiments to attain accurate information about the bound state conformations.

4.1.7 Defining the The ligand’s interaction with its receptor molecules is already Deviations in Diffusion known to have a significant contribution in ligand’s diffusion Coefficient Values of CPP rates through the solution [121]. And interestingly, the change in Molecules upon Interaction the diffusion coefficient values of those bound ligand can directly with BBB Membrane Model be figured out through a modern NMR technique, pulsed field gradient spin echo (PFGSE), also termed as diffusion ordered spectroscopy (DOSY) NMR [122, 123]. Earlier it was widely used to separate various components from the mixture by their inherent physical properties, i.e., diffusion rate. However, it may also be proven worthwhile for supervising the effect of molecular interaction between CPP and BBB membrane model on the diffu- sion rate of CPP. In addition, what makes them more convenient for use is the hydrodynamic radius and molecular weight of the CPP-BBB membrane model complex; both can be derived using the diffusion coefficient value. Surprisingly, two-dimensional DOSY NMR also seemed to have an advantage in the characteriza- tion of LUVs upon interaction with CPP molecules involving size distributions, etc. [124].

Precautions While performing the experiment, few parameters should be seri- ously taken into consideration including the delay time between each applied gradient pulses (when the diffusion of mixture com- ponents takes place), gradient length, and gradient power. Further- more, the temperature, 90 pulse length, and pulse power should also be optimized during the experiment. However, once the data acquisition is made for both the free and receptor-bound CPP, the data processing, calculation, and analysis can be executed using appropriate equation.

4.2 Exploring the Apart from defining ligand alignment, conformation, and dynamics CPP-Induced Effects in the membrane atmosphere, the understanding of ligand- on BBB Mimicking mediated changes in various membrane parameters has also drawn Model Membranes at significant attention so as to gain complete insights into ligand- 31 Atomic Resolution membrane interactions [125–127]. P NMR is one such widely used technique that directly assists in exploring lipid membrane simply by using 31P of lipid phosphate head groups (i.e., the NMR-active nuclei) as a magnetic probe [128]. Here, relying on the chemical shift deviation as well as peak broadening of 31P NMR spectra, explaining the change in orientation of phospholipid head groups has become much easier [129]. Besides, what raises the interest of using 31P NMR more is their broad application in drug-membrane interactions along with studying other ligand- membrane interactions. For example, the degree of CPP effects Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 131

on BBB membrane model surface, their distribution, and mobility within the membrane, all these can be successfully monitored through the changes in 31P NMR spectra of LUVs, provided that they interacted with each other. However one can also perform paramagnetic quenching NMR experiments using a paramagnetic quencher, Mn2+ ions, to study whether the CPP under analysis has any role in BBB membrane pore formation or not, as there is always a chance of 31P signal broadening for inner membrane leaflet as a consequence of paramagnetic quenching, if pore formation has ever taken place. However no such 31P resonance broadening or decreased intensity now could be seen for the same part of the membrane. Overall, 31P NMR seems to be useful enough for analyzing membrane property in details.

5 Conclusions

In summary, this chapter discusses mitigation strategies and ther- anostics against BBB-directed neurodegenerative diseases using cell-penetrating peptides. It highlights the need for high-resolution structural analysis of CPPs in model membranes to further expand the utility of CNS therapeutics, the rationale being that the biological function of any biomolecule can never be understood thoroughly without the complete analysis of its structure and dynamic features at an atomic level. NMR techniques are always preferred over other classical methods for acquiring atomic resolu- tion information from dynamic biological samples. By implement- ing the advanced NMR techniques that are discussed here, it should be possible to obtain detailed information about the 3D conforma- tions, binding pockets, orientations, backbone dynamics, and relax- ation rates of CPP molecules in their membrane-bound states (i.e., BBB membrane model). These robust atomic-level details of CPP’s interaction in various lipid environments may be helpful for the rational design of new-generation therapeutic drugs for treatment of deadly neurodegenerative diseases.

Acknowledgments

This research was supported by Institutional fund (Plan Project-II), CSIR, and Indo-Swedish (DST-VR) research grant. S.B. thanks CSIR-UGC, Govt. of India, for senior research fellowship. A.B. thanks Prof. Kalipada Pahan, Rush University Medical Center, for in vivo data. 132 Swapna Bera and Anirban Bhunia

References

1. Daneman R, Prat A (2015) The blood-brain 17. Tarasoff-Conway JM et al (2015) Clearance barrier. Cold Spring Harb Perspect Biol 7(1): systems in the brain-implications for Alzhei- a020412 mer disease. Nat Rev Neurol 11(8):457–470 2. Saunders NR et al (2016) The biological sig- 18. Ulrich JD, Huynh TP, Holtzman DM (2015) nificance of brain barrier mechanisms: help or Re-evaluation of the blood-brain barrier in hindrance in drug delivery to the central ner- the presence of Alzheimer’s disease pathology. vous system? F1000Res 5:F1000 Faculty Rev- Neuron 88(2):237–239 313 19. Masserini M (2013) Nanoparticles for brain 3. Abbott NJ (2005) Dynamics of CNS barriers: drug delivery. ISRN Biochem 2013:238428 evolution, differentiation, and modulation. 20. Choonara YE et al (2016) Improving drug Cell Mol Neurobiol 25(1):5–23 delivery technology for treating neurodegen- 4. Zenaro E, Piacentino G, Constantin G (2017) erative diseases. Expert Opin Drug Deliv 13 The blood-brain barrier in Alzheimer’s dis- (7):1029–1043 ease. Neurobiol Dis 107:41–56 21. Oller-Salvia B et al (2016) Blood-brain barrier 5. Patel MM, Patel BM (2017) Crossing the shuttle peptides: an emerging paradigm for blood-brain barrier: recent advances in drug brain delivery. Chem Soc Rev 45 delivery to the brain. CNS Drugs 31 (17):4690–4707 (2):109–133 22. Chen H et al (2014) A new brain drug deliv- 6. Antoniou X, Borsello T (2010) Cell perme- ery strategy: focused ultrasound-enhanced able peptides: a promising tool to deliver neu- intranasal drug delivery. PLoS One 9(10): roprotective agents in the brain. e108880 Pharmaceuticals (Basel) 3(2):379–392 23. Copolovici DM et al (2014) Cell-penetrating 7. Savelieff MG, Lee S, Liu Y, Lim MH (2013) peptides: design, synthesis, and applications. Untangling amyloid-β, tau, and metals in Alz- ACS Nano 8(3):1972–1994 heimer’s disease. ACS Chem Biol 8 24. Lindgren M et al (2000) Cell-penetrating (5):856–865 peptides. Trends Pharmacol Sci 21 8. Marques F et al (2013) Blood-brain-barriers (3):99–103 in aging and in Alzheimer’s disease. Mol Neu- 25. Lindgren M, Langel U (2011) Classes and rodegener 8:38 prediction of cell-penetrating peptides. Meth- 9. Bowman GL, Quinn JF (2008) Alzheimer’s ods Mol Biol 683:3–19 disease and the blood-brain barrier: past, pres- 26. Kauffman WB et al (2015) Mechanism mat- ent and future. Aging Health 4(1):47–55 ters: a taxonomy of cell penetrating peptides. 10. Hamley IW (2012) The amyloid beta peptide: a Trends Biochem Sci 40(12):749–764 chemist’s perspective. Role in Alzheimer’s and 27. Green M, Loewenstein PM (1988) Autono- fibrillization. Chem Rev 112(10):5147–5192 mous functional domains of chemically 11. Provias J, Jeynes B (2014) The role of the synthesized human immunodeficiency virus blood-brain barrier in the pathogenesis of tat trans-activator protein. Cell 55 senile plaques in Alzheimer’s disease. Int J (6):1179–1188 Alzheimers Dis 2014:191863 28. Frankel AD, Pabo CO (1988) Cellular uptake 12. Brightman MW (1977) Morphology of of the tat protein from human immunodefi- blood-brain interfaces. Exp Eye Res 25 ciency virus. Cell 55(6):1189–1193 (Suppl):1–25 29. Prochiantz A (1999) Homeodomain-derived 13. Schlosshauer B (1993) The blood-brain bar- peptides. In and out of the cells. Ann N Y rier: morphology, molecules, and neurothelin. Acad Sci 886:172–179 BioEssays 15(5):341–346 30. Joliot A et al (1991) Antennapedia homeobox 14. Patel MM et al (2009) Getting into the brain: peptide regulates neural morphogenesis. Proc approaches to enhance brain drug delivery. Natl Acad Sci U S A 88(5):1864–1868 CNS Drugs 23(1):35–58 31. Pooga M et al (1998) Cell penetration by 15. Stewart PA (2000) Endothelial vesicles in the transportan. FASEB J 12(1):67–77 blood-brain barrier: are they related to perme- 32. Elliott G, O’Hare P (1997) Intercellular traf- ability? Cell Mol Neurobiol 20(2):149–163 ficking and protein delivery by a herpesvirus 16. Banks WA (2016) From blood-brain barrier structural protein. Cell 88(2):223–233 to blood-brain interface: new opportunities 33. Chaloin L et al (1998) Design of carrier for CNS drug delivery. Nat Rev Drug Discov peptide-oligonucleotide conjugates with 15(4):275–292 rapid membrane translocation and nuclear Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 133

localization properties. Biochem Biophys Res 47. Futaki S, Hirose H, Nakase I (2013) Commun 243(2):601–608 Arginine-rich peptides: methods of transloca- 34. Oehlke J et al (1998) Cellular uptake of an tion through biological membranes. Curr alpha-helical amphipathic model peptide with Pharm Des 19(16):2863–2868 the potential to deliver polar compounds into 48. Wu X, Gehring W (2014) Cellular uptake of the cell interior non-endocytically. Biochim the Antennapedia homeodomain polypeptide Biophys Acta 1414(1–2):127–139 by macropinocytosis. Biochem Biophys Res 35. Snyder EL, Dowdy SF (2004) Cell Commun 443(4):1136–1140 penetrating peptides in drug delivery. Pharm 49. Zahid M, Robbins PD (2015) Cell-type spe- Res 21(3):389–393 cific penetrating peptides: therapeutic pro- 36. Derossi D et al (1994) The third helix of the mises and challenges. Molecules 20 Antennapedia homeodomain translocates (7):13055–13070 through biological membranes. J Biol Chem 50. Orange JS, May MJ (2008) Cell penetrating 269(14):10444–10450 peptide inhibitors of nuclear factor-kappa 37. Derossi D et al (1996) Cell internalization of B. Cell Mol Life Sci 65(22):3564–3591 the third helix of the Antennapedia homeo- 51. Sayers EJ et al (2014) Distal phenylalanine domain is receptor-independent. J Biol Chem modification for enhancing cellular delivery 271(30):18188–18193 of fluorophores, proteins and quantum dots 38. Deshayes S et al (2005) Cell-penetrating pep- by cell penetrating peptides. J Control Release tides: tools for intracellular delivery of thera- 195:55–62 peutics. Cell Mol Life Sci 62(16):1839–1849 52. Stewart KM, Horton KL, Kelley SO (2008) 39. Ramsey JD, Flynn NH (2015) Cell- Cell-penetrating peptides as delivery vehicles penetrating peptides transport therapeutics for biology and medicine. Org Biomol Chem into cells. Pharmacol Ther 154:78–86 6(13):2242–2255 40. Heitz F, Morris MC, Divita G (2009) Twenty 53. Yang Y et al (2014) PEGylated liposomes with years of cell-penetrating peptides: from NGR ligand and heat-activable cell- molecular mechanisms to therapeutics. Br J penetrating peptide-doxorubicin conjugate Pharmacol 157(2):195–206 for tumor-specific therapy. Biomaterials 35 41. Bahnsen JS et al (2013) Antimicrobial and (14):4368–4381 cell-penetrating properties of penetratin ana- 54. Raucher D, Ryu JS (2015) Cell-penetrating logs: effect of sequence and secondary struc- peptides: strategies for anticancer treatment. ture. Biochim Biophys Acta 1828 Trends Mol Med 21(9):560–570 (2):223–232 55. Bolhassani A (2011) Potential efficacy of cell- 42. Alves ID et al (2009) The interaction of cell- penetrating peptides for nucleic acid and drug penetrating peptides with lipid model systems delivery in cancer. Biochim Biophys Acta and subsequent lipid reorganization: thermo- 1816(2):232–246 dynamic and structural characterization. J 56. Dissanayake S et al (2017) Recent develop- Pept Sci 15(3):200–209 ments in anticancer drug delivery using cell 43. Alves ID et al (2011) Relationships between penetrating and tumor targeting peptides. J membrane binding, affinity and cell internali- Control Release 250:62–76 zation efficacy of a cell-penetrating peptide: 57. Ueda Y et al (2012) Induction of autophagic penetratin as a case study. PLoS One 6(9): cell death of glioma-initiating cells by cell- e24096 penetrating D-isomer peptides consisting of 44. Jobin ML, Alves ID (2014) On the impor- Pas and the p53 C-terminus. Biomaterials 33 tance of electrostatic interactions between cell (35):9061–9069 penetrating peptides and membranes: a path- 58. Araki D et al (2010) Cell-penetrating D-iso- way toward tumor cell selectivity? Biochimie mer peptides of p53 C-terminus: long-term 107(Pt A):154–159 inhibitory effect on the growth of bladder 45. Patel LN, Zaro JL, Shen WC (2007) Cell cancer. Urology 75(4):813–819 penetrating peptides: intracellular pathways 59. Snyder EL et al (2004) Treatment of terminal and pharmaceutical perspectives. Pharm Res peritoneal carcinomatosis by a transducible 24(11):1977–1992 p53-activating peptide. PLoS Biol 2(2):E36 46. Mayor S, Pagano RE (2007) Pathways of 60. Aroui S et al (2009) Efficient induction of clathrin-independent endocytosis. Nat Rev apoptosis by doxorubicin coupled to cell- Mol Cell Biol 8(8):603–612 penetrating peptides compared to 134 Swapna Bera and Anirban Bhunia

unconjugated doxorubicin in the human 74. Lee J et al (2005) Transdermal delivery of breast cancer cell line MDA-MB 231. Cancer interferon-gamma (IFN-gamma) mediated Lett 285(1):28–38 by penetratin, a cell-permeable peptide. Bio- 61. Aroui S et al (2010) Cytotoxicity, intracellular technol Appl Biochem 42(Pt 2):169–173 distribution and uptake of doxorubicin and 75. Kamada H et al (2007) Creation of novel cell- doxorubicin coupled to cell-penetrating pep- penetrating peptides for intracellular drug tides in different cell lines: a comparative delivery using systematic phage display tech- study. Biochem Biophys Res Commun 391 nology originated from tat transduction (1):419–425 domain. Biol Pharm Bull 30(2):218–223 62. Cheng H et al (2015) Activable cell- 76. Rothbard JB et al (2000) Conjugation of argi- penetrating peptide conjugated prodrug for nine oligomers to cyclosporin A facilitates tumor targeted drug delivery. ACS Appl topical delivery and inhibition of inflamma- Mater Interfaces 7(29):16061–16069 tion. Nat Med 6(11):1253–1257 63. Lo¨fgren K et al (2008) Antiprion properties of 77. Kim YC, Ludovice PJ, Prausnitz MR (2007) prion protein-derived cell-penetrating pep- Transdermal delivery enhanced by magainin tides. FASEB J 22(7):2177–2184 pore-forming peptide. J Control Release 122 64. Ezzat K et al (2011) PepFect 14, a novel cell- (3):375–383 penetrating peptide for oligonucleotide deliv- 78. Gautam A et al (2016) Topical delivery of ery in solution and as solid formulation. protein and peptide using novel cell Nucleic Acids Res 39(12):5284–5298 penetrating peptide IMT-P8. Sci Rep 65. Garibotto FM et al (2011) Penetratin analo- 6:26278 gues acting as antifungal agents. Eur J Med 79. Lim S et al (2015) dNP2 is a blood-brain Chem 46(1):370–377 barrier-permeable peptide enabling ctCTLA- 66. Gomarasca M et al (2017) Bacterium-derived 4 protein delivery to ameliorate experimental cell-penetrating peptides deliver gentamicin autoimmune encephalomyelitis. Nat Com- to kill intracellular pathogens. Antimicrob mun 6:8244 Agents Chemother 61(4). https://doi.org/ 80. Lin T et al (2016) Nose-to-brain delivery of 10.1128/AAC.02545-16 macromolecules mediated by cell-penetrating 67. Delcroix M, Riley LW (2010) Cell- peptides. Acta Pharm Sin B 6(4):352–358 penetrating peptides for antiviral drug devel- 81. Keeler J (2005) Understanding NMR spec- opment. Pharmaceuticals (Basel) 3 troscopy. Wiley, Chichester (3):448–470 82. Levitt HM (2008) Spin dynamics: basics of 68. Moulton JD, Jiang S (2009) Gene knock- nuclear magnetic resonance, 2nd edn. Wiley, downs in adult animals: PPMOs and vivo- Chichester morpholinos. Molecules 14(3):1304–1323 83. M€aler L, Gr€aslund A (2011) NMR studies of 69. Bright R et al (2004) Protein kinase C delta three-dimensional structure and positioning mediates cerebral reperfusion injury in vivo. J of CPPs in membrane model systems. Meth- Neurosci 24(31):6880–6888 ods Mol Biol 683:57–67 70. Vaslin A et al (2011) Excitotoxicity-induced 84. Maniti O et al (2010) Distinct behaviour of endocytosis mediates neuroprotection by the homeodomain derived cell penetrating TAT-peptide-linked JNK inhibitor. J Neuro- peptide penetratin in interaction with differ- chem 119(6):1243–1252 ent phospholipids. PLoS One 5(12):e15819 71. Hirt L et al (2004) D-JNKI1, a cell- 85. M€aler L (2013) Solution NMR studies of cell- penetrating c-Jun-N-terminal kinase inhibi- penetrating peptides in model membrane sys- tor, protects against cell death in severe cere- tems. Adv Drug Deliv Rev 65(8):1002–1011 bral ischemia. Stroke 35(7):1738–1743 86. Pourmousa M, Karttunen M (2013) Early 72. Kilic E, Kilic U, Hermann DM (2006) TAT stages of interactions of cell-penetrating pep- fusion proteins against ischemic stroke: cur- tide penetratin with a DPPC bilayer. Chem rent status and future perspectives. Front Phys Lipids 169:85–94 Biosci 11:1716–1721 87. Breton RC, Reynolds WF (2013) Using 73. Xu W, Zhou M, Baudry M (2008) Neuropro- NMR to identify and characterize natural pro- tection by cell permeable TAT-mGluR1 pep- ducts. Nat Prod Rep 30(4):501–524 tide in ischemia: synergy between carrier and 88. Castan˜ar L, Parella T (2015) Recent advances cargo sequences. Neuroscientist 14 in small molecule NMR: improved HSQC (5):409–414 Cell-Penetrating Peptides as Theranostics Against Impaired Blood-Brain... 135

and HSQMBC experiments. Annu Rep NMR 102. Ma L et al (2007) In vitro and in vivo evalua- Spectrosc 84:163–232 tion of Alexa Fluor 680-bombesin[7-14] 89. Kwan AH et al (2011) Macromolecular NMR NH2 peptide conjugate, a high-affinity fluo- spectroscopy for the non-spectroscopist. rescent probe with high selectivity for the FEBS J 278(5):687–703 gastrin-releasing peptide receptor. Mol Imag- 90. Schanda P, Kupce E, Brutscher B (2005) ing 6(3):171–180 SOFAST-HMQC experiments for recording 103. Bhunia A, Bhattacharjya S, Chatterjee S two-dimensional heteronuclear correlation (2012) Applications of saturation transfer dif- spectra of proteins within a few seconds. J ference NMR in biological systems. Drug Biomol NMR 33(4):199–211 Discov Today 17(9–10):505–513 91. Favier A, Brutscher B (2011) Recovering lost 104. Mayer M, Meyer B (1999) Characterization magnetization: polarization enhancement in of ligand binding by saturation transfer differ- biomolecular NMR. J Biomol NMR 49 ence NMR spectroscopy. Angew Chem Int (1):9–15 Ed 38:1784–1788 92. Bhunia A et al (2009) NMR structural studies 105. Pereira A et al (2009) Functional cell-based of the Ste11 SAM domain in the dodecyl screening and saturation transfer double- phosphocholine micelle. Proteins 74 difference NMR have identified haplosamate (2):328–343 A as a cannabinoid receptor agonist. ACS 93. Bhunia A et al (2011) NMR structures and Chem Biol 4(2):139–144 interactions of temporin-1Tl and temporin- 106. Haselhorst T, Lamerz AC, Itzstein M (2009) 1Tb with lipopolysaccharide micelles: mecha- Saturation transfer difference NMR spectros- nistic insights into outer membrane permea- copy as a technique to investigate protein- bilization and synergistic activity. J Biol Chem carbohydrate interactions in solution. Meth- 286(27):24394–24406 ods Mol Biol 534:375–386 94. Clore GM, Gronenborn AM (1969) Theory 107. Herfurth L et al (2005) Comparative epitope and applications of the transferred nuclear mapping with saturation transfer difference overhauser effect to the study of the confor- NMR of sialyl Lewis(a) compounds and deri- mations of small ligands bound to proteins. J vatives bound to a monoclonal antibody. J Magn Reson 48(3):402–417 Med Chem 48(22):6879–6886 95. Wang Z et al (1993) Membrane-bound con- 108. Farrow NA et al (1994) Backbone dynamics formation of a signal peptide: a transferred of a free and phosphopeptide-complexed Src nuclear Overhauser effect analysis. Biochem- homology 2 domain studied by 15N NMR istry 32(50):13991–13999 relaxation. Biochemistry 33(19): 96. Meyer B, Peters T (2003) NMR spectroscopy 5984–6003 techniques for screening and identifying 109. Jaremko Ł et al (2015) The quest for simplic- ligand binding to protein receptors. Angew ity: remarks on the free-approach models. J Chem Int Ed Engl 42(8):864–890 Phys Chem B 119(36):11978–11987 97. Calle LP, Can˜ada FJ, Jime´nez-Barbero J 110. Bera S et al (2016) Biophysical insights into (2011) Application of NMR methods to the the membrane interaction of the core study of the interaction of natural products amyloid-forming Aβ40 fragment K16-K28 with biomolecular receptors. Nat Prod Rep and its role in the pathogenesis of Alzheimer’s 28(6):1118–1125 disease. Phys Chem Chem Phys 18:16890 98. Kneller DGK, Irwin D (1993) UCSF Sparky 111. Franco R et al (2017) Probing conforma- an NMR display, annotation and assignment tional exchange dynamics in a short-lived pro- tool. J Cell Biochem 53(S17C):254 tein folding intermediate by real-time 99. Wu¨thrich K (1986) NMR of proteins and relaxation-dispersion NMR. J Am Chem Soc nucleic acids. Wiley, New York 139(3):1065–1068 100. Wishart DS et al (1995) 1H, 13C and 15N 112. Loria JP, Rance M, Palmer AG (1999) A relaxation-compensated random coil NMR chemical shifts of the com- À À À mon amino acids. I. Investigations of nearest- Carr Purcell Meiboom Gill sequence for neighbor effects. J Biomol NMR 5(1):67–81 characterizing chemical exchange by NMR spectroscopy. J Am Chem Soc 121 101. Bera S et al (2016) Structural elucidation of (10):2331–2332 the cell-penetrating penetratin peptide in model membranes at the atomic level: prob- 113. Chiliveri SC, Deshmukh MV (2016) Recent ing hydrophobic interactions in the blood- excitements in protein NMR: large proteins brain barrier. Biochemistry 55 and biologically relevant dynamics. J Biosci 41 (35):4982–4996 (4):787–803 136 Swapna Bera and Anirban Bhunia

114. Yamamoto K et al (2010) Use of a copper- by PFG-NMR using band-selective RF pulses. chelated lipid speeds up NMR measurements Eur Biophys J 43(6–7):331–339 from membrane proteins. J Am Chem Soc 123. Ghosh A et al (2016) NMR structure and 132(20):6929–6931 binding of esculentin-1a (1-21)NH2 and its 115. Battiste JL, Wagner G (2000) Utilization of diastereomer to lipopolysaccharide: correla- site-directed spin labeling and high- tion with biological functions. Biochim Bio- resolution heteronuclear nuclear magnetic phys Acta 1858(4):800–812 resonance for global fold determination of 124. Hinton PD, Johnson SC (1993) Diffusion large proteins with limited nuclear overhau- ordered 2D NMR spectroscopy of phospho- ser effect data. Biochemistry 39(18): lipid vesicles: determination of vesicle size dis- 5355–5365 tributions. J Phys Chem 97(35):9064–9072 116. Clore GM, Iwahara J (2009) Theory, practice, 125. Dufourc EJ et al (1992) Dynamics of phos- and applications of paramagnetic relaxation phate head groups in biomembranes. Com- enhancement for the characterization of tran- prehensive analysis using phosphorus-31 sient low-population states of biological nuclear magnetic resonance lineshape and macromolecules and their complexes. Chem relaxation time measurements. Biophys J 61 Rev 109(9):4108–4139 (1):42–57 117. Tjandra N, Bax A (1997) Direct measurement 126. Burnell EE, Cullis PR, de Kruijff B (1980) of distances and angles in biomolecules by Effects of tumbling and lateral diffusion on NMR in a dilute liquid crystalline medium. phosphatidylcholine model membrane Science 278(5340):1111–1114 31P-NMR lineshapes. Biochim Biophys Acta 118. de Alba E, Tjandra N (2004) Residual dipolar 603(1):63–69 couplings in protein structure determination. 127. Schubert R et al (1986) Structural changes in Methods Mol Biol 278:89–106 membranes of large unilamellar vesicles after 119. Deschamps M, Campbell ID, Boyd J (2005) binding of sodium cholate. Biochemistry 25 Residual dipolar couplings and some specific (18):5263–5269 models for motional averaging. J Magn Reson 128. Holzgrabe U, Wawer I, Diehl B (2008) NMR 172(1):118–132 spectroscopy in drug development and analy- 120. Chen K, Tjandra N (2012) The use of residual sis. Wiley. WILEY-VCH Verlag GmbH, D- dipolar coupling in studying proteins by 69469 Weinheim (Federal Republic of NMR. Top Curr Chem 326:47–67 Germany) 121. Johnson SC Jr (1999) Diffusion ordered 129. Traı¨kia M et al (2000) Formation of unilamel- nuclear magnetic resonance spectroscopy: lar vesicles by repetitive freeze-thaw cycles: principles and applications. Prog Nucl Magn characterization by electron microscopy and Reson Spectrosc 34:203–256 31P-nuclear magnetic resonance. Eur Bio- 122. Yao S et al (2014) Measuring translational phys J 29(3):184–195 diffusion coefficients of peptides and proteins Chapter 8

Microbial Translocation of the Blood-Brain Barrier

Charles T. Spencer and Mireya G. Ramos Muniz

Abstract

Microbial infections can transit from their peripheral sites of entry into the brain and central nervous system. Once there, they can wreak havoc on the control center for the entire body. The presence of culturable microbes in the brain and spinal cord is fairly easy to detect. Even unculturable microbes in the central nervous system can be detected by PCR and other genomic assays. However, determining the mechanism by which these microbes cross the blood-brain barrier requires a large collection of data. Herein, we delineate protocols for in vivo, ex vivo, and in vitro analyses to determine the mechanism by which microbes cross the blood-brain barrier. These experimental techniques include flow cytometric analyses of brain extracts, histological analyses of brain sections, and in vitro transwell co-culture systems that mimic the blood-brain barrier.

Key words Microbial translocation, Blood-brain barrier, Paracellular, Transcellular, Trojan horse

1 Introduction

For centuries, histological sections of brains from infected patients and/or animals have been examined under the microscope for centuries to identify both structures and abnormalities. Like- wise, identifying the presence of a microbial agent inside of the brain parenchyma is relatively straightforward. Outgrowth of the microbe demonstrates the presence of live microbes and was used by clinicians to analyze biopsies and cerebral spinal fluid (CSF). More recently, microscopy allows for the ability to visualize micro- bial invasion in the brain. Yet still, discriminating microbes from normal brain constituents in tissue sections required the increased sensitivity and resolution provided by fluorescent, electron, and confocal microscopy. While this achievement allows one to deter- mine the presence of a pathogen, it does not determine how that pathogen invaded the brain. More dynamic measurements are nec- essary to determine how microbes translocate across the blood- brain barrier.

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_8, © Springer Science+Business Media, LLC, part of Springer Nature 2019 137 138 Charles T. Spencer and Mireya G. Ramos Muniz

Three mechanisms are recognized by which microbes cross the blood-brain barrier: transcellular translocation, paracellular translo- cation, and the Trojan horse method of translocation [1, 2]. Bacteria and viruses are reported to utilize one, or multiple, of these path- ways to invade the brain [1, 3]. Determining the mechanism of translocation of a microbe across the blood-brain barrier requires the accumulation of data in support or rejection of each potential translocation pathway. Experiments to probe these pathways are generally broken down into either ex vivo (Sect. 3.1) or in vitro (Sect. 3.3). However, we advocate the inclusion of a third experimental design, not yet well received in the field (Sect. 3.2). The ex vivo protocol detailed here uses the power of flow cytometry for analysis of hundreds of thousands of cells. Depend- ing on the microbial load in the infection, finding the relatively few infected endothelial or neural cells in traditional histology sections can be a monumental undertaking. Flow cytometry, however, is designed to identify such rare events with statistical power. More- over, imaging flow cytometry combines the visual representation of histology with the analytical capability of flow cytometry. This combinatorial analysis provides powerful visualization of infected cells. In vitro analysis of microbial transport utilizes a transwell cul- ture system lined with endothelial cells to mimic the blood-brain barrier. Various in vitro cell culture models have been reported including static and dynamic models of monoculture and co-culture. Culture systems of brain endothelial cells with or with- out astrocyte co-culture can be established in vitro to mimic the blood-brain barrier. These have been expertly reviewed in [4] and will not be discussed here. The use of an endothelial cell mono- layer approximates the wall of the blood vessels in the brain. The tight junctions between blood vessel cells regulate the permeabil- ity of the blood-brain barrier allowing translocation of blood constituents into the brain. In vivo, this permeability is regulated and, as such, different permeabilities of the in vitro endothelial monolayer must also be considered. In vitro permeability is inversely correlated with electrical resistance measurements [5] and can be modulated by addition of GSK inhibitors to stimulate the Wnt/β-catenin pathways and tighten the junctions between endothelial cells [6]. Recognizing that in vitro assays may not entirely represent in vivo conditions, we advocate for similar analyses in vivo using dual labeled microbes. However, while direct ex vivo analyses of the presence of infected immune cells or endothelial cells suggest the Trojan horse or transcellular pathways, respectively, in vivo support for paracellular transport is difficult to determine. We maintain that concomitant infection with live microbes and administration of dead microbes can demonstrate paracellular transport as the live microbe interacts with the blood-brain barrier causing loosen- ing of endothelial tight junctions allowing passage of the dead Microbial Translocation of the Blood-Brain Barrier 139

microbe into the brain. Paracellular transport can further be sup- ported by co-incubation of tight endothelial cell monolayers with live and dead microbes. We have combined the multiple analytical techniques described herein to analyze bacterial translocation into the brain following peripheral inoculation. Throughout these protocols, we are focused on characterization of the bacterial infection Francisella tularensis. However, modifications to the specifics of these protocols would make them applicable to the analysis of other microbial species.

2 Materials

2.1 Cell Lines HMVEC-d Adenovirus-transduced Dermal Microvascular Endo- thelial Cells (Lonza, catalog #CC-2543). bEnd.3 middle T antigen-transformed Brain endothelialpolyoma (ATCC, catalog #CRL-2299). J virus immortalized C57BL/6 bone marrow-derived murine mac- rophage line (BEI Resources, catalog #NR-9456).

2.2 Materials 70 μm filter (Fisher Scientific, catalog #22363548). Dounce homogenizers coarse (A) and fine (B) (Kimble Kontes, catalog #8853000007). Positively charged glass slides (Thermo Scientific, catalog #6776214). Transwell chamber and transwell inserts (Costar, 0.4 μm catalog #3470, 3 μm catalog #3472, 8 μm catalog #3464). TEER unit and appropriate electrode (STX01 or STX03 for Multi- well 6-, 12-, and 24-well inserts/plates; STX00 for Multiwell 96-well cell culture plates). Perfusion needle.

2.3 Reagents EGM-2 BulletKit (Lonza, catalog #CC-3162). HBSS (w/o Ca2+,Mg2+). RPMI (w/o phenol red). Percoll (GE Healthcare, catalog #17-0891-02). Collagen 3.47 mg/mL (Corning, catalog #354236). GSK-3 inhibitor (BIO, Sigma-Aldrich, catalog #B1686). Mounting medium (O.C.T.). Hexanes (Sigma-Aldrich, catalog #227064). Triton X-100 (Fisher Scientific, catalog #BP151). Lysine (Sigma-Aldrich, catalog #L5501). Glutaraldehyde (Sigma-Aldrich, catalog #3802). Normal Donkey Serum (Abcam, catalog #ab7475). 140 Charles T. Spencer and Mireya G. Ramos Muniz

2.4 Animals All animal procedures should be approved by your Institutional Animal Care and Use Committee (IACUC) prior to experimenta- tion involving animals. It is important that you select an animal model that is appropriate for your experimental system. Most com- mon animal models in microbial research include, but are not limited to, BALB/c and C57BL/6. C57BL/6J mice (Jackson Laboratories, catalog #000664).

2.5 Antibodies/ BacLight Green (Molecular Probes/Thermo Fisher, catalog Stains #B-35000). BacLight Red (Molecular Probes/Thermo Fisher, catalog #B-35001). Dextran (Thermo Fisher, 3kD-TxRed—catalog #D3329, 10kD- TMR—catalog #D1816, 40kD-FITC—catalog #D1844). Wheat germ agglutinin (WGA, Invitrogen, catalog #W11263). Rabbit α-mouse Iba-1 (Wako, catalog #019-19741). Biotinylated donkey α-rabbit (Abcam, catalog #ab97062). Streptavidin AlexaFluor 488 (Thermo Fisher, catalog #S11223). DAPI (Invitrogen, catalog #D3571). α-Mouse CD16/CD32 (Tonbo Bioscience, catalog #40-0161, clone: 2.4G2). FITC α-mouse CD45 (Tonbo Bioscience, catalog #35-0451, clone: 30-F11). PE-CF594 α-mouse B220/CD45R (BD Bioscience, catalog #562290, clone: RA3-6B2). Pacific Blue α-mouse CD11b (Tonbo Bioscience, catalog #75-0112, clone: M1/70). BV510 α-mouse SiglecF (BD Bioscience, catalog #740158, clone: E50-2440). PerCP α-mouse CD3 (Tonbo Bioscience, catalog #40-0161, clone: 2.4G2). A647 α-mouse CD31 (BD Bioscience, catalog #563608, clone: 390). PE α-mouse Tie-2 (Abcam, catalog #ab95722, clone: TEK4). PE α-mouse GFAP (BD Bioscience, catalog #561483, clone: 1B4).

2.6 Preparation of 1. Decontaminate external surfaces of all vials, including the Medium and Buffers medium bottle, with ethanol or isopropanol.

2.6.1 Complete EGM-2 2. To formulate endothelial growth medium, transfer the entire Growth Media, per contents of the EGM-2 SingleQuots kit Supplements and Manufacturer’s Instructions Growth Factors containing hydrocortisone solution, GA-100, fetal bovine serum, human fibroblast growth factor basic (hFGFb), human vascular endothelial growth factor Microbial Translocation of the Blood-Brain Barrier 141

(hVEGF), analog of human insulin-like growth factor-1, long R3-IGF-1, ascorbic acid solution, human epidermal growth factor (hEGF), and heparin to basal medium with a pipette, and rinse each vial with medium (see Note 1). 3. Transfer the label provided with each kit to the basal medium bottle being supplemented. Use it to record the date and amount of each supplement added. Store at 4 C and use within 1 month. Do not freeze medium.

2.6.2 Fluorescent 1. Allow a vial of BacLight bacterial stain to warm to room Staining of Microbe Prior temperature before opening. to Infection (see Note 2) 2. Prepare the 1 mM stock solution of dye by dissolving the vial contents in 90 μL DMSO. 3. Prepare the 100 μM working solution of the BacLight bacterial stain by adding 2 μL of the 1 mM stock solution to 18 μLof DMSO in a microcentrifuge tube and mix well. 4. To prepare the bacteria sample, add 1 μL of the working dye solution to 1 mL of the bacteria sample. 5. Incubate for 15 min at room temperature. 6. Wash with buffer to remove excess dye.

2.6.3 Flow Cytometric 1. 1Â PBS pH 7.6 with 4% heat-inactivated fetal bovine serum Staining Buffer (FBS). 2. Add 2 mM EDTA if clumping of cells is an issue in cell suspen- sions (optional). 3. Filter through 0.2 μm filter and store at 4 C.

2.6.4 Antibody Staining For every 106 cells, combine 2 μg of each of the following anti- Cocktail for Flow Cytometry bodies in a final volume of 100 μL flow cytometry staining buffer (see Note 3) (above): α-mouse CD45-FITC. α-mouse B220/CD45R-PE-CF594. α-mouse CD11b-Pacific Blue. α-mouse SiglecF-BV510. α-mouse CD3-PerCP. α-mouse CD31-AlexaFluor647. α-mouse Tie-2-PE.

2.6.5 Fixation Buffer for 2% Paraformaldehyde or formalin in flow cytometric staining Flow Cytometry buffer. 142 Charles T. Spencer and Mireya G. Ramos Muniz

2.6.6 Prepare Isotonic ISP: 4.5 mL Percoll + 0.5 mL 10Â HBSS (5 mL total volume). Percoll (ISP) Gradients 70% ISP: 1.75 mL ISP + 0.625 mL 1Â HBSS (2.5 mL total volume (Volumes per Sample) per sample).

2.6.7 Postfix Solution 4% Paraformaldehyde + 12 g sucrose in PBS.

2.6.8 Cryoprotectant 30% Ethylene glycol + 20% glycerol in PBS. Solution

3 Methods

3.1 Basic Protocol 1. Inoculate C57BL/6J mice with 1 Â 106 cfu Francisella tularensis A: Flow Cytometry LVS in 50 μm sterile PBS via intradermal route (see Note 4). of Brain Homogenates 2. Allow the infection 3 days for the infection to spread from the Following Infection site of inoculation into the brain. This must be determined empirically for each infection, but a general rule of thumb is 2–5 days post-inoculation. 3. Euthanize the animal according to approved veterinary prac- tices (e.g., isoflurane, carbon dioxide overdose) 4. Perfuse the animal with 50 mL saline to displace blood and blood components, i.e., free microbes, from the brain vascula- ture (see Sect. 3.4.1 through step 6). 5. Extract the intact brain from the cranium (see Sect. 3.4.2 through step 7). 6. Place the brain and 5 mL medium in a 70 μm filter set in a 60 cm sterile petri dish. 7. Using the rubber stopper from a 5 mL syringe plunger, gently press the brain tissue through the 70 μm filter using a grinding motion (see Note 5). 8. Pipet the resultant material back through the filter to remove any large clumps or debris. 9. Transfer the supernatant containing single cells into a 15 mL conical tube and incubate on ice for 5 min. During this time, fatty neuronal cells will float toward the top while heavy clumps and debris will settle to the bottom. If the microbe might directly infect neurons and their analysis is desired include this top layer in the medium transferred below. 10. Transfer the single cell-containing medium between the float- ing cells and sunken debris into a clean 15 mL conical tube and centrifuge for 10 min and 500 Â g. 11. Discard the supernatant, resuspend the cells in 5 mL of staining buffer, and count the number of cells. Adjust the concentration to 1 Â 107 cells/mL. Microbial Translocation of the Blood-Brain Barrier 143

3.1.1 Alternate Protocol 1. Perfuse the euthanized infected animal with PBS using 12 mL A: Percoll Gradient syringe with 23G needle. Purification 2. Dissect out the brain and place it in 3 mL of HBSS w/ 10 mM HEPES. 3. Use Dounce homogenizer to grind the brain tissue (ten strokes with the coarse pestle, then ten strokes with the fine pestle). 4. Adjust final volume to 7 mL with RPMI. 5. Add 3 mL Isotonic Percoll (ISP) to the brain homogenate to make a 30% ISP solution. 6. In a fresh 15 mL conical tube, add 2.5 mL 70% ISP. 7. Slowly overlay 30% ISP-brain homogenate using a transfer pipette without causing drops to form and disrupt the interface. 8. Centrifuge at 500 Â g for 30 min at room temperature with NO break. 9. Using suction, remove most of the top layer leaving 1 mL from the interface. 10. Transfer the interface to a new tube and wash by adding 8–10 mL of HBSS w/ HEPES. 11. Centrifuge at 500 Â g for 10 min at 4 C and resuspend the pellet in 1 mL staining buffer. 12. Count the number of cells and adjust the concentration to 1 Â 107 cells/mL.

3.1.2 Specific Protocol 1: Two of the mechanisms by which microbes transit the blood-brain Standard Flow Cytometric barrier involve the infection of specific cell types. Transcellular Analysis of Invasive and translocation relies on either (1) the ability of the microbe to infect Resident Immune Cells endothelial cells lining the blood vessels at the blood-brain barrier or (2) the ability of those same cells to take up the microbe. In addition, the Trojan horse model relies on the migration of infected peripheral immune cells, particularly macrophages, into the brain carrying the microbe with them. Therefore, the presence of infected endothelial cells or peripheral immune cells provides evi- dence for these two pathways, respectively. 1. Aliquot 100 μL of the single-cell suspension generated in Protocol A into individual wells of a 96-well plate; no lid is necessary (see Note 6). 2. Add 2 μg anti-CD16/CD32 antibodies to block Fc receptors present on numerous murine cell subsets (see Note 7). 3. Incubate at 4 C, protected from light for 15–20 min. 4. Add 100 μL staining cocktail of fluorescently labeled antibodies (see Notes 3, 8, and 9). Be sure that the antibody panel, includ- ing the pre-stained microbe (if used), can be detected by your 144 Charles T. Spencer and Mireya G. Ramos Muniz

available instrument as not all instruments or instrument con- figurations can detect all fluorophores. 5. Incubate at 4 C, protected from light for 30–60 min. 6. Add 200 μL staining buffer and centrifuge for 2 min at 900 Â g. 7. Repeat wash. 8. Remove medium and resuspend pellet in 250 μL of either staining buffer for immediate analysis or fixation buffer for later analysis (see Note 10). 9. Run samples on flow cytometric analyzer available to you. 10. For data analysis, we are particularly interested in two cell populations (see Note 11) defined as: (a) Endothelial cells: CD31pos Tie-2pos CD45neg. (b) Peripheral immune cells: CD45hi and subdivided as follows: – Macrophages: CD45hi B220neg CD11bhi SiglecFneg. – B cells: CD45hi B220pos CD3neg CD11bneg SiglecFneg. – T cells: CD45hi B220neg CD3pos CD11bneg SiglecFneg. 11. After gating on the individual populations, query the fluores- cence intensity of the microbe as a histogram. It may be useful to overlay the histograms to determine which cell type(s), if any, contain the microbe.

3.1.3 Specific Protocol 2: It can be difficult to determine which cell type(s) contain greater Imaging Flow Cytometry to numbers of fluorescent microbial burden than others using stan- Determine Cellular Microbe dard flow cytometry. In this case, visualization may provide added Burden information (Fig. 1). This protocol is designed for a similar analysis as Specific Protocol 1 but using imaging flow cytometry (e.g., ImageStream or FlowSight) to visualize the microbial burden in each cell type. Alternatively, cells can be examined on a slide, e.g., using a Cyto-spin or FACS sorter (see Note 12).

Fig. 1 Imaging flow cytometry allows distinction of infected cells with similar levels of infection as well as intracellular localization of the microbe. Imaging flow cytometry combines standard flow cytometric analysis of populations (CD45+ CD11b+ peripheral macrophages) with confocal imaging providing (DIC and fluorescence images) Microbial Translocation of the Blood-Brain Barrier 145

1. Aliquot 100 μL single-cell suspension generated above into individual wells of a 96-well plate; no lid is necessary (see Note 6). 2. Add 2 μg anti-CD16/CD32 antibodies to block Fc receptors present on numerous murine cells (see Note 7). 3. Incubate at 4 C, protected from light for 15–20 min. 4. Add 100 μL staining cocktail of fluorescently labeled antibodies (see Notes 3, 8, and 9). Be sure that the antibody panel, includ- ing the pre-stained microbe (if used), can be detected by your available instrument as not all instruments or instrument con- figurations can detect all fluorophores. 5. Incubate at 4 C, protected from light for 30–60 min. 6. Add 200 μL staining buffer and centrifuge for 2 min at 900 Â g. 7. Repeat wash. 8. Remove medium and resuspend pellet in 250 μL of either staining buffer for immediate analysis or fixation buffer for later analysis (see Note 10). 9. Run samples on imaging flow cytometer. 10. For data analysis, we are particularly interested in two cell populations (see Note 11) defined as: (a) Endothelial cells: CD31pos Tie-2pos CD45neg. (b) Peripheral immune cells: CD45hi and subdivided as follows: – Macrophages: CD45hi B220neg CD11bhi SiglecFneg. – B cells: CD45hi B220pos CD3neg CD11bneg SiglecFneg. – T cells: CD45hi B220neg CD3pos CD11bneg SiglecFneg. 11. After gating on the individual populations, visualize the fluo- rescent signal from the microbe and compare across cell types.

3.2 Basic Protocol B: Paracellular translocation across the blood-brain barrier is generally Co-injection of Mice considered a passive mechanism. However, certain microbes can with Live and Inactive actively degrade or weaken the endothelial tight junctions at the Microbes blood-brain barrier allowing the microbe to slip between the inter- cellular clefts and accumulate in the brain. Passive paracellular translocation of microbes is typically the result of breakdown of the blood-brain barrier during inflammation and loosening of endothelial tight junctions through which microbes then are “washed” into the brain. This protocol utilizes differentially labeled microbes, one to induce the inflammation and another that pas- sively enters the brain. 1. Label the microbes with two stains/dyes that are detectable by the microscopic configuration available to you (e.g., green and red) per manufacturer’s instructions (see Note 13). 146 Charles T. Spencer and Mireya G. Ramos Muniz

2. Inactivate one of the labeled microbes (e.g., green) using an appropriate condition for the microbe (see Note 14). For F. tularensis, bacteria are heat-killed by incubation at 65 C for 2 h. The live microbe (red) will be used to stimulate inflam- mation and potentially damage the blood-brain barrier. The inactivated microbe (green) will be used as a probe for paracel- lular translocation. 3. Verify inactivation of the microbes by inoculating appropriate culture medium; there should be no growth of the microbe after inactivation. 4. Inoculate C57BL/6J mice with 1 Â 106 cfu Francisella tular- ensis LVS in 50 μm sterile PBS via intradermal route (see Note 4). 5. On the third day after inoculation (see Note 15), inject the volume equivalent for 107 cfu inactivated green microbes intravenously. 6. Sacrifice a subset of animals kinetically every 6 h after injection of the inactivated green microbe (see Note 16). 7. Perfuse the animals through the heart with sterile PBS followed by 4% paraformaldehyde to fix the tissue (see Sect. 3.4.1). 8. Extract the brain and incubate in 12% sucrose/4% paraformal- dehyde solution overnight at 4 C, and flash freeze tissue in cooled hexanes (see Sect. 3.4.2). Tissues can be stored indefi- nitely at À80 C. 9. Alternatively, unfixed perfused brain tissue can be extracted and dissociated as in Sect. 3.1 for culture of the microbe. This culture mechanism must be capable of distinguishing between the labeled live (red) and inactivated (green) microbes. 10. For data analysis, the live red microbe serves as a control to ensure the functional invasion of live microbe into the brain. If no live red microbes are visible in tissue sections, discard the sample. Samples containing live red microbes should be analyzed for the presence of inactivated green microbes. Extra- cellular green microbes would support the paracellular translo- cation model, particularly if present around blood vessels, since inactive microbes cannot actively degrade the tight junction and any engulfed by immune or endothelial cells are likely to be degraded. Conversely, the presence of only red microbes may support either paracellular or transcellular translocation path- ways that require the microbe to be functional.

3.3 Basic Protocol C: Our lab utilizes a static monolayer of endothelial cells in transwell In Vitro Co-culture inserts to model the blood-brain barrier allowing for determination Systems of microbial translocation, modified based on [7]. The choice of transwell plates can be unique to given conditions; we utilize a 3 and 0.4 μm pore size to differentiate between the Trojan horse and para/transcellular movement of microbes, respectively. Microbial Translocation of the Blood-Brain Barrier 147

Distinguishing between these requires confluency in the endothe- lial monolayer. As the monolayer approaches confluency sealing the edges of individual cells with gap junction proteins, the permeabil- ity of the monolayer decreases. This decrease in permeability is correlated with an increase in electrical resistance [5]. Low electrical resistance allows the free movement of ions between chambers; however, a confluent monolayer impedes ion movement and resis- tance climbs. For in vitro transwell models of the blood-brain barrier, it is imperative to monitor the permeability of the endothe- lial monolayer. 1. Sterilize a pair of forceps in 70% ethanol for 5 min and allow to air-dry or use a bead sterilizer. 2. Open the transwell chamber to access the transwell inserts and transfer any inserts not to be used in this experiment to a separate 24-well plate for storage. 3. Add 150 μL collagen coating solution to the membrane of the insert to facilitate endothelial cell attachment. 4. Incubate for 30 min at 37 C. 5. Aspirate excess collagen solution being careful not to puncture or tear the membrane. 6. Coated inserts should be kept at 4 C and used within 4 days. If not using immediately, add 250 μL PBS to the insert and 600 μL to the lower compartment to prevent drying of the collagen. 7. Use sterile forceps to transfer needed inserts into the 24-well receiving plate. 8. Seed 5 Â 104 HVMEC or bEnd.3 cells per insert in 250 μL growth medium and add 500 μL growth medium to the lower compartment. 9. Incubate the inserts in 24-well receiving plate at 37 C, 5% CO2, until confluent. 10. Confluency is determined by transendothelial electrical resis- tance (TEER). There are multiple TEER units available, every- thing from a single-electrode unit done outside of the incubator to multiple-electrode units that monitor resistance at programma- ble time points while still inside the incubator. Due to its afford- ability, we use a single-electrode model. 11. For TEER measurement, sterilize the electrode with 70% etha- nol for 15 min and allow to air-dry for 15 s. 12. Rinse the electrode in sterile cell culture medium. 13. Allow the cells to come to room temperature. 14. Set the MODE switch to ohms and turn the power on. 15. Immerse the electrode so that the shorter tip is in the insert and the longer tip is in the lower chamber. The shorter tip should not contact cells growing on the membrane and the longer tip should 148 Charles T. Spencer and Mireya G. Ramos Muniz

Fig. 2 Transendothelial electrical resistance measurement identifies confluency of the endothelial cell growth at the time of the resistance plateau and is correlated with the tightness of the endothelial cell junctions

just miss the bottom of the lower chamber. To ensure stable and reproducible results, make sure that the electrode is held steady and at 90 angle to the plate insert. 16. Measure the resistance of blank wells without cells. 17. Measure the resistance in the sample wells. 18. Measure the resistance of the blank wells once more. 19. Average the resistance of the two blank measurements. 20. Subtract this blank resistance from the sample-well resistance measurement. 21. Confluency is generally attained between 24 and 96 h after seeding, although this can vary depending on the cell line used. 22. Monitor TEER every 8 h after the first 24 h to establish the rise in resistance. Continue monitoring every 4 h between 24 and 96 h to determine when the plateau is reached (see Note 16). 23. Electrical resistance may be increased by addition of a GSK-3 inhibitor [7] (Fig. 2). 24. For this, 2.5 μM BIO is added during seeding of the insert. This tighter junction may better represent the intact blood-brain barrier while the uninhibited condition may represent an injured or “leaky” blood-brain barrier (see Note 17).

3.3.1 Alternate Protocol In the absence of a TEER measurement device, diffusion of fluo- C: Fluorescent Markers rescent markers can be used to determine confluency of the monolayer. 1. Prepare working solution of 10 μM fluorescent marker, e.g., dextran. We recommend using a mixture of multiple size of dextrans (e.g., 3, 10, 40 kDa) labeled with different fluorophores. Microbial Translocation of the Blood-Brain Barrier 149

Fig. 3 Paracellular translocation is defined as movement of free microbes between endothelial cells of the blood-brain barrier. This can be mimicked in vitro in a transwell dish lined with a monolayer of endothelial cells if the microbe does not penetrate the cells but instead passes between the cells

2. Remove 200 μL of medium from the insert and replace with 200 μL of marker-containing medium. 3. Take 100 μL samples from the lower chamber at 1, 2, and 3 h and 100 μL from the insert at the last time point. 4. Measure fluorescence in a black 96-well analytical plate in a fluorometer. 5. Calculate the ratio of fluorescent signal in the lower (L) medium compared to the terminal signal in the insert (I) as 100 Â L/I. In the absence of cells, this ratio should start low and approach 100% as diffusion reaches equilibrium. Confluency prevents diffusion of the markers so the L/I ratio should remain low.

3.3.2 Specific Protocol 1: Paracellular translocation is the movement of microbes between Paracellular Translocation endothelial cells comprising the blood-brain barrier (Fig. 3). In order to mimic this in vitro, we use transwell inserts with a 0.4 μm pore size preventing migration of cells but allowing diffu- sion of microbes. However, with the lack of blood pressure in this static model, microbial migration can take several days to detect. In addition, when analyzing this mode of translocation, we use both the “injured” blood-brain barrier model (cells alone) and the “intact” blood-brain barrier model (+GSK-3 inhibitor) as described in Sect. 3.3. 1. To provide a large differential between the insert and lower chamber media, we use 1 Â 106 cfu/mL of bacteria in the insert. Therefore, dilute microbial concentration to 1 Â 107/ mL in growth medium (see Note 18). 2. Replace the medium from the insert with 100 μL of the microbe-containing medium. 3. Every 8 h sample 100 μL of medium from the lower chamber (see Note 16). At the final time point, remove 100 μL from the 150 Charles T. Spencer and Mireya G. Ramos Muniz

insert to determine the number of microbes that have not translocated. Replenish with fresh growth medium as needed in the insert and lower chamber. 4. Using an appropriate method for the microbe, determine the number of viable microbes in the 100 μL sample. For F. tular- ensis, tenfold serial dilutions are plated on Chocolate agar plates supplemented to 1% isovitalex solution. 5. For analysis, compare the number and rate of microbial trans- location of the endothelial cell boundary in the presence and absence of GSK-3 inhibitor. Since GSK-3 inhibitor “tightens” the gap junctions between cells, a noticeable decline in micro- bial translocation should be observed in the presence of GSK-3.

3.3.3 Specific Protocol 2: Transcellular translocation is the movement of microbes through Transcellular Translocation endothelial cells comprising the blood-brain barrier (Fig. 4). In order to mimic this in vitro, we use transwell inserts with a 0.4 μm pore size preventing migration of cells but allowing passage of microbes. It is critical that microbes NOT be allowed to pass between cell and so we recommend use of the GSK-3 inhibitor to “tighten” the gap junctions. As a control, we fix GSK-3 inhibitor- treated endothelial cell monolayers to prevent internalization of the microbes and export to the basolateral side. 1. For the control conditions, use sterile forceps to transfer inserts to a separate 24-well plate with 500 μL PBS in the lower chamber. 2. Carefully aspirate growth medium from the insert and add 50 μL 0.15% glutaraldehyde. 3. Incubate at room temperature for 30 s. 4. Neutralize glutaraldehyde by adding 200 μL 0.2 M lysine.

Fig. 4 Transcellular translocation is defined as movement of free microbes through the endothelial cells of the blood-brain barrier. This can be mimicked in vitro in a transwell dish lined with a monolayer of endothelial cells if the microbe can be found inside of the endothelial cells priot to being deposited on the other side Microbial Translocation of the Blood-Brain Barrier 151

5. Wash the insert twice with growth medium being careful not to disturb the cell monolayer or puncture the transwell membrane. 6. Using sterile forceps, transfer the inserts back to the experi- mental plate. 7. Compared with paracellular translocation, a large differential between the insert and lower chamber media is not desirable since the microbes should be actively transported through the cells. Therefore, 5 Â 105 fluorescently labeled microbes are added to the insert so adjust bacterial concentration to 5 Â 106 cfu/mL in growth medium. 8. Replace the growth medium in the insert with 100 μL of the microbe-containing medium. 9. Every 8 h sample 100 μL of medium from the lower chamber (see Note 16). At the final time point, remove 100 μL from the insert to determine the number of microbes that have not translocated. Replenish with fresh growth medium as needed in the insert and lower chamber. 10. Using an appropriate method for the microbe, determine the number of viable microbes in the 100 μL sample. For F. tular- ensis, tenfold serial dilutions are plated on Chocolate agar plates supplemented to 1% isovitalex solution. 11. For analysis, compare the number and rate of microbial trans- location of live and fixed endothelial cells. Since transcellular translocation depends on endothelial cell membrane dynamics, an absence of microbial translocation should be observed follow- ing fixation. 12. To confirm the presence of microbes within the endothelial cells, inserts can be stained and imaged during or at the con- clusion of the experiment (see Sect. 3.4.7).

3.3.4 Specific Protocol 3: Trojan horse translocation is the movement of microbes across the Trojan Horse blood-brain barrier internalized by an immune cell, generally a phagocytic macrophage or dendritic cell (Fig. 5). Here, infected immune cells are placed in the upper chamber of the transwell plate and given the opportunity to extravasate into the lower chamber carrying with them internalized microbes. In order to mimic this in vitro, we use transwell inserts with a 3 μm pore size to allow the extravasation of the immune cell. When analyzing this mode of translocation, we primarily use the “injured” blood-brain barrier model (cells with no GSK-3 inhibitor) as described in Sect. 3.3 to facilitate the extravasation of immune cells. However, inclusion of the “intact” blood-brain barrier model (+GSK-3 inhibitor) can also provide additional evidence. Since extravasation of infected immune cells is dependent upon their ability to change shape, as a control, we include glutaraldehyde-fixed infected immune cells. 152 Charles T. Spencer and Mireya G. Ramos Muniz

Fig. 5 The Trojan horse method of translocation is defined as movement of microbes across the blood-brain barrier inside of a carrier cell, e.g., macrophage. This can be mimicked in vitro in a transwell dish lined with a monolayer of endothelial cells if infected immune cells are added to the upper chamber and allowed to extravasate across the endothelial monolayer into the lower chamber

This treatment prevents the cells from extravasating to the lower chamber. For these experiments, free microbes will not be used; instead, microbe-infected immune cells will be solely used. We use a bacteria-infected macrophage cell line for this analysis. 1. Seed a 12-well plate with 7.5 Â 106 macrophages in 1 mL growth medium (see Note 18). 2. Inoculate the well with MOI ¼ 40, i.e., 3 Â 108 cfu/well (see Note 18). 3. Incubate at 37 C for 2 h (see Note 18). 4. Wash three times by centrifuging at 500 Â g for 10 min and resuspending in 5 mL of media each time. 5. For the fixation control, transfer 5 Â 104 cells to a 96-well plate or sterile microcentrifuge tube and pellet the cells at 500 Â g for 10 min. 6. Resuspend in 100 μL 0.15% glutaraldehyde. 7. Incubate at room temperature for 30 s. 8. Neutralize glutaraldehyde by adding 500 μL 0.2 M lysine. 9. Wash the cells twice with PBS or growth medium and recount. 10. It is important not to crowd the endothelial layer with infected immune cells. Therefore, adjust the cell suspension to 1 Â 106 cells/mL. 11. Replace the medium from the insert with 100 μL of the cell suspension. It may be necessary to add a chemoattractant to the lower chamber in order to establish a chemotactic gradient, thereby inducing the immune cells to extravasate. We recommend initial testing without chemoattractant in order to determine baseline mobility of your infected immune cells. Subsequent experiments can include an appropriate chemoattractant in the Microbial Translocation of the Blood-Brain Barrier 153

600 μL of growth medium in the lower chamber, e.g., 100 ng/mL MCP-1 for macrophage migration [8]. 12. Every 8 h sample 100 μL of medium from the lower chamber (see Note 16). At the final time point, remove 100 μL from the insert to determine the number of cells that have not translo- cated. Replenish with fresh growth medium as needed in the insert and lower chamber. 13. Count the number of viable immune cells in the lower chamber. 14. For analysis, compare the number and rate of immune cell translocation of the endothelial cell boundary in the presence and absence of GSK-3 inhibitor. Since GSK-3 inhibitor “tight- ens” the gap junctions between cells, a noticeable decline in microbial translocation should be observed in the presence of GSK-3. In addition, the fixation control should have negligible translocation. 15. It is advisable to lyse and culture the immune cells, or otherwise measure (e.g., via fluorescent microscopy) the amount of microbe carried across the membrane.

3.4 Supporting 1. After mouse has been euthanized, make an incision below the Protocols rib cage at about the level of the liver.

3.4.1 Support Protocol: 2. Continue to dissect the skin along the right and left sides of the Perfusion rib cage by cutting through the skin, muscle, and bones. 3. Excise the flap created by cutting the diaphragm and above the sternum just below the clavicle. At this point, the lungs, heart, and liver will be exposed in the inner chest cavity. 4. Use a clamp to hold the heart and create a small incision in the left ventricular wall. 5. Quickly position the perfusion needle in the incision and use a hemostat clamp to secure heart and needle placement. It may be necessary to also cut the aorta to provide easier exsanguination. 6. Begin pushing saline solution to clear blood out. Maintain a clean workspace. At least 50 mL of saline solution is necessary to clear out entire blood volume; however, more can be used. For applications requiring live cells stop here and do not proceed to step 7. 7. Once the liver has visually cleared of blood, switch to 4% paraformaldehyde (PFA). At least 50 mL of PFA is necessary for a successful cross-linking. 8. The mouse carcass will harden into a rigid position indicating successful fixation. As before, more PFA may be necessary. If the liver did not entirely clear of blood with PBS flush, it will continue to clear with the administration of PFA. 154 Charles T. Spencer and Mireya G. Ramos Muniz

3.4.2 Support Protocol: 1. Decapitate the carcass. Brain Harvesting 2. Cut the skin from the base of the skull toward the nose along the cranial ridge. Continue to dissect all of the muscle and connective tissue until a handle-like structure is formed. It is important to remove all connective tissue from the skull. 3. Fold the skin flap over the nose to create a handle. 4. Cut the skull from the base of the head toward the nose being careful not to damage brain. 5. Cut two lateral incisions at the base and two lateral incisions near the eyes. The longer the incisions are, the easier it will be to remove the top of the skull. 6. Remove the cranium to expose brain. The olfactory bulb may be severed or preserved as needed. 7. Collect the brain by scooping out of skull and clipping the brain stem. For applications requiring live cells stop here and do not proceed to step 8. 8. Transfer brain to postfix solution and store at 4 C for 16–18 h. 9. Working inside fume hood, cool hexanes with dry ice. 10. Remove the brain from the postfix solution and blot it dry. 11. Place the brain into the cooled hexanes and allow it to freeze for 3–5 min. 12. Retrieve brain, blot it, and place in a plastic vial for storage at À80 C.

3.4.3 Support Protocol: 1. Prepare the microtome per manufacturer’s instructions. Tissue Sectioning 2. Fill 24-well plates with 2 mL cryoprotectant medium and place on cold plate. 3. For mouse brain, tissue section thickness of 30 μm is preferred. 4. Break off dry ice chips, fill the pockets of the stage, wait until stage is frozen, create a square with PBS in the middle, and allow it to freeze as well. 5. Shave off PBS to create a flat surface on which to place the brain. 6. With a clean utility blade, create a flat surface on the brain such that when placed onto the stage it will be straight and oriented properly to desired plane of section. 7. To mount brain onto stage, place a few drops of PBS onto the flattened-frozen PBS and quickly place the brain onto those drops. The PBS drops will rapidly freeze so work quickly but accurately as once frozen it is very difficult to reposition. 8. Cover the mounted brain with dry ice dust keeping the top clear of dust. It is important to keep the dry ice dust loosely packed on top of the brain; if packed too tightly it can dull or damage the blade. 9. Bring blade to level with the top of the brain and begin to section. Microbial Translocation of the Blood-Brain Barrier 155

10. Collect tissue section with a small damp brush and carefully place it in appropriate well. When collecting tissue, it is best to collect in anatomical order. 11. Once desired brain sectioning has been completed, store at 4 C for up to 1 week or at À20 C for years.

3.4.4 Support Protocol: 1. Wash tissue for 5 min using PBS at room temperature on a belly Tissue Staining dancer shaker at moderate speed, five exchanges. (see Note 19) 2. Make a block solution (49 mL PBS + 1 mL normal donkey serum + 50 μL Triton X). 3. Incubate tissue in block solution at room temperature for 3 h on a belly dancer shaker, covered. 4. Prepare primary antibody solution at optimized dilution, gen- erally 1:500–1:5000 in blocking solution. 5. Place blocked tissue in 500 μL primary antibody solution in a 24-well non-tissue culture-treated flat-bottom plate. Incubate at 4 C for 16–20 h on a belly dancer shaker at moderate speed, covered. 6. Wash tissue for 5 min using PBS at room temperature on a belly dancer shaker at moderate speed, five exchanges. 7. Prepare secondary antibody solution at optimized dilution, generally 1:100–1:1000 in blocking solution. 8. Place washed tissue in 500 μL secondary antibody solution in a 24-well non-tissue culture-treated flat-bottom pate. Incubate for 5–6 h at room temperature on a belly dancer shaker at moderate speed, covered. 9. Wash tissue for 5 min using PBS at room temperature on a belly dancer shaker at moderate speed, five exchanges. 10. Prepare fluorophore solution at optimized dilution, generally 1:500–1:5000 in blocking solution. If desired, counterstain with DAPI (1:4000) or Neurotrace (1:250) for 1 h at room temperature on a belly dancer shaker, covered. 11. Wash tissue for 5 min using PBS at room temperature on a belly dancer shaker at moderate speed, five exchanges. 12. Store in PBS for up to 1 week at 4 C.

3.4.5 Support Protocol: 1. Submerge the tissue section in a petri dish containing PBS and Tissue Mounting flatten using a fine brush (size 0 or 00). 2. Submerge a positively charged slide into the petri dish and transfer the tissue section onto the glass slide with as little handling as possible. 3. Carefully remove the glass slide from the buffer and allow it to air-dry before adding another tissue section. 156 Charles T. Spencer and Mireya G. Ramos Muniz

3.4.6 Support Protocol: 1. When initially plating endothelial cells from cryopreservation, Subculturing of HVMEC or the recommended density is 1.25 Â 104 cells in multiple T-25 bEnd.3 flasks.

Initiation and Maintenance 2. Add 5 mL growth medium to the T-25 flasks and allow the  of Culture vessels to equilibrate in a 37 C, 5% CO2, 90% humidity incu- bator for at least 30 min. 3. Wipe the cryovial with ethanol or isopropanol before opening. 4. In a sterile field, briefly twist the cap a quarter turn to relieve pressure and then retighten. 5. Quickly thaw the cryovial in a 37 C water bath being careful not to submerge the entire vial. Watch your cryovial closely; when the last sliver of ice melts, remove it. Thawing cells for longer than 2 min results in less than optimal results. 6. Carefully mix the cell suspension using a micropipette (see Note 20) and dispense cells directly into the T-25 flasks (see Note 21). 7. Gently rock the culture vessel to evenly distribute the cells and return to the incubator. 8. Change the growth medium 16–24 h after seeding and every other day (every 48 h) thereafter. 9. When cell confluency is 25–45%, increase media volume to 7.5 mL. 10. When cell confluence is greater than 45%, increase the media volume to 10 mL (see Note 22).

Subculturing (for Each T-25 1. Subculture the cells when they are 70–85% confluent. Flask) 2. Bring 2 mL of trypsin/EDTA, 7–10 mL of HEPES buffered saline solution (HEPES-BSS), 5 mL of trypsin-neutralizing solution, or serum-containing medium, and growth medium to room temperature. 3. In a sterile field, aspirate the medium from one culture vessel. Subculture one flask at a time. These cells can be finicky and adjustments may need to be made for the subculturing of the other flasks. 4. Rinse the cells with 5 mL of room-temperature HEPES-BSS. DO NOT forget this step. The medium contains complex proteins and calcium that neutralize the trypsin. 5. Aspirate the HEPES-BSS from the flask. 6. Cover the cells with 2 mL of trypsin/EDTA solution. 7. Place the culture vessels into a 37 C humidified incubator for 3–5 min. 8. Periodically examine the cell layer microscopically and check for cell detachment. Microbial Translocation of the Blood-Brain Barrier 157

9. Allow the trypsinization to continue until approximately 90% of the cells are rounded up. 10. At this point, tap the flask against the palm of your hand to release the majority of cells from the culture surface (see Note 23). If only a few cells detach, you may not have let them trypsinize long enough. Wait for 30 s and tap again. If cells still do not detach, wait and tap every 30 s thereafter. This entire process should take no more than 5 min. 11. After cells are released, neutralize the trypsin in the flask with 5 mL of trypsin-neutralizing solution or serum-containing medium at room temperature. 12. Quickly transfer the detached cells to a sterile 15 mL centrifuge tube. 13. Rinse the flask with a final 2 mL of HEPES-BSS to collect remaining cells and add this rinse to the centrifuge tube. 14. Examine the harvested flask under the microscope to make sure that the harvest was successful by looking at the number of cells left behind; this should be less than 5%. 15. Centrifuge the harvested cells at 200 Â g for 5 min to pellet the cells. 16. Aspirate most of the supernatant, except for 100–200 μL. 17. Flick the tube with your finger to loosen the pellet. 18. Dilute the cells to a final volume of 2–3 mL of growth medium and count. 19. The recommended seeding density when subculturing endo- thelial cells is 1.25 Â 104 cells per T-25 flask or 3.75 Â 104 cells per T-75. 20. Bring to a volume of 5 mL in the T-25 flask with growth medium.  21. Place the new culture vessels into a 37 C, 5% CO2, 90% humidity incubator.

Cryopreservation 1. Prepare cryopreservation media by adding 80% EGM-2 + 10% DMSO + 10% FBS, sterile filter using a 0.2 μm filter, and chill to 4 C. 2. Harvest and centrifuge cells according to steps 1–17 of Sect. 3.4.6.2. 3. Resuspend cells in cold cryopreservation media at 5 Â 105–2 Â 106 cells/mL (see Note 24). 4. Pipet 1 mL aliquots into freezing vials or ampoules and seal. 5. Insulate aliquots in Styrofoam or propanol freezing canisters. 6. Store cells at À80 C overnight. 158 Charles T. Spencer and Mireya G. Ramos Muniz

7. Within 12–24 h, place cells in liquid nitrogen for long-term storage. Cells will be compromised by long-term storage at À80 C.

3.4.7 Support Protocol: 1. Using sterile forceps, transfer the inserts to a fresh 24-well plate Staining of Endothelial filled with 500 μL blocking buffer. Cells in Transwell Inserts 2. Aspirate growth medium and wash the inserts twice with PBS. 3. Add 100 μL blocking buffer to the inserts. 4. Incubate for 1 h at room temperature. 5. Prepare 100 μL of membrane dye for each insert to be stained; we prefer 5 μg/mL wheat germ agglutinin. 6. Aspirate blocking buffer from the insert and add 100 μLof membrane stain (see Note 19). 7. Incubate for 10 min at 37 C. 8. Wash inserts twice with PBS. 9. If desired, counterstain with DAPI nuclear dye diluted 1:1000 in PBS; add 100 μL per insert. 10. Incubate for 10 min at 37 C. 11. Wash twice with PBS. 12. Place a drop of mounting medium on a glass slide. 13. Using sterile forceps, remove the inserts and gently blot dry with a KimWipe. 14. Invert the insert on a work surface and gently cut out the insert membrane using a sharp scalpel blade. 15. Place the insert cell side up onto the mounting medium and cover with a coverslip. 16. Allow the mounting medium to dry. 17. Image using an inverted confocal fluorescent microscope. 18. If microbes are being translocated transcellularly, they should be located inside of the boundaries of the endothelial cells. Microbes carried into the brain via the Trojan horse method should be located inside the boundaries of peripheral immune cells.

4 Notes

1. When preparing BulletKit Media, it may not be possible to recover the entire volume listed for each vial. Small losses (up to 10%) should not affect cell growth characteristics of the supplemented medium. 2. The signal-to-noise ratio is greatly increased in later steps if the microbe is fluorescently labeled prior to infection. Therefore, if Microbial Translocation of the Blood-Brain Barrier 159

a method exists for the microbe of interest for it to be fluores- cently labeled in such a way to maintain that label throughout the course of analysis, we highly recommend to do so. However, be certain that the chosen fluorescent dye/signal does not interfere with antibody combinations used in later steps. 3. This antibody panel is designed for a ten-color flow analyzer now generally available in state-of-the-art facilities. If such an instrument is not available for your use, adjust the staining combinations to suit your instrument. This panel is also designed to be used with samples containing a pre-labeled microbe. If you are infecting with a pre-labeled microbe include a stain control for the microbe. The panel can also be subdivided into two panels, each of which is then used to stain the same sample. 4. The route of inoculation will vary depending upon the microbe of interest. Generally, these include intradermal, intraperito- neal, intramuscular, oral, gavage, intravenous, or aerosol. Each route has a limit on the volume for injection that must be followed to prevent injury to the mouse: i.d.—50 μL, s.c.— 200 μL, i.v.—200 μL, i.m.—50–100 μL and i.p.—1 mL. 5. While other methods exist to generate single-cell suspensions from tissue (e.g., grinding between frosted slides, mechanical disruption, enzymatic digestion), we have found that the method described is the most cost and time effective. When analyzing directly ex vivo, it is always important to minimize time between harvest and analysis during which things can change in your sample. 6. We generally stain in a 96-well round-bottom sample plate; however, other plates or tubes (e.g., 200 μL tube, 5 mL round-bottom tube, 15 mL conical tube) can be used. Staining in minimal volume minimizes the amount of antibodies neces- sary and sample lost during washing steps. 7. If using a species that does not express Fc receptors to the extent mice do, this step can be omitted, but it is a good idea anyway. 8. The volumes and centrifugation steps listed are for 96-well plates using the antibody panel described in Sect. 2 for a ten-color instrument. If staining in tubes, increase the centri- fugation time to 10 min and the wash buffer to 1–2 mL. 9. Some protocols will wash away unbound anti-CD16/CD32 antibodies prior to addition of monoclonal antibodies; how- ever, we find no interference between the two. 10. Analyze flow cytometric samples within 36 h of fixation, pref- erably within 12 h. 160 Charles T. Spencer and Mireya G. Ramos Muniz

11. It is possible that resident neural cells, e.g., microglia and astrocytes, engulf microbes; however, these do not add data to microbial translocation. If these are of interest, they can be defined by addition of other antibodies to the staining cocktail, e.g., IBA-1 or GFAP. These particular antibodies bind internal proteins and, therefore, require cell fixation and permeabiliza- tion prior to staining. 12. If these instruments are unavailable for your use, it may also be possible to purify the indicated subsets using a Cyto-Spin or FACS sorter and visualize them using epi- or confocal fluores- cent microscopy. 13. For this protocol, we will consider bacterium labeled with a FITC or APC fluorescent amine-reactive dye such as BacLight (Thermo Fisher Scientific). However, numerous other fluoro- phores and formats are also available, including transgenic expression of fluorescent proteins or conjugated proteins. 14. Inactivation protocols vary by microbial species. For our experiments, bacteria are heat-killed by incubation at 65 C for 2 h. Care must be taken with chemical inactivation that the chemical is fully removed before use in animal models. 15. Refinement of this time series may be necessary depending on the microbe under investigation. For our bacterium, peak inflammation is reached at 3 days post-inoculation and invasion is observed within 48 h. 16. After establishing the baseline for your particular system, these monitoring time points may be altered. 17. Use of GSK-3 inhibitor may be particularly important when analyzing paracellular and transcellular modes of transport. During microbial infection, the blood-brain barrier is often compromised resulting in a leaky phenotype that may be more consistent with in vitro models in the absence of this inhibitor. 18. These steps will need to be optimized for each endothelial- microbe culture combination. 19. When staining tissue, you may use one or several primary and secondary antibodies at the same time. You may also use one or several fluorophores adjusting the panel to what technology is available to you. 20. Endothelial cells tend to more strongly adhere to the cryovial than other cell types. Additional and/or more forceful pipet- ting may be necessary to remove all cells. 21. Centrifugation should not be performed to remove cells from cryoprotectant cocktail. This action is more damaging than the effects of DMSO residue in the culture. Microbial Translocation of the Blood-Brain Barrier 161

22. Warm an appropriate amount of medium to 37 C in a sterile container. Remove the medium and replace it with the warmed, fresh medium and return the flask to the incubator. Avoid repeated warming and cooling of the medium. If the entire contents are not needed for a single procedure, transfer and warm only the required volume in a sterile secondary container. 23. If the majority of cells do not detach within 5 min, the trypsin is either not warm enough or not active enough to release the cells. Harvest the culture vessel as described and either re-trypsinize with fresh, warm trypsin/EDTA solution or rinse with trypsin-neutralizing solution or serum-containing medium and then add fresh, warm growth medium to the culture vessel. Return to the incubator until fresh trypsiniza- tion reagents are available. 24. Work quickly. Once exposed to the DMSO, cells become very fragile.

References

1. Kim KS (2008) Mechanisms of microbial tra- Wolburg H, Fruttiger M, Taketo MM, von versal of the blood-brain barrier. Nat Rev Micro- Melchner H, Plate KH, Gerhardt H, Dejana E biol 6(8):625–634 (2008) Wnt/beta-catenin signaling controls 2. Santiago-Tirado FH, Onken MD, Cooper JA, development of the blood-brain barrier. J Cell Klein RS, Doering TL (2017) Trojan horse tran- Biol 183(3):409–417. https://doi.org/10. sit contributes to blood-brain barrier crossing of 1083/jcb.200806024 a eukaryotic pathogen. mBio 8(1):e02183-16. 7. Czupalla CJ, Liebner S, Devraj K (2014) In vitro https://doi.org/10.1128/mBio.02183-16 models of the blood–brain barrier. In: Milner R 3. Kim K. How pathogens penetrate the blood- (ed) Cerebral angiogenesis: methods and proto- brain barrier. Microbe Magazine. https://doi. cols. Springer, New York, NY, pp 415–437. org/10.1128/microbe.9.487.1 https://doi.org/10.1007/978-1-4939-0320- 4. He Y, Yao Y, Tsirka SE, Cao Y (2014) Cell- 7_34 culture models of the blood-brain barrier. Stroke 8. Green TD, Park J, Yin Q, Fang S, Crews AL, Jones 45(8):2514–2526. https://doi.org/10.1161/ SL, Adler KB (2012) Directed migration of mouse STROKEAHA.114.005427 macrophages in vitro involves myristoylated 5. Crone C, Olesen SP (1982) Electrical resistance alanine-rich C-kinase substrate (MARCKS) pro- of brain microvascular endothelium. Brain Res tein. J Leukoc Biol 92(3):633–639. https://doi. 241(1):49–55 org/10.1189/jlb.1211604 6. Liebner S, Corada M, Bangsow T, Babbage J, Taddei A, Czupalla CJ, Reis M, Felici A, Chapter 9

Transport Across the Choroid Plexus: How to Culture Choroid Plexus Cells and Establish a Functional Assay System

Sen Takeda and Keishi Narita

Abstract

Choroid plexus epithelial cells (CPECs) contribute to the production of cerebrospinal fluid (CSF), which plays an important role in maintaining the milieu inte´rieur of the central nervous system. To elucidate the function of CPECs, in vitro primary culture is an ideal system as the choroid plexus (CP) is situated deep in the brain ventricular system in situ. This location makes detailed analysis of these cells difficult. Moreover, its highly undulating nature prevents quantitative study using molecular and cell biological tools. The protocols herein describe primary culture of CPECs in a differentiated state, in a well-integrated monolayer sheet that recapitulates the in vivo blood-cerebrospinal fluid barrier, to enable study of fluid transcytosis through the cytoplasm.

Key words Choroid plexus (CP), Choroid plexus epithelial cell (CPEC), Transcytosis, Cilia, Cere- brospinal fluid (CSF), Autocrine, G-protein-coupled receptor (GPCR)

1 Introduction

The choroid plexus (CP) is a unique vascular tissue comprising the pia mater, with capillaries enwrapped by choroid plexus epithelial cells (CPECs). During the development of the brain, part of pia mater facing the neuroepithelial monolayer starts to invaginate into brain ventricles, followed by the differentiation of neuroepithelial cells into CPECs [1], which exhibit a grapelike appearance. CPECs are simple cuboidal cells that have been regarded as a major site for producing the cerebrospinal fluid (CSF). The capillaries embedded in the CP are fenestrated, allowing passage of water into the inter- stitial space. However, because of the presence of tight junctions between each CPEC [2] an efficient transport system that facilitates the secretion of water into the ventricular cavity is required [3]. Additionally, some water transport may take place through water channels between the CPECs [4].

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_9, © Springer Science+Business Media, LLC, part of Springer Nature 2019 163 164 Sen Takeda and Keishi Narita

CPECs possess bundles of cilia, which are different from motile ependymal cilia [5] as they assume a 9 + 0 axonemal configuration like the mouse nodal cilia that determine the left–right asymmetry of the body [6]. While they are motile during the perinatal stage, they gradually lose their motility [7] and start to express receptors for modulating the secretion of CSF [3, 8]. One of the receptors expressed in CPEC cilia is the receptor for neuropeptide FF (NPFF), a member of the G-protein-coupled receptor (GPCR) family. CPECs produce and secrete NPFF, which exerts its auto- crine effects via the specific receptors on the primary cilia. To study this, we have established a primary monolayer culture of CPECs in transwells, in which transcytosis mimicking the in situ CP can be reproduced in vitro. Considering that disruption of CSF homeostasis results in development of hydrocephalus [9], this experimental design is useful for elucidating the pathogenesis of this disorder. This chapter describes the general procedure used for primary culture of CPECs from pigs (basic protocol), the method for assaying transcytosis (Sect. 3.4.1) and measuring the transepithelial resistance (Sect. 3.4.2). Because of limited availability of CP from mice due to their small size, our protocol takes advantage of the large size of pigs and the consequently large quantity of CP. However, application of the current protocol to other species such as mice would be possible, with modification. The cultured CP cells differentiate well into a ciliated monolayer sheet, as visualized by detecting expression of zona occludens-1 (ZO-1), witnessing the formation of tight junctions [3]. Moreover, these cells express a specific molecular marker of CPECs, transthyretin [10], validating the homogeneity of the culture. The major disadvantage of using porcine tissue lies in the species per se, because genetic information is much sparser for pigs than mice. However, the pig database is being increasingly augmented and updated, making it possible to more easily conduct genetic manipulation [11]. Although this protocol is applicable to rodents, such as mice and rats, by increasing the number of animals used for the experiment, the cells will not expand to fill the entire surface of the transwells and therefore will not be as confluent as cultures using pig tissue (see Note 1). Regarding the techniques for assaying transcytosis, we employ fluorescent dextran and spectrom- etry or western blotting of apical media in the upper compartment of the transwells [8].

2 Materials

Prepare all solutions for the experiment described here using cell culture-grade or sterilized reagents. Autoclaved Milli-Q water is used for all procedures using water throughout this protocol. Culture of Choroid Plexus Cells 165

Regarding disposable plasticware, we do not have any preferences for a specific supplier. Transwells with a 0.4 μm pore polyester membrane from Corning International (Tokyo, Japan; Cat#3450), refrigerators, freezers (À20 C), deep freezers (À80 C), and refrigerated centrifuges are required.

2.1 Surgical 1. Dissection saw for cutting the skull. Instruments 2. Dissection chisels (with a 20 mm wide blade) and a bone mallet for Dissecting the Pig for splitting the skull. Brain 3. Large scissors for scalp and dura mater resection. 4. Small scissors for preparing the choroid plexus. 5. Fine watchmaker’s forceps. These tools should be autoclaved in a sterile bag.

2.2 Reagents All stock solutions are filter sterilized, aliquoted, and stored at À20 C. For stock solutions, the concentration and dilution to be used are in parentheses. 1. Culture Medium For preparing 50 mL, DMEM/HAM’s F-12 medium is sup- plemented with 5 mL of fetal bovine serum (FBS), 4 mM L- glutamine (200 mM, 1/50), 5 μg/mL insulin (1.5 mg/mL, 1/300), 200 ng/mL hydrocortisone (40 μg/mL, 1/200), 30 nM sodium selenite (150 μM, 1/5000), 10 ng/mL epider- mal growth factor (100 μg/mL, 1/10,000), 20 μM cytosine arabinoside (0.2 M, 1/10,000), and 500 μL of antibiotic/ antimycotic (Â100 stock solution, Thermo Fisher Scientific). 2. Fetal bovine serum. 3. Serum-free medium For preparing 50 mL, the above culture medium is used but the FCS is excluded. 4. Phosphate-buffered saline (PBS) supplemented with Ca2+ and Mg2+

To 50 mL of PBS, add 250 μL of 0.2 M CaCl2 and 50 μLof 1 M MgCl2. 5. Trypsin Dissolve in Ca2+-, Mg2+-free Hanks’ balanced salt solution (HBSS) to 2.5% as a stock solution and store at À80 C. For usage, dilute it 1/10 with HBSS or serum-free medium. There is no preference for a specific manufacturer. 6. Matrigel Matrigel (Corning, New York) is thawed at 4 C and diluted 1/50 with culture medium. Make sure not to produce bubbles 166 Sen Takeda and Keishi Narita

and avoid clumps. Add the diluted solution at 37 C and incubate overnight to coat the transwells or coverslips. 7. Chemical deciliation (see Note 2) For chemical deciliation [12], cultured CPECs are incubated with culture medium supplemented with 4 mM chloral hydrate (1 M, 1/250). After 24 h, cilia are completely removed.

2.3 Assay For quantifying transepithelial transport, fluorescently labeled tra- for Transcytosis cers such as Oregon Green 488-labeled dextran (average molecular and Transepithelial weight 70 kDa; Thermo Scientific, Carlsbad, CA, USA; Transport Cat#D7173), AlexaFluor 488-conjugated biocytin (Thermo Scien- tific; Cat#A12924), or AlexaFluor 488-conjugated bovine serum albumin (Thermo Scientific; Cat#A13100) are added to the medium of the lower (basolateral) compartment.

2.4 Transepithelial A Millicell-ERS (Millipore, Billerica, MA, USA) is used for measur- Resistance ing the transepithelial resistance.

3 Methods

Basic protocol: This is the fundamental protocol for culturing the porcine CPECs to obtain a flat monolayer with apical ciliation.

3.1 Day Before 1. Overlay 100 μL of diluted Matrigel on the plastic dishes, glass Starting the Culture coverslips, or transwells and incubate in a CO2 incubator at  (Day 0) 37 C. 2. Dilute 0.5 mL of 2.5% trypsin stock solution 1/10 in a 50 mL tube and refrigerate at 4 C. Prepare one tube per animal. 3. Aliquot 50 mL of HBSS for rinsing the CP and store at 4 C.

3.2 Cell Culture Decapitated pigs’ heads are available from local slaughterhouses (see (Day 1) Note 3). Bring the following items (Sect. 2.1) to the slaughter- house for sequences 1–4:

l 0.25% Trypsin solution on ice. l Autoclaved dissection instruments. l A spray bottle with 70% ethanol. l Lab coat and plastic gloves. 1. Dissect the brain from a pig’s head. Make incisions in the scalp, initially on the median line spanning the nose to the nape of the neck, and then on the line between the bilateral orbits and the ears (Fig. 1a). Peel off the scalp to expose the surface of the skull. Make sure to remove the periosteal covering to facilitate sawing (see Note 4). Remove the temporalis muscle and Culture of Choroid Plexus Cells 167

Fig. 1 Dissection sequence for collecting the porcine choroid plexus (CP). The most difficult part is trepanation of skull because of its extremely thick occipital bone (panel c). Dissected CP from the lateral ventricle is usually 10–12 cm in length (panel i)

muscles attached around the foramen magnum to facilitate sawing (Fig. 1b). 2. Using a saw, make an incision on the parietal and frontal bone at the line between the orbits, just posterior to the scalp inci- sion to avoid the region where fibrous components remain (Fig. 1c). Subsequently, make an oblique incision bridging the lateral aspects of the temporal bone to the upper rim of the foramen magnum. Since this is the thickest region of the skull, make sure to saw to just above the dura mater (see Note 5). To help detachment of the roof of the skull (calvaria) from the base of skull, use a chisel to break the bone effectively (Fig. 1d). 3. Detach the calvaria and expose the dura mater. Spray 70% ethanol on the exposed dura mater. To minimize the risk of contamination, change dissection tools to clean ones. After making incisions in the dura mater to expose the brain (Fig. 1e), carefully insert a pair of forceps and scissors into the cerebral longitudinal fissure to visualize the corpus callosum 168 Sen Takeda and Keishi Narita

(Fig. 1f). Cut the corpus callosum using the back of the scissor blade and expose the greater cerebral vein of Galen (Fig. 1g), to which a pair of posterior choroid plexus veins drain via internal cerebral veins. This is a hallmark to identify the CP of the lateral ventricles (see Note 6). 4. Dissect the CP from the lateral ventricles by gently pulling on the greater cerebral vein (Galen, Fig. 1h). Pinch the vein to pull out the CP (Fig. 1i) and put into HBSS briefly to wash out the blood. Then, transfer the tissue to a 15 mL tube containing trypsin solution. 5. Incubate the CP in 0.25% trypsin solution for 2.5 h at 4 C(see Note 7), followed by warming at 37 C for 30 min. 6. Stop trypsin digestion by adding 1 mL of fetal bovine serum and then scratch the tissue in the tube using a Pasteur pipette to detach the epithelial cell layer. Remove the undigested tissue debris. 7. Centrifuge at 250 Â g for 5 min and discard the supernatant. From one pair of CPs, approximately 800 μL of pellet (mostly erythrocytes) should be obtained. 8. Resuspend the pellet in 10 mL of culture medium and transfer all of it to a Matrigel-coated 90 mm dish. Incubate the culture overnight to allow epithelial cells to adhere to the dish.

3.3 Incubation 1. Rinse the culture several times with HBSS to remove blood to Confluence (Days cells, and then add fresh culture medium. Typically, approxi- 2–6) mately 50–60% of the surface of the dish will be covered by small patches of epithelial cells at this stage. 2. Change the culture medium every 2 days.

3.4 Maintenance 1. Check the status of the culture every day to confirm the forma- of Cells for Assays tion of a monolayer of cobblestone appearance. Around (After Day 6) 4–6 days are required to obtain a confluent culture (~1 Â 106 cells/dish; ~1.5 Â 104 cells/cm2). When plating in transwells, seed the cells at a density of >3.0 Â 104 cells/cm2 to ensure that the epithelial cells form a completely sealed monolayer (Fig. 2a, b). At confluency (see Note 8), CPECs display multi- ple cilia (see Note 9) on their apical surface (Fig. 2c). These are functionally active sensory cilia harboring G-protein-coupled receptors (see Note 10) and transient receptor potential (TRP) cation channels, such as TRPV4 (see Note 11). 2. After reaching confluency, change the media to serum-free medium and change every day to maintain the culture. Culture of Choroid Plexus Cells 169

Fig. 2 Cultured CPECs form unequivocal epithelial sheets. (a) A Coomassie Brilliant Blue-stained confluent culture of CPECs shows a cobblestone appearance. (b) Immunofluorescence image of (a) probed with anti-ZO- 1 and phalloidin reveals polygonal cells with centered nuclei (stained with DAPI), suggesting the existence of specific intercellular complexes such as tight junctions. (c) Scanning electron microscopy (SEM) of (b) reveals tiny projections from the apical membrane. (d) These structures are primary cilia, with an axonemal configuration of 9 + 0

3.4.1 Specific Protocol A: For performing the transcytosis assay (see Note 12), prepare the Transcytosis Assay confluent monolayers in the transwells. After obtaining a confluent culture, follow the instructions described below (Fig. 3). 1. Rinse the cell monolayer three times gently with serum-free medium.

2. Pre-equilibrate the culture by incubating for 1 h in a CO2 incubator. 3. Add fluorescently labeled tracer to the lower (basolateral) medium at concentrations of 1.0–10 μM. 170 Sen Takeda and Keishi Narita

Fig. 3 Schematic representation of the transcytosis assay. (a) In this panel, fluorescent dextran is added to the lower chamber to monitor transcytosis from the basolateral to apical direction. (b) Recovery of fluid from the upper chamber at regular time intervals shows the kinetics of fluid transport across the CPECs

4. Take two 10 μL samples from the upper (apical) medium every hour and transfer to a clear 96-well plate. 5. Measure the fluorescence of the tracer using a plate reader.

3.4.2 Specific Protocol B: To examine functionally whether the structural integrity of the Measuring epithelial monolayer has established and matured, measure the the Transepithelial transepithelial resistance of the CPEC sheet on the transwell Resistance according to the following procedures. 1. Test the meter according to the manufacturer’s user guide. 2. Test the electrodes using phosphate-buffered saline. 3. Sterilize the electrodes by immersing in 70% ethanol for 15 min. Culture of Choroid Plexus Cells 171

4. Using an empty transwell without cells, read the blank resis- tance by immersing the shorter and longer electrodes in the upper and lower medium, respectively. 5. Immerse the electrodes in the sample media. Typical values for transepithelial electrical resistance of CPECs are 55 Æ 5 Ω [8].

4 Notes

1. If you would like to establish a culture of rodent CPECs, CPs from one littermate of neonates work well for immunocyto- chemical purposes. However, as far as we have gone through, they will not generate monolayer epithelial sheets functionally qualified enough for the transcytosis assay. Therefore, it will be necessary to optimize the conditions to achieve monolayer cultures for that purpose. 2. We attempted to genetically knock down molecules responsible for ciliogenesis but it was difficult to balance viability and effects of treatment. Lipofection was the worst method to deciliate the CPECs as it devastated the cellular integrity, while lentiviral infection did not work well. 3. We isolate the choroid plexus tissue in a dissection room at the same premises, to prepare the fresh tissue as soon as possible. In particular, during the hot and humid summer season in Japan, transportation of decapitated heads at ambient temperatures for an hour severely decreases the quality of CP, such that it cannot expand to confluency. 4. Fibrous residue attached to the skull makes it difficult to saw the bone because of adherence of tissue to the instruments. An elevator, as used for orthopedic surgery, is useful for scraping away the periosteum and residual connective tissues. 5. The thickest part of porcine occipital bone reaches 5 cm or more and can take quite a long time to saw through. 6. Regarding CPs from other brain ventricles, those in the fourth ventricle are easily identified from the dorsolateral aspect of the cerebellum, where the lateral foramina of the fourth ventricle (Luschka) open to communicate with the subarachnoid space. To dissect the CP from the fourth ventricle, careful detachment of the cerebellum at the cerebellar peduncles facilitates its harvest. 7. During this step, we carry the sample back to the lab. Keep the samples in 0.25% trypsin on ice during transportation. The tissues tolerate these conditions well for 2.5 h. 172 Sen Takeda and Keishi Narita

8. If desired, formation of intercellular junctions between adja- cent cells can be assessed by immunocytochemistry, such as using an antibody for ZO-1 [3]. 9. For visualizing the multiple cilia, we recommend the use of anti-acetylated tubulin antibody (6-11B-1; Merck) for immu- nocytochemistry. Scanning electron microscopy (SEM) is an alternative to check for ciliogenesis but SEM cannot localize specific molecular identities, such as GPCRs. 10. Localization of specific receptors on cilia such as GPCRs can be visualized by double staining with an anti-acetylated tubulin antibody, as above. For example, we stain the cells with an anti- neuropeptide FF (NPFF) antibody. Functional activity of ciliary GPCRs can be assessed by ligand administration to the monolayer sheets cultured on transwells (Fig. 3b [3]). 11. Functionally active TRPV4 can be demonstrated by assessment of Ca2+ dynamics using ratiometry [8]. 12. In the transcytosis assay, tracers can be added to either side of the medium, according to the experimental design. When adding the tracers to the medium, make sure that they are evenly distributed. A small volume of concentrated tracers may not diffuse easily and may damage the cell monolayer, particularly when added to the upper medium.

Acknowledgments

We thank Osamu Kutomi, PhD, for his help in preparing Fig. 1. This study is supported by a Grant-in-Aid for Scientific Research from MEXT (17K08511) to S.T., and by the Japan Spina Bifida and Hydrocephalus Research Foundation to K.N. We thank Ann Turn- ley, PhD, from Edanz Group (www.edanzediting.com/ac) for edit- ing a draft of this manuscript.

References

1. Wolburg H, Wolburg-Buchholz K, Mack AF, 4. Papadopoulos MC, Verkman AS (2013) Aqua- Reichenbach A (2010) Ependymal cells. In: porin water channels in the nervous system. Squire LR (ed) Encyclopedia of neuroscience, Nat Rev Neurosci 14:265–277 Academic Press, pp 1133–1140. 5. Takeda S, Narita K (2011) Structure and func- 2. Brightman MW, Reese TS (1969) Junctions tion of vertebrate cilia, towards a new taxon- between intimately apposed cell membranes in omy. Differ Res Biol Divers 83:1–8. https:// the vertebrate brain. J Cell Biol 40:648–677. doi.org/10.1016/j.diff.2011.11.002 https://doi.org/10.1083/jcb.40.3.648 6. Takeda S, Yonekawa Y, Tanaka Y et al (1999) 3. Narita K, Kawate T, Kakinuma N, Takeda S Left-right asymmetry and kinesin superfamily (2010) Multiple primary cilia modulate the protein KIF3a: new insights in determination fluid transcytosis in choroid plexus epithelium. of laterality and mesoderm induction by KIF3A Traffic Copenhagen Denmark 11:287–301 (À/À) mice analysis. J Cell Biol 145:825–836. https://doi.org/10.1083/jcb.145.4.825 Culture of Choroid Plexus Cells 173

7. Nonami Y, Narita K, Nakamura H et al (2013) 10. Baehr C, Reichel V, Fricker G (2006) Choroid Developmental changes in ciliary motility on plexus epithelial monolayers—a cell culture choroid plexus epithelial cells during the peri- model from porcine brain. Cerebrospinal natal period. Cytoskeleton 70:797–803. Fluid Res 3:13. https://doi.org/10.1186/ https://doi.org/10.1002/cm.21132 1743-8454-3-13 8. Narita K, Sasamoto S, Koizumi S et al (2015) 11. Dawson HD, Chen C, Gaynor B et al (2017) The TRPV4 regulates the integrity of the blood- porcine translational research database: a manu- cerebrospinal fluid barrier and modulates trans- ally curated, genomics and proteomics-based epithelial protein transport. FASEB J research resource. BMC Genomics 18:643. 29:2247–2259. https://doi.org/10.1096/fj. https://doi.org/10.1186/s12864-017-4009-7 14-261396 12. Chakrabarti A, Schatten H, Mitchell KD et al 9. Banizs B (2005) Dysfunctional cilia lead to (1998) Chloral hydrate alters the organization altered ependyma and choroid plexus function, of the ciliary basal apparatus and cell organelles and result in the formation of hydrocephalus. in sea urchin embryos. Cell Tissue Res Development 132:5329–5339. https://doi. 293:453–462. https://doi.org/10.1007/ org/10.1242/dev.02153 s004410051137 Chapter 10

Drosophila as a Model to Study the Blood-Brain Barrier

Cameron R. Love and Brigitte Dauwalder

Abstract

The Drosophila blood-brain barrier (BBB) has been shown to be largely analogous in structure and function to the vertebrate BBB. Thanks to the genetic tools available for this organism, Drosophila is uniquely suited to study bbb physiology and function, with high relevance for mammalian function. In this chapter we discuss targeting strategies to specifically mark and manipulate BBB cells, how to test BBB integrity, and methods to isolate single-BBB cells.

Key words Drosophila, Blood-brain barrier, Subperineurial (SPG) cells, Gal4/UAS/Gal80ts, Cell dissociation

1 Introduction

Like in vertebrates, the insect blood-brain barrier protects the brain from components of the circulating fluid and allows selective uptake of nutrients and other important molecules. A big difference to vertebrates lies in the fact that insects do not have blood vessels but an open circulatory system. The hemolymph, the circulating fluid, is moved through the body of the animal by the pumping action of the heart and bathes all organs—except for the nervous system that is protected by the BBB. The insect hemolymph con- tains particularly high K+ concentrations that would be detrimental to neuronal function. Due to the absence of blood vessels, the blood-brain barrier surrounds the entire brain like a tight “cap” (Fig. 1). The insect BBB is best studied in Drosophila where it has been shown to be analogous in structure and function to the vertebrate BBB. In contrast to vertebrates, the insect BBB is formed by two layers of glial cells, an outer layer called perineurial glial (PG) cells and the inner layer, the subperineurial glial cells (SPG) with the septate junctions that form the tight barrier. The SPG barrier forms early in development and its cells do not divide in later stages anymore. Therefore, the number of SPG cells is low. To adapt to the growing brain, the cells flatten and become

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_10, © Springer Science+Business Media, LLC, part of Springer Nature 2019 175 176 Cameron R. Love and Brigitte Dauwalder

Fig. 1 Isolated adult Drosophila brains with BBB visualized. Dorsal is on top. (a) Nuclei of SPG cells are labeled by expression of dsRed. Genotype: Mdr-Gal4/+; UAS-dsRed/+. SPG cells are large and flat, and their number is low. (b) Both layers of the BBB (PG and SPG cells) express a genomically encoded indy-GFP fusion protein, visualized by immunohistochemistry (green). DNA is shown in blue (the majority of the nuclei seen are neuronal). An optical confocal section is shown to illustrate the tight barrier layer that surrounds the entire brain. The opening inside the brain with a BBB layer is the esophageal foramen

polyploid [1]. Despite differences between the insect and the mam- malian BBB, it has been shown that components that form the barrier (for example neurexins and forms of claudins) as well as many of the functional properties are shared between species. Sev- eral excellent reviews and recent papers discuss these aspects in detail [2–7]. Given the unparalleled genetic tools available in Dro- sophila, this tractable organism is uniquely suited to study BBB physiology and function, with high relevance for mammalian func- tion [8]. These studies go beyond mere barrier function, with novel insights starting to emerge about physiological processes inside BBB cells that influence neuronal development, adult neuronal function, and sex-specific behavior [9–12]. In this chapter we describe several methods that are being used to examine and manipulate Drosophila BBB function. We discuss the specific labeling of BBB cells using the Gal4/UAS/Gal80ts system, the dissociation and selection of BBB cells, and a method to examine BBB integrity.

2 BBB Labeling and Manipulation Using the Gal4/UAS System

One of the most powerful tools in Drosophila research, the binary Gal4/UAS system, allows the manipulation of cells in a tissue- specific and temporally restricted manner (Fig. 2a)[13, 14]. It allows the expression of desired molecules in the cells of choice and is efficiently used to label cells with fluorescent molecules, or to manipulate cells by expression of interfering/silencing RNAs (RNAi) or any transgene of interest. The Gal4/UAS system makes use of the yeast transcription factor Gal4 that binds to the UAS sequence (upstream-activating sequence) and activates Drosophila Blood Brain Barrier 177

A Tubulin promoter Gal80ts

18°C Bbb promoter Gal4 UAS Transcript of interest

32°C Bbb promoter Gal4 UAS Transcript of interest

Mature animal 32°C 12-16hrs B 18°C Development Eclosion Mature Day 7 Assay animal

Fig. 2 The Gal4/UAS/Gal80ts system allows temporally controlled expression of sequences of choice in the cells of choice. (a) Gal4 expression is directed to the BBB by a BBB-specific promoter. There, Gal4 binds to UAS and leads to the transcription of the downstream sequence. However, at the same time, a ubiquitous tubulin promoter guides direction of the temperature-sensitive Gal80ts, an inhibitor of Gal4. At 18 C, Gal80ts is active and inhibits Gal4 activity and transcription is blocked. To release this block, animals are shifted to 32 C. At this temperature Gal80ts becomes inactive, and Gal4 can initiate transcription from the UAS promoter. (b) Protocol to manipulate BBB cells specifically in mature adult flies using the approach described in (a)

transcription of sequences downstream of UAS. Gal4 and the UAS constructs are introduced into the flies as independent transgenes, and strains containing either element can be crossed as desired. The system allows expression of sequences of choice (such as fluorescent proteins, or interfering RNAs) in any cells of choice, as long as a specific promoter sequence is known that directs expression to the targeted cells. Temporal control of expression is achieved by the simultaneous presence of a transgene that ubiquitously expresses a temperature-sensitive inhibitor of Gal4, Gal80ts [14] (Fig. 2b). When the animals are kept at 18 C, Gal80ts represses Gal4 and no expression occurs from the UAS target. Upon shifting of the flies to temperatures between 29 and 32 C, Gal80ts is inactivated and Gal4 can begin transcription of the sequences downstream of UAS. The use of a conditional expression system to manipulate transcript levels at defined times allows the study of the temporal requirement of a gene. This is also valuable for genes that might have functions in both development and adulthood. 178 Cameron R. Love and Brigitte Dauwalder

2.1 Materials A critical feature for specificity of expression is the promoter used to drive the expression of Gal4. Several bbb-Gal4 transgenic lines have been generated and described. Among them, the moody-Gal4 (also called SNG-Gal4 or SPG-Gal4) lines generated by Bainton et al. [9] have been widely used [2, 3, 6, 7, 9]. Several sublines exist that contain the SNG-Gal4 insertion at different chromosomal location and that may vary in the degree to which they express in a few other cells outside of the SPG. As a standard procedure Gal4 lines should be crossed to a UAS-fluorescent protein to examine the expression pattern of a particular Gal4 line. Our lab has recently generated a Mdr65-Gal4 line using the promoter of the SPG-specifically expressed Mdr65 gene [6], a P-glycoprotein homolog (Fig. 2a; unpublished). In comparison, a genomic fusion construct that leads to the expression of the indy protein fused to GFP (indy- GFP) is expressed in both SPG and PG cells (as shown in Fig. 1b). Line 9-137-GAL4 (Ulrike Heberlein, Janelia Farm Research Cam- pus, VA) is expressed in both layers of the BBB [2]. A different transgenic line carrying indy-Gal4 has recently been described that is specifically expressed in PG cells [15]. These fly lines can be obtained from the labs that created them.

3 Assessment of Barrier Integrity

Like in mammalian systems, the integrity of the BBB is tested by injection of small-molecular-weight molecules into the circulatory system and assessment of their exclusion from the brain. 10 kDa Dextran coupled to fluorescent Texas Red (Dextran-TR) is effec- tively excluded from the brain in flies with an intact BBB. Since flies have an open circulatory system, the dye can be injected into the fly’s abdomen. Following injection the dye circulates throughout the body of the animal and will accumulate at the BBB, excluded from the brain, where its accumulation can be visualized following brain dissection. This system has also successfully been used to screen for new mutants that affect BBB integrity, be it in develop- ment or in adults. When the BBB is leaky, the dye will not be excluded and enter the brain and fluorescence can be seen accumu- lating in the eyes of intact flies when flies with unpigmented eyes (white eyes) are used [6]. Protocols for this assay have been devel- oped by Schwabe et al. and Bainton et al. [9, 16]. Figure 3a shows accumulation of 10 kDa Dextran-TR at the BBB surrounding the brain in wild-type flies. In contrast, flies with a leaky BBB (such as moody mutant flies [9]) have a defective barrier and the dye diffuses through the BBB as shown in Fig. 3b. Drosophila Blood Brain Barrier 179

bbb bbb TR neuronal TR nuclei neuronal nuclei

Wildtype moody mutant with leaky bbb

Fig. 3 An intact BBB is not permeable to 10 kDa-TR. (a) Following injection of 10 kDa Dextran marked with the Texas-Red (TR) fluorophore, Dextran-TR circulates in the hemolymph and accumulates at SPG cells. The optical confocal section shows accumulation of Dextran-TR (red) at the barrier. Neuronal cell bodies inside the brain are marked with DAPI. (b) moody mutant flies have a leaky BBB [9] and Dextran-TR can be seen entering the brain (picture reproduced with permission from Hoxha et al. [10])

3.1 Materials 1. 2.5 mM 10 kDa Texas-Red-conjugated Dextran, fixable (Invi- trogen D-1863) in H2O. 2. 4% Paraformaldehyde (EM grade, Polysciences Inc. #00380- 250) 3. Microinjector. 4. Razor blades. 5. Microscope slides and coverslips. 6. Double-sided tape. 7. Vectashield Antifade Mounting Media with DAPI (Vector Laboratories # H-1200). Brain dissections: Several YouTube videos are available with good instructions on how to dissect brains from fly heads. We like https://www.youtube.com/watch?v¼j4rVa7JCzdg [17]. It is important to clean the brains up as thoroughly as possible (i.e., remove fat body and trachea). The secret to intact and cleanly dissected brains is a LOT of practice! 180 Cameron R. Love and Brigitte Dauwalder

3.2 Approach 1. Anesthetize adult flies on ice. 2. Microinject a small amount (20–50 nL) of 2.5 mM 10 kDa Texas-Red-conjugated Dextran in H2O in between sternites or under the scutellum. 3. After injection allow flies to recover in regular food vials overnight. 4. Anesthetize flies on ice. 5. Remove fly heads with a razor blade and drop them into 4% paraformaldehyde in PBS. Fix heads for 30 min at room tem- perature (RT). 6. Remove the proboscis (mouth part) for enhanced penetration of the fixative and incubate for an additional 5 min at RT. 7. Dissect out the brain. Wash in 1Â PBS three times for 30 min each. 8. Mount the brains on a slide with Vectashield mounting media containing DAPI to stain DNA. After a few hours, seal the coverslip with nail polish. 9. View under a confocal microscope. DAPI-stained cell nuclei are visualized at 405 nm, and Texas Red Dextran at 633 nm.

3.3 Notes To mount fly brains on a slide without them being “smashed” by the coverslip, put a small square of double-sided tape onto the slide. With a razor blade cut a small window into the tape into which the brains will be put like in a basket. Add mounting media to the “window,” add the brains, and cover with a coverslip. A little “canal” can be cut from the square so that extra mounting media can drain. The tape will allow the brains some room while the height of the slide plus coverslip is still compatible with imaging under an upright microscope.

4 Isolation of BBB Cells

SPG cells can be isolated either manually or by FACS sorting. Both approaches require that the BBB cells are labeled by fluorescence (by using the Gal4/UAS expression system, for example, as described above) and that brains are dissected prior to isolation of the cells. It is important that surrounding tissues are removed carefully. Below we will describe the sorting of SPG cells following a protocol that was developed by DeSalvo et al. [2], the most efficient protocol developed to date. Figure 4 illustrates the pro- gression from SPG cells on the brain to isolated cells. It is worth noting that since SPG cells are large and very flat, once removed from their neighbors, their shape changes. Drosophila Blood Brain Barrier 181

Isolated brains before Brain after Isolated SPG cells Collagenase treatment Collagenase treatment A B C

D E F

Fig. 4 Dissociation of SPG cells from isolated brains. SPG cells were labeled by SPG-specific expression of cytoplasmic GFP (a–c), or by nuclear dsRed (d–f) using the Mdr-Gal4 driver. The whole-cell labeling in (a) illustrates the “cap-like” structure of the BBB. Isolated fly brains are shown (a, d). Following treatment with collagenase, the brains have lost most of the marked SPG cells (b, e). The brain in (b) is situated on the filter that is used to collect dissociated cells. (c) Dissociated GFP-labeled SPG cells after sorting. (f) An isolated SPG cell with dsRed expression in the nucleus

4.1 Isolation Cells can be sorted when marked with GFP or dsRed (or other of Fluorescently fluorescent proteins). The marker protein can either be cytoplas- Marked BBB Cells by mic, nuclear, or membrane bound, depending on the choice of the FACS Sorting UAS line and the protein localization signals attached to the pro- tein. A large variety of fly strains with UAS-fluorescent-protein transgenes are available from the Bloomington Stock Center (https://bdsc.indiana.edu/). Figure 3 illustrates the removal of BBB cells from brains in which BBB cells are labeled either by whole-cell and membrane-bound expression of GFP or by nuclear expression of dsRed.

4.1.1 Materials 1. 50 mg/mL Collagenase A in ddH2O (Millipore Sigma # 10103578001).

2. 50 mg/mL DNase I in ddH2O (NEB, #M0303S). 3. Schneider’s culture medium (BD Biosciences). 4. BSA (5 or 10%). 5. 0.5 M EDTA. 6. Thermomixer R (Eppendorf). 7. Dissecting forceps. 8. Eppendorf tubes. 182 Cameron R. Love and Brigitte Dauwalder

9. Ice. 10. 100 μm Filter unit that fits on top of 50 mL Falcon tube (Falcon Filters, # 352360).

4.1.2 Approach 1. Prepare the collagenase solution on the day of use. 2. Preheat thermomixer to 37 C: Fill slots with water, and check the temperature with a thermometer. 3. Dissect fly brains in cold-filtered Schneider’s medium contain- ing 1% BSA in batches of 10–15 per Eppendorf tube and keep on ice. Dissect for 2 h or less. 4. Coat 50 mL Falcon tubes with Schneider’s/BSA solution by adding 1 mL to the bottom of the tube, swirl around to coat the bottom, and remove. Continue with the same solution to coat all required 50 mL Falcons (one per sample). Place a 100 μm filter onto the Falcon tube. 5. Pre-coat a 1 mL tip with Schneider’s/BSA. Use it to remove the medium from the brain sample by holding the tube up to light to ensure that you don’t remove the brains from the bottom (be aware of any floating brains). 6. Add 1 mL Schneider’s/BSA to wash the samples. Remove solution and replace with 220 μL Schneider’s/BSA. 7. Add 10 μL collagenase and 5 μL DNase to the side of the tube and flick gently to mix. 8. Immediately insert tubes into the 37 C thermomixer and shake at 500 rpm for 5 min. Return samples to ice immediately. 9. Add 2.5 μL 0.5 M EDTA to each sample to inactivate the enzymes. Mix, then remove the Schneider’s/BSA + brains, and add to the appropriate 50 mL Falcon filter. Pipette any drops on the underside of the filter and add to the filtrate (be careful not to add bubbles). 10. Keep the tubes on ice until ready for FACS sorting. We have found it ok to keep cells on ice for transport to FACS sorting facility for 1–2 h after dissociation. 11. Prepare one sample with non-labeled BBB cells as a control for FACS sorting. 12. Coat the FACS tube and all pipettes to be used with Schnei- der’s medium/1% BSA. FACS sort the cells with a 100 μm nozzle into medium/1% BSA or RNA isolation buffer. 13. Nonfluorescent cells are sorted first to determine the window in which they appear. A fair amount of autofluorescence was observed in the control calibration experiment. This control was used to define the window for the collection of GFP-positive cells. 14. Samples can be processed for RNA extraction immediately or stored for later use at À80 C. Drosophila Blood Brain Barrier 183

4.1.3 Notes 1. We recommend following the procedures recommended by your FACS core for sorting. 2. It is helpful to have a dissecting scope with a UV source for dissection to check progress, but not necessary. 3. In our experience, there is a fairly large number of autofluores- cing cells and some cell debris. We have set a stringent cutoff for fluorescent-positive cells. 4. Yield from about 30–50 brains has been around 500–700 SPG cells. It is possible that this number can be increased when cells can be sorted sooner after dissociation. 5. Collagenase A: We have found that different lots of collage- nases can vary widely in their efficiency to dissociate BBB cells (even when ordered from the same supplier under the same order number). Therefore, new batches need to be optimized. Incubate samples at 750 rpm for 5 min if collagenase is ineffec- tive at 500 rpm for 5 min. We found one batch that was unable to remove the cells. A comparison of the composition of differ- ent lots with the help of the supplier showed that while colla- genase amounts were similar, the preparation contains other proteases whose amounts can vary widely. It is possible that they contribute to the dissociation of the cells and the varia- bility among lots. 6. If cells are not sorted, but the goal is to just enrich for SPG cells, following dissociation and filtration the cells can be pel- leted and dissolved in the desired solution. For example, the cells can be placed on a slide for visualization. If the volume containing the isolated cells is larger than desired, the cells can be pelleted by centrifugation at 4 C at 5000 rpm for 10 min and resuspended in the volume and medium of choice.

4.2 Remove Cells by We have found that due to the coherence of BBB cells (due to their Dissection septate junctions) it is possible to remove the cells in clusters under and Forceps a dissecting microscope with a UV light source using fine forceps. These collections will contain some non-BBB cells and will not be as clean as FACS-sorted preparations.

4.2.1 Materials 1. Flies in which SPG cells have been marked by expression of a fluorescent protein such as GFP or dsRed. 2. Dissecting microscope with UV light. 3. Small petri dish filled half with 1.5% agarose, covered with 1Â PBS, for dissection. 4. Ice. 5. Forceps: Either ultrafine (Dumont #5SF Forceps, order #11252-00) or fine (Dumont #5 fine forceps for dissection, straight, #11254-20); both can be used. 6. Dry ice. 184 Cameron R. Love and Brigitte Dauwalder

4.2.2 Methods 1. Anesthetize flies on ice. 2. Dissect flies on ice in a small petri dish half-filled with agarose covered with cold PBS. 3. Dissect fluorescent cells under a stereomicroscope with a UV source. 4. To transfer the cells to a solution such as Trizol, we have found it useful to freeze a droplet of a couple of microliters of Trizol in a weigh boat on top of dry ice. Touching of the little “frozen ball” with the forceps while the cells are still attached to the dissecting forceps causes the cells to “jump over” to the ice droplet. Several batches of cells can be accumulated on one droplet which can subsequently be frozen at À80 C for later processing.

Acknowledgments

This research has been supported by grants from the National Science Foundation (NSF) to B.D.

References

1. Unhavaithaya Y, Orr-Weaver TL (2012) Poly- 7. Stork T, Engelen D, Krudewig A, Silies M, ploidization of glia in neural development links Bainton RJ, Klambt C (2008) Organization tissue growth to blood-brain barrier integrity. and function of the blood-brain barrier in Dro- Genes Dev 26(1):31–36. https://doi.org/10. sophila. J Neurosci 28(3):587–597. https:// 1101/gad.177436.111 doi.org/10.1523/JNEUROSCI.4367-07. 2. DeSalvo MK, Hindle SJ, Rusan ZM, Orng S, 2008 Eddison M, Halliwill K, Bainton RJ (2014) 8. Hindle SJ, Munji RN, Dolghih E, Gaskins G, The Drosophila surface glia transcriptome: Orng S, Ishimoto H, Soung A, DeSalvo M, evolutionary conserved blood-brain barrier Kitamoto T, Keiser MJ, Jacobson MP, processes. Front Neurosci 8:346. https://doi. Daneman R, Bainton RJ (2017) Evolutionarily org/10.3389/fnins.2014.00346 conserved roles for blood-brain barrier xenobi- 3. DeSalvo MK, Mayer N, Mayer F, Bainton RJ otic transporters in endogenous steroid parti- (2011) Physiologic and anatomic characteriza- tioning and behavior. Cell Rep 21 tion of the brain surface glia barrier of Dro- (5):1304–1316. https://doi.org/10.1016/j. sophila. Glia 59(9):1322–1340. https://doi. celrep.2017.10.026 org/10.1002/glia.21147 9. Bainton RJ, Tsai LT, Schwabe T, DeSalvo M, 4. Hindle SJ, Bainton RJ (2014) Barrier mechan- Gaul U, Heberlein U (2005) moody encodes isms in the Drosophila blood-brain barrier. two GPCRs that regulate cocaine behaviors Front Neurosci 8:414. https://doi.org/10. and blood-brain barrier permeability in Dro- 3389/fnins.2014.00414 sophila. Cell 123(1):145–156 5. Limmer S, Weiler A, Volkenhoff A, Babatz F, 10. Hoxha V, Lama C, Chang PL, Saurabh S, Klambt C (2014) The Drosophila blood-brain Patel N, Olate N, Dauwalder B (2013) barrier: development and function of a glial Sex-specific signaling in the blood-brain barrier endothelium. Front Neurosci 8:365. https:// is required for male courtship in Drosophila. doi.org/10.3389/fnins.2014.00365 PLoS Genet 9(1):e1003217. https://doi.org/ 6. Mayer F, Mayer N, Chinn L, Pinsonneault RL, 10.1371/journal.pgen.1003217 Kroetz D, Bainton RJ (2009) Evolutionary 11. Chell JM, Brand AH (2010) Nutrition- conservation of vertebrate blood-brain barrier responsive glia control exit of neural stem chemoprotective mechanisms in Drosophila. J cells from quiescence. Cell 143(7): Neurosci 29(11):3538–3550. https://doi. 1161–1173. https://doi.org/10.1016/j.cell. org/10.1523/JNEUROSCI.5564-08.2009 2010.12.007 Drosophila Blood Brain Barrier 185

12. Speder P, Brand AH (2014) Gap junction pro- Perineurial barrier glia physically respond to teins in the blood-brain barrier control nutrient- alcohol in an Akap200-dependent manner to dependent reactivation of Drosophila neural promote tolerance. Cell Rep 22 stem cells. Dev Cell 30(3):309–321. https:// (7):1647–1656. https://doi.org/10.1016/j. doi.org/10.1016/j.devcel.2014.05.021 celrep.2018.01.049 13. Brand AH, Perrimon N (1993) Targeted gene 16. Schwabe T, Bainton RJ, Fetter RD, expression as a means of altering cell fates and Heberlein U, Gaul U (2005) GPCR signaling generating dominant phenotypes. Develop- is required for blood-brain barrier formation in ment 118(2):401–415 drosophila. Cell 123(1):133–144 14. McGuire SE, Roman G, Davis RL (2004) Gene 17. Wu JS, Luo L (2006) A protocol for dissecting expression systems in Drosophila: a synthesis of Drosophila melanogaster brains for live imaging time and space. Trends Genet 20(8):384–391 or immunostaining. Nat Protoc 1 15. Parkhurst SJ, Adhikari P, Navarrete JS, (4):2110–2115. https://doi.org/10.1038/ Legendre A, Manansala M, Wolf FW (2018) nprot.2006.336 Chapter 11

Zebrafish (Danio rerio) as a Viable Model to Study the Blood-Brain Barrier

Tianzhi Yang and Shuhua Bai

Abstract

As the blood-brain barrier (BBB) is essential for maintaining brain homeostasis and protecting the brain from exogenous substances, impermeability of the BBB is a major obstacle for drug delivery into the brain. Under pathological conditions, the integrity of the BBB is susceptible to disruption and can be broken down in severe brain diseases. Therefore, the understanding of intrinsic complexity as well as modulation of the BBB is critical to discover potential therapeutics for the treatment of brain diseases. Zebrafish (Danio rerio) have emerged as a suitable animal model in studying pathology of diseases and screening leading compounds in the drug development and discovery because of their highly conserved nature in both genetics and cell biology as higher vertebrates. Importantly, due to their small body size, ease of care, rapid development, and transparency in the early embryo stage, zebrafish allow researchers to study the BBB and carry out high-throughput screening of potential therapeutics with cost-effectiveness. We thus aim to provide a technical overview of the procedures that can be used to analyze BBB integrity and functionality in zebrafish. Low permeability and strong tight junction-based BBB in zebrafish are very similar to those of higher vertebrates. Zebrafish could be an excellent experimental model organism for studying the develop- ment and maintenance of the BBB, defining disease pathway, and discovering specific and powerful therapies for the treatment of brain diseases.

Key words Zebrafish, Blood-brain barrier, Permeability, Efflux, Drug delivery, Microinjection

1 Introduction

Blood-brain barrier (BBB) is the key to maintain brain homeostasis and proper neuronal activities as well as protect the brain from pathogens and harmful materials circulating in the blood. As an interface between the blood and brain, the BBB limits the transit of molecules based on size, charge, hydrophobicity, and/or affinity to carriers [1]. With these spontaneous hurdles in the delivery of most therapeutic drugs, successful BBB permeability is a prerequisite for the development of pharmaceuticals into the brain for the treat- ment of brain diseases [2]. The BBB is also a complex and dynamic interface responding to physiological changes. Many neurological diseases such as multiple sclerosis and cerebral ischemia often lead

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_11, © Springer Science+Business Media, LLC, part of Springer Nature 2019 187 188 Tianzhi Yang and Shuhua Bai

to disruption and breakage of the barrier [3, 4]. In order to develop highly needed therapeutics for the treatment of brain diseases, it is very important to understand regulatory mechanisms on drug delivery involved in the maintenance and modulation of the BBB under normal and pathological conditions [3]. Regarding the BBB study, in vitro cell culture models have shown many advantages compared to in vivo or ex vivo animal models, including (1) easy permeability assessment of potential drugs; (2) opportunity to elucidate the mechanism of drug trans- port; (3) quick evaluation of strategies for improving drug absorp- tion; (4) possible performance of studies on human cells; and (5) great potential to minimize time-consuming and expensive animal studies [5]. Although the basic knowledge gained from cell culture models provides a strong foundation for the design of more complex in vivo experiments, no adequate in vitro tissue culture displays the full anatomical and physiologic permeability and enzymatic barrier characteristics of the mammalian BBB sys- tem. The genetics, cell biology, and functionality are not well con- served when compared to higher vertebrates, including the endothelial tight junctions at the BBB along with the association of astrocytes, pericytes, and neurons [1]. While considerable studies have been carried out over the previous two decades in establishing useful in vitro cultures for the BBB, no model is able to fully mimic the high electrical resistance, low paracellular permeability, and selective influx/efflux characteristics of the in vivo BBB situation. In vertebrates, the BBB consists of tight junctions between adjacent endothelial cells and restrict paracellular hydrophilic mate- rials [6]. An endothelial cell monolayer with high transendothelial electrical resistance (TEER, greater than 120 Ω cm2) and low solute À permeability (on the order of 10 4 cm/min) indicates the well- formed tight junctions among the cells [6]. So far, great advances have been made in determining possible molecular compositions forming the tight junction at the BBB [7, 8]. The major compo- nents of tight junction at the BBB include tetraspanning integral membrane proteins such as claudins as well as associated cytoplas- mic anchoring proteins such as ZO-1 [7, 9, 10]. Notably, claudin- family proteins are involved in establishing tight junctions and regulating size selectivity [9]. Furthermore, ATP-binding cassette (ABC) transporters are also highly expressed at the BBB [11] (Fig. 1). These ATP-driven efflux pumps with remarkably broad substrates are responsible for the inability of many xenobiotics to enter the brain [12]. Penetration of therapeutics is limited by the efflux transporters and this leads to insufficient drug distributions in the brain (Fig. 2). As a complex and dynamic interface, the BBB is naturally responding to physiological changes. Tight-junction proteins and efflux transporters are affected by and can even be promoted by brain diseases. Alterations in their expression have Zebrafish BBB Model 189

Fig. 1 Permeability study of markers across the BBB in zebrafish, including labeled common vein (injection size) in zebrafish embryo (a), microinjection manipulator (b), and confocal images of transgenic Tg(fli1:GFP) zebrafish (c) with green blood vessels and distribution of red injected doxorubicin in the vasculature and brain at 5 dpf. Fluorescent trackers retained in the restricted brain vessels and crossed out of the body vessels after the injection

Fig. 2 Expression of tight-junction protein and multidrug resistance protein in the zebrafish brain. (a) Multidrug resistance 1 (MDR-1) was detected by western blots in both wild AB zebrafish embryo and adult. (b) Tg(fli1: GFP) zebrafish embryos at 5 dpf were stained with monoclonal claudin-5 antibody and confocal image of whole-mount embryos showed the expression of claudin-5 (blue) at the BBB 190 Tianzhi Yang and Shuhua Bai

been identified in many brain disorders, including epilepsy, brain cancer, spinal cord injury, Alzheimer’s disease, and Parkinson’s disease [12]. Zebrafish (Danio rerio) has been widely used in the drug devel- opment and discovery over the past few decades, ranging from drug screening, identification and target confirmation to toxicity assess- ment [13, 14]. The use of zebrafish essentially owns unique fea- tures from both cell culture and animal models. With large-scale and high-throughput advantages just like cell studies, zebrafish provide a distinct in vivo system to validate the BBB with the full anatomical and physiologic permeability and tight junction char- acteristics of systems [15, 16]. Multiple advantages come from zebrafish models on the BBB studies including their high fecundity, rapid development, transparency during embryonic and larval stages, available genetic editing tools, pharmacological manipula- tions, and cost-effectiveness [17]. The nervous system is formed at 1 day postfertilization (dpf) and by 3 dpf the BBB is observed in zebrafish embryos [15, 18]. About 3 mm of larvae at 5 dpf and 3 cm size of adult zebrafish enable large numbers of these verte- brates to be maintained in a relatively small laboratory space [19]. The transparency in zebrafish embryos allows living track of microstructure of BBB without complicated surgery and histology. These types of dynamic studies have never been exploited in in vitro, in cell, or in vivo higher animal models [13]. Performing studies in zebrafish would greatly improve our understanding of the regulatory network and function for the BBB.

2 Materials

2.1 Zebrafish Wild-type AB and transgenic Tg(fli1:EGFP) zebrafish were obtained from the Zebrafish International Resource Center (Uni- versity of Oregon, Eugene, OR).

2.2 Reagents 60 μg/mL Sea salts of egg water.

Hanks’ solution with 0.137 M NaCl, 5.4 mM KCl, 0.25 mM Na2H PO4, 0.44 mM KH2PO4, 1.3 mM CaCl2, 1.0 mM MgSO4, and 4.2 mM NaHCO3. Zebrafish system water. 1% Evans blue dye. Agarose. Injection needles. Clark borosilicate standard wall capillaries with filament. Mineral oil. 0.04% Tricaine (3-aminobenzoic acid ethyl ester, MS-222). Zebrafish BBB Model 191

Doxorubicin. ® BODIPY 564/570-labeled paclitaxel. Pierce protein BCA assay kit. Cell lysis buffer. 1-Phenyl 2-thiourea (PTU).

3% H2O2/0.5% KOH medium. 4% Paraformaldehyde. Western lightening chemiluminescence reagents. Primary and secondary antibodies. 0.125% Trypsin. 1% BSA, 3% normal horse serum, and 0.4% Triton X-100 in 1Â PBS blocking solution.

2.3 Preparation 4 g of agarose was added in 100 mL of embryo medium and heated of Agarose-Lined in a microwave to be dissolved. The clear agarose solution was first Plates slowly poured to one half of 60 mm petri dish. A lined mold was gently placed down starting with the mold at one end and slowly lowering the whole body onto the agarose (Note 1). When the agarose was cooled and hardened at room temperature, the lined mold was removed. The plate was finally sealed with Parafilm and stored in a 4 C refrigerator.

2.4 Preparation Borosilicate glass microcapillary injection needles (Harvard Appa- and Loading ratus, Holliston, Massachusetts, 1 mm OD Â 0.78 mm) were of Injection Needles prepared using a micropipette Flaming/Brown p-97 puller device (Sutter Instrument Inc., Novato, CA, USA) with the setting for a long tip of air pressure 500, heat 510, pull 100, velocity 200, and time 60. The needle tip was carefully broken and a tip opening of 5 μm was obtained. Pulled micropipette was first backfilled with mineral oil using a 30 G Â 200 needle and syringe ensuring no air bubbles before the attachment to an injector. At the home position, the micropipette needle was placed in the Nanoject IITM Auto- Nanoliter Injector connected to a pneumatic microinjection pump (Drummond Scientific Company, Broomall, PA, USA).

2.5 Zebrafish Zebrafish pigmentation begins in the retinal epithelium and mela- Pigment Inhibition nophore pigment cells and develops rapidly within hours of embryogenesis. In order to improve signal detection by whole- 2.5.1 PTU Method mount in situ hybridization or confocal microscopy, embryos can be treated with 1-phenyl 2-thiourea (PTU) during embryogenesis to inhibit pigmentation. PTU inhibits melanogenesis by blocking tyrosinase-dependent steps in the melanin pathway, but can be toxic at high concentrations [20]. To inhibit pigment formation, embryos were treated with 0.1 mM PTU (Sigma-Aldrich, St. Louis, MO, USA). At this concentration, PTU-treated fish did not show a 192 Tianzhi Yang and Shuhua Bai

significant difference in experimental results compared to untreated embryos [20]. Zebrafish were treated at the 28-cell stage and the embryos remained transparent as long as the PTU treatment was continued (Note 2).

2.5.2 Oxidization Method Although PTU prevents the formation of melanin pigments and greatly facilitates visualization of the final signal, a reduction of cell viability in catecholaminergic neuronal cells has been reported after PTU treatment [20]. As PTU is also an inhibitor of tyrosinase, an enzyme required for melanin synthesis, the biosynthesis of dopa- mine in catecholaminergic neurons regulated by tyrosine hydroxy- lase has been inhibited. An alternative endpoint method using hydrogen peroxide to remove the zebrafish pigment can avoid the above side effects. After treatments, embryos were first euthanized by 0.4% overdose of tricaine (3-aminobenzoic acid ethyl ester, MS-222, Sigma-Aldrich, St. Louis, MO, USA) and then fixed in 4% paraformaldehyde in 1Â PBS overnight at 4 C(Note 3). Fixed embryos were placed in a 24-well plate up to 20 embryos and incubated at room temperature in a 3% H2O2/0.5% KOH medium with slight shaking for 1 h until pigmentation completely disap- peared. There could be a lot of bubbles formed in the medium (Note 4). After pigment removal, the embryos were washed for 5 min in 1Â PBS to remove the H2O2 and stop the bleaching reaction. The treated embryos were further dehydrated in 25%, 50%, and 100% methanol in PBS for 15 min at room temperature, respectively. The final dehydration for embryos was placed at À20 C in 100% methanol for at least 2 h.

3 Methods

3.1 Zebrafish Care Zebrafish eggs, embryos, and adults were raised and bred according to standard methods [21]. All zebrafish eggs were kept in 60 μg/mL sea salts of egg water and stored in the incubator at 28.5 C. 100 eggs were divided in each petri dish (100 Â 15 mm) filled with 20 mL of warm egg water (Note 5). After 2 dpf (day postfertilization), zebrafish embryos were moved into embryo medium (Hanks’ solution with 0.137 M NaCl, 5.4 mM KCl, 0.25 mM Na2HPO4, 0.44 mM KH2PO4, 1.3 mM CaCl2, 1.0 mM MgSO4, and 4.2 mM NaHCO3). If needed, zebrafish were removed from the egg (dechorionate) by gently pipetting and released to the medium. All chemical salts were purchased from VWR (Radnor, PA, USA). 50 zebrafish embryos per petri dish were kept in 28.5 C incubator. 1% Evans blue dye (Sigma-Aldrich, St. Louis, MO, USA) in water prepared fresh was dosed in embryo medium for the suppression of fungal growth. The fish were checked and all dead fish were removed daily. Adult zebrafish were maintained at Zebrafish BBB Model 193

28.5 C on a 14-h light/10-h dark cycle in standard system water (pH 7.0–7.2, Aquatic Habitats, Apopka, FL, USA). Five adult fish could be placed in 1 L of water and be fed daily at regular intervals. For mating, a ratio of one male to two females was established in one tank for maximum embryo production at the appropriate time.

3.2 Permeability In order to test the permeability of marker agents, the formulation of Markers Across of a fluorescent marker can be administered by incubating embryos the BBB in Zebrafish in medium or by microinjection administration. The incubation method is much easier and less time consuming compared to the microinjection. However, treated marker agents can be absorbed in the zebrafish through skin, gill, and/or GI tract. It does not guar- antee accurate doses in the study. A microinjection technique is provided here to demonstrate the permeability of injected fluores- cent marker examples, such as Rhodamine 123 or anticancer drugs, across the BBB and entering the brain. Although it needs certain time and skills for the administration injection, the protocol below provides a more precise distribution study of injected agents in the zebrafish brain as well as body.

3.2.1 Micropipette Once the micropipette was tightly secured to the collet of micro- Calibration manipulator, excess oil was expelled by pressing and holding the “empty” button. After that, the injected formulation solution was retracted by pressing and holding the “fill” button and about 5 μL of sample was drawn at the tip of the pipette. Injection volume was determined according to a previous report [22]. Briefly, a mineral droplet was mounted on the scale of the object micrometer. The arising vehicle sphere was measured with the scale of the object micrometer. Each vehicle was administered several times into the oil droplet until the target injection volume was achieved. According to the sphere volume formula (V ¼ 1/6πd3), a sphere diameter of 1 bar corresponded to an injection volume of 0.5 nL. Two bars corresponded to 4.6 nL.

3.2.2 Microinjection Zebrafish embryos were anesthetized with 0.04% tricaine (MS-222) for microinjection and imaging. Single embryos were transferred and aligned into the trenches of an agarose mold-type injection plate. Autofluorescent anticancer drug doxorubicin or fluorescence-labeled paclitaxel (1 mg/ml in PBS) was injected into the common cardinal vein of embryos using a Nanoject IITM Auto-Nanoliter (Drummond Scientific Company, Broomall, PA, USA) with Â20 magnification under an inverted microscope (Leica M80, Leica Microsystems Inc. Buffalo Grove, IL).

3.2.3 Living Image The living embryos were washed three times in 1Â PBS and placed in 1% agarose containing tricaine. Live embryos/larvae or immu- nohistochemical samples were mounted in 1% low-melting agarose. 194 Tianzhi Yang and Shuhua Bai

Samples were imaged at 18 h postinjection using an Olympus LSM 1000 confocal microscope with FluoView 10 software (Olympus Corp., Central Valley, PA, USA). Green (GFP) and red (doxorubi- cin and BODIPY® 564/570-labeled paclitaxel) channels were excited using an argon/krypton and helium/neon laser, and emis- sions were detected and analyzed using filters set by the Olympus Confocal Software. After injection, living embryos were returned to embryo medium and incubated at 28.5 C.

3.3 Protein In order to better assess the functional permeability and integrity of Expression at the BBB the BBB in zebrafish, it is crucial to characterize and monitor in Zebrafish changes in permeability and functionality-related proteins. Immu- nohistochemical staining can easily localize specific molecular com- ponent markers of functional BBB in the zebrafish brain microvasculature including the tight-junction proteins zonula occludens-1 and claudin-5 [18]. In addition, electron micrographs indicated that the junctions formed between overlapping edges of brain endothelial cells and astrocytes in adult zebrafish brains [23]. Studies have also demonstrated the restricted permeability of brain blood vessels and ultrastructural similarities of the BBB among zebrafish, rodents, and man. This chapter discussed the application of immunochemical staining and western blotting methods to analyze the expression of functional protein markers as they permit comparisons of barrier integrity and expression of barrier-related proteins.

3.3.1 Protein Extraction Heads from 20 zebrafish embryos or 1 brain from 1 adult fish were added in cold 1Â PBS. The samples were then centrifuged and all supernatant solution was removed as much as possible. A commer- cial cell lysis buffer (Cell Signaling Technology Inc., Boston, MA, USA) was added in the samples and all samples were homogenized for 30 s. Protein was isolated from samples and the concentrations of isolated protein were determined by a Pierce BCA assay kit (Sigma-Aldrich, St. Louis, MO, USA).

3.3.2 Western Blotting Functional proteins in zebrafish were analyzed by a western blot- ting method according to previously published procedures [24, 25]. Proteins were separated and electrophoretically trans- ferred to a polyvinylidene difluoride (PVDF) membrane. After the membranes were treated with multidrug resistance protein 1 (MDR1) or ATP-binding cassette subfamily B member 1 (ABCB1) antibody as an example of functional protein, a second- ary antibody (Life Technologies, Grand Island, NY, USA) was added. Signals for MDR1 efflux transporter protein were detected by Western Lightening Chemiluminescence Reagents (Amersham Biosciences, Inc., Piscataway, NJ, USA). The protein levels were quantified from the densitometric intensity of each radiographic band using a Bio-Rad Quantity One software (version 4.5.2, Zebrafish BBB Model 195

Bio-Rad Laboratories, Hercules, CA, USA). Results were normal- ized to the total protein loading confirmed by BCA protein assay and expressed as a percentage of the band volume (the product of western band intensity and band area) of treatment compared to that derived from the buffer control.

3.3.3 Whole-Mount After three washes in 1Â PBS, the zebrafish embryo samples were Immunohistochemical digested with 0.125% trypsin (Invitrogen, Grand Island, NY, USA) Staining for 20 min at room temperature. The samples were incubated in a blocking solution containing 1% BSA, 3% normal horse serum, and 0.4% Triton X-100 in 1Â PBS with mouse anti-claudin-5 (Life Technologies, Grand Island, NY, USA) at 4 C for 8 h. After thorough wash with 0.4% Triton X-100 in 1Â PBS, the samples were incubated in the blocking solution with the appropriate secondary antibody (Invitrogen, Grand Island, NY, USA) at room temperature for 1 h. Images were taken with the samples mounted in 1% agarose using an Olympus LSM 1000 confocal microscope with FluoView 10 software (Olympus Corp., Central Valley, PA, USA).

4 Notes

1. Make sure that there are no bubbles in the agarose. 2. PTU treatment must be initiated before the initial pigmenta- tion because it does not remove already formed pigment. 3. Paraformaldehyde solution should be fresh and not older than 2 days. 4. Hydrogen peroxide solution should be prepared fresh immedi- ately before use. Wear gloves to avoid skin irritation from KOH medium. 5. Eggs should be checked and any debris, dead eggs (whitish/ opaque), and unfertilized eggs (very small in comparison to others) should be removed daily. Altogether, cerebral microvessels are impermeable to fluores- cent markers possibly because tight junctions and efflux transpor- ters detected in zebrafish are concomitant with restriction. Additional real-time analysis of fluorescent tracers in embryonic zebrafish suggests that they may be used as an in vivo model for the study of drug delivery across the BBB.

References

1. Wilhelm I, Fazakas C, Krizbai IA (2011) In 2. De Rosa G, Salzano G, Caraglia M, Abbruzz- vitro models of the blood-brain barrier. Acta ese A (2012) Nanotechnologies: a strategy to Neurobiol Exp (Wars) 71:113–128 overcome blood-brain barrier. Curr Drug Metab 13:61–69 196 Tianzhi Yang and Shuhua Bai

3. Pardridge WM (2005) The blood-brain bar- 92:567–570. https://doi.org/10.1038/clpt. rier: bottleneck in brain drug development. 2012.168 NeuroRx 2:3–14. https://doi.org/10.1602/ 16. Xie J, Farage E, Sugimoto M, Anand-Apte B neurorx.2.1.3 (2010) A novel transgenic zebrafish model for 4. Pardridge WM (2007) Blood-brain barrier deliv- blood-brain and blood-retinal barrier develop- ery. Drug Discov Today 12:54–61. https://doi. ment. BMC Dev Biol 10:76. https://doi.org/ org/10.1016/j.drudis.2006.10.013 10.1186/1471-213X-10-76 5. Nicolazzo JA, Charman SA, Charman WN 17. Santoriello C, Zon LI (2012) Hooked! Mod- (2006) Methods to assess drug permeability eling human disease in zebrafish. J Clin Invest across the blood-brain barrier. J Pharm Phar- 122:2337–2343. https://doi.org/10.1172/ macol 58:281–293. https://doi.org/10. JCI60434 1211/jpp.58.3.0001 18. Jeong JY et al (2008) Functional and develop- 6. Reichel A, Begley DJ, Abbott NJ (2003) An mental analysis of the blood-brain barrier in overview of in vitro techniques for blood-brain zebrafish. Brain Res Bull 75:619–628. barrier studies. Methods Mol Med https://doi.org/10.1016/j.brainresbull.2007. 89:307–324. https://doi.org/10.1385/1- 10.043 59259-419-0:307 19. Chakraborty C, Hsu CH, Wen ZH, Lin CS, 7. Davis TP, Abbruscato TJ, Egleton RD (2015) Agoramoorthy G (2009) Zebrafish: a complete Peptides at the blood brain barrier: knowing animal model for in vivo drug discovery and me knowing you. Peptides 72:50–56. https:// development. Curr Drug Metab 10:116–124 doi.org/10.1016/j.peptides.2015.04.020 20. Karlsson J, von Hofsten J, Olsson PE (2001) 8. Keaney J, Campbell M (2015) The dynamic Generating transparent zebrafish: a refined blood-brain barrier. FEBS J 282:4067–4079. method to improve detection of gene expres- https://doi.org/10.1111/febs.13412 sion during embryonic development. Mar Bio- 9. Haseloff RF, Dithmer S, Winkler L, technol (NY) 3:522–527. https://doi.org/10. Wolburg H, Blasig IE (2015) Transmembrane 1007/s1012601-0053-4 proteins of the tight junctions at the blood- 21. Avdesh A et al (2012) Regular care and main- brain barrier: structural and functional aspects. tenance of a zebrafish (Danio rerio) laboratory: Semin Cell Dev Biol 38:16–25. https://doi. an introduction. J Vis Exp e4196. https://doi. org/10.1016/j.semcdb.2014.11.004 org/10.3791/4196 10. Kniesel U, Wolburg H (2000) Tight junctions 22. Sive HL, Grainger RM, Harland RM (2010) of the blood-brain barrier. Cell Mol Neurobiol Calibration of the injection volume for micro- 20:57–76 injection of Xenopus oocytes and embryos. 11. Mahar Doan KM et al (2002) Passive perme- Cold Spring Harb Protoc 2010:pdb ability and P-glycoprotein-mediated efflux dif- prot5537. https://doi.org/10.1101/pdb. ferentiate central nervous system (CNS) and prot5537 non-CNS marketed drugs. J Pharmacol Exp 23. Eliceiri BP, Gonzalez AM, Baird A (2011) Zeb- Ther 303:1029–1037. https://doi.org/10. rafish model of the blood-brain barrier: mor- 1124/jpet.102.039255 phological and permeability studies. Methods 12. Mahringer A, Ott M, Reimold I, Reichel V, Mol Biol 686:371–378. https://doi.org/10. Fricker G (2011) The ABC of the blood-brain 1007/978-1-60761-938-3_18 barrier - regulation of drug efflux pumps. Curr 24. Yang T et al (2015) Exosome delivered anti- Pharm Des 17:2762–2770 cancer drugs across the blood-brain barrier for 13. Rubinstein AL (2003) Zebrafish: from disease brain cancer therapy in Danio rerio. Pharm Res modeling to drug discovery. Curr Opin Drug 32:2003–2014. https://doi.org/10.1007/ Discov Devel 6:218–223 s11095-014-1593-y 14. Lieschke GJ, Currie PD (2007) Animal models 25. Yang T et al (2017) Delivery of small interfering of human disease: zebrafish swim into view. Nat RNA to inhibit vascular endothelial growth fac- Rev Genet 8:353–367. https://doi.org/10. tor in zebrafish using natural brain endothelia 1038/nrg2091 cell-secreted exosome nanovesicles for the 15. Umans RA, Taylor MR (2012) Zebrafish as a treatment of brain cancer. AAPS J model to study drug transporters at the blood- 19:475–486. https://doi.org/10.1208/ brain barrier. Clin Pharmacol Ther s12248-016-0015-y Chapter 12

Evans Blue-Albumin as a Marker to Evaluate Blood-Brain Barrier Integrity in Neonatal and Adult Rodents

Fabricia Petronilho, Julia L. Goldman, and Tatiana Barichello

Abstract

There has been an increase in the study of brain barriers and the several roles that they may play in neurological disorders. This type of research requires appropriate models and markers to demonstrate the integrity of the interface between the blood and the brain. Historically, dyes have been used to analyze the blood-brain barrier (BBB) mechanisms and for measurements of plasma volume. Despite some limitations, Evans blue is still the most commonly used marker for investigations of brain barrier integrity in in vivo applications. In this book chapter, we describe a simple and reproducible method for the evaluation of BBB integrity using Evans blue as a marker. Our protocol is focused on the evaluation of BBB integrity in neonatal and adult rodents, and the methods are divided into two protocols.

Key words Blood-brain barrier, Evans blue, Rodent

1 Introduction

The term blood-brain barrier (BBB) describes the structural, phys- iological, and molecular mechanisms that control the entry and exit of molecules between the blood and the brain [1]. Important nutrients penetrate brain barriers by passive diffusion or on the polarity of the proteins [2, 3]. At the same time, several substances that can damage the brain are unable to cross the BBB, which maintains brain homeostasis [2]. At the interface between systemic circulation and the central nervous system (CNS), the BBB is composed of highly specialized and polarized endothelial cells with tight junctions sealing the intercellular clefts, basement membranes, pericytes, and astrocyte end-feet with anchoring transmembrane proteins, and it establishes communication with the neurons in the neurovascular unit [4]. These structures and molecules result in the characteristically stable internal environment of the brain, both during development and in adults [5].

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_12, © Springer Science+Business Media, LLC, part of Springer Nature 2019 197 198 Fabricia Petronilho et al.

Changes in any of these structures directly affect the concen- tration ratios and thus the apparent permeability. In response to stressful events such as trauma, burn, or infection, the character- istics of these barriers can be altered, leading to edema and recruit- ment of inflammatory cells into the brain parenchyma [6]. However, it is often not clear whether barrier dysfunction is involved in the primary pathology or if barrier dysfunction occurs as a consequence of the pathology, perhaps exacerbating the effects of the disorder [7, 8]. In this sense, the BBB can play an important role in the pathophysiology of neurological damage, and this notion has prompted a renewed interest in studies of its function and integrity [9]. Such studies require suitable models and markers to demonstrate the integrity of the interfaces between the blood and the brain. Dyes have a venerable history in studies of BBB mechanisms in both the developing and the adult brain [1]. The first use of dyes for this purpose was by Ehrlich in the mid-nineteenth century or by Goldmann (1909–1913) in the early twentieth century [10]. One dye in particular, Evans blue, is still the most commonly used marker for studies of BBB integrity. It has been widely claimed that Evans blue binds tightly and exclusively to plasma albumin and that visualization and/or quan- titation can be used to define increases in BBB permeability to albumin [11–13]. The techniques used to assess BBB integrity in vivo in neonatal (Protocol A) and adult (Protocol B) rodents with the Evans blue assay are described in this chapter, followed by observations concerning the limitations of Evans blue.

2 Materials

2.1 Animals 1. Neonatal rats or mice, 3–4 days postnatal (15–20 g and 1–5 g of body weight, respectively) (Sect. 3.1). 2. Adult rats or mice, 60 days old (250–350 g and 20–26 g of body weight, respectively) (Sect. 3.2).

2.2 Reagents 1. Evans blue dye. 2. Phosphate-buffered saline (PBS), pH 7.4. 3. Saline (0.9%, w/v) NaCl (sterile). 4. Ethyl alcohol. 5. Fifty percent trichloroacetic acid solution. Evans Blue-Albumin as a Marker to Evaluate Blood-Brain Barrier Integrity... 199

2.3 Preparation Prepare a 1% (10 mg/mL) solution of Evans blue dye in PBS of Evans Blue Solution solution (Sect. 3.1) or a 2% (20 mg/mL) solution of Evans blue dye in PBS solution (Sect. 3.2/Note 1).

2.4 Preparation Prepare 50% trichloroacetic acid solution powder in distilled water, of 50% Trichloroacetic and mix to dissolve completely (Note 2). Acid Solution

2.5 Equipment 1. Molecular device and a multimode microplate reader. 2. Perfusion apparatus.

3 Methods

3.1 Protocol A 1. Intraperitoneal administration is commonly used in rats and (Neonatal Model) mice and results in a faster absorption into the vasculature. Anesthesia is not required. 2. Aspirate 40 (mice) or 160 (rats) μL of 1% Evans blue dye solution into a 1 mL syringe. Please avoid introducing air bubbles into the syringe. 3. Immobilize the neonatal rodent with the right hind limb immobilized and the head and body tilted downward. Hold onto the tail with the nondominant hand between the thumb and forefinger. 4. Insert the needle (small gauge, 26–30) at a 10–15 angle into the peritoneal cavity in the caudal right abdominal quadrant through the skin and the abdominal wall, thereby avoiding injection into the cecum or the stomach on the left side. Slowly inject 1% Evans blue solution (80 mg/kg) into the rat experi- mental model or the mouse experimental model. 5. Place the rodent back into its cage and wait until the extremities turn blue, which indicates that the dye has moved in the rodent’s circulation. The rodent needs to be euthanized within an hour to avoid physiological clearance of the Evans blue from the system [14] (Fig. 1).

3.2 Protocol B 1. Aspirate 0.1 (mice) or 1.2 (rats) mL of 2% Evans blue dye (Adult Model) solution into a 1 mL syringe. 2. Intravenously inject Evans blue into the lateral tail vein in mice or rats. The lateral tail-vein injection is the preferred technique for vascular access in mice. Place the rodent into a restraining device such that the rodent is not freely mobile but the tail can be handled. Insert the needle (28–30 gauge needle for mice and 25–27 gauge needle for rats) into the caudal vein toward the direction of the head. Keep the needle and syringe parallel to the tail and slowly inject the Evans blue solution. Anesthesia is not required for this procedure. 200 Fabricia Petronilho et al.

Fig. 1 Blood-brain barrier perfusion. After anesthesia, the rodents are placed on their backs, and their feet are pinned on a dissection board. The abdominal and thoracic cavities are then opened using surgical scissors to expose the thoracic area. The heart is then flushed slowly with 25–50 mL of sterile saline in mice and 200–250 mL in rats through the left ventricle at 110 mm Hg until colorless perfusion fluid is obtained from the cut right atrium

3. Alternatively, intravenously inject Evans blue into the femoral vein in mice or rats. Anesthetize the rodents with a mixture of ketamine (80–100 mg/kg) and xylazine (10 mg/kg), given intraperitoneally. Shave the fur from the surgical region (inner leg region). Using 70% ethanol scrub, scrub the shaved surgical regions starting in the center and making a circular sweep outward. Place the animal on a sterile surface, and place a sterile drape over the surgical areas. Ensure that all surgical tools for the procedure have been sterilized. Place the rat onto its back (supine position), and make an incision of approximately ½ in. (12 mm) in the inguinal area along the natural angle of the hind leg. Blunt dissect to separate the connective tissue until the femoral artery and vein are exposed (e.g., with blunt-tipped scissors, hemostats, cotton swabs). This separation is typically accomplished by holding the blunt-tipped scissors and/or cot- ton swabs at a 45 angle to ensure the easier localization of the region of interest. The vein is dark red in color, and the artery is clearer and brighter than the vein. The nerve that runs along the artery is whitish in color. Insert the needle (28–30 gauge needle for mice and 25–27 gauge needle for rats) into the femoral vein toward the direction of the head. Keep the needle and syringe parallel to the vein, and slowly inject the Evans blue solution (2% wt/vol in normal saline) (4 mL/kg). 4. The rodent should be euthanized within 1 h to avoid physio- logical clearance of the Evans blue from the system [15] (Fig. 1). Evans Blue-Albumin as a Marker to Evaluate Blood-Brain Barrier Integrity... 201

3.3 Brain Collection, 1. Anesthetize the rodents by intraperitoneal administration of a Extraction, mixture of ketamine (80–100 mg/kg) and xylazine (5–10 mg/ and Quantification kg) [15] or isoflurane. Isoflurane should be delivered as a of Evans Blue Dye known percentage (we recommend 4–5% for deep anesthesia during this procedure) in oxygen from a precision vapor- izer. Ensure deep anesthesia by testing for reflexes (i.e. toe pinch response) before proceding to the next step. 2. Place the rodent on their back and pin its feet on a dissection board. Open the abdominal and thoracic cavity using surgical scissors to expose the thoracic and abdominal organs. 3. Flush the heart slowly with 25–50 mL of saline in mice and 200–250 mL in rats through the left ventricle at 110 mm Hg until a colorless perfusion fluid is obtained from the cut right atrium (see Fig. 1). 4. After perfusion, remove the head using a guillotine. Make a midline incision along the integument from the neck to the nose, and expose the skull. Place the sharp end of a pair of iris scissors into the foramen magnum on one side, and carefully slide the scissors along the inner surface of the skull. Next, make a cut extending to the distal edge of the posterior skull surface. Make an identical cut on the contralateral side. Care- fully slide the scissors along the inner surface of the skull as the tip travels from the dorsal distal posterior corner to the distal frontal edge of the skull while lifting the blade as you cut to prevent damage to the brain. Repeat for the opposite side. Using a spatula, sever the olfactory bulbs and nervous connec- tions along the ventral surface of the brain. Gently remove the brain away from the head, trimming any dura that still connects the brain to the skull (see Fig. 2). Place the brain in an Eppen- dorf tube (1.5 or 2.0 mL).

Fig. 2 Whole brain perfused to evaluate BBB permeability. (a) Perfused brain from the healthy control group and (b) perfused brain after 24 h from adult rats subjected to pneumococcal meningitis 202 Fabricia Petronilho et al.

5. Weigh an empty tube and bring the balance value to zero. Transfer the brain and weigh it. Repeat for all brain samples. 6. Add 500 μL of 1:3 (wt:vol) trichloroacetic acid solution to the sample, and homogenize the sample. 7. Centrifuge the sample for 20 min at 10,000 rpm to pellet any remaining tissue fragments. 8. Dilute the Evans blue in the supernatant of each sample by adding ethanol (1:4) to each tube. 9. Measure the fluorescence (excitation at 620 nm and emission at 680 nm) in 100 μL of each sample, and an external standard (62.5–500 ng/mL) in the same solvent using a 96-well plate reader. 10. Calculate the dye concentration as the ratio of absorbance rela- tive to the amount of tissue. Calibration factor (CF) ¼ ng/ mL EB Ä absorbance of external standard. Absorbance of the sample  mean CF Ä weight of sample

4 Notes

1. To prepare the Evans blue solution, filter-sterilize the solution to remove any particulate matter that has not dissolved, and stir constantly. 2. The trichloroacetic solution should be stored at 4C and protected from light.

References

1. Bentivoglio M, Kristensson K (2014) Tryps brain and Neurotoxicology. Neurotoxicology and trips: cell trafficking across the 100-year- 33(3):586–604 old blood-brain barrier. Trends Neurosci 37 6. de Wit NM, Vanmol J, Kamermans A, (6):325–333 Hendriks J, de Vries HE (2016) Inflammation 2. Rochfort KD, Cummins PM (2015) The at the blood-brain barrier: the role of liver X blood-brain barrier endothelium: a target for receptors. Neurobiol Dis 107:57–65 pro-inflammatory cytokines. Biochem Soc 7. Danielski LG, Giustina AD, Badawy M, Trans 43(4):702–706 Barichello T, Quevedo J, Dal-Pizzol F et al 3. Saunders NR, Habgood MD, Mollgard K, (2017) Brain barrier breakdown as a cause Dziegielewska KM (2016) The biological sig- and consequence of neuroinflammation in sep- nificance of brain barrier mechanisms: help or sis. Mol Neurobiol 55(2):1045–1053 hindrance in drug delivery to the central ner- 8. Saunders NR, Liddelow SA, Dziegielewska vous system? F1000Res 5. https://doi.org/10. KM (2012) Barrier mechanisms in the devel- 12688/f1000research.7378.1 oping brain. Front Pharmacol 3:46 4. Quaegebeur A, Lange C, Carmeliet P (2011) 9. Saunders NR, Ek CJ, Habgood MD, Dziegie- The neurovascular link in health and disease: lewska KM (2008) Barriers in the brain: a molecular mechanisms and therapeutic impli- renaissance? Trends Neurosci 31(6):279–286 cations. Neuron 71(3):406–424 10. Saunders NR, Dreifuss JJ, Dziegielewska KM, 5. Ek CJ, Dziegielewska KM, Habgood MD, Johansson PA, Habgood MD, Mollgard K Saunders NR (2012) Barriers in the developing et al (2014) The rights and wrongs of Evans Blue-Albumin as a Marker to Evaluate Blood-Brain Barrier Integrity... 203

blood-brain barrier permeability studies: a 13. Yen LF, Wei VC, Kuo EY, Lai TW (2013) walk through 100 years of history. Front Distinct patterns of cerebral extravasation by Neurosci 8:404 Evans blue and sodium fluorescein in rats. 11. Manaenko A, Chen H, Kammer J, Zhang JH, PLoS One 8(7):e68595 Tang J (2011) Comparison Evans blue injec- 14. Barichello T, Dagostim VS, Generoso JS, tion routes: intravenous versus intraperitoneal, Simoes LR, Dominguini D, Silvestre C et al for measurement of blood-brain barrier in a (2014) Neonatal Escherichia coli K1 meningi- mice hemorrhage model. J Neurosci Methods tis causes learning and memory impairments in 195(2):206–210 adulthood. J Neuroimmunol 272(1–2):35–41 12. Saunders NR, Dziegielewska KM, Mollgard K, 15. Belayev L, Busto R, Zhao W, Ginsberg MD Habgood MD (2015) Markers for blood-brain (1996) Quantitative evaluation of blood-brain barrier integrity: how appropriate is Evans blue barrier permeability following middle cerebral in the twenty-first century and what are the artery occlusion in rats. Brain Res 739 alternatives? Front Neurosci 9:385 (1–2):88–96 Chapter 13

Experimental Tools to Study the Regulation and Function of the Choroid Plexus

Isabel Gonc¸alves, Telma Quintela, Ana Catarina Duarte, Peter Hubbard, Grac¸a Baltazar, Christian Schwerk, Andrea Carmine Belin, Joana Toma´s, and Cecı´lia Reis A. Santos

Abstract

There is an increasing recognition of the choroid plexus’ (CP) functional relevance for brain homeostasis, and its malfunction has been associated with neurologic diseases, in newborns, young adults, and the elderly, like kernicterus, multiple sclerosis, and Alzheimer’s disease. Yet, the CP still remains an overlooked organ requiring further investigation. The minute size of the CP, particularly in rodent models, increases the difficulties associated with the implementation of suitable protocols to address the ever-increasing research questions. In recent years we have implemented fundamental methods to study gene expression and function in the CP. These include CP epithelial cell (CPEC) primary cultures; use of CP explants for expression analysis, and electrophysiol- ogy and bioluminescence assays; Ca2+ imaging; gene silencing in CP epithelial cell lines; and transport studies across blood-cerebrospinal fluid barrier (BCSFB) in vitro models. This chapter describes these protocols aiming to attract more researchers willing to enhance the current knowledge on CP functions and the relevance of its malfunction to the central nervous system pathophysiology.

Key words Choroid plexus, Calcium imaging, Primary cultures, siRNA transfection, Explant electro- physiology, Blood-cerebrospinal fluid barrier, Choroid plexus epithelial cells, Cell culture filter inserts, Transepithelial electrical resistance, Bioluminescence assay

1 Introduction

The choroid plexuses (CPs), located in the ventricles of the brain (Fig. 1a), are formed by single layers of cuboidal epithelial cells lying on a basement membrane. The apical cell membranes face the cerebrospinal fluid (CSF). Extensive infolding at the basement membrane contacts the underlying connective tissue, where fene- strated capillaries enable nutrients, oxygen, and signaling molecules to the CPs’ epithelial cells (CPEC) and remove excreted molecules originating from CPEC detoxification processes and from the brain metabolism [1–3]. CPEC are connected by tight junctions,

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_13, © Springer Science+Business Media, LLC, part of Springer Nature 2019 205 206 Isabel Gonc¸ alves et al.

Fig. 1 The choroid plexus. (A) Location of the CP in the brain ventricular system. (B) Main biological functions of the choroid plexuses. CP choroid plexuses, CSF cerebrospinal fluid, CPEC choroid plexuses epithelial cell, TJ tight junction, AJ adherens junction, ZO zonula occludens protein, JAM junctional adhesion molecules, Cyp450 cytochrome P450, GPX glutathione peroxidase, UGT UDP glucuronosyltransferase [18]

adherens junctions, and desmosomes, forming a sealed barrier that prevents paracellular movement of substances into and out of the brain, thereby forming the blood-CSF barrier (BCSFB) (Fig. 1b). Besides functioning as a chemical and biological barrier, other well- recognized functions of CPs are CSF formation [4], nutrient and hormone supply to the CSF and brain, clearance and detoxification of waste products from brain metabolism [5–7], immune surveil- lance [1, 8], amyloid clearance [9–11], and neurogenesis [12–14]. More recently identified functions of the CPs are chemical surveillance as depicted from the presence of the taste and odorant transduction pathways in CPEC [15, 16], and the potential func- tion of the CP as an extra-suprachiasmatic nucleus circadian clock [17, 18]. Due to its wide range of action, CPs’ dysfunctions often lead to neuronal diseases, and several alterations in the CPs’ function have been reported in neuronal diseases from newborns, young adults, and the elderly raising the interest in the study of these epithelia even further [19]. In newborns, kernicterus, a severe neuronal condition resulting from the accumulation of unconjugated biliru- bin during the neonatal period, is associated with the impaired uptake of this compound at the apical membrane of the choroidal tissue, combined to a reduced efflux into the blood via the baso- lateral MRP1 pump [20]. In young adults, the CPs of patients with multiple sclerosis display T lymphocytes in vessels and stroma, vascular cell adhesion molecule-1 expression on endothelia, and intense HLA-DR (human leukocyte antigen-antigen D related) immunostaining in CP and epiplexus cells [21]. Additionally, the upregulation of intercellular adhesion molecule-1 in the CPs’ Choroid Plexus Protocols 207

epithelium and the de novo expression of mucosal addressin cell adhesion molecule-1 in the experimental autoimmune encephalo- myelitis (EAE) mice model of MS sustains the relevance of the CP for the entrance of the cells of the immune system into the brain [22]. In the elderly, the CP epithelial cells present a general atrophy, and the blood vessel walls nourishing these epithelia become thicker [23]. As a consequence the CP synthetic capacity decreases and CSF production diminishes, as well as the clearance of CSF out of the brain, compromising the excretion of toxic compounds. This includes the CPs’ capacity to eliminate amyloid beta (Aβ)[9] enhancing its accumulation as observed in the brains of Alzheimer’s disease patients. Conversely, the accumulation of Aβ in the CP further enhances the disruption of the BCSFB [24], with potential impact on neurodegeneration [19]. In spite of the overall increasing recognition of the CPs’ func- tional relevance and the impact of CP-associated dysfunctions in neurologic diseases, the CP still remains as an overlooked organ requiring further investigation at all life stages from the early embryonic development till adulthood and aging. The minute size of the CP, particularly in rodent models, increases the difficulties associated with the implementation of suitable protocols to address relevant research questions. In recent years we have implemented indispensable methods to study gene expression and function of CP cells. These include CPEC primary cultures [25]; use of CP explants for expression analysis, and elec- trophysiology and bioluminescence assays; Ca2+ imaging and gene silencing in CP epithelial cell lines [15]; and transport studies across BCSFB in vitro models [26]. These protocols are described in detail in this chapter and intend to attract more researchers to enhance the current knowledge on CP functions and relevance of its distur- bance to central nervous system pathophysiology.

2 Materials

2.1 Primary Cultures A. Animals and Equipment of Rat Choroid Plexus 1. Neonatal (2–5 days old) Wistar Han rats. Epithelial Cells (CPEC) 2. Sterile dissection instruments: scissors, fine and curved forceps, disposable scalpels and blades, spoons (Fine Science Tools). 3. Dissection stereomicroscope. 4. Sterile Petri dishes, serological pipettes, pipette tips, 1.5 mL tubes, 50 mL tubes. 5. Sterile 12-well culture plates. 6. Filtration units for the preparation of sterile solutions (0.22 μm). 208 Isabel Gonc¸ alves et al.

B. Reagents 1. Phosphate-buffered saline (PBS): 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4. Adjust the pH to 7.4 and autoclave. Store at room temperature (RT). 2. Digestion solution: Dissolve 2 mg of Pronase (Fluka, Seelze, Germany) in 1 mL of PBS and filter using a 0.22 μm filter. The solution should be freshly prepared and kept at 4 C until use. 3. Dulbecco’s modified Eagle’s medium (DMEM; Sigma D5523). 4. Fetal bovine serum (FBS; Biochrom AG, Berlin) heat inactivated. 5. Penicillin-streptomycin (Pen/Strep; Sigma): with 10,000 units penicillin and 10 mg streptomycin/mL. 6. Epidermal growth factor (EGF; Sigma) Human, Recombinant. 7. Insulin from bovine pancreas (Sigma). 8. Cytosine β-D-arabinofuranoside (ara-C; Sigma). 9. DMEM (for 100 mL): Combine 10 mL FBS, 1 mL Pen/- Strep, 0.37 g sodium bicarbonate, and 89 mL DMEM. Filter and store at 4 C. 10. Supplemented DMEM (for 100 mL): Combine 10 mL FBS, 1 mL Pen/Strep, 0.37 g sodium bicarbonate, 500 μL ara-C, 50 μL insulin, 10 μL EGF, and 88.44 mL DMEM. Filter and store at 4 C.

2.2 Use of Ex Vivo CP A. Animals and Equipment Explants for 1. Neonatal (5–6 days old) Wistar Han rats. Experimental 2. Sterile scalpel, scissors, and forceps. Procedures 3. 24-Well culture plates. 2.2.1 Ex Vivo CP Explants 4. Petri dishes, serological pipettes, pipette tips, 1.5 mL tubes, for Immunohistochemistry 15 mL tubes. 5. Glass slides and coverslips. 6. Confocal laser scanning microscope LSM 710 (Zeiss, Germany). B. Reagents 1. Phosphate-buffered saline (PBS) (see Sect. 2.1-B). 2. Paraformaldehyde (PFA) 4% in PBS: Store at 4 C. 3. 30% Sucrose solution in PBS: Store at 4 C. 4. Blocking and antibody dilution solution: 2.5% Bovine serum albumin (BSA), 0.2% Triton X-100 in PBS. Store at 4 C and prepare fresh every week. Choroid Plexus Protocols 209

5. Washing solution: PBS with 0.1% Tween 20. Store at 4 C. 6. Primary and secondary antibodies for target proteins. 7. Hoechst 33342 (Molecular Probes) for nuclear staining. 8. Mounting medium (Dako).

2.2.2 Ex Vivo CP Explants A. Animals and Equipment for PERIOD2::LUCIFERASE 1. Two- to four-month-old transgenic knock-in PERIOD2:: Bioluminescence Assay LUCIFERASE (PER2::LUC) [27] female mice housed in a 12-h light/12-h dark cycle. 2. 40 mm Sterile cover glasses (40 mm circle no.1). 3. High Vacuum Grease (Dow Corning). 4. 35 mm Petri dishes. 5. Culture membranes (Milli-CM 0.4 μm, Millipore, Bedford, MA). 6. Micro-sample Osmometer (Fiske 210). 7. Photomultiplier tube (PMT) device (LumiCycle). 8. LumiCycle, Actimetrics Inc. Software. 9. Origin, OriginLab, Northampton, MA, USA. B. Reagents 1. Air-buffered culture medium (for 1000 mL): Combine Dulbecco’s modified Eagle’s medium-low glucose (Sigma; D2902) lacking phenol red (phenol red reduces light signal penetration), 20 mL B-27 Supplement (50Â) (Gibco), 4.7 mL sodium bicarbonate 7.5% solution (Gibco), 10 mL HEPES 1 M buffer solution (Gibco), 2.5 mL Pen/Strep 10,000 U/mL (Gibco), and 3.5 g D-glucose powder (Sigma). Add and mix all the components to 800 mL auto- claved Milli-Q H2O. Adjust pH to 7.2 and bring the volume up to 1000 mL with autoclaved Milli-Q water. Adjust the osmolality between 285 and 315 mOsm/Kg (closer to 300 mOsm/kg is ideal). Filter sterilize the culture medium, store at 4 C protecting from light, and pre-warm to 37 C before use. 2. Hanks’ balanced salt solution (HBSS; for 1000 mL): Com- bine 100 mL HBSS 10Â (Gibco), 10 mL Pen/Strep 10,000 U/mL (Gibco), 5 mL sodium bicarbonate 7.5% solution (Gibco), and 10 mL HEPES 1 M buffer solution (Gibco). Add and mix all the components to 800 mL and bring the volume up to 1000 mL with autoclaved Milli-Q water. Filter sterilize the HBSS and store at 4 C. 3. Luciferin medium: Add 10 μL of luciferin (final concentra- tion 0.1 mM, beetle luciferin, potassium salt, Promega Co., Madison, WI) to 10 mL of air-buffered culture medium (use 210 Isabel Gonc¸ alves et al.

approximately 1.2 mL of luciferin medium per culture dish). Prepare and place luciferin medium in a dark heating cham- ber (37 C) just before the luciferase activity assay.

2.3 Use of HIBCPP A. Cells and Equipment Cells on Cell Culture 1. Cells: Human CP Papilloma Cell Line HIBCPP [28]. Inserts for 2. Cell culture filter inserts (pore diameter 3.0 μm, pore den- Experimental sity 2.0 Â 106 pores per cm2, membrane surface area Procedures 0.33 cm2; Greiner Bio-One, Frickenhausen, Germany). 3. Millicell-ERS2 with an STX-01 electrode (Merck- Millipore, Schwalbach, Germany). 4. 24-Well cell culture plates (CYTOONE, Starlab, Hamburg, Germany). 5. 12-Well cell culture plates (CYTOONE, Starlab, Hamburg, Germany); see Note 1. B. Reagents 1. Dulbecco’s modified Eagle’s medium: Nutrient F-12 Ham (DMEM/F12 (Ham); Gibco, Thermo Fisher, Darmstadt, Germany) supplemented with 10% fetal calf serum (FCS; Life Technologies, Thermo Fisher, Darmstadt, Germany), insulin (5 μg/mL; Sigma, Deisenhofen, Germany), and pen- icillin/streptomycin (100 U/mL/100 μg/mL; Thermo Fisher, Darmstadt, Germany); store at 4 C and pre-warm to 37 C before use. Use medium without phenol red when permeability measurements using fluorescein isothiocyanate- labeled inulin (FITC-inulin) are intended to avoid interfer- ence with the fluorescence of the FITC-inulin. 2. Phosphate-buffered saline (PBS; Gibco, Thermo Fisher, Darmstadt, Germany); pre-warm to 37 C before use. 3. Trypsin-ethylenediaminetetraacetic acid (trypsin-EDTA; Gibco, Thermo Fisher, Darmstadt, Germany); store at 4 C and pre-warm to RT before use. 4. Fluorescein isothiocyanate-labeled inulin (FITC-inulin; Sigma, Deisenhofen, Germany); generate a 5 mg/mL stock solution in cell culture medium [DMEM/F12 (Ham)] and store at À80 C; pre-warm to RT before use. 5. 80% Ethanol (EtOH).

2.4 Single-Cell A. Cells and Equipment Calcium Imaging in CP 1. Cells: Human CP Papilloma Cell Line HIBCPP [28] and Epithelial Cells Murine Choroidal epithelial Cell line Z310 [29]. Microscope and accessories: Inverted fluorescence microscope (Axio Imager A1, Carl Zeiss, Germany); Lambda DG4 apparatus (Sutter Instruments, Novato, CA, USA); Bandpass filter Choroid Plexus Protocols 211

(Carl Zeiss, Germany); AxioVision camera and software (Carl Zeiss, Germany). 2. Fiji software (MediaWiki, USA, and Germany). B. Reagents 1. “10% Total culture medium”: Dulbecco’s modified Eagle’s medium DMEM/HAM’s F12 1:1 (Gibco) supplemented with 4 mM L-glutamine (Gibco), 10% heat-inactivated fetal calf serum (FCS, Invitrogen), 5 μg/mL insulin (Invitrogen), and antibiotics(100 U/mL penicillin and 100 μg/mL streptomycin). 2. 25% Trypsin (Gibco) and 1 mM EDTA. 3. I-Slide 8-well ibiTreat chambers (Ibidi, Germany). 4. Tyrode’s solution: 140 mM NaCl, 5 mM KCl, 1.0 mM MgCl2, 2 mM CaCl2, 10 mM Na-pyruvate, 10 mM glucose, 10 mM HEPES, 5 mM NaHCO3 (pH 7.4). 5. Fura-2 AM (Molecular Probes, Thermo Fisher Scientific). 6. Odorant stimuli: Spermine, spermidine, cadaverine, and putrescine (Sigma, UK). 7. Bitter receptor agonists: Denatonium benzoate, D-salicin, 6-propyl-3-thiouracil, N-phenylthiourea (Sigma, UK). 8. Bitter receptor blocker: Probenecid (Sigma, UK).

2.5 Silencing Gene A. Reagents Expression in CP 1. siPORT amine transfection agent (Ambion, Thermo Fisher Epithelial Cells Scientific). 2. siRNA Gαolf (GNAL, s128137; Ambion, Thermo Fisher Scientific). ® 3. Scramble siRNA: Silencer Select Negative Control No.1 siRNA (Ambion, Thermo Fisher Scientific). 4. Gαolf antibody (Santa Cruz Biotechnology, Dallas, TX, USA). 5. ImageJ software: http://imagej.nih.gov/ij/.

2.6 Ex Vivo A. Tissue and Equipment Electrophysiological 1. Whole brains taken from mice sacrificed by cervical disloca- Studies with the CP tion with the CP exposed by cutting into the ventricles (with a scalpel) from the top of the cerebrum. 2. Basic electrophysiological setup, including D.C. amplifier (e.g., for recording intracellular potentials), filters, and data acquisition system (i.e., PC running appropriate software and with analogue-digital converter). Small tissue chamber (e.g., Petri dish with rubber support for the brain). Ideally, record- ing should take place within a Faraday cage. 212 Isabel Gonc¸ alves et al.

3. Dissecting or stereomicroscope. 4. Micromanipulators: These can be relatively course (e.g., Nar- ishige M-152). 5. Stimulus delivery system: This consists of plastic tubing (~1 mm i.d.) and glass tubes (e.g., hematocrit tubes) for positioning the artificial cerebrospinal fluid (aCSF) flow onto the CP, and a tap or valve to introduce stimulus- containing aCSF into this flow. B. Reagents 1. Artificial cerebrospinal fluid (aCSF: 148 mM NaCl, 3.0 mM KCl, 1.5 mM CaCl2, 1.0 mM MgCl2, 1.4 mM NaHPO4, 5.56 mM glucose, pH 7.4). 2. Chemical (odorant) stimuli, including positive control (e.g., cadaverine) dissolved in aCSF. 3. Modulators of intracellular transduction pathways (e.g., SQ22536 or ion channel blockers); the tools for investigat- ing the CP response dissolved in aCSF.

3 Methods

3.1 Primary Cultures To ensure the most sterile environment possible, all the protocol is of Rat CP Epithelial carried out inside the laminar flow hood. Cells (CPEC) Place the animals in a paper or other barrier material on crushed ice for up to 15 min for hypothermia anesthesia. After decapitation, 3.1.1 CP Isolation remove the skin with the help of two forceps and cut the skull using scissors from the base of the head to the mid-eye area. Peel the skull away and carefully remove the brain. Place the brain into a Petri dish containing cold PBS. Under a stereomicroscope, position the brain at the midline. Using fine forceps and a disposable scalpel, make three incisions (one parallel and the others perpendicular to the midline of the brain) and pull the cortex away, exposing the lateral ventricle. Pull gently the CPs and place them in 1 mL of cold (4 C) PBS.

3.1.2 CP Cell Remove carefully the PBS solution with a sterile pipette so as not to Dissociation and Culture disturb the CPs that have settled at the bottom of the tube. Add 1 mL of digestion solution, gently invert the tube, and incubate at 37 C for 5 min. Centrifuge the cell suspension at 500 Â g for 2 min at RT. Remove the supernatant and wash the pellet with 1 mL DMEM. Mechanically dissociate cells by slowly passing the tissue through a 1 mL pipette tip. Repeat this washing step, removing the medium and replacing it with fresh DMEM (pipette up and down). Finally remove the supernatant and replace with 12 mL of supple- mented DMEM. Plate the cells onto 12-well plate culture wells (approximately two CPs per well). Gently agitate the plate in several Choroid Plexus Protocols 213

directions to disperse the cells and to ensure coverage of the well,  and place it in a humidified incubator in 95% air-5% CO2 at 37 C. Replace the growth medium 1 day after the initial seeding, and every 2 days thereafter. Cells become ready for experiments within 4–5 days after seeding.

3.1.3 Characterization The protocol for primary cultures of CPEC developed in our of CP Epithelial Cells laboratory turned out to be a simple, economic, and consistent method. Optimal results were obtained with partially dissociated tissue, with cell aggregates showing adhesion and proliferation in non-coated culture wells. Cell adhesion was optimal after 1 day in culture, with about 50–60% of the dissociated cells attached to the culture plates [25]. In isolation of CP epithelial cells, contamina- tion with fibroblasts is a common problem. In our protocol, con- tamination with non-CPEC was controlled with ara-C, and homogeneous cultures of highly enriched CPEC were obtained. The cells displayed predominantly the characteristic polygonal epi- thelial morphology, and could be maintained in culture for 1 week [25]. Once CPEC have been successfully isolated and grown in culture, typically 100% of the cells are labeled for transthyretin (TTR; Fig. 2), a thyroxine transport protein known to be exclu- sively produced by the choroidal epithelia in the central nervous system [30]. Western blot analyses further confirm the production and secretion of TTR by these cells (Fig. 3).

Fig. 2 Detection of TTR in CPEC obtained from newborn rats by immunocytochemistry. TTR is stained green, and localizes in the cytosol of CPEC [25] 214 Isabel Gonc¸ alves et al.

Fig. 3 Comparison of TTR levels in CPEC obtained from newborn and adult rats by Western blot [25]

These results indicate that primary cultures of CPEC display the properties of choroidal epithelia and may be used for blood- CSF barrier research.

3.2 Use of Ex Vivo Immunohistochemistry protocols allow the staining and localiza- CP Explants tion of a protein in a sectioned tissue. However, the CP sectioning for Experimental can introduce artifacts or even induce structural modifications. Our Procedures approach overcomes the classical methodology since it enables the direct imaging of the whole CP tissue, such as the imaging of the 3.2.1 Ex Vivo CP Explants CP three-dimensional morphology or the localization of taste as a Model to pathway-related proteins [16]. This protocol can be followed to Immunohistochemistry study the expression and localization of any protein in the CP tissue Studies with or without incubation with various stimuli (e.g., hormones, odorants, peptides). A. Collection of CP Explants Collect the CP from the lateral ventricles of euthanized Wistar Han rats (postnatal days 5–6), previously anesthetized by hypo- thermia. Using scissors, separate the head from the body, and with dissection scissors and forceps gently remove the skull and the cranium to expose the brain. Remove the brain and place into a Petri dish with cold (4 C) PBS. Next, with a scalpel perform two perpendicular cuts in one of the brain hemispheres. Carefully, using forceps “open” the cut to expose the lateral ventricle and collect the CP. Repeat the process in the other brain hemisphere. B. Immunostaining Protocol After collection, place immediately each CP explant into a well of a 24-well plate with 300 μL of PFA 4%, for 30 min at RT, to fix the tissue. Use a plate shaker during incubations and washing steps to obtain best results. Then, aspirate the PFA 4% and wash three times with 300 μL of PBS. Use a Pasteur pipette with a white tip to aspirate carefully. Incubate the CP explants for 2 h with a 30% sucrose solution and wash three times with PBS. During 4 h, incubate the tissues with 300 μL of blocking solution. For primary antibody treatment dilute the antibody to the desired concentration in the same solution, and incubate the CP explants with 150 μL, overnight at 4 C. Next, wash five times with 300 μL of PBS-T 0.1% and incubate for 3 h, at RT, Choroid Plexus Protocols 215

with 150 μL of the secondary antibody (typically a 1/10000 dilution). Perform this and the following steps in the dark. Wash again five times with 300 μL of PBS-T 0.1% and, for nuclei staining, incubate the CP explants with 150 μL of Hoechst 33342 dye (1:1000, Molecular Probes) for 20 min. Wash three more times with 300 μL of PBS-T 0.1%. Mount the CP explants on glass slides using forceps and a paintbrush, and place some mounting media (Dako) on the CP explants and a cover- slip. Observe the slides using a confocal microscope LSM 710 (Zeiss, Germany). Select some regions in each CP explant and acquire Z-stack images at 0.5 μm thickness (63Â magnifi- cation). Using the Zen software (Zeiss, Germany) analyze the protein localization through orthogonal view. Also, 3D images can be reconstructed using the Zen software.

3.2.2 Circadian Luciferase reporter technology is used widely in circadian rhythm Regulation of the PER2:: studies, to measure the expression and function of clock genes/ LUC Bioluminescence proteins. In general, luciferases, enzymes that catalyze biolumines- Rhythm in Ex Vivo CP cence reactions, are used as reporters. Several studies have described Explants the advantage of transgenic knock-in mice that contain a fusion protein of PERIOD 2 (PER2) and the firefly enzyme LUCIFER- ASE. Those studies also characterized the dynamics of biolumines- cence rhythms using in vitro cultures or explants cultured ex vivo [31]. However, this technique was never described using ex vivo CP explants from PER2::LUC mice, which was a challenge due to the small size of the CP. The protocol that we describe has been an alteration of the previously published study that used organotypic SCN cultures for real-time bioluminescence recording [32]. A. Preparation for Bioluminescence Recording Fill 5 mL disposable syringes with vacuum grease. Protect the syringe with aluminum foil and autoclave it. Right before CPs collection prepare 35 mm Petri dishes, covering the edge of each Petri dish with the autoclaved vacuum grease, and the luciferin medium. Expose all instruments and the previously prepared Petri dishes to UV. B. CP Isolation and Culture Euthanize animals by cervical dislocation at approximately 8 h after lights on. Remove and place the brain dorsal side up in a Petri dish. Using fine forceps and a disposable scalpel make three incisions (one parallel and the others perpendicular to the mid- line of the brain) and pull the cortex away from the dorsal hippocampus, exposing the lateral ventricle. Pull gently the CPs (wavy red line) from the lateral ventricle and place them in chilled HBSS. Repeat the procedure on the contralateral hemisphere to obtain the CP from the other lateral ventricle. Fill the 35 mm Petri dish with 1.2 mL of luciferin medium and 216 Isabel Gonc¸ alves et al.

place a culture membrane on the base of the culture dish (make sure that there are no bubbles between the culture dish and the membrane). Transfer the CPs’ explants (four CPs/culture dish) to the culture membrane using a 1 mL pipette tip. Cover culture dishes with 40 mm cover glasses and seal with more vacuum grease to prevent evaporation. C. Bioluminescence Recording and Data Analysis Place the sealed culture into a 32-channel luminometer device (LumiCycle, Actimetrics, Inc.), which is kept inside an incuba-  tor set at 36–37 C, 0% CO2. Start photomultiplier tube (PMT) measure for ~70 s at 10-min intervals and record as counts/s for up to 7 days. Analysis of circadian parameters (including phase, period, and amplitude) is made using LumiCycle and OriginLab software. D. Bioluminescence Analysis of CP Explants Bioluminescence recording is a technique that allows the study and characterization of diurnal variations in gene expression, during several days. This system is widely used in suprachias- matic nucleus cultures, but was never used in ex vivo CP explants. Our protocol steps proceed very much in the same way as indicated in Savelyev et al. [32], with some modifica- tions, namely in the culture preparation. With this experiment we verify that ex vivo CP explants are viable in culture for at least 6–7 days, with a decrease of the amplitude of circadian rhythm along this period (Fig. 4). Thus, these results indicate that the

400

300

200

100 Bioluminescence (Counts/ sec)

123456789 10 Time (days)

Fig. 4 Bioluminescence analysis of CP explants from Per2:LUC knock-in mice [54] Choroid Plexus Protocols 217

ex vivo CP explants can be used in blood-CSF barrier research and particularly in pharmacological experiments to measure the effect of in vivo treatments on blood-CSF barrier circadian function.

3.3 Use of HIBCPP The morphological correlate of the BCSFB is constituted by the Cells on Cell Culture epithelial cells of the CP, which are connected to each other by tight Inserts for junctions. A functional in vitro model of the BCSFB would have to Experimental reproduce typical properties as the formation of a strong barrier Procedures function and a low permeability for macromolecules [33]. Until recently no functional human model of the BCSFB was 3.3.1 Background available. A human CP epithelial cell line derived from a CP papil- loma of the right lateral ventricle of a Japanese woman was described by Ishiwata and colleagues and termed HIBCPP cells. HIBCPP cells grow in an irregular pattern and can also form papilloma-like structures [34]. Subsequently, the properties of HIBCPP cells in terms of their characteristics of CP epithelial cells were determined. HIBCPP cells express tight-junction (claudin-1, -2, and -3, occludin, ZO-1) and adherens-junction [E-cadherin (Ecad)] components and present continuous tight-junction strands. They display a distinct polarity with basolateral surface receptors (Ecad, Met). Furthermore the expression and function of ABC transporters in HIBCPP cells have been shown [28, 35, 36]. As schematically depicted in Fig. 5, using cell culture filter inserts HIBCPP cells can be cultivated on the upper side of the filter membrane (standard culture model) as well as on the bottom side of the membrane (inverted culture model). When grown on cell culture filter inserts HIBCPP cells develop a strong transepithe- lial electrical resistance (TEER) that is mirrored by a low perme- ability for macromolecules [28].

3.3.2 Cultivation A. Trypsinization of HIBCPP Cells of HIBCPP Cells The cell culture medium (DMEM/F12 (Ham) supplemented with 10% FCS, 5 μg/mL insulin, 100 U/mL penicillin, and 100 μg/mL streptomycin) and PBS are pre-warmed in a water bath to 37 C. After aspiration of old cell culture medium from HIBCPP cells grown in a T75 cell culture flask to 80% con- fluency, the cells are washed two times with 10 mL PBS pre-warmed to 37 C. Subsequently, 3 mL of 0.25% trypsin- EDTA is added to the cells. The flask is gently swirled and incubated for approximately 20 min in an incubator at 37 C and 5% CO2. Check the cells by microscopy for rounding and detachment from the growth surface (see Note 2). Stop the trypsinization by addition of 17 mL cell culture medium and resuspend the cells by pipetting up and down with a 20 mL plastic pipette. After resuspension the cells are counted with a hemocytometer. 218 Isabel Gonc¸ alves et al.

Fig. 5 Schematic depiction of the growth of HIBCPP cells on cell culture filter inserts. (A) Standard model: HIBCPP cells are seeded into medium containing cell culture filter inserts placed into the 24 wells of a 24-well plate. The next day cell culture medium is also added to the bottom compartment. The enlargement at the right side shows the HIBCPP cells grown on the upper side of the filter membrane. (B) Inverted model: Cell culture filter inserts are placed upside down into 12 wells of a 12-well plate. The medium that is filled into the 12 wells touches the membrane of the filter insert. HIBCPP cells are seeded on the top of the inverted filter insert. The next day the filter inserts are placed in the correct orientation into the wells of a 24-well plate. Medium is added to the upper and lower compartments. The enlargement at the right side shows the HIBCPP cells grown on the bottom side of the filter membrane

B. Seeding of HIBCPP Cells Subsequently to counting, the HIBCPP cells are centrifuged in a 50 mL Falcon tube at 50 Â g and RT for 10 min. The supernatant is discarded, the cells are carefully resuspended by pipetting up and down in medium (adjusted to a total volume of 1 mL), and the cell concentration is adjusted to the desired amount (e.g., 1 Â 106 cells/mL). Between 1 and 6 Â 106 HIBCPP cells in overall 10 mL cell culture medium are seeded into a T75 cell culture flask for maintenance. During culture the medium should be changed every 2 days. Although HIBCPP cells can be passaged multiple times, passages older than 38 should not be used for experiments.

3.3.3 Growth of HIBCPP Cell culture filter inserts (pore diameter 3.0 μm, pore density Cells on Cell Culture Filter 2.0 Â 106 pores per cm2, membrane surface area 0.33 cm2) are Inserts in the Standard placed into the wells of a 24-well plate. Fill the upper compartment Model (filter insert) with 350–400 μL of cell culture medium pre-warmed Choroid Plexus Protocols 219

to 37 C. Leave the lower compartment empty until the next day. Generate a suspension of 1 Â 106 trypsinized HIBBCP cells per mL as described under Sect. 3.3.2. Seed 1.0–1.5 Â 105 cells (100–150 μL of the 1 Â 106 cells/mL suspension; final volume in the upper compartment is 500 μL) into the upper compartment, cover the 24-well plate with the lid, and place the plate into an  incubator (37 C, 5% CO2). The following day add 1 mL of cell culture medium pre-warmed to 37 C to the lower compartment (well). Replace the cell culture medium with fresh medium every 2 days. Switch to 1% FCS-containing cell culture medium, when the cell culture inserts display a TEER value above 70 Ω cm2 (see Note 3).

3.3.4 Growth of HIBCPP A. Preparation of the Filter Inserts Cells on Cell Culture Filter Cell culture filter inserts (pore diameter 3.0 μm, pore density Inserts in the Inverted 2.0 Â 106 pores per cm2, membrane surface area 0.33 cm2) are Model placed upside down into the wells of a 12-well cell culture plate. Subsequently, the filter inserts are filled from below with cell culture medium pre-warmed in a water bath to 37 C. The cell culture medium has to touch the membrane of the filter insert; see Note 4. Rinse the top of the filter membrane with a drop of cell culture medium. Wait until the drop has passed through the filter membrane, cover the 12-well plate with the lid, and place  the plate into an incubator (37 C, 5% CO2) until seeding of HIBCPP cells occurs. B. Seeding of HIBCPP Cells Generate a suspension of 1 Â 106 trypsinized HIBBCP cells per mL as described under Sect. 3.3.2. Seed 0.8–1.0 Â 105 cells (80–100 μL of the 1 Â 106 cells/mL suspension) on the top of the prepared inverted filter inserts, which will represent the bottom of the filter membrane when the insert is placed in standard orientation. Wait until the liquid of the cell suspension has passed through the filter membrane; the cells will stay on top of the membrane. Cover the 12-well plate with the lid and place  the plate into an incubator (37 C, 5% CO2). C. Growth of HIBCPP Cells The day following seeding of the cells the filter inserts are transferred to 24 wells. For this, use sterile forceps to lift the cell culture filter inserts from the 12-well and place the inserts in the correct orientation into the 24 wells of a 24-well plate. The cells are now located on the bottom of the filter membrane oriented toward the bottom compartment (well). Immediately fill the bottom compartment (well) with 1 mL and the upper compartment (filter insert) with 500 μL of the preconditioned 220 Isabel Gonc¸ alves et al.

medium from the 12-well. Cover the 24-well plate with the lid  and place the plate into an incubator (37 C, 5% CO2). Replace the cell culture medium with fresh medium every 2 days. Switch to 1% FCS-containing cell culture medium, when the cell cul- ture inserts display a TEER value above 70 Ω cm2 (see Note 3).

3.3.5 Determination of Disinfect the STX-01 electrode of the Millicell-ERS2 system by the TEER immersing the electrode tips for 15 min in 80% EtOH. Remove the electrode from the ethanol, let it dry, and equilibrate it shortly in pre-warmed cell culture medium. To determine the TEER of each filter insert place the electrode with the longer arm into the bottom compartment (well) and with the shorter arm into the upper compartment (filter insert). After finishing the TEER mea- surements (see Note 5) disinfect the electrode again in 80% EtOH.

3.3.6 Determination of A. Preparing the Cell Culture Filter Inserts the FITC-Labeled Prepare a 5 mg/mL stock solution of FITC-inulin in cell culture Inulin Flux medium or thaw a frozen aliquot (see Note 6). Further dilute the stock 1:5 in cell culture medium and add 50 μL of the stock solution into the upper compartment (filter insert) to obtain a concentration of 100 μg/mL (the final volume in the upper compartment is 500 μL). Cover the 24-well plate containing the filter inserts with the lid and place the plate into an incubator  (37 C, 5% CO2) for the desired experimental period (e.g., 2 h). To determine the percentage of FITC-inulin flux from the upper (filter insert) to the lower (well) compartment, serial 1:2 dilutions of a 50 μg/mL FITC-inulin solution (corresponding to 100%), which cover 50–0.2% of the 50 μg/mL solution, as well a solvent-only control (0%), are required. It is recommend- able to prepare the 50 μg/mL FITC-inulin solution (100%) used for serial dilutions in parallel by adding the inulin to the filter inserts and incubate it under the same conditions like the  filter inserts (in the incubator at 37 C, 5% CO2). B. Measuring the FITC-Inulin Flux At the end of the experimental period separately collect the cell culture medium from each of the lower (well) compartment of the experimental setup. Prepare serial dilution series (1:2) of your preincubated solution corresponding to 100% and use 0.2 mL of the dilutions and the samples for inulin measurement. Measurements are performed in duplicates. The percentage of FITC-inulin flux is calculated from the fluorescence values measured in microplate reader. Choroid Plexus Protocols 221

3.4 Single-Cell A new paradigm in physiology suggests that several classes of Calcium Imaging in CP understudied receptors [olfactory receptors (ORs), taste receptors Epithelial Cells to (TR), and orphan G-protein-coupled receptors (GPCRs)] play key Quantify roles in a variety of tissues, including the CP epithelial cells “Chemosensory” [15, 16], where they may serve as sensitive chemoreceptors. 2+ 2+ Receptors’ Responses The fine control of cytosolic concentration of free Ca ([Ca ] i) has long been recognized as a fundamental mechanism of cell 3.4.1 Background activation [37]. So, our group has monitored the responses in the [Ca2+]i of the CP epithelial cells by using Ca2+ imaging of individ- ual cells to test the effect of odorant and taste stimuli. A direct knowledge of [Ca2+]i is possible in living cells due to the development of intracellularly trapped fluorescent indicators, as Fura-2 among many others [38]. Ca2+ indicators bind and interact only with freely diffusible Ca2+ ions. Fura-2 is a radiometric indica- tor, what means that it shifts the peak wavelength of either their excitation or emission curve upon binding Ca2+. This class of indicators enables a very accurate quantification of Ca2+ concentra- tion that is corrected for uneven dye loading, dye leakage, photo- bleaching, and changes in cell volume, but at the cost of increased spectral bandwidth. To quantify the CP epithelial cells’ responses to olfactory and taste stimuli, we use a chemical modification of Fura- 2, Fura-2 AM (acetoxymethyl). AM dyes are sufficiently hydropho- bic in that they are membrane permeable and can be passively loaded into cells simply by adding them to the medium. Intracellu- lar esterases then cleave the AM group and trap the dye inside the cells.

3.4.2 Intracellular A. Choice of Cells Calcium Concentration 2+ We measure changes in [Ca ]i levels in primary cultures of CP Evaluation epithelial cells and in the epithelial CP cell lines. We routinely use the following epithelial CP cell lines: the murine Z310 and the human HIBCPP. Below we describe our protocol using HIBCPP cell line, which can be easily applied to the other cell types. B. Preparation of Cells Warm the “10% total culture medium” in an incubator to 37 C. Thaw one vial of frozen cells in a 37 C water bath. Gently mix and transfer the cells from the vial to a 15 mL tube with 6 mL pre-warmed medium, put the suspension into a T75 flask, and  incubate the cells in a CO2 incubator at 37 C. The day after, change the medium to remove the DMSO remnants. Then grow the cells until confluence, changing the medium every 2 days (this will take 3–4 weeks). 222 Isabel Gonc¸ alves et al.

C. Splitting and Plating Cells Aspirate the medium from confluent cell flasks and wash the cells with 10 mL PBS. Add 3 mL of 0.25% trypsin-EDTA and incubate during 22 min at 37 C. Stop reaction with 10 mL “10% total culture medium.” Put the cell suspension into a 50 mL Falcon tube and centrifuge with 500 rpm for 10 min at RT. Resuspend the cells in an appropriate amount of “10% total culture medium” (~20 mL) and count the cells. Apply the desired amount of cells and proper amount of medium to glass-bottom 8-well dishes: 75,000 cells in 250 μL of medium in each of the 8 wells. The cells will grow until 75% confluence, which usually takes approximately 4 days. D. Dye Loading and Single-Cell Ca2+ Imaging Aspirate the medium and load the cells with 5 μM Fura-2 AM, 0.1% BSA, and 0.02% pluronic acid F-12 7 for 1 h at 16 C with Tyrode’s solution (see Note 7). Wash the cells with Tyrode’s and incubate in the same solution for 30 min before acquisition. Put the 8-well chambers on an inverted fluorescence microscope. To perform stimuli, each compound has to be dissolved in Tyrode’s at 100 mM. Before each stimulus basal fluorescence should be recorded for 5 min and no alterations should be observed. The following procedure was performed with poly- amine olfactory stimuli [15] and with the bitter receptor ago- nists denatonium benzoate, D-salicin, 6-propyl-3-thiouracil (PROP), and N-phenylthiourea (PTC) [16]. For each com- pound, stimulate the cells in a dose-response between 0.001 and 5 mM, for 3 min. Carefully, apply the stimuli using a micropipette, after recording the baseline, allowing instanta- neous stimulation by focal application. To perform receptor blockage, as for example for the bitter taste receptor blocker probenecid, incubate cells with a concentration range between 0.01 and 1 mM at RT, 30 min prior to application of the stimulus. Evaluate the intracellular calcium concentration by quantifying the ratio of the fluorescence emitted at 520 nm following alternate excitation at 340 nm and 380 nm, using a Lambda DG4 apparatus and a 520 nm bandpass filter, under a 40Â objective, with an AxioVision camera and software. Pro- cess the data using the Fiji software. Results should be pre- sented as an average of the changes in intracellular calcium levels of 15–20 cells from three or more independent experiments.

3.4.3 Silencing the Once the CP epithelial cell responses have been achieved, it is “Chemosensory” Receptor important to determine if there is a connection between the Responses responses to the stimuli and the respective signaling pathway (in the case of the olfactory stimuli, the olfactory signaling Choroid Plexus Protocols 223 pathway). The most direct way of achieving this is to transfect the epithelial CP cells with siRNAs against one of the proteins involved in the signaling pathway. To confirm protein knocking down, it is possible to measure the expression of the target protein in the transfected cells (by ICQ, for example) or to analyze the differences in the “sensory” responses by Ca2+ imaging. A. Transfect the cells Transfect the cells with siRNA using siPORT amine transfection agent according to the manufacturer’s instructions. In order to knock down the olfactory pathway, we used siRNA against Gαolf in the HIBCPP cell line, which can be easily applied to the other siRNAs and cell lines. The procedure is the following: Plate 75,000 cells in 250 μL of “10% total culture medium” in each of the wells of the 8-well ibiTreat chamber. The cells will grow to 35–40% confluence in 2 days. 24 h before transfection, change medium to “5% total culture medium” without antibio- tics. Aspirate the medium and apply to the cells 250 μL of the transfection mixture containing 10 nM of siRNA targeting the Gαolf and 5 μL of siPORT amine transfection agent in “10% total culture medium,” for 4 h. Replace the medium with “10%  of total culture medium” and incubate at 37 CinCO2 for 48 h. Use a control transfection with the scramble siRNA at the same conditions of the target siRNA. The cells are harvested to determine the knockdown levels, by one of the procedures below. B. Determine Protein Knockdown Using Immunocytochemistry In this case, confirm the Gαolf protein knockdown by immuno- fluorescence. Quantify staining intensity, using Image J. Normalize the background against the cell vicinity to adjust for possible exposure differences. Calculate the mean and SEM for both control and transfected cultures and compare. C. Determine Silencing of the “Sensory” Responses with Ca2+ Imaging Dye load the HIBCPP cells transfected with Gaolf siRNA, in 8-well ibiTreat chambers (as described in Sect. 3.4.2) and trans- fer them to an inverted fluorescence microscope. Apply the olfactory stimuli in the same concentrations described above (Sect. 3.4.2-D). Perform the single-cell Ca2+ imaging exactly as described above. Compare the intracellular Ca2+ responses of the siRNA-transfected cells with the responses from the non-transfected ones. The scrambled siRNA-transfected con- trol cells have also to be analyzed by Ca2+ imaging. 224 Isabel Gonc¸ alves et al.

3.5 Ex Vivo The CP of mammals, in common with other organs such as the Electrophysiological kidney [39, 40] and brain [41], expresses olfactory receptors [42], Studies in the CP members of the G-protein-coupled receptor (GPCR) superfamily (reviewed in [43, 44]). The function of such “ectopic” olfactory 3.5.1 Background receptors is not yet clear. A major step in understanding their function would be to identify the ligands that they detect, presum- ably in the CSF. We have recorded large (several millivolts) extra- cellular DC field potentials from the CP of mice ex vivo in response to the polyamines, cadaverine, putrescine, spermine, and spermi- dine [15]. Which other classes of compounds are detected by these receptors remains to be investigated; we believe that the method developed to record such electrophysiological responses from the CP will enable both rapid screening of many different potential ligands and the investigation of the cellular transduction pathways and physiological basis for the observed response.

3.5.2 Odorants in Aquatic A major difference between aquatic and terrestrial olfaction is in the and Terrestrial Media compounds that are detected; a potential odorant in the aquatic medium needs only to be water soluble, whereas in the terrestrial environment it needs (with some exceptions) to be both water soluble and volatile. Thus, common fish odorants such as amino acids and bile acids are not detected by the olfactory system of mammals. However, one class of organic compounds—the poly- amines—can be detected by the olfactory system of both fish [45–47] and mammals [48]. Furthermore, levels of polyamines in the CSF are altered in neurodegenerative diseases such as Alzhei- mer’s [49], Parkinson’s [50], and multiple sclerosis [51]. As the CP is an epithelium continuously bathed in an aqueous solution—the CSF—the method for recording electrophysiological responses was based on one designed to record electrophysiological responses from another epithelium continuously bathed in aqueous solution, the olfactory epithelium of fish. The underwater electro- olfactogram, or EOG, is an extracellular DC field potential recorded immediately above the olfactory epithelium, and is used extensively to assess the olfactory sensitivity to a given odorant or odor [52]. It involves a negative voltage change, generally accepted to be due to the inward current of Na+ and Ca2+ ions. The CP response is usually positive (i.e., positive voltage change), so the ionic basis is likely to be different.

3.5.3 Recording DC Field We recorded extracellular DC field potentials from the CP in Potentials from the CP response to polyamines using a novel ex vivo setup consisting of the whole mouse brain in which the CP is exposed (Fig. 6). Mice should be euthanized by cervical dislocation and the whole brain immediately removed. The CPs are exposed by two parallel inci- sions to open both lateral ventricles with the CPs remaining attached to the ventricle walls and the brain placed in a small tissue chamber. We found that it was unnecessary to fix the brain in Choroid Plexus Protocols 225

Fig. 6 Photograph of the ex vivo brain preparation for DC field potential recording from the choroid plexus. A recording electrode, B reference electrode, C tube for delivering aCSF and test stimulus, D isolated brain with CP exposed, and E guide for excess aCSF removal position. The ventricle and CPs are superfused with aCSF, via a glass tube (~1 mm i.d.; hematocrit tubes are ideal), under gravity at a flow rate of about 1 mL/min. The overflow from the tissue chamber should not be allowed to drip into the waste collector; a simple method to avoid this is to “guide” the flow with a tissue paper strip into waste. Experiments should take place within a Faraday cage to reduce interference; most electrophysiology rigs have one. The chemical stimulus, dissolved in aCSF, can be intro- duced into the flow delivered to the CPs using a three-way solenoid valve, or three-way tap, or even injected into the flow with a syringe and needle. The recording method—DC field potential—is based on that developed for the underwater electro-olfactogram in fish (for example, see [53] with the modifications required for the brain–CP setup; the low flow of aCSF referred to above, earthed via a silver–silver chloride pellet held in contact with the cerebrum using a micromanipulator, and using slightly different micropip- ettes). These are made with conventional borosilicate glass tubes on a puller (any microelectrode puller will do), but cut at the sharp end (with a diamond pen) to give a relatively “blunt” tip (100–200 mm diameter) and fill with 1 M NaCl in 1% agar. The agar solution can contain food dye to make it more visible under the dissecting microscope. The recording electrode is placed near the CP using a micromanipulator at the position that gives the strongest response to the positive control (e.g., 5 mM cadaverine), and the reference 226 Isabel Gonc¸ alves et al.

Fig. 7 Schematic diagram of the recording setup for DC field potentials from the CP as described in this chapter. A typical response recorded from the CP in response to cadaverine (horizontal bar) is shown in inset. Note the large amplitude of response, but the relatively slow time course

electrode in contact with the cerebrum, similarly with a microma- nipulator (Fig. 6). The voltage signal can be recorded using a DC amplifier of the type used for intracellular recording and filtered with a low pass of 50 Hz; the sampling frequency does not need to be high as the response is slow (Fig. 7). Our experiments were carried out at room temperature; the effects of temperature varia- tion were not investigated. Similar to the EOG, the health and stability of the preparation can be assessed by periodic exposure to the positive control; we found that stable recordings could be obtained for 2 or 3 h.

4 Notes

1. Using these 12-well plates has the advantage that the lid will not touch the top of the inverted cell culture filter insert when placed as cover onto the 12-well plate. 2. The cells do not necessarily separate completely into single cells. After detachment from the flask surface often small agglomerates are present additionally to single cells. This degree of separation is sufficient for seeding the cells into a new flask or onto the cell culture filter insert membranes. 3. The switch to medium containing low amounts of FCS is not absolutely necessary. The cells will develop a strong TEER also Choroid Plexus Protocols 227

in cell culture medium containing 10% FCS, although a certain increase of the TEER in medium containing low amounts of FCS has been shown. Low amounts of FCS might be preferable in certain experimental setups, e.g., when analyzing interac- tions with bacterial pathogens. 4. This is best done as follows: Using the serological pipette needed to fill the bottom compartment the upside-down-ori- ented filter insert is tilted a few degrees and fixed against the wall of the well. Now the well containing the filter insert is filled with cell culture medium until the upper level of the medium touches the membrane of the filter insert. Afterwards remove as much medium from outside of the inverted insert until the upper level of the medium is about 1 cm below the membrane of the insert. Directly below the membrane (inside of the inverted insert) a medium column touching the filter mem- brane will be maintained due to adhesion forces. It is necessary that the medium touches the membrane; otherwise the cells seeded in the next step onto the top of the inverted insert will not be fed. 5. For determination of the TEER a blank value has to be estab- lished by measuring a filter, which does not contain cells. To calculate the TEER, this blank value is subtracted from the measured value. The result is multiplied with the filter surface area (0.33 cm2) to obtain the TEER in Ω cm2. 6. When a frozen aliquot is thawed warm the solution to RT and dilute 1:5 in cell culture medium (dilution can be stored at 4 C). Before adding to the cell culture filter inserts, pre-warm the dilution to 80 C and pellet potential precipitates by a 2-min centrifuge spin at maximum speed. Use the supernatant to avoid introducing precipitates into the experimental setup. 7. A major disadvantage of the chemical Ca2+ indicators is that they tend to compartmentalize and are extruded from the cell during long recording experiments [38]. During our CP epi- thelial cell experiments we observed this compartmentalization problem and the simple and successful strategy to combat this trickiness was lowering the cell incubation temperature of the dye from 37 to about 15–16 C.

Acknowledgments

This work was supported by the Portuguese Foundation for Sci- ence and Technology (FCT, Portugal—http://www.fct.pt) project grants (PTDC/SAU-NEU/114800/2009, project UID/Multi/ 04326/2013 and project UID/Multi/00709/2013), and FEDER funds through the POCI—COMPETE 2020— 228 Isabel Gonc¸ alves et al.

Operational Programme Competitiveness and Internationalization in Axis I—Strengthening research, technological development and innovation (Project POCI-01-0145-FEDER-007491), the Swed- ish Research Foundation, the Swedish Brain Foundation, A˚ ke Wibergs stiftelse, and Karolinska Institutet Research Funds. T Quintela is a recipient of a FCT fellowship (SFRH/BPD/70781/ 2010). AC Duarte is a recipient of ICON. Joana Toma´s was sup- ported by a grant from CENTRO-07-ST24-FEDER-002015. This work and AC Duarte were supported by “Programa Operacional do Centro, Centro 2020” through the funding of the ICON project (Interdisciplinary Challenges On Neurodegeneration; CENTRO-01-0145-FEDER-000013).

References

1. Strazielle N, Khuth ST, Ghersi-Egea JF (2004) opportunities for diagnostics and treatment. Detoxification systems, passive and specific Neurobiol Dis 72(Pt A):54–60 transport for drugs at the blood-CSF barrier 10. Pascale CL, Miller MC, Chiu C et al (2011) in normal and pathological situations. Adv Amyloid-beta transporter expression at the Drug Deliv Rev 56:1717–1740 blood-CSF barrier is age-dependent. Fluids 2. Redzic ZB, Segal MB (2004) The structure of Barriers CNS 8:21 the choroid plexus and the physiology of the 11. Krzyzanowska A, Carro E (2012) Pathological choroid plexus epithelium. Adv Drug Deliv alteration in the choroid plexus of Alzheimer’s Rev 56:1695–1716 disease: implication for new therapy 3. Ghersi-Egea JF, Monkkonen KS, Schmitt C approaches. Front Pharmacol 3:75 et al (2009) Blood-brain interfaces and cerebral 12. Falcao AM, Marques F, Novais A et al (2012) drug bioavailability. Rev Neurol () The path from the choroid plexus to the sub- 165:1029–1038 ventricular zone: go with the flow! Front Cell 4. Damkier HH, Brown PD, Praetorius J (2013) Neurosci 6:34 Cerebrospinal fluid secretion by the choroid 13. Johansson PA (2014) The choroid plexuses plexus. Physiol Rev 93:1847–1892 and their impact on developmental neurogen- 5. Johanson C, Stopa E, McMillan P et al (2011) esis. Front Neurosci 8:340 The distributional nexus of choroid plexus to 14. Lun MP, Monuki ES, Lehtinen MK (2015) cerebrospinal fluid, ependyma and brain: toxi- Development and functions of the choroid cologic/pathologic phenomena, periventricu- plexus-cerebrospinal fluid system. Nat Rev lar destabilization, and lesion spread. Toxicol Neurosci 16:445–457 Pathol 39:186–212 15. Goncalves I, Hubbard PC, Tomas J et al 6. Richardson SJ, Wijayagunaratne RC, D’Souza (2016) ‘Smelling’ the cerebrospinal fluid: DG et al (2015) Transport of thyroid hor- olfactory signaling molecules are expressed in mones via the choroid plexus into the brain: and mediate chemosensory signaling from the the roles of transthyretin and thyroid hormone choroid plexus. FEBS J 283:1748–1766 transmembrane transporters. Front Neurosci 16. Tomas J, Santos CR, Quintela T et al (2016) 9:66 “Tasting” the cerebrospinal fluid: another 7. Spector R, Keep RF, Robert Snodgrass S et al function of the choroid plexus? Neuroscience (2015) A balanced view of choroid plexus 320:160–171 structure and function: Focus on adult 17. Quintela T, Sousa C, Patriarca FM et al (2015) humans. Exp Neurol 267:78–86 Gender associated circadian oscillations of the 8. Schwartz M, Baruch K (2014) The resolution clock genes in rat choroid plexus. Brain Struct of neuroinflammation in neurodegeneration: Funct 220:1251–1262 leukocyte recruitment via the choroid plexus. 18. Santos CR, Duarte AC, Quintela T et al (2017) EMBO J 33:7–22 The choroid plexus as a sex hormone target: 9. Pahnke J, Langer O, Krohn M (2014) Alzhei- functional implications. Front Neuroendocri- mer’s and ABC transporters – new nol 44:103–121 Choroid Plexus Protocols 229

19. Marques F, Sousa JC, Brito MA et al (2017) clock rhythms of gene expression using lucifer- The choroid plexus in health and in disease: ase bioluminescence reporters. J Vis Exp. dialogues into and out of the brain. Neurobiol https://doi.org/10.3791/4234 Dis 107:32–40 32. Savelyev SA, Larsson KC, Johansson AS et al 20. Brito MA, Silva RFM, Brites D (2006) Cell (2011) Slice preparation, organotypic tissue response to hyperbilirubinemia: a journey culturing and luciferase recording of clock along key molecular events. In: Chen FJ gene activity in the suprachiasmatic nucleus. J (ed) New trends in brain research. Nova Sci- Vis Exp. https://doi.org/10.3791/2439 ence Publishers, Inc., New York, pp 1–38 33. Strazielle N, Ghersi-Egea JF (2011) In vitro 21. Vercellino M, Votta B, Condello C et al (2008) models of the blood-cerebrospinal fluid barrier Involvement of the choroid plexus in multiple and their use in neurotoxicological research. sclerosis autoimmune inflammation: a neuro- NeuroMethods 56:161–184 pathological study. J Neuroimmunol 34. Ishiwata I, Ono I, Kiguchi K et al (2005) 199:133–141 Establishment and characterization of a 22. Engelhardt B, Wolburg-Buchholz K, Wolburg human thyroid carcinoma cell line (HOTHC) H (2001) Involvement of the choroid plexus in producing colony stimulating factor. Hum Cell central nervous system inflammation. Microsc 18:163–169 Res Tech 52:112–129 35. Grundler T, Quednau N, Stump C et al (2013) 23. Marques F, Sousa JC, Sousa N et al (2013) The surface proteins InlA and InlB are interde- Blood-brain-barriers in aging and in Alzhei- pendently required for polar basolateral inva- mer’s disease. Mol Neurodegener 8:38 sion by Listeria monocytogenes in a human 24. Vargas T, Antequera D, Ugalde C et al (2010) model of the blood-cerebrospinal fluid barrier. Gelsolin restores A beta-induced alterations in Microbes Infect 15:291–301 choroid plexus epithelium. J Biomed Biotech- 36. Bernd A, Ott M, Ishikawa H et al (2015) nol 2010:805405 Characterization of efflux transport proteins 25. Quintela T, Alves CH, Goncalves I et al (2008) of the human choroid plexus papilloma cell 5Alpha-dihydrotestosterone up-regulates line HIBCPP, a functional in vitro model of transthyretin levels in mice and rat choroid the blood-cerebrospinal fluid barrier. Pharm plexus via an androgen receptor independent Res 32:2973–2982 pathway. Brain Res 1229:18–26 37. Rasmussen H (1986) The calcium messenger 26. Dinner S, Borkowski J, Stump-Guthier C et al system (1). N Engl J Med 314:1094–1101 (2016) A choroid plexus epithelial cell-based 38. Paredes RM, Etzler JC, Watts LT et al (2008) model of the human blood-cerebrospinal fluid Chemical calcium indicators. Methods barrier to study bacterial infection from the 46:143–151 basolateral side. J Vis Exp. https://doi.org/ 39. Pluznick JL, Zou DJ, Zhang X et al (2009) 10.3791/54061 Functional expression of the olfactory signaling 27. Yoo SH, Yamazaki S, Lowrey PL et al (2004) system in the kidney. Proc Natl Acad Sci U S A PERIOD2::LUCIFERASE real-time reporting 106:2059–2064 of circadian dynamics reveals persistent circa- 40. Rajkumar P, Aisenberg WH, Acres OW et al dian oscillations in mouse peripheral tissues. (2014) Identification and characterization of Proc Natl Acad Sci U S A 101:5339–5346 novel renal sensory receptors. PLoS One 9: 28. Schwerk C, Papandreou T, Schuhmann D et al e111053 (2012) Polar invasion and translocation of Neis- 41. Grison A, Zucchelli S, Urzi A et al (2014) seria meningitidis and Streptococcus suis in a Mesencephalic dopaminergic neurons express novel human model of the blood-cerebrospinal a repertoire of olfactory receptors and respond fluid barrier. PLoS One 7:e30069 to odorant-like molecules. BMC Genomics 29. Zheng W, Zhao Q (2002) Establishment and 15:729 characterization of an immortalized Z310 cho- 42. Quintela T, Goncalves I, Carreto LC et al roidal epithelial cell line from murine choroid (2013) Analysis of the effects of sex hormone plexus. Brain Res 958:371–380 background on the rat choroid plexus tran- 30. Herbert J, Wilcox JN, Pham KT et al (1986) scriptome by cDNA microarrays. PLoS One Transthyretin: a choroid plexus-specific trans- 8:e60199 port protein in human brain. The 1986 S. Weir 43. Foster SR, Roura E, Thomas WG (2014) Mitchell award. Neurology 36:900–911 Extrasensory perception: odorant and taste 31. Ramanathan C, Khan SK, Kathale ND et al receptors beyond the nose and mouth. Phar- (2012) Monitoring cell-autonomous circadian macol Ther 142:41–61 230 Isabel Gonc¸ alves et al.

44. Kang N, Koo J (2012) Olfactory receptors in 50. Paik MJ, Ahn YH, Lee PH et al (2010) Poly- non-chemosensory tissues. BMB Rep amine patterns in the cerebrospinal fluid of 45:612–622 patients with Parkinson’s disease and multiple 45. Da Silva JP, Choudhury R, Porel M et al system atrophy. Clin Chim Acta (2014) Synthetic versus natural receptors: 411:1532–1535 supramolecular control of chemical sensing in 51. Noga MJ, Dane A, Shi S et al (2012) Metabo- fish. ACS Chem Biol 9:1432–1436 lomics of cerebrospinal fluid reveals changes in 46. Michel WC, Sanderson MJ, Olson JK et al the central nervous system metabolism in a rat (2003) Evidence of a novel transduction path- model of multiple sclerosis. Metabolomics way mediating detection of polyamines by the 8:253–263 zebrafish olfactory system. J Exp Biol 52. Scott JW, Scott-Johnson PE (2002) The elec- 206:1697–1706 troolfactogram: a review of its history and uses. 47. Rolen SH, Sorensen PW, Mattson D et al (2003) Microsc Res Tech 58:152–160 Polyamines as olfactory stimuli in the goldfish 53. Hubbard PC, Barata EN, Canario AV (2002) Carassius auratus. J Exp Biol 206:1683–1696 Possible disruption of pheromonal communi- 48. Pinel JP, Gorzalka BB, Ladak F (1981) Cadav- cation by humic acid in the goldfish, Carassius erine and putrescine initiate the burial of dead auratus. Aquat Toxicol 60:169–183 conspecifics by rats. Physiol Behav 27:819–824 54. Quintela T, Albuquerque T, Lundkvist G et al 49. Inoue K, Tsutsui H, Akatsu H et al (2013) (2018) The choroid plexus harbors a circadian Metabolic profiling of Alzheimer’s disease oscillator modulated by estrogens. Chronobiol brains. Sci Rep 3:2364 Int 35(2):270–279 Chapter 14

Techniques for Evaluating Efflux Transport of Radiolabeled Drugs and Compounds from the Cerebrospinal Fluid Across the Blood-Cerebrospinal Fluid Barrier

Shin-ichi Akanuma, Yoshiyuki Kubo, and Ken-ichi Hosoya

Abstract

Choroid plexus epithelial cells are known to play a role as the blood-cerebrospinal fluid (CSF) barrier (BCSFB), which separates the compartments of the CSF in the cerebroventricles from the circulating blood. Recent reports have identified the molecular based efflux transport systems at the BCSFB. Because these transport systems participate in the elimination of compounds/drugs from the CSF, these experi- mental findings about the systems are of great importance to increase our knowledge of the homeostasis of compound concentration in the brain and CSF. There are many reports of in vivo and in vitro methods to examine BCSFB-mediated organic compound efflux transport. In this section, we describe the in vivo intracerebroventricular administration technique to evaluate carrier-mediated elimination of compounds from the CSF in rats. As the in vitro methods, the transport studies using choroid plexus prepared from rat cerebroventricles and a conditionally immortalized rat choroid plexus epithelial cell line, TR-CSFB3 cells, are described in detail. The information obtained from these studies will help us to understand the molecular mechanisms of compound efflux transport across the BCSFB.

Key words Blood-cerebrospinal fluid barrier, Cerebrospinal fluid, Choroid plexus epithelial cells, Immunohistochemistry, Intracerebroventricular administration, Organic anion transporter, Organic cation transporter, Transport, Uptake studies

1 Introduction

The blood-cerebrospinal fluid (CSF) barrier (BCSFB) separates the compartments of the CSF in the cerebroventricles from the circu- lating blood and this barrier is located at the choroid plexus (ChP). The ChP has a vascularized veil structure, and is located in the lateral, third, and fourth ventricles of the brain. It is known that the ChP is responsible for the secretion of CSF [1]. In addition, the nonselective exchange of compounds between the circulating blood and the CSF is restricted by the ChP which consists of epithelial cells with villi and fenestrated blood vessels. Since the choroid plexus epithelial cells (CPECs) form the intercellular tight

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_14, © Springer Science+Business Media, LLC, part of Springer Nature 2019 231 232 Shin-ichi Akanuma et al.

Fig. 1 Schematic diagram of carrier-mediated efflux transport across the BCSFB. CF carboxyfluorescein, ChP choroid plexus, CRT creatine transporter, CSF cerebrospinal fluid, E217βG estradiol 17β-glucuronide, GAA guanidinoacetate, ISF interstitial fluid, MPP+ 1-methyl-4-phenylpyridinium, OAT organic anion transporter, Oatp organic anion transporting polypeptide, OCT organic cation transporter, PGT prostaglandin transporter, PMAT plasma membrane monoamine transporter

junctions [2], this nonselective compound exchange is responsible for the CPECs. In summary, CPECs in the ChP play the main role in the BCSFB. Recent reports have shown that the BCSFB possesses “selec- tive” transport mechanism(s) at the BCSFB. In particular, the CSF compartment has been recognized as a “sink” for waste in the brain [3], so that compound elimination from the CSF across the BCSFB has been studied widely (Fig. 1). For example, the protein localiza- tion of several kinds of organic anion transporters, such as organic anion transporter (OAT) 1 (solute carrier (SLC) 22A6), OAT3 (SLC22A8), organic anion transporting polypeptide (Oatp) 1a5 (Slco1a5), and prostaglandin transporter (Pgt/Slco2a1) on the apical membrane of the BCSFB has been thoroughly investigated [4–7]. It has been found that these transporters take part in the elimination of compounds/drugs, such as 6-carboxyfluorescein [5], benzylpenicillin [4], estradiol 17β-glucuronide [8], and pros- taglandin D2 [7]/E2 [9], from the CSF. As the efflux transport system(s) for organic cationic and zwitterionic compounds, the In vivo/vitro Analyses for BCSFB-Mediated Efflux 233 efflux transport activities of guanidinoacetate, creatinine, 1-methyl- 4-phenylpyridinium (MPP+), and histamine in the BCSFB have been studied [10–13]. As the molecules which are involved in the efflux transport at the BCSFB, the expression and functional con- tribution of creatine transporter (CRT/SLC6A8), organic cation transporter 3 (OCT3/SLC22A3), and plasma membrane mono- amine transporter (PMAT/SLC29A4) have been examined [10–13]. In addition, it has been reported that mRNAs of seroto- nin transporter (SLC6A4), OCT1-2 (SLC22A1-2), organic cat- ion/carnitine transporter 1-2 (OCTN1-2/SLC22A4-5), and multidrug and toxin extrusion 1-2 (MATE1-2/SLC47A1-2) are expressed [11, 14]. To summarize these points, the BCSFB pos- sesses several organic compound transport system(s) which take part in CSF-to-blood efflux transport across the BCSFB. In order to increase our knowledge of the homeostasis of compound con- centration in the brain and CSF, methods to test BCSFB-mediated efflux transport of compounds/drugs via transporters are helpful. So far, several techniques for evaluating the elimination of compounds from the CSF and BCSFB-mediated compound trans- port in animals have been established. As an in vivo method, a ventriculo-cisternal perfusion technique has been applied in several animals, such as rats, rabbits, dogs, and cats [15–19], and bolus injection has also been used to monitor elimination from the CSF in mice and rats [20, 21]. Regarding ex vivo and in vitro analyses, transport studies using radiolabeled compounds and/or fluores- cent substrates in experimental animals such as mice [22, 23] and rats [24–28] have been reported. Several choroid plexus epithelial cell lines have also been established. For instance, mouse immorta- lized choroid plexus cells, such as ECPC-4 cells, were established and used to study cellular proliferation [29] and for proteomics [30]. The immortalized and conditionally immortalized rat cho- roid plexus epithelial cell lines, which are named as Z310 cells [31] and TR-CSFB3 cells, respectively [32], have also been used for transport studies [12, 13, 20, 33, 34]. CPC-2 cells are also known to be derived from carcinoma of the human choroid plexus [35, 36]. By applying these in vivo and in vitro experimental tools, it is expected that we will be able to obtain a full picture of compound/drug transport mechanisms in the BCSFB. In this chapter, the bolus intracerebroventricular administra- tion technique in rats as an in vivo method for evaluating carrier- mediated radiolabeled compound elimination from the rat CSF is described. To discuss the involvement of this carrier-mediated compound elimination across the BCSFB based on the results of bolus administration, in vitro experimental approaches involving rats are useful. Hence, we have also introduced two in vitro studies for assessing the transport of radiolabeled compounds using cho- roid plexus isolated from rats and TR-CSFB3 cells [32]. 234 Shin-ichi Akanuma et al.

2 Materials

2.1 For Intracerebro- 1. Male adult Wistar rats (150–200 g; Japan SLC, Hamamatsu, ventricular Bolus Japan). Administration 2. Artificial CSF (aCSF; 122 mM NaCl, 25 mM NaHCO3,3mM KCl, 1.4 mM CaCl2, 1.2 mM MgSO4, 0.4 mM K2HPO4, 10 mM D-glucose, 10 mM 2-[4-(2-hydroxyethyl)-1-piperazi- nyl]ethanesulfonic acid (HEPES)-NaOH; pH 7.4; see Note 1). 3. 3H-labeled test compound and 14C-labeled D-mannitol (ARC0127; American Radiolabeled Chemicals, St. Louis, MO, USA) or inulin carboxyl (MC-1464; Moravek Biochem- icals, Brea, CA, USA) (see Note 2). 4. Injectable anesthetics (i.e., pentobarbital sodium solution, 50 mg/mL). 5. Evaporator. 6. Temperature-controlled heating pad or hot plate maintained at 37 C. 7. Stereotaxic frame (SR-5R, Narishige, Tokyo, Japan; or permis- sible substitutes). 8. Permanent marker (a black or blue one is recommended). 9. Electric dental drill. 10. Electric shaver. 11. 29G needle. 12. Polyethylene tubing (inside, ~0.28 mm diameter; outside, ~0.61 mm diameter). 13. Silicon tubing (inside, ~0.58 mm diameter; outside, ~1.0 mm diameter). 14. Silk thread. 15. Sanitary cotton. 16. Microsyringe (20 μL; Hamilton, Reno, NV, USA). ® 17. Disposable syringe and needle (i.e., 27G Myjector , TER- UMO, Tokyo, Japan). 18. Surgical instruments for rats (scalpels, scissors, forceps, etc.). 19. Glass vials (20 mL volume, clear). 20. Appropriate shielding for β-emissions. 21. Liquid scintillation counter (LSC-8000, Hitachi-Aloka Medi- cal, Tokyo, Japan). 22. Scintillant: Commercially available scintillation cocktail (i.e., Ultima Gold™; PerkinElmer, Boston, MA, USA). In vivo/vitro Analyses for BCSFB-Mediated Efflux 235

2.2 For Isolated 1. Male adult Wistar rats (150–200 g; Japan SLC, Hamamatsu, Choroid Plexus Japan). Transport Studies 2. 3H-labeled test compound and 14C-labeled n-butanol (ARC0190; American Radiolabeled Chemicals, St. Louis, MO, USA) (see Note 3). 3. Inhalation anesthetics (i.e., isoflurane). 4. aCSF. 5. Temperature-controlled heating pad or hot plate maintained at 37 C. 6. Refrigerated microcentrifuge (i.e., MX-107, TOMY, Tokyo, Japan). 7. Block incubator (for 1.5 mL tubes) maintained at 37 C (i.e., BI-516S with TM-15, ASTEC, Fukuoka, Japan). 8. Mineral oil (d ¼ 0.84; 1.5 g)–silicone oil (d ¼ 1.05; 8.1 g) mixture. 9. 0.4 mL Sampling tube (i.e., ST-004PE; INA OPTIKA, Osaka, Japan). 10. 1.5 mL Conventional plastic tubes. 11. 3 M Potassium hydroxide (KOH). 12. Microforceps. 13. Conventional surgical instruments for rats (scalpels, scissors, forceps, etc.). 14. Glass dishes on the crushed ice. 15. Glass vials (20 mL volume, clear). 16. Appropriate shielding for β-emissions. 17. Liquid scintillation counter (LSC-8000). 18. Scintillant: Alkali-compatible scintillation cocktail (i.e., Hionic-Fluor; PerkinElmer).

2.3 For Studies Using 1. TR-CSFB3 cells. Conditionally 2. Collagen I-coated culture plates for maintaining the cells. Immortalized Rat ® 3. Collagen I-coated 24-well culture plate (i.e., Corning Bio- Choroid Plexus Coat™ Collagen I 24-well plate; CORNING, Corning, NY, Epithelial Cells USA). 4. aCSF. 5. Temperature-controlled heating pad or hot plate. 6. Water bath warmed to 37 C. 7. 1 M Sodium hydroxide (NaOH). 8. 1 M Hydrochloric acid (HCl). 9. Ice box. 10. Glass vials (~7 mL volume, clear). 236 Shin-ichi Akanuma et al.

11. Appropriate shielding for β-emissions. 12. Reagents for the determination of protein concentration (i.e., Bio-Rad DC Protein Assay Kit II; BIO-RAD, Hercules, CA, USA). 13. Liquid scintillation counter (LSC-8000). 14. Scintillant: Commercially available scintillation cocktail (i.e., Ultima Gold™; PerkinElmer).

3 Methods

3.1 Intracerebro- 1. Cut 29G needle (~1.0 cm from the top). ventricular 2. Mark the position 4.0 mm from the top by a permanent Administration marker. 3.1.1 Preparation of the 3. Set polyethylene tubing to the bottom of the needle. Syringe for the 4. Cut the polyethylene tubing at the bottom of the needle. Intracerebro- 5. Cut the silicone tubing (length, ~20 cm) and set the polyethyl- ventricular Administration ene tube (with a needle). 6. Tightly fix the silicone tubing-polyethylene tubing-needle with the silk thread (named as the tube-needle complex; Fig. 2).

3.1.2 Preparation of 1. Place 3H-labeled test compounds and 14C-labeled reference Tracer (See Note 4) compound into a 1.5 mL sampling tube. 2. Dry the solvent using an evaporator. 3. Add an appropriate volume of aCSF. 4. Collect 0.5–1.0 μL of the solution (triplicate) and check the 3H/14C-radioactivities (the radioactivities (dpm/μL) are used to calculate the injected dose).

3.1.3 Anesthesia 1. Record the weight of the rat. 2. Administer the injectable anesthetics. In the case of pentobar- bital sodium solution, inject the solution intraperitoneally (50 mg/kg weight). 3. Keep the rat on the heat pad at 37 C and assess the depth of anesthesia regularly.

Fig. 2 Structure of a tube-needle complex. aCSF artificial cerebrospinal fluid In vivo/vitro Analyses for BCSFB-Mediated Efflux 237

3.1.4 Preparation for the 1. Shave the hair at the back of the neck. Intracerebroventricular 2. Fix the anesthetized rat to the stereotaxic frame. Administration 3. Shave off the hair of the head and cut the scalp with scalpels. 4. Check the position of the bregma. 5. Check the surface of the left lateral ventricle (X ¼ 1.5 mm, Y ¼À0.5 mm from the bregma) and mark the position with a permanent marker (see Note 5). 6. At the position, make a small hole in the skull with the dental drill (Fig. 3a). If blood leaks from the hole, use a sanitary cotton bud to stop it. ® 7. Collect 1 mL aCSF in the 27G Myjector . ® 8. Set the Myjector to the tube-needle complex. 9. Fulfill the aCSF in the tube-needle complex. 10. Fix the tube-needle complex to the manipulator equipped with the stereotaxic frame. 11. Insert the syringe to a depth of 4.0 mm from the surface of the head skull (Fig. 3b). 12. Collect 10 μL tracer solution in the microsyringe. ® 13. Remove the Myjector from the tube-needle complex and immediately set the microsyringe (you can see the compart- ment of air in the tube). 14. Push the 10 μL tracer into the tube. 15. Remove the microsyringe from the tube-needle complex and ® immediately set the Myjector (you can see the compartment of air in the tube).

Fig. 3 Position of intracerebroventricular administration. (a) Surface of the left lateral ventricle (X, 1.5 mm; Y, À0.5 mm from the bregma). (b) Stereotaxic image of the site of administration into the left lateral ventricle. The brain map is from [46] 238 Shin-ichi Akanuma et al.

16. Gently push the tracer solution into the cerebroventricles via the tube-needle complex (Fig. 2). 17. 30 s before the chosen time, remove the tube-needle complex from the cerebroventricle and place the rat in a position to collect the CSF easily. ® 18. Insert a new 27G Myjector into the cisternal magna, and collect the CSF at the chosen time (see Note 6; after this collection, inject an excess of anesthetic to euthanize the animal). 19. Add ~50 μL CSF and the scintillant to the glass vial. 20. Mix well and measure the 3H- and 14C-radioactivities.

3.1.5 Data Analysis The residual compound concentration in rat CSF normalized by the injected compound can be obtained by using Eq. 1: Residual CSF concentrationðÞ%dose=mL CSF Compound concentration in CSFðÞ dpm=mL ¼ Â 100 ð1Þ Injectate concentrationðÞÂ dpm=mL 0:01 mL To obtain the clearance of compound from the CSF, the data of time course of the residual CSF concentration of 3H-labeled test compound and 14C-labeled reference compound, a marker of CSF bulk flow, was fitted to Eq. 2 based on one-compartmental kinetics using nonlinear least-squares regression analysis:

Residual CSF concentration ðÞt ðÞCCSFðÞt =Dose;%dose=mL CSF exp ðÞÀk  t ¼ el  100 V d,CSF ð2Þ

where t, CCSF(t), kel, and Vd,CSF represent the chosen time, the compound concentration in the CSF at the chosen time, the appar- ent elimination rate constant from the CSF, and the distribution volume of the compound in the cerebroventricles/CSF. By multi- plying kel by Vd,CSF, the apparent compound elimination clearance from the CSF (CLCSF) can be obtained.

3.2 Transport 1. Collect 3H/14C-labeled test compounds into a 1.5 mL Studies Using Isolated sampling tube. Rat Choroid Plexus 2. Dry the solvent using an evaporator. 3 3.2.1 Preparation of 3. Add an appropriate amount of a radiolabeled marker ([ H] 14 Tracer (See Note 4) water or [ C]n-butanol) of the volume of the ChP used in the study. 4. Add an appropriate volume of aCSF. 5. Collect 1.0 μL of the solution (triplicate) and check the 3H/14C ratio (the radioactivities (dpm/μL) are used to calcu- late the [3H]/[14C] ratio in the tracer). In vivo/vitro Analyses for BCSFB-Mediated Efflux 239

Fig. 4 Schematic diagram of reaction-stopping tube. ChP choroid plexus

3.2.2 Prepare the 1. Place the 0.4 mL sampling tube in the appropriate tube stand. Reaction-Stopping Tube 2. Add 50 μL 3 M KOH and centrifuge (alkali layer). 3. Add 100 μL mineral oil–silicone oil mixture and spin down (oil layer; Fig. 4). 4. Keep the reaction-stopping tube at room temperature.

3.2.3 Isolation of Choroid 1. Anesthetize the rats by inhalation anesthetics. Plexus from Rats and 2. Euthanize the rats and collect the brain. Uptake Reaction 3. Move the brain to the dishes on the ice. 4. Collect ChPs from the left and right lateral ventricles by microforceps. 5. Immediately transfer the ChPs to the 1.5 mL tube and then add 50 μL37C-warmed aCSF. 6. Incubate for 1 min at 37 C in the block incubator. 7. Add 50 μL tracer solution (see Note 7) and incubate at 37 C. 8. 20 s before the chosen time, collect the ChPs and solution by the pipet and then transfer it to the reaction-stopping tube. 9. Place the reaction-stopping tube in the microcentrifuge. 10. At the chosen time, centrifuge the tube (10,000 Â g,20C, 2 min). 11. See that ChPs are in the alkali layer. 240 Shin-ichi Akanuma et al.

12. Remove the tracer solution and oil layer in the reaction- stopping tube. 13. Incubate (at room temperature, for more than 2 h) until the ChPs have dissolved. 14. Transfer ChP-lysate into the glass vial and then add the scintillant. 15. Mix well and measure the 3H- and 14C-radioactivities.

3.2.4 Data Analysis Tissue/medium ratio is defined as the distribution volume of com- pounds in ChPs per the volume of ChPs (Eq. 3): Tissue=medium ratioðÞμL=μL ChP ðÞ= = : ðÞ=μ ¼ Test compound in ChP dpm sample Test compound conc in tracer dpm L ChP volume ðÞμL ChP=sample ð3Þ Because the actual ChP volume in this study is expressed as the ratio of the ChP volume marker between the ChP lysate sample and tracer, this tissue/medium ratio is determined by following Eq. 4: Tissue=medium ratioðÞμL=μL ChP ðÞ= = : ðÞ=μ ¼ Test compound in ChP dpm sample Test compound conc in tracer dpm L Volume marker in ChPðÞ dpm=sample =Volume marker conc:in tracerðÞ dpm=μL ð4Þ

3.3 Transport 1. Culture and maintain the TR-CSFB3 cells by referring to pre- Studies Using vious reports [32, 37]. Conditionally 2. Collect trypsinized TR-CSFB3 cells into a 15 mL plastic tube. Immortalized Rat  3. Centrifuge (200  g,4 C, 5 min) and remove the supernatant. Choroid Plexus  5 Epithelial Cells 4. Dilute the cells to a density of 2.0 10 cells/mL using regular culture medium. 3.3.1 Preparation 5. Add 500 μL cell suspension to each well of a collagen I-coated of TR-CSFB3 Cells on the 24-well plate. 24-Well Plate  6. Culture the cells for 48 h at 33 Cin5%CO2/air.

3.3.2 Preparation 1. Collect radiolabeled test compounds in a 1.5 mL of Uptake Solution sampling tube. 2. Dry the solvent using an evaporator. 3. Add an appropriate volume of aCSF (for the cells cultured onto a 24-well plate). 4. Collect 5.0 μL the solution (triplicate) and check the radioac- tivity in the solution (the radioactivity (dpm/μL) is used to calculate the uptake activities). In vivo/vitro Analyses for BCSFB-Mediated Efflux 241

3.3.3 Uptake Reaction 1. Place a TR-CSFB3 cell-cultured 24-well plate on the temperature-controlled heating pad (see Note 8). 2. Aspirate the culture medium and rinse the cells with 1 mL 37 C pre-warmed aCSF three times. 3. Apply 200 μL uptake solution. 4. At a chosen time, aspirate the uptake solution. 5. Immediately rinse the cells with 1 mL ice-cold aCSF three times. 6. Add 400 μL 1 M NaOH and solubilize the cells by incubating at room temperature for more than 12 h. 7. Neutralize the lysate by adding 400 μL 1 M HCl. 8. Use reagents for the determination of protein concentration to obtain the amount of protein in the cellular lysate. 9. Transfer 500 μL lysate to the counting glass vial and then add the appropriate volume of scintillant. 10. Mix well and measure the radioactivity.

3.3.4 Data Analysis The uptake activities of the test compound in TR-CSFB3 cells are expressed as the distribution volume of cell/medium ratio (Eq. 5): Cell=medium ratioðÞμL=mg protein ðÞ= ¼ Radioactivities in lysate dpm sample Radioactivity conc:in uptake bufferðÞÂ dpm=μL Protein amountðÞ mg=sample ð5Þ

3.4 Example of the para-Tyramine ( p-tyramine), known as a trace amine, supports Application of These neuronal actions and binds to trace amine-associated receptor Methods [37] 1, which is considered to be one of the therapeutic targets for schizophrenia. In dogs, the concentration of p-tyramine in the CSF is reported to be ~2.6-fold lower than that in plasma [38]. Thus, it is possible that the efflux transport systems at the BCSFB take part in the concentration gradient between the CSF and plasma. We have investigated the carrier-mediated elimination of p-tyramine from the CSF in vivo. Moreover, the characteristics of p-tyramine transport at the BCSFB were examined by two in vitro methods.

3.4.1 In Vivo p-Tyramine [3H]p-tyramine (0.4 μCi/10 μL) and [14C]D-mannitol Elimination from Rat CSF (0.005 μCi/10 μL) were administered into the lateral ventricles of adult Wistar rats (6 weeks old, ~160 g body weight). The residual concentration in CSF of [3H]p-tyramine after cerebroventricles at 1.0, 1.5, and 2.0 min was significantly lower than that of [14C]D- 3 mannitol (Fig. 5a). The kel, Vd,CSF, and CLCSF of [ H]p-tyramine were found to be 0.243 Æ 0.072/min, 159 Æ 16 μL, and 242 Shin-ichi Akanuma et al.

Fig. 5 In vivo elimination of [3H]p-tyramine from the CSF. (a) Time profile of the residual CSF concentration (% dose/mL Â 100) of [3H]p-tyramine (open circles; 0.40 μCi/injection) and [14C]D-mannitol (closed squares; 0.005 μCi/injection) after administration into rat lateral ventricles and sampled from the cisternal magna. The solid line was obtained by nonlinear least-squares regression analysis. Each point represents the mean Æ SEM (n ¼ 3–4). *p < 0.05, significantly different from the respective value for [14C]D-mannitol. (b) The residual CSF concentration of [3H]p-tyramine at 2 min normalized to that of [14C]D-mannitol after intracerebroventricular administration of [3H]p-tyramine and [14C]D-mannitol in the absence (control) or presence of unlabeled 75 mM p- tyramine. Each column represents the mean Æ SEM (n ¼ 3–5). *p < 0.05, significantly different from the control. This figure was prepared by reference to Akanuma S., Yamazaki Y., Kubo Y., Hosoya K. Role of cationic drug-sensitive transport systems at the blood-cerebrospinal fluid barrier in para-tyramine elimination from rat brain. Fluids Barriers CNS., 15:1 (2018) [37]

3 38.6 Æ 12.0 μL/min, respectively. Because the CLCSF of [ H]p- tyramine was 3.5-fold greater than that of [14C]D-mannitol, a CSF flow marker (10.9 Æ 8.3 μL/min), it is suggested that the elimina- tion pathway(s) except for CSF bulk flow participate in p-tyramine clearance from the CSF. In addition, the concentration of [3H]p- tyramine in the CSF relative to that of [14C]D-mannitol was increased by the co-administration of 75 mM unlabeled p-tyramine (Fig. 5b). Taking these lines of evidence into consideration, it is suggested that carrier-mediated transport process(es) are involved in p-tyramine clearance from the CSF.

3.4.2 In Vitro Transport To test the transport properties of p-tyramine at the BCSFB, uptake Studies of p-Tyramine studies using isolated rat choroid plexus and TR-CSFB3 cells were Using Isolated Rat Choroid performed. Time-dependent [3H]p-tyramine uptake was found in Plexus and TR-CSFB3 Cells the isolated rat choroid plexus (Fig. 6a) and TR-CSFB3 cells (Fig. 6b). The [3H]p-tyramine uptake by TR-CSFB3 cells for 2 min at 4 C was significantly reduced compared with that at In vivo/vitro Analyses for BCSFB-Mediated Efflux 243

Fig. 6 In vitro analyses of [3H]p-tyramine uptake. (a) Time-dependent uptake of [3H]p-tyramine uptake by isolated rat choroid plexus. The choroid plexus was incubated with [3H]p-tyramine (1 μCi/sample) and [14C]n- butanol (0.05 μCi/sample) at 37 C. The solid line was obtained by nonlinear least-squares regression analysis. Each point represents the mean Æ SEM (n ¼ 3–6). (b) Time and temperature dependence of [3H] p-tyramine uptake by TR-CSFB3 cells. [3H]p-tyramine uptake (0.15 μCi/well) was measured at 37 C (open circles) and 4 C (closed square) for indicated times. The solid line was obtained by nonlinear least-squares regression analysis. Each point represents the mean Æ SEM (n ¼ 3). **p < 0.01, significantly different from [3H]p-tyramine uptake at 37 C for 2 min. (c) Concentration-dependent p-tyramine uptake over the concen- tration range 20 nM–60 mM at 37 C for 2 min by TR-CSFB3 cells. The data were subjected to Eadie- Scatchard analysis in addition to Michaelis-Menten kinetics (inset). The solid, dashed, and dotted lines represent overall, saturable, and non-saturable transport, respectively. Each point represents the mean Æ SEM (n ¼ 3). This figure was adapted from Akanuma S., Yamazaki Y., Kubo Y., Hosoya K. Role of cationic drug- sensitive transport systems at the blood-cerebrospinal fluid barrier in para-tyramine elimination from rat brain. Fluids Barriers CNS., 15:1 (2018) [37] 244 Shin-ichi Akanuma et al.

37 C. In addition, p-tyramine uptake by TR-CSFB3 cells showed saturable kinetics with an apparent Km, Vmax, and Kd of 3.48 Æ 0.83 mM, 7.26 Æ 1.61 nmol/(min mg protein), and 0.978 Æ 0.066 μL/(min mg protein), respectively (Fig. 6c). This result indicates that carrier-mediated process(es) contribute to p- tyramine transport at the BCSFB, at least in part. We also examined the effect of several substrates/inhibitors for transporters to check the involvement of well-known transporters (Table 1). [3H]p-tyramine uptake by isolated rat choroid plexus was significantly inhibited by propranolol, pyrilamine, and amantadine at 10 mM in addition to imipramine and verapamil at 1 mM. These drugs also inhibited [3H]p-tyramine uptake by TR-CSFB3 cells. In addition, 10 mM nicotine and 10 mM desipramine inhibited [3H] p-tyramine uptake by TR-CSFB3 cells. Recently, it has been reported that unidentified transport system(s) which recognize these drugs involve cationic substrate transport in various tissues and blood-central nervous system barriers [39–45]. Therefore, it is considered that the transport system(s) are expressed at the BCSFB and involved in p-tyramine efflux transport at the BCSFB. Regarding the known transporters, the contribution to p-tyra- mine transport at the BCSFB is considered to be low from the inhibition study (Table 1). Choline inhibited [3H]p-tyramine uptake by isolated rat choroid plexus, but did not inhibit [3H]p- tyramine uptake by TR-CSFB3 cells. This nonidentical effect of choline on [3H]p-tyramine uptake suggests that choline-sensitive transporters including OCT1-2 do not significantly contribute to p-tyramine transport at the BCSFB. In addition, [3H]p-tyramine uptake by both isolated rat choroid plexus and TR-CSFB3 cells was not significantly reduced in the presence of serotonin (a substrate of serotonin transporter), 1-methyl-4-phenylpyridinium (MPP+;a typical substrate of OCT and PMAT), tetraethylammonium (TEA; a substrate of OCT, OCTN, and MATE), and p-aminohip- purate (PAH; a substrate of OAT). Moreover, L-carnitine (a substrate of OCTN), norepinephrine (a substrate of norepineph- rine transporter), cimetidine (a substrate of OCT and MATE), tyrosine (a substrate of large amino acid transporter), and pyri- methamine (a substrate of MATE) did not reduce [3H]p-tyramine uptake by TR-CSFB3 cells. Taking these lines of evidence into consideration, it is suggested that known organic cation transport systems and monoamine transport systems are not involved in p- tyramine efflux transport at the BCSFB. In conclusion, these in vivo and in vitro studies demonstrate the involvement of unidentified carrier-mediated transport system (s) at the BCSFB in p-tyramine elimination from the CSF. So, it is expected that these in vivo and in vitro methods can be applied to identify the molecular mechanisms of compound elimination from the CSF across the BCSFB. In vivo/vitro Analyses for BCSFB-Mediated Efflux 245

Table 1 Effect of several compounds on [3H]p-tyramine uptake by isolated rat choroid plexus and TR-CSFB3 cells

Isolated rat choroid plexus TR-CSFB3 cells

Compounds Conc. (mM) % of control Conc. (mM) % of control

Control 100 Æ 9 100 Æ 1 p-Tyramine 10 65.9 Æ 4.6** 10 51.8 Æ 3.7** Propranolol 10 61.0 Æ 5.4** 10 29.0 Æ 4.2** Pyrilamine 10 61.2 Æ 0.8** 10 65.4 Æ 10.8** Amantadine 10 61.4 Æ 0.6** 10 45.3 Æ 2.4** Imipramine 1 67.4 Æ 1.1** 10 31.5 Æ 2.9** Verapamil 1 70.8 Æ 1.1** 3 69.6 Æ 6.7* Choline 10 72.3 Æ 4.6** 10 150 Æ 6** Nicotine n.d. 10 52.2 Æ 3.8** Desipramine n.d. 10 63.4 Æ 1.9** Serotonin 10 79.6 Æ 2.2 10 114 Æ 5 MPP+ 10 84.7 Æ 6.3 10 155 Æ 19** TEA 10 93.0 Æ 3.1 10 168 Æ 13** PAH 10 95.4 Æ 5.1 10 110 Æ 6 L-Carnitine n.d. 10 75.0 Æ 5.7 Norepinephrine n.d. 10 96.1 Æ 7.3 Cimetidine n.d. 10 115 Æ 2 Tyrosine n.d. 10 140 Æ 17** Control (1% DMSO) n.d. 100 Æ 5 Pyrimethamine (1% DMSO) n.d. 0.2 91.0 Æ 7.3 [3H]p-tyramine uptake by isolated rat choroid plexus (1 μCi/sample) and TR-CSFB3 cells (0.15 μCi/well) at 37 C was performed for 0.75 min and 2 min, respectively, in the absence (control) or presence of unlabeled compounds. Each value represents the mean Æ SEM (n ¼ 3–8). *p < 0.05 and **p < 0.01, significantly different from control. n.d. not determined, MPP+ 1-methyl-4-phenylpyridinium, TEA tetraethylammonium, PAH p-aminohippurate. This table was prepared by reference to Akanuma S., Yamazaki Y., Kubo Y., Hosoya K. Role of cationic drug-sensitive transport systems at the blood-cerebrospinal fluid barrier in para-tyramine elimination from rat brain. Fluids Barriers CNS., 15:1 (2018) [37]

4 Notes

1. The use of analytical grade reagents and ultrapure water is recommended. 2. 14C-labeled D-mannitol and inulin carboxyl are known to be markers of CSF bulk flow [21, 27]. If a 14C-labeled test 246 Shin-ichi Akanuma et al.

compound needs to be assessed, 3H-labeled marker com- pounds can be used, such as 3H-labeled D-mannitol (NET101; PerkinElmer, Boston, MA, USA). 3. 14C-labeled n-butanol is used as a marker of the volume of the choroid plexus [25]. If a 14C-labeled test compound needs to be tested, 3H-labeled water (PerkinElmer) can be used as the reference compound. 4. The ratio of 3H and 14C in tracer solution depends on the capacity of the liquid scintillation counter. 5. The position of the lateral ventricle depends on the strain and body size of the rats used in the study. Before the experiment, a check of the injection position using a rat brain atlas such as [46] in the brain by a preliminary injection test using dye solution, such as 0.4% trypan blue, is recommended. ® 6. If you cannot easily insert Myjector into the cisternal magna, you can cut the back of the neck to see the surface of the cisternal magna using an electric surgical knife (i.e., KN-301A-TSB; Natsume, Tokyo, Japan) before fixation of the rat in the stereotaxic frame. 7. Because the tracer concentration is twofold diluted for this procedure, the 3H- and 14C-radioactivities in the tracer which are measured in the process of Sect. 3.2.1, step 5, should be divided by 2 in the data analyses. 8. Before the experiment, the temperature of the heat pad or hot plate should be checked to place the solution in the 24-well plate at 37 C.

Acknowledgment

This research, especially the in vivo and in vitro p-tyramine trans- port study, was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI [Grant Numbers JP16H05110 and JP16K08365] and the Research Grant from the Smoking Research Foundation.

References

1. de R, Ames A 3rd, Nesbett FB, Hofmann HF 3. Lun MP, Monuki ES, Lehtinen MK (2015) (1960) Fluid formed by choroid plexus; a tech- Development and functions of the choroid nique for its collection and a comparison of its plexus-cerebrospinal fluid system. Nat Rev electrolyte composition with serum and cister- Neurosci 16:445–457 nal fluids. J Neurophysiol 23:485–495 4. Nagata Y, Kusuhara H, Endou H, Sugiyama Y 2. Brightman MW, Reese TS (1969) Junctions (2002) Expression and functional characteriza- between intimately apposed cell membranes in tion of rat organic anion transporter 3 (rOat3) in the vertebrate brain. J Cell Biol 40:648–677 the choroid plexus. Mol Pharmacol 61:982–988 In vivo/vitro Analyses for BCSFB-Mediated Efflux 247

5. Nagle MA, Wu W, Eraly SA, Nigam SK (2013) transporting protein of guanidinoacetate, an Organic anion transport pathways in antiviral endogenous convulsant. J Neurochem handling in choroid plexus in Oat1 (Slc22a6) 107:768–778 and Oat3 (Slc22a8) deficient tissue. Neurosci 14. Wakayama K, Ohtsuki S, Takanaga H, Lett 534:133–138 Hosoya K, Terasaki T (2002) Localization of 6. Ohtsuki S, Takizawa T, Takanaga H, Hori S, norepinephrine and serotonin transporter in Hosoya K, Terasaki T (2004) Localization of mouse brain capillary endothelial cells. Neu- organic anion transporting polypeptide rosci Res 44:173–180 3 (oatp3) in mouse brain parenchymal and 15. Levin E, Nogueira GJ, Garcia Argiz CA (1966) capillary endothelial cells. J Neurochem Ventriculo-cisternal perfusion of amino acids in 90:743–749 cat brain. I. Rates of disappearance from the 7. Tachikawa M, Tsuji K, Yokoyama R, perfusate. J Neurochem 13:761–767 Higuchi T, Ozeki G, Yashiki A, Akanuma S, 16. Levin E, Garcia Argiz CA, Nogueira GJ (1966) Hayashi K, Nishiura A, Hosoya K (2012) A Ventriculo-cisternal perfusion of amino acids in clearance system for prostaglandin D2, a cat brain. II. Incorporation of glutamic acid, sleep-promoting factor, in cerebrospinal fluid: glutamine and GABA into the brain paren- role of the blood-cerebrospinal barrier trans- chyma. J Neurochem 13:979–988 porters. J Pharmacol Exp Ther 343:608–616 17. Baethmann A, Steude U, Horsch S, Brendel W 8. Ohtsuki S, Takizawa T, Takanaga H, (1970) The thiosulphate (35S) space in the Terasaki N, Kitazawa T, Sasaki M, Abe T, CNS of rats after ventriculo-cisternal perfu- Hosoya K, Terasaki T (2003) In vitro study of sion. Pflugers Arch 316:51–63 the functional expression of organic anion 18. Bradbury MW, Brondsted HE (1973) Sodium- transporting polypeptide 3 at rat choroid dependent transport of sugars and iodide from plexus epithelial cells and its involvement in the cerebral venticles of the rabbit. J Physiol the cerebrospinal fluid-to-blood transport of 234:127–143 estrone-3-sulfate. Mol Pharmacol 63:532–537 19. Suzuki H, Sawada Y, Sugiyama Y, Iga T, 9. Tachikawa M, Ozeki G, Higuchi T, Hanano M (1985) Saturable transport of Akanuma S, Tsuji K, Hosoya K (2012) Role cimetidine from cerebrospinal fluid to blood of the blood-cerebrospinal fluid barrier trans- in rats. J Pharmacobiodyn 8:73–76 porter as a cerebral clearance system for prosta- glandin E(2) produced in the brain. J 20. Fujiyoshi M, Tachikawa M, Ohtsuki S, Ito S, Neurochem 123:750–760 Uchida Y, Akanuma S, Kamiie J, Hashimoto T, Hosoya K, Iwatsubo T, Terasaki T (2011) 10. Usui T, Nakazawa A, Okura T, Deguchi Y, Amyloid-beta peptide(1-40) elimination from Akanuma S, Kubo Y, Hosoya K (2016) Hista- cerebrospinal fluid involves low-density lipo- mine elimination from the cerebrospinal fluid protein receptor-related protein 1 at the across the blood-cerebrospinal fluid barrier: blood-cerebrospinal fluid barrier. J Neurochem involvement of plasma membrane monoamine 118:407–415 transporter (PMAT/SLC29A4). J Neurochem 139:408–418 21. Ogawa M, Suzuki H, Sawada Y, Hanano M, Sugiyama Y (1994) Kinetics of active efflux via 11. Okura T, Kato S, Takano Y, Sato T, choroid plexus of beta-lactam antibiotics from Yamashita A, Morimoto R, Ohtsuki S, the CSF into the circulation. Am J Phys 266: Terasaki T, Deguchi Y (2011) Functional char- R392–R399 acterization of rat plasma membrane mono- amine transporter in the blood-brain and 22. Sweet DH, Miller DS, Pritchard JB, blood-cerebrospinal fluid barriers. J Pharm Sci Fujiwara Y, Beier DR, Nigam SK (2002) 100:3924–3938 Impaired organic anion transport in kidney and choroid plexus of organic anion trans- 12. Tachikawa M, Kasai Y, Takahashi M, porter 3 (Oat3 (Slc22a8)) knockout mice. J Fujinawa J, Kitaichi K, Terasaki T, Hosoya K Biol Chem 277:26934–26943 (2008) The blood-cerebrospinal fluid barrier is a major pathway of cerebral creatinine clear- 23. Sykes D, Sweet DH, Lowes S, Nigam SK, ance: involvement of transporter-mediated Pritchard JB, Miller DS (2004) Organic anion process. J Neurochem 107:432–442 transport in choroid plexus from wild-type and organic anion transporter 3 (Slc22a8)-null 13. Tachikawa M, Fujinawa J, Takahashi M, mice. Am J Physiol Renal Physiol 286: Kasai Y, Fukaya M, Sakai K, Yamazaki M, F972–F978 Tomi M, Watanabe M, Sakimura K, Terasaki T, Hosoya K (2008) Expression and 24. Lowes S, Sykes D, Breen CM, Ragone LJ, possible role of creatine transporter in the brain Miller DS (2005) Multiple components of and at the blood-cerebrospinal fluid barrier as a 2,4-dichlorophenoxyacetic acid uptake by rat 248 Shin-ichi Akanuma et al.

choroid plexus. J Pharmacol Exp Ther radiation therapy and chemotherapy for huge 315:136–143 choroid plexus carcinoma in an infant – case 25. Suzuki H, Sawada Y, Sugiyama Y, Iga T, report. Neurol Med Chir (Tokyo) 36:179–184 Hanano M (1986) Transport of cimetidine by 36. Szmydynger-Chodobska J, Pascale CL, Pfeffer the rat choroid plexus in vitro. J Pharmacol Exp AN, Coulter C, Chodobski A (2007) Expres- Ther 239:927–935 sion of junctional proteins in choroid plexus 26. Kitazawa T, Hosoya K, Takahashi T, epithelial cell lines: a comparative study. Cere- Sugiyama Y, Terasaki T (2000) In-vivo and brospinal Fluid Res 4:11 in-vitro evidence of a carrier-mediated efflux 37. Akanuma S, Yamazaki Y, Kubo Y, Hosoya K transport system for oestrone-3-sulphate across (2018) Role of cationic drug-sensitive trans- the blood-cerebrospinal fluid barrier. J Pharm port systems at the blood-cerebrospinal fluid Pharmacol 52:281–288 barrier in para-tyramine elimination from rat 27. Akanuma S, Sakurai T, Tachikawa M, Kubo Y, brain. Fluids Barriers CNS 15:1 Hosoya K (2015) Transporter-mediated L-glu- 38. Faraj BA, Camp VM, Ansley JD, Scott J, Ali tamate elimination from cerebrospinal fluid: FM, Malveaux EJ (1981) Evidence for central possible involvement of excitatory amino acid hypertyraminemia in hepatic encephalopathy. J transporters expressed in ependymal cells and Clin Invest 67:395–402 choroid plexus epithelial cells. Fluids Barriers 39. Fukao M, Ishida K, Horie A, Taguchi M, CNS 12:11 Nozawa T, Inoue H, Hashimoto Y (2014) 28. Akanuma S, Shimada H, Kubo Y, Hosoya K Variability of bioavailability and intestinal (2017) Involvement of carrier-mediated trans- absorption mechanisms of metoprolol. Drug port at the blood-cerebrospinal fluid barrier in Metab Pharmacokinet 29:162–167 spermine clearance from rat brain. Biol Pharm 40. Ishida K, Horie A, Nishimura M, Taguchi M, Bull 40:1599–1603 Fujii N, Nozawa T, Inoue H, Hashimoto Y 29. Barkho BZ, Monuki ES (2015) Proliferation of (2013) Variability of bioavailability and intesti- cultured mouse choroid plexus epithelial cells. nal absorption characteristics of bisoprolol. PLoS One 10:e0121738 Drug Metab Pharmacokinet 28:491–496 30. Takano M, Otani M, Kaji T, Sano K, Hamada- 41. Tega Y, Akanuma S, Kubo Y, Hosoya K (2015) Kanazawa M, Matsuyama S (2016) Proteomic Involvement of the H+/organic cation anti- analysis of mouse choroid plexus cell line porter in nicotine transport in rat liver. Drug ECPC-4 treated with lipid A. Inflamm Res Metab Dispos 43:89–92 65:295–302 42. Tega Y, Yuzurihara C, Kubo Y, Akanuma S, 31. Zheng W, Zhao Q (2002) Establishment and Ehrhardt C, Hosoya K (2016) Functional characterization of an immortalized Z310 cho- expression of nicotine influx transporter in roidal epithelial cell line from murine choroid A549 human alveolar epithelial cells. Drug plexus. Brain Res 958:371–380 Metab Pharmacokinet 31:99–101 32. Kitazawa T, Hosoya K, Watanabe M, 43. Tega Y, Akanuma S, Kubo Y, Terasaki T, Hos- Takashima T, Ohtsuki S, Takanaga H, oya K (2013) Blood-to-brain influx transport Ueda M, Yanai N, Obinata M, Terasaki T of nicotine at the rat blood-brain barrier: (2001) Characterization of the amino acid involvement of a pyrilamine-sensitive organic transport of new immortalized choroid plexus cation transport process. Neurochem Int epithelial cell lines: a novel in vitro system for 62:173–181 investigating transport functions at the blood- 44. Okura T, Hattori A, Takano Y, Sato T, cerebrospinal fluid barrier. Pharm Res Hammarlund-Udenaes M, Terasaki T, Degu- 18:16–22 chi Y (2008) Involvement of the pyrilamine 33. Song H, Zheng G, Liu Y, Shen XF, Zhao ZH, transporter, a putative organic cation trans- Aschner M, Luo WJ, Chen JY (2016) Cellular porter, in blood-brain barrier transport of oxy- uptake of lead in the blood-cerebrospinal fluid codone. Drug Metab Dispos 36:2005–2013 barrier: novel roles of Connexin 43 hemichan- 45. Yamazaki M, Fukuoka H, Nagata O, Kato H, nel and its down-regulations via Erk phosphor- Ito Y, Terasaki T, Tsuji A (1994) Transport ylation. Toxicol Appl Pharmacol 297:1–11 mechanism of an H1-antagonist at the blood- 34. Gu H, Zhong Z, Jiang W, Du E, Dodel R, brain barrier: transport mechanism of mepyra- Farlow MR, Zheng W, Du Y (2014) The role mine using the carotid injection technique. of choroid plexus in IVIG-induced beta-amy- Biol Pharm Bull 17:676–679 loid clearance. Neuroscience 270:168–176 46. Paxinos G, Watson C (2007) The rat brain in 35. Kumabe T, Tominaga T, Kondo T, stereotaxic coordinates in stereotaxic coordi- Yoshimoto T, Kayama T (1996) Intraoperative nates. Elsevier, Amsterdam Chapter 15

In Vivo Analysis to Study Transport Across the Blood-Retinal Barrier

Yoshiyuki Kubo, Shin-ichi Akanuma, and Ken-ichi Hosoya

Abstract

We present our detailed, standardized in vivo protocol for the evaluation of influx and efflux transport at the blood-retinal barrier (BRB). Discovering specific transport of drugs and nutrients is essential to improve our understanding of the pharmacological and physiological roles of the BRB. This will contribute to the development of novel systemic drug delivery system that will help with safer and more efficient therapy of severe retinal diseases, such as diabetic retinopathy and age-related macular degeneration. This report describes protocols for analyzing influx transport across the BRB, including integration plot and retinal uptake index (RUI) methods, providing the in vivo influx clearance and RUI value, respectively. In addition, microdialysis, a protocol for analyzing efflux transport across the BRB, provides the apparent first-order rate constant during the terminal phase. Furthermore, in vivo inhibition analyses using RUI and microdialysis provide the data which will help in investigating the involvement of carrier-mediated transport process.

Key words Blood-retinal barrier, Integration plot, Retinal uptake index (RUI), Microdialysis, Influx transport, Efflux transport, Transporter

1 Introduction

The retina plays a pivotal role in vision, and its function is adversely influenced by severe eye diseases, including glaucoma, age-related macular degeneration, and diabetic retinopathy, that are major causes of blindness [1], suggesting that the safe and efficient drug delivery into the retina is essential for maintaining the quality of life (QOL) of patients. However, the retina is located in the most posterior region of eye, and the topical drug administration (eye drop) is not enough to produce therapeutic concentrations in the retina because of the longer diffusional distance and counter- directional intraocular convection from the ciliary body to Schlemm’s canal. Therefore, drug therapy for eye diseases requires

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_15, © Springer Science+Business Media, LLC, part of Springer Nature 2019 249 250 Yoshiyuki Kubo et al.

Fig. 1 Structure of blood-retinal barrier (BRB). The BRB has two barrier structures, the inner BRB and the outer BRB, of which the responsible cells are the retinal capillary endothelial cells and the retinal pigment epithelial (RPE) cells. While these cells form tight junction to strictly restrict the nonspecific paracellular transport at the BRB, the contribution of influx and efflux transporters has been suggested to the selective transport across the BRB

invasive administration using intravitreal and intrascleral delivery to the retina, and intravitreal delivery with implants and direct injec- tions carries a high risk of serious side effects, such as postoperative endophthalmitis, hemorrhage, and retinal detachment. In addition, side effects including mild pain have been reported as a disadvan- tage of intrascleral delivery with direct injections, suggesting that developing more suitable drug delivery system to the retina is an important challenge to improve the therapy of eye diseases. In the eye, the neural retina and the circulating blood are separated by the blood-retinal barrier (BRB), and nonspecific trans- port between them is severely restricted [2]. The BRB has two barrier structures, the inner and outer BRB, and the responsible cells are retinal capillary endothelial cells and retinal pigment epi- thelial (RPE) cells, respectively (Fig. 1). Recent studies have revealed the expression of influx and efflux transporters in these responsible cells, and it is suggested that these are involved in supplying nutrients and eliminating xenobiotics and metabolites across the inner and outer BRB. Cumulative evidence suggests that the application of transporters is a promising method for In vivo Analysis of BRB Transport 251

systemic drug delivery to the retina that will be helpful for newly developed drugs for the treatment of retinal diseases [2]. This chapter describes the in vivo analytical methods used in rats for evaluation of the permeability across the BRB.

2 Materials

2.1 Animals For in vivo analysis of the BRB permeability, including integration plot, retinal uptake index (RUI), and microdialysis, male Wistar rats (6–8 weeks, 150–250 g) were obtained from Japan SLC (Hama- matsu, Japan). Prior to in vivo analyses, the rats were anesthetized by intraperitoneal administration of pentobarbital (50 mg/kg) (see Note 1).

2.2 Reagents The reagents and equipment shown below are essential for carrying and Equipment out in vivo analyses in rats. Ringer-HEPES buffer (pH 7.4) was prepared by dissolving 141 mM NaCl, 4 mM KCl, 2.8 mM CaCl, and 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) in double-distilled water, and 5 M NaOH was added to adjust the pH. Saline was also prepared by dissolving 0.9 w/v% NaCl in double-distilled water (see Note 2). For integration plot, an injection solution (1–5 μCi/400 μL) was prepared by dissolving a radiolabeled test compound in Ringer- HEPES buffer. Clear 1.5 mL tubes and 25-gauge syringe needles were flushed with heparin (1000 units/mL), and stored at 4 C. Surgical cotton and gauze were thoroughly moistened with plenty of saline in a glass dish [3, 4]. For RUI, the injection solution (injectate) was prepared by dissolving a radiolabeled test compound and freely diffusible refer- ence compound ([3H]water or [14C]n-butanol) in Ringer-HEPES buffer (see Note 3). Heavy paper with vinyl tape on one side was cut into small pieces (3 mm  1 cm), and a 29-gauge syringe needle was bent at a 120–140 angle using forceps (Fig. 2)[5–7]. For microdialysis, the injection solution (injectate) was prepared by dissolving radiolabeled test compound and a bulk flow marker ([3H]D-mannitol or [14C]D-mannitol) in Ringer- HEPES buffer. The injectate was put into a microsyringe, and its radioactivity was measured to allow a preliminary calculation of the dosage (dpm/μL  1 μL/apply) (see Note 4). When an in vivo inhibition analysis is planned, an inhibitor should be added to the perfusate and delivered to the probe as a perfusate (see Note 5). A customized microdialysis probe (TEP-50) was designed to allow insertion into the rat eye ball for several hours (Fig. 3). Polyethyl- ene tubing (inner diameter, 0.4 mm; length, ~60 cm), with a 1 mL syringe for perfusate filling, was connected to the inlet tube of the 252 Yoshiyuki Kubo et al.

Fig. 2 Retinal uptake index (RUI). 29-Gauge syringe needle was bent at a 120–140 angle by means of forceps in order to facilitate the administration of test compound. A piece of heavy paper with vinyl tape on one side (3 mm  1 cm) was inserted under the isolated common carotid artery of each rat, and care should be taken to avoid the interruption of blood flow

probe, and polyethylene tubing (inner diameter, 0.2 mm; length, ~20 cm) was connected to the outlet tube. In addition, a syringe (1 mL) for perfusate filling was connected to a 23-gauge needle, and filled with perfusate. This syringe was placed on an infusion pump (Model 11), and the perfusate was continuously delivered to the probe at 2.2 μL/min. After filling the polyethylene tubing and microdialysis probe with Ringer-HEPES buffer, the tubing was inserted into a clean sampling vial (see Note 6)[8–10].

2.2.1 Reagents Heparin (1000 unit/mL). HCl (2 M). Scintillation cocktail. NaOH (2 M). Pentobarbital. Surgical glue. 2% Xylocaine (lidocaine). In vivo Analysis of BRB Transport 253

Fig. 3 Microdialysis. Customized microdialysis probe was designed. While the BRB-impermeable compound, such as D-mannitol, is eliminated only through bulk flow, several BRB-permeable compounds are eliminated through efflux transport at the BRB in addition to bulk flow. BRB blood-retinal barrier, ID inner diameter, OD outer diameter, RPE retinal pigment epithelium

2.2.2 Equipment Balance. Block incubator. Bone scissors. Centrifuge. Curved and tapered tweezers. 120–140 Angle by means of forceps. Fine ophthalmic scissors. Fine tweezers. 22-Gauge syringe needle. 23-Gauge syringe needle. 25-Gauge syringe needle. 29-Gauge syringe needle. Gauze. Heavy paper. 254 Yoshiyuki Kubo et al.

Infusion pump (Model 11, Harvard, Holliston, MA). Kite string. Liquid scintillation counter. Microdialysis probe (TEP-50, Eicom, Kyoto, Japan). 10 μL Microsyringe (Hamilton, Reno, NE). Polyethylene tubing (SP-19, Natsume, Tokyo, Japan). Polyethylene tubing (inner diameter, 0.2 mm, Natsume, Tokyo, Japan). Polyethylene tubing (inner diameter, 0.4 mm, Natsume, Tokyo, Japan). Ring tweezers. Rodent guillotine. Sampling glass vial. Stereotaxic instrument (SR-5R, Narishige, Tokyo, Japan). Surgical cotton. Surgical sutures. Syringe (1 mL). 1.5 mL Tube. Vinyl tape.

3 Methods

3.1 Integration Plot Integration plot analysis was performed to evaluate the in vivo influx clearance at the BRB [3, 4]. Radiolabeled test compound was administered via the femoral vein of the rat, and blood samples were collected from the jugular vein at 15–30 s prior to the selected time. The rat was decapitated to collect the retina at the designated time, followed by determination of the radioactivity. In data analy- sis, the values of the retinal distribution volume of the compound at time ta (Vd), the plasma concentration of the compound at time ta (Cp(ta)), and the area under the plasma concentration-time curve from time 0 to ta (AUCp(ta)) were used to determine the retinal influx clearance of the compound (CLinfux,retina), which is one of the key values for influx transport across the BRB.

3.1.1 Administration 1. Wistar rats were anesthetized by intraperitoneal administration of Test Compound of pentobarbital (50 mg/kg) (see Note 7). 2. The femoral vein and jugular vein of the rats were exteriorized (see Note 8). 3. The injection solution containing radiolabeled test compound was given via the femoral vein. In vivo Analysis of BRB Transport 255

4. Blood samples (400 μL) were collected from the jugular vein at 15–30 s before the designated time. The collected blood sam- ples were then centrifuged for 10 min at 5000 Â g and 4 C, and the supernatant was stored as blood plasma. 5. Rats were decapitated at the designated time.

3.1.2 Collection of Retina 1. The bone of the eye socket was cut using bone scissors, and the eye ball was picked up using ring tweezers. 2. The eye ball was collected by cutting the optic nerve in the posterior region with small scissors. 3. The collected eye ball was placed on gauze moistened with Ringer-HEPES buffer. 4. Each eye ball was cut at the boundary of the anterior (transpar- ent color) and posterior parts (yellowish-while color) using a fine ophthalmic scissors. 5. The posterior part of eye ball was collected. 6. The sclera was removed from the retina using a fine tweezers (see Note 9). 7. The retina was collected for weighting in a clear 1.5 mL tube.

3.1.3 Radioactivity 1. 2 M NaOH (1 mL) was added to the retina in a 1.5 mL tube. Determination 2. The retina was lysed by incubating for 2 h at 65 C. 3. 2 M HCl was added to a 1.5 mL tube in order to neutralize the lysed sample. 4. The neutralized sample was mixed with commercially available scintillation cocktail (i.e., Ultima Gold™, PerkinElmer, Wal- tham, MA) (see Note 10). 5. The radioactivity of the sample was measured in a liquid scin- tillation counter.

3.1.4 Calculation of Vd 1. The measured radioactivity in the retina and blood plasma was divided by their weight (g) and volume (mL), respectively. The results were expressed as the amount of compound in the retina at time ta (Xretina,tot(ta), dpm/g retina) and Cp(ta) (dpm/mL), respectively.

2. The value of Vd (mL/g retina) was calculated using Eq. 1:

V d ¼ X retina,totðÞt a =CpðÞta ð1Þ

3.1.5 Calculation of AUCp 1. The values of Cp (dpm/mL) were plotted at designated times on a single logarithmic chart, and a suitable pharmacokinetic model, such as linear one compartment model, was selected. 256 Yoshiyuki Kubo et al.

2. The initial value of Cp(t) was calculated, and the optimum expression for Cp(t) was developed using a nonlinear least- squares regression analysis program (see Note 11).

3. The resultant expression for Cp(t) was put into Eq. 2, and the value of AUCp(ta) (dpm·min/mL) was calculated: ð ta AUCpðÞ¼t a CpðÞt dt ð2Þ 0

3.1.6 Calculation of In 1. AUCp(ta)/Cp(ta) and Vd were plotted on the horizontal and Vivo Influx Clearance vertical axes, respectively, and the value of CLinfux,retina was determined as the slope of Eq. 3 (see Note 12):

X retina,totðÞt a =CpðÞ¼t a V d ¼ CLinflux,retina  AUCpðÞt a =CpðÞþt a V i ð3Þ

3.2 Retinal Uptake The retinal uptake index (RUI) was determined to evaluate the Index (RUI) in vivo influx transport across the BRB, and the permeability of test compound into the retina was expressed as a percentage of the permeability of the reference ([3H]water or [14C]n-butanol) to the retina [5–7]. The radiolabeled test compound was given via the common carotid artery, and the rat was decapitated 15 s after administration. The radioactivity of the collected retina was measured in a liquid scintillation counter, followed by calculating the value of RUI which is also one of the key values for influx transport across the BRB. The in vivo inhibition analysis with the RUI is helpful to evaluate the contribution of a carrier-mediated process at the BRB. The working process of the RUI is closely similar to that of the brain uptake index (BUI), an in vivo analytical method to evaluate the in vivo influx transport across the blood- brain barrier (BBB), and the BUI can be assessed concurrently with the RUI [6].

3.2.1 Rat Treatment 1. Wistar rats were anesthetized by the intraperitoneal administra- tion of pentobarbital (50 mg/kg) (see Note 7). 2. The anesthetized rats were fixed by tying their forepaws with vinyl tape or kite strings. 3. The common carotid artery of the rats was exteriorized, and isolated with curved and tapered tweezers (see Note 13). 4. The heavy paper (3 mm  1 cm) was placed under the isolated common carotid artery of the rats (see Note 14) (Fig. 2). 5. Each rat was partially unfixed, and moved to a place close to the rodent guillotine (see Note 15). In vivo Analysis of BRB Transport 257

3.2.2 Administration and 1. At 0 s, the injection solution was given via the common carotid Decapitation artery (see Note 16). 2. At 12 s, clean surgical cotton was held in the nondominant hand while the syringe needle was held in the dominant hand. 3. At 13 s, surgical cotton was placed on the injection part, and the syringe needle was put down or removed. 4. At 14 s, the rat was moved to the rodent guillotine with the dominant hand. 5. At 15 s, the rat was decapitated. 6. The eye ball on the injection side was enucleated, and the retina was collected in a clean 1.5 mL tube (see Note 17). The retina can be efficiently collected in accordance with “Collection of retina” described in the Sect. 3.1.

3.2.3 Radioactivity 1. 2 M NaOH (1 mL) was added to the retina in 1.5 mL tube. Determination 2. The retina was lysed by incubating for 2 h at 65 C. 3. 2 M HCl was added to a 1.5 mL tube in order to neutralize the lysed sample. 4. The neutralized sample was mixed with commercially available scintillation cocktail (i.e., Ultima Gold™, PerkinElmer) (see Note 10). 5. The radioactivity of the sample was measured in a liquid scin- tillation counter.

3.2.4 Calculation of RUI 1. The RUI value (%) was calculated by assigning the measured Value radioactivity (dpm) to Eq. 4 or 5: ½Test compound in the retinaðÞ dpm = Test compound in injectateðފ dpm RUIðÞ¼% 100  ½Reference compound in the retinaðÞ dpm = Reference compound in injectateðފ dpm ð4Þ

½ŠTest=Reference ratio in the retinaðÞ dpm RUIðÞ¼% 100  ½ŠTest=Reference ratio in injectateðÞ dpm ð5Þ

3.3 Microdialysis Microdialysis was performed to evaluate the in vivo efflux transport across the BRB by means of a customized microdialysis probe [8–10]. The injection solution (1 μL) was administered to the vitreous body of the rat (Fig. 3). Microdialysis was performed for 180 min by using an infusion pump connected to the microdialysis probe, and the dialysate was collected every 10 min, followed by 258 Yoshiyuki Kubo et al.

Fig. 4 Microdialysis in the study of anionic drug transport. The efflux transport of [3H]benzylpenicillin was assessed by means of microdialysis. (a) Time course of concentrations of [3H]benzylpenicillin (closed triangle) and [14C]D-mannitol (open circle) in the dialysate. Each point represents the mean Æ S.E.M. (n ¼ 6). (b) The effect of several compounds on [3H]benzylpenicillin elimination was analyzed in the in vivo inhibition analysis. Each column represents the mean Æ S.E.M. (n ¼ 3–16). *P < 0.01, significantly different from the control. PAH p-aminohippuric acid. Cd, concentration of the compound in dialysate; λ2, apparent first-order rate constant during terminal phase. Figures were adapted from Hosoya K, Makihara A, Tsujikawa Y, Yoneyama D, Mori S, Terasaki T, Akanuma S, Tomi M, Tachikawa M. Roles of inner blood-retinal barrier organic anion transporter 3 in the vitreous/retina-to-blood efflux transport of p-aminohippuric acid, benzylpenicillin, and 6-mercaptopurine. J Pharmacol Exp Ther 329:87–93 (2009) with permission from The American Society for Pharmacology and Experimental Therapeutics (ASPET) [10]

radioactivity determination in a liquid scintillation counter. After calculating the concentration of the compound in the dialysate (Cd, % dose/mL), the time course of the compound concentration in dialysate (Cd(t)) was fitted to a biexponential equation (Fig. 4a), and the apparent first-order rate constant during the terminal phase (λ2) was obtained. When the value of λ2 for the test compound was greater than that of the bulk flow marker (D-mannitol), the test compound is suggested to be eliminated through facilitative efflux transport across the BRB in addition to elimination via Schlemm’s canal and/or the uveoscleral outflow route. In the in vivo inhibi- tion analysis with microdialysis, the relative Δλ2 change is helpful for evaluating the contribution of the carrier-mediated process to efflux transport at the BRB (Fig. 4b).

3.3.1 Calculation The probe recovery ratio (%) of test compound should be prelimi- of Recovery narily assessed by probe perfusion in tube filled with test solution, and can be calculated by Eq. 6, where CT (dpm/mL) and CV (dpm/mL) are the concentration of the dialysate solution and the concentration in the test solution (or isolated vitreous humor), respectively. In our previous report, the recovery for [3H]spermine and [14C]D-mannitol was 5.22% and 8.69%, respectively, and was In vivo Analysis of BRB Transport 259

constant over 180 min [9]. When an inconstant recovery was observed for the test compound, the composition of the perfusate should be subjected to suitable modification, such as the addition of bovine serum albumin (BSA): C Recovery ratioðÞ¼% 100 Â T ð6Þ CV

3.3.2 Administration 1. The anesthetized rat was fixed on a stereotaxic instrument to Vitreous Body and (SR-5R). Dialysate Collection 2. The eyelid of rat was locally anesthetized by instillation of 2% xylocaine (lidocaine), and the eyelid was fixed with surgical sutures to prevent blinking. 3. A 22-gauge needle was inserted approximately 0.5 mm below the corneal scleral limbus through the pars plana at a depth of 3.0 mm, and the discharge was mopped up with clean gauze. 4. After removing the needle, 1 μL injection solution was admi- nistered to the vitreous body by means of a 10 μL microsyringe at a depth of 3.0 mm from the surface of the eye (Fig. 3)(see Note 18). 5. After removing the microsyringe, the microdialysis probe was immediately inserted at a depth of 3.0 mm from the surface of the eye (see Note 19). 6. The infusion pump was started promptly upon completing insertion of the probe (time 0), and dialysate began to be collected in sample vials (see Note 20). 7. The dialysate was collected in sample vials every 10 min. 8. The weight of sampling vials with dialysate was measured. 9. The dialysate was mixed with commercially available scintilla- tion cocktail (i.e., Ultima Gold™, PerkinElmer). 10. The sample radioactivity was measured in a liquid scintillation counter.

3.3.3 Calculation 1. The value of Cd (% dose/mL) was calculated by Eq. 7, where Z, of Apparent First-Order V, and Dinjectate are the radioactivity in collected dialysate Rate Constants During (dpm/vial), the volume of collected dialysate (μL/vial), and the Terminal Phase dosage amount (dpm), respectively (see Note 21): ÀÁ Cd ¼ fgZ =ðÞV =1000 Â 100=Dinjectate ð7Þ 2. Using nonlinear least-squares regression analysis program, such as MULTI, the time course of Cd(t) was fitted to a biexponential equation formed by the initial and terminal phases (Eq. 8 and Fig. 4a). The constants C1 and C2 were intercepts on the y-axis for each exponential segment of the 260 Yoshiyuki Kubo et al.

curve in Eq. 8. The constants λ1 and λ2 were the apparent first- order rate constants for the initial and terminal phases, respec- tively (see Note 22):

Cd ¼ C1  expðÞþÀλ1  t C2  expðÞðÀλ2  t 8Þ

Δλ λ 3.3.4 Calculation of 2 1. In the in vivo inhibition analysis, the difference in the 2 values change of the test compound and D-mannitol (Δλ2 ¼ λ2,test com- pound À λ2,D-mannitol) was used in Eq. 9:

ðÞΔλ2with inhibitor Relative Δλ2ðÞ¼% 100 Â ð9Þ ðÞΔλ2without inhibitor

2. The resultant relative Δλ2 (% of control) was used for evaluating the effect of inhibitor (Fig. 4b).

4 Actual Cases in the Study of Drug Transport at the BRB

4.1 Integration Plot It has been reported that the BRB permeability of compounds, and RUI in a Study undergoing passive diffusion, is well correlated with their lipophi- of Verapamil Transport licity, showing a lipophilicity trend line (r2 ¼ 0.807, Eq. 10), when at the BRB the lipophilicity was expressed as the n-octanol/Ringer-HEPES buffer (pH 7.4) distribution coefficient (DC) [6]: RUI ¼ 46:2 Â expðÞ 0:515 Â log DC ð10Þ However, at the same time, the BRB permeability of com- pounds undergoing carrier-mediated transport was not consistent with a prediction based on their lipophilicity, and the substrate of influx transporters exhibited a higher RUI value than the predicted one while the substrate efflux transporters, such as P-glycoprotien, exhibited a lower value than predicted one. In our previous study, among the typical substrates of P-gp, [3H]verapamil exhibited a greater RUI value than the one pre- dicted by Eq. 10 although [3H]digoxin and [3H]vincristine exhib- ited lower values (Fig. 5a), suggesting blood-to-retina transport of verapamil at the BRB [6]. In the in vivo inhibition analysis, the RUI value of [3H]verapa- mil was increased in the presence of quinidine, a P-gp substrate, while no effect was shown by choline. Interestingly, the RUI value was decreased in the presence of pyrilamine, an H1 receptor antag- onist (Fig. 5b), while the BUI value of [3H]verapamil was increased and unchanged in the presence of quinidine and pyrilamine, respectively [11]. In the integration plot analysis of [3H]verapamil, the influx clearance of verapamil at the BRB was calculated to be 614 Æ 40 mL/(min g retina) that is 4.7-fold greater than the influx clearance of verapamil in the BBB where the expression of P-gp was In vivo Analysis of BRB Transport 261

Fig. 5 Integration plot and RUI in the study of cationic drug transport. The transport of [3H]verapamil was assessed by means of in vivo analyses. (a) RUI values of compounds undergoing carrier-mediated transport at the BRB were actually determined. Closed circles (●) express the substrates of solute carrier (SLC) transporter, and squares (■, □) express the substrates of ATP-binding cassette (ABC) transporter. The lipophilicity trend line indicates the correlation between the RUI and the log DC of the compounds undergoing passive diffusion across the BRB. (b) The effect of compounds on the RUI of [3H]verapamil was analyzed in the in vivo inhibition analysis. Each column represents the mean Æ S.E.M. (n ¼ 3–16). *P < 0.01, significantly different from the control. (c) The influx clearance of [3H]verapamil was assessed by means of an integration plot. Each point represents the mean Æ S.E.M. (n ¼ 3–4). RUI retinal uptake index, Vd distribution volume of compound at time t, AUCp(t) area under the plasma concentration-time curve from time 0 to t, Cp(t) plasma concentration of the compound at time t. Figures were adapted from Hosoya K, Yamamoto A, Akanuma S, Tachikawa M. Lipophilicity and transporter influence on blood-retinal barrier permeability: a comparison with blood-brain barrier permeability. Pharm Res 27:2715–2724 (2010), and Kubo Y, Kusagawa Y, Tachikawa M, Akanuma S, Hosoya K. Involvement of a novel organic cation transporter in verapamil transport across the inner blood-retinal barrier. Pharm Res 30:847–856 (2013) with permission from Springer [6, 11]

reported (Fig. 5c). These results suggest the involvement of a pyrilamine-sensitive transport system in the blood-to-retina trans- port of verapamil across the BRB, and the in vitro analyses with a conditionally immortalized retinal capillary endothelial cell line (TR-iBRB2 cells) supported the contribution of novel organic cation transporter to the influx transport of verapamil at the inner BRB [11].

4.2 Microdialysis The involvement of organic anion transporter in the elimination of in the Study benzylpenicillin from the retina was assessed by means of micro- of Benzylpenicillin dialysis [10]. In the analysis, [3H]benzylpenicillin and [14C]D- Transport mannitol were injected into vitreous body, and their concentration in dialysate was determined over time. In addition, perfusate and injectate with or without an inhibitor were used in the in vivo inhibition analysis. The concentration of inhibitor in the retina was determined using high-performance liquid chromatography (HPLC), after perfusion. The microdialysis analysis exhibited biexponential elimination of [3H]benzylpenicillin from the vitreous humor with an apparent elimination constant (λ2) during terminal phase of 262 Yoshiyuki Kubo et al.

1.56 Â 10À2 Æ 0.08 Â 10À2/min that was 1.7-fold higher than that of [14C]D-mannitol used as a bulk flow marker (Fig. 4a). In the in vivo inhibition analysis, the value of Δλ2 was reduced by 78% in the presence of unlabeled benzylpenicillin, and also by 35% in the presence of probenecid, an inhibitor of Oat1-3 and organic anion transporting polypeptide (Oatps). Furthermore, the Δλ2 was reduced by 49% in the presence of p-aminohippuric acid, an inhibitor of Oat1-3, while no effect was shown by digoxin which inhibits Oatp1a4 (Fig. 4b). These results suggest the contri- bution of Oat1-3 to the retina-to-blood transport of benzylpeni- cillin across the BRB and, in particular, the importance of Oat3 at the inner BRB was supported since it is the only Oat isoform the expression of which was detected at the inner BRB.

5 Notes

1. All analyses using animals should be in accordance with the Association for Research in Vision and Ophthalmology (ARVO) Statement and approved by the institutional animal care committee. In addition, rats were subjected to fasting for 18–24 h when the BRB permeability of endogenous com- pounds was investigated. 2. Saline can also be used as a substitute for Ringer-HEPES buffer. 3. The radioactivity in the injectate should be confirmed with a liquid scintillation counter, and the radioactivity for [3H]- labeling set more than fivefold the radioactivity for [14C]-label- ing in both injectate and collected tissue samples to maintain the reliability of dual measurement with a liquid scintillation counter. In addition, when [125I]-labeling was analyzed as a test compound, a gamma counter should be used to confirm the radioactivity for [125I]-labeling. In addition, the energy range of measurement should be altered to prevent overestima- tion of the radioactivity of the reference compound when the measurement range for [125I]-labeling occasionally overlaps with that of [3H]- or [14C]-labeling. 4. The radioactivity for [3H]-labeling should be approximately fivefold greater than that for [14C]-labeling. For example, [3H]E17βG(5μCi) and [14C]D-mannitol (1 μCi) were previ- ously dissolved as test compound and bulk flow marker, respec- tively [8]. In addition, when [125I]-labeled was analyzed as a test compound, a gamma counter should be used to confirm the radioactivity for [125I]-labeling. In addition, the energy range of measurement should be altered to prevent overestima- tion of the radioactivity of the reference compound when the In vivo Analysis of BRB Transport 263

measurement range for [125I]-labeling occasionally overlaps with that of [3H]- or [14C]-labeling. 5. It should be noted that the retinal concentration of inhibitor is usually one-twenty-fifth that of the diluted injection solution and perfusate at least. The retinal concentration may be experi- mentally determined if necessary. 6. The runoff perfusate from polyethylene tubing connected to the outlet tube should be confirmed visually. In addition, the weight of sampling vials should be measured beforehand. 7. Supplemental anesthesia (~20 mg/kg) is performed by verify- ing the condition of the rats. 8. The exteriorized tissue should be covered with surgical cotton moistened with saline to prevent drying. In addition, when blood sampling is planed over time, the femoral artery should be cannulated with polyethylene tubing containing heparin (100 unit/mL) in Ringer-HEPES buffer. 9. The vitreous body, crystal lens, and inner limiting membrane should be removed from the posterior part of the eye using fine tweezers. 10. When alkali-compatible scintillation cocktail (i.e., Hionic- fluor, PerkinElmer) is used, neutralization of the sample is not essential.

11. The initial of Cp(t) can be calculated using spreadsheet soft- ware, such as Microsoft Excel. As a nonlinear least-squares regression analysis program, we have used MULTI. 12. The mass balance equation of the compound can be expressed as in Eq. 11 during a short period where the metabolism and efflux transport of the compound may be negligible. In addi- tion, CLinfux,retina was regarded as the slope of Eq. 12, when AUCp(ta), the integral of the plasma concentration of the compound from time 0 to ta, was plotted on the horizontal axis, and Xretina(ta), the amount of the compound taken up by the retina, was plotted on the vertical axis. Xretina,tot(ta) can be experimentally obtained, and is not strictly equal to Xretina(ta), since Xretina,tot(ta) is the value obtained in the radioactivity measurement of the retina, and is assumed to be equal to the sum of Xretina(ta) and Xvas(ta) that are the amount of the compound taken up by the retina and the residual amount of the compound in the retinal capillaries, respectively, as shown in Eq. 13. Then, Eq. 12 can be converted to Eq. 14. Further- more, the distribution volume of the compound (Vi) in the compartment where the compounds equilibrate rapidly within the retinal blood vessels can be expressed by Eq. 15, and this leads to the conversion of Eq. 14 to Eq. 3: 264 Yoshiyuki Kubo et al.

dX retina=dt ¼ CLinflux,retina Á Cp ð11Þ ð t a X retinaðÞ¼t a CLinflux,retina  CpðÞt dt 0

¼ CLinflux,retina Á AUCpðÞta ð12Þ

X retina,totðÞ¼ta X retinaðÞþt a X vasðÞt a ð13Þ

X retina,totðÞ¼t a CLinflux,retina Á AUCpðÞþt a X vasðÞt a ð14Þ

X vasðÞ¼t a CpðÞÂta V i ð15Þ

13. Care should be taken not to injure nerve fiber. 14. Care should be taken not to obstruct the blood flow. 15. Create an enabling environment for carrying the rat in one hand. 16. For only a moment, the translucent-becoming blood vessel may be observed, and the syringe needle should be held at the injection site for 13 s. 17. For example, the eye ball on the right side should be enucleated when the test compound was administered from the right common carotid artery. In addition, a delay of 1–2 h often has no significant influence on the radioactivity of [3H]water and [14C]n-butanol in our experience, and the retinas can all be collected at one time, when you must analyze multisamples. 18. The microsyringe should be removed from the eye 30 s after administration. 19. Care should be taken to prevent the probe from contacting other sites. 20. The probe should be attached to the conjunctiva with surgical glue (i.e., Aron Alpha A “Sankyo,” Daiichi Sankyo, Tokyo, Japan). The moment when the probe is inserted should be taken as time 0, and microdialysis should be carried out for 180 min. 21. The values of V should be calculated by subtracting the weight of the clean sampling vial from the weight of the sampling vial filled with dialysate, supposing the density of the dialysate to be 1 mg/μL.

22. Cd(t) can be fitted to a biexponential equation in many cases, and the terminal phase is assumed reasonable for analyzing the elimination of the compound from the vitreous humor. When a significant difference was observed between the λ2 values of the test compound and the bulk flow marker (D-mannitol), it is In vivo Analysis of BRB Transport 265

suggested that the elimination of the test compound from the vitreous humor is mediated by efflux transport at the BRB in addition to passage from the vitreous humor to Schlemm’s canal and/or the uveoscleral outflow route (Figs. 3 and 4a).

Acknowledgments

This study was financially supported in part by the Japan Society for the Promotion of Science (JSPS) KAKENHI (grant number JP16H05110 and JP17K08409), JSPS Core-to-Core Program (B. Asia-Africa Science Platforms), and Research Grants from the Smoking Research Foundation and the Takeda Science Foundation.

References

1. Zhang K, Zhang L, Weinreb RN (2012) Oph- 7. Alm A, To¨rnquist P (1981) The uptake index thalmic drug discovery: novel targets and method applied to studies on the blood-retinal mechanisms for retinal diseases and glaucoma. barrier. I. A methodological study. Acta Physiol Nat Rev Drug Discov 11:541–559 Scand 113:73–79 2. Hosoya K, Tomi M, Tachikawa M (2011) Stra- 8. Katayama K, Ohshima Y, Tomi M, Hosoya K tegies for therapy of retinal diseases using sys- (2006) Application of microdialysis to evaluate temic drug delivery: relevance of transporters the efflux transport of estradiol 17-beta glucu- at the blood-retinal barrier. Expert Opin Drug ronide across the rat blood-retinal barrier. J Deliv 8:1571–1587 Neurosci Methods 156:249–256 3. Hosoya K, Minamizono A, Katayama K, 9. Kubo Y, Tomise A, Tsuchiyama A, Akanuma S, Terasaki T, Tomi M (2004) Vitamin C transport Hosoya K (2014) Involvement of the carrier- in oxidized form across the rat blood-retinal bar- mediated process in the retina-to-blood trans- rier. Invest Ophthalmol Vis Sci 45:1232–1239 port of spermine at the inner blood-retinal bar- 4. Blasberg RG, Fenstermacher JD, Patlak CS rier. Exp Eye Res 124:17–23 (1983) Transport of alpha-aminoisobutyric 10. Hosoya K, Makihara A, Tsujikawa Y, acid across brain capillary and cellular mem- Yoneyama D, Mori S, Terasaki T, Akanuma S, branes. J Cereb Blood Flow Metab 3:8–32 Tomi M, Tachikawa M (2009) Roles of inner 5. Tachikawa M, Takeda Y, Tomi M, Hosoya K blood-retinal barrier organic anion transporter (2010) Involvement of OCTN2 in the trans- 3 in the vitreous/retina-to-blood efflux trans- port of acetyl-L-carnitine across the inner port of p-aminohippuric acid, benzylpenicillin, blood-retinal barrier. Invest Ophthalmol Vis and 6-mercaptopurine. J Pharmacol Exp Ther Sci 51:430–436 329:87–93 6. Hosoya K, Yamamoto A, Akanuma S, Tachi- 11. Kubo Y, Kusagawa Y, Tachikawa M, kawa M (2010) Lipophilicity and transporter Akanuma S, Hosoya K (2013) Involvement of influence on blood-retinal barrier permeability: a novel organic cation transporter in verapamil a comparison with blood-brain barrier perme- transport across the inner blood-retinal barrier. ability. Pharm Res 27:2715–2724 Pharm Res 30:847–856 Chapter 16

Increasing BBB Permeability via Focused Ultrasound: Current Methods in Preclinical Research

Dallan McMahon, Charissa Poon, and Kullervo Hynynen

Abstract

The tightly regulated permeance of the blood-brain barrier (BBB) greatly limits the range of therapeutic treatment options for central nervous system (CNS) diseases. The use of focused ultrasound (FUS), in conjunction with circulating microbubbles, is a unique approach whereby the transcranial application of acoustic energy, focused within targeted brain areas, can be used to induce a noninvasive, transient, and targeted increase in BBB permeability. This can provide an avenue for the delivery of therapeutic agents from the systemic circulation into the brain. While this approach continues to show great promise and has entered clinical testing, there remains a need for preclinical research to investigate the long-term effects of single and repeated FUS treatment on cerebrovascular health and neurological function, as well the pharmacokinetics of specific drugs following FUS. Additionally, there is a need for improved monitoring strategies that can precisely predict resulting bio-effects. This will allow the continued development of control algorithms that can further increase the safety profile of FUS. Here we will describe two approaches to study FUS-mediated increases in BBB permeability in rodent models: MRI-guided FUS and in vivo two-photon fluorescence microscopy FUS experiments. The goal of this chapter is to outline each proce- dure, present options for experimental design, and highlight important considerations for the collection and interpretation of data.

Key words Focused ultrasound, Blood-brain barrier, Microbubbles, Drug delivery, Two-photon microscopy, Magnetic resonance imaging

1 Introduction

The BBB plays an integral role in maintaining a tightly regulated environment within the central nervous system, essential for proper cerebral function. The physical barrier created by tightly linked endothelial cells limits passive paracellular diffusion to small (<400–500 Da), hydrophilic molecules [1]. Movement of sub- stances from the vascular lumen into the brain parenchyma via a transcellular route can be inhibited by a variety of endothelial cell

Dallan McMahon and Charissa Poon contributed equally to this work.

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_16, © Springer Science+Business Media, LLC, part of Springer Nature 2019 267 268 Dallan McMahon et al.

efflux pumps and enzymes, as well as a reduced rate of vesicle- mediated transport. Additionally, astrocytes, microglia, and peri- cytes all have roles in maintaining proper barrier function. While the BBB is crucial for preserving homeostasis in the CNS, it severely limits the ability to deliver therapeutic agents to the brain [1]. There are numerous methods of circumventing the BBB for drug delivery currently under investigation. These include intrana- sal delivery [2], intra-arterial administration of hyperosmotic solu- tions [3], and radio frequency heating of magnetic nanoparticles [4], among others [5–10]. In general, the majority of methods designed to enhance delivery of therapeutic agents to the brain are limited by poor spatial resolution, invasive methodology, and/or require complex biochemical design. In contrast, FUS, used in conjunction with intravenous microbubble (MB) administration, is a noninvasive way to transiently increase the permeability of the BBB in targeted areas [11]. Following FUS treatment, therapeutic agents can be introduced into systemic cir- culation and penetrate the BBB in therapeutically relevant concen- trations at the intended targets [12–23]. This technique has the potential to provide clinicians with a powerful tool to treat a variety of neuropathologies. The use of FUS, without MBs, as a noninva- sive method of ablating small volumes of brain tissue has under- gone considerable research and has shown a high degree of success in the treatment of essential tremors [24, 25]. The feasibility of using FUS and MBs to increase BBB permeability was first demon- strated in 2001 by Hynynen et al. [11]. Nearly 17 years later, this technique has now entered human testing with several ongoing (ClinicalTrials.gov identifiers: NCT03321487, NCT02343991, NCT03608553, NCT03616860, NCT03626896) and completed trials [26, 27]. While much work has shown that FUS can be used to tran- siently increase the permeability of the BBB with a high degree of safety and in a targeted manner, there is considerable room for further preclinical research. Questions that remain to be answered include determining the mechanisms behind increased BBB perme- ability following sonication, characterizing long-term effects of repeated treatment, designing strategies to modulate the degree of FUS-mediated BBB permeability, and fully assessing the biological events which follow FUS. The goal of this chapter is to describe the details of how FUS can be used to increase BBB permeability in rodents, as well as how this can be used in conjunc- tion with MRI or two-photon fluorescence imaging to visualize the in vivo effects of the technique. FUS-Mediated Increase in BBB Permeability 269

2 Principles of FUS-Mediated Increases in BBB Permeability

Research conducted in the 1990s suggested that the cavitation and thermal effects of FUS could be used to increase BBB permeability without gross tissue damage. However, inconsistencies with the technique raised concerns with respect to its safety and efficacy [28]. The addition of MBs to FUS treatments allowed a reduction in the acoustic power required to influence BBB permeability by lowering the cavitation threshold within the vasculature. This reduction in acoustic energy significantly reduces skull heating effects and permits safe transcranial ultrasound administration. As the MBs travel through the circulatory system they are dispersed and eventually pass through blood vessels within the focus of the ultrasound. Here, they begin to expand and contract in response to the cycles of rarefaction and compression of the propagating ultrasound wave. The mechanical stimulation of vessel walls caused by the behavior of MBs is believed to drive the resul- tant increase in BBB permeability [29]. Physical interactions between MBs, FUS, and vasculature believed to contribute signifi- cantly to this effect include shear stress caused by the microstream- ing of stably oscillating MBs [30, 31], radiation force on MBs in the direction of the propagating acoustic wave [32], and inertial cavi- tation [33]. It is important to note that inertial cavitation has been associated with tissue damage and hemorrhage, but is not necessary to achieve FUS-mediated increases in BBB permeability [34, 35]. The biological mechanisms by which these physical forces induce increased BBB permeability have yet to be fully elucidated. Following sonication, there is an increased number of vesicles, vacuoles, fenestrations, and transcellular channels in capillary endo- thelial cells, as well as dye leakage through the paracellular space, past the tight junctions [36]. Changes in tight junction integrity are mirrored at the protein level, with a significant reduction in the immunoreactivity of occludin, claudin-5, and zona occludens-1 at 1 and 2 h following FUS. A return to baseline levels is observed at 4 h, corresponding to the return of tight junction integrity [37]. The observed increase in the number of endothelial vesicles, predominantly found in arterioles, is also supported by evidence at the protein level. An upregulation of caveolin-1 has been measured by immunohistochemistry and Western blot analyses at 1 h follow- ing FUS [38].

3 Applications

While FUS may prove to have a wide range of applications, preclin- ical research involving FUS-induced increases in BBB permeability have thus far largely focused on the delivery of therapeutic agents. Given the scarcity of drugs which permeate the BBB at 270 Dallan McMahon et al.

concentrations sufficient to achieve clinically relevant effects [1], work demonstrating the ability to deliver substances to the brain using FUS has thus far shown great promise. Examples of thera- peutics that have been shown to cross the BBB following FUS include chemotherapeutics [12–16, 39–42], viral vectors [17–20, 43–45], antibodies [21–23], natural killer cells expressing Her2 chimeric antigen receptor [46], iron-labeled green fluorescent protein-expressing neural stem cells [47], therapeutic magnetic nanoparticles [48], and brain-derived neurotrophic factor [49]. Much work remains to be done in characterizing the pharma- cokinetics, pharmacodynamics, optimal treatment schedule, among others, of these compounds. In addition to increasing the efficiency of delivering therapeutic agents, FUS-induced increases in BBB permeability may have other clinically relevant applications. In 2013, Jorda˜o et al. first demon- strated that FUS treatment, without delivery of therapeutics, was sufficient to significantly decrease mean Aβ plaque size and total surface area in transgenic mice model of Alzheimer’s disease (AD) [50]. Subsequent work showed that three weekly treatments of FUS improved AD-like pathology on both cellular and behav- ioral levels [51]. Similar results have been reported by a separate research group using different murine models of AD [52–54], and a naturally aged canine model [55]. Further investigation is required to elucidate the mechanisms behind and kinetics of how FUS alters AD-like pathology. FUS has also been reported to induce neurogenesis in the dentate gyrus [56]. Recent work suggests that this effect is depen- dent on increasing the permeability of the BBB, as sonication at low pressures with MBs, as well as at high pressures in the absence of MBs, both failed to have an effect on the density of BrdU-positive cells in the dentate gyrus [57]. While the impact of FUS-stimulated neurogenesis on hippocampal function is unclear, the link between reduced hippocampal volumes and decreased rates of neurogenesis in mood and neurodegenerative disorders [58–61] may suggest a role for FUS in their treatment. Future work should focus on characterizing the behavioral impact of FUS-induced neurogenesis in animal models of depression. In recent years, there has been debate regarding the magnitude and duration of neuroinflammation that follows sonication [62–67]. There is evidence that FUS causes an acute inflammatory response [50, 52, 62, 65, 68], and that a variety of parameters, including MB dose, are likely to influence this response [65, 66]. A wide range of tools have been used to assess the impact of FUS-mediated increases in BBB permeability on brain health, including imaging [55, 65, 69–74], behavioral testing [51, 52, 55, 70, 74–76], and histological assessment [34, 39, 44, 53, 75, 77–80]. From these studies, it appears that it is possible to FUS-Mediated Increase in BBB Permeability 271

transiently increase the permeability of the BBB with minimal impact on short-term brain health and no detectable impact on long-term brain health, provided that appropriate sonication para- meters are used.

4 FUS System

A FUS system used for BBB applications must have the capability of generating short, controlled bursts of ultrasound and should have the flexibility to adjust acoustic parameters (peak negative pressure, pulse repetition frequency, burst length, sonication duration, etc.) to obtain the desired result. The range of sonication frequencies used to effectively increase the permeability of the BBB is limited, as higher frequencies are accompanied by increased bone attenuation [81] and lower frequencies result in large focal spots. The optimal frequency for clinical transcranial FUS application is likely to be in the range of 0.2–1.5 MHz [82, 83]. Communication between software and a function generator allows appropriate signals to be sent to an RF power amplifier based on the chosen sonication parameters. Impedance of the electrical current coming from the power amplifier needs to be matched to that of the transducer; this is achieved with a matching circuit (Fig. 1). Monitoring the forward and reflected RF-power (using a power meter) can reveal telling information on the coupling of the transducer with the animal. For example, if the reflected power is high, it is advisable to check for air bubbles under the transducer (which may be reflecting the ultrasound). The safety and repeatability of FUS-mediated enhancement of BBB permeability depends on the cavitation activity of MBs. Stable cavitation of MBs in blood vessels can cause a transient increase in paracellular and transcellular transport. FUS treatments in various brain regions and animal models have shown that a controlled level of stable cavitation results in minimal to no detectable impact on brain health, as discussed above. In contrast, the occurrence of inertial cavitation can result in hemorrhage, edema, and damage to endothelial cells [84, 85]. Although the sonication parameters required to cause increased BBB permeability differ based on factors such as driving frequency, species (predominately effecting skull attenua- tion), and MB dose, McDannold et al. showed that when expressed as a function of mechanical index (MI), the threshold for BBB opening is approximatelypffiffiffi 0.46 [86]. From the equation for MI (MI ¼ PNP= f , where PNP ¼ estimated in situ peak negative pressure in MPa, and f is the center frequency of the ultrasound wave), an appropriate PNP can be estimated, how- ever, validation for each experimental protocol is advisable to ensure the desired effects are achieved. 272 Dallan McMahon et al.

Fig. 1 FUS system. Sonication parameters (e.g., frequency, duration) are entered into computer software, which sends information to the function generator. The signal is then amplified and measured by the power meter, before feeding into the transducer. MB activity can be measured by the addition of a hydrophone. This figure shows the setup for a FUS two-photon microscopy experiment; however, the hardware controlling MRI-guided FUS experiments follows the same configuration. The transducer is adhered onto a coverslip, which is coupled to the brain with saline. The connection between the transducer and brain surface is stabilized by 1% agarose

Another method to determine an appropriate PNP is to cali- brate this parameter based on MB activity, monitored during soni- cation with a hydrophone. Based on acoustic emissions, used to infer the type of MB activity present, a control algorithm can be used to adjust PNP in real time. One strategy that has shown to be effective in producing consistent increases in BBB permeability is to ramp up PNP in small increments with each subsequent burst. Once sub- or ultraharmonic emissions, indicative of nonlinear oscillations, are detected, the PNP is reduced by half for the remainder of FUS treatment [85]. The magnitude of second har- monic emissions has also been shown to have value for adjusting PNP to yield consistent increases in BBB permeability [39].

5 Important Factors to Consider for FUS Experiments

There are several important details to be aware of when conducting FUS experiments. Here, we review four crucial components that are relevant in all FUS experiments: animal preparation, water, MBs, and anesthesia. FUS-Mediated Increase in BBB Permeability 273

5.1 Animal While animal preparation for MRI-guided FUS and two-photon Preparation fluorescence imaging of FUS differ considerably, there remain sev- eral principles and procedural steps that are common to both techniques; this section will highlight the commonalities before detailing the specifics of each technique in subsequent sections. Rodent preparation for any FUS procedure must accomplish three basic goals: (1) achieve a sufficient plane of anesthesia, (2) pro- vide a means to introduce agents into the systemic circulation, and (3) provide a route for ultrasound to propagate efficiently. A commonly used rodent anesthetic protocol in our lab con- sists of induction with 5% isoflurane at 1 L/min with oxygen as a carrier gas, followed by maintenance at 1–2% isoflurane at 1 L/min with medical air. This protocol has the advantage of being tunable to the status of the animal and having a quick recovery time. As long as procedure times are not abnormally long (>30 min), mice and rats tolerate the anesthetic well and consistent increases in BBB permeability are seen following FUS. During the FUS procedure, several agents (MBs, contrast agents, therapeutic agents) must be administered into the systemic circulation. While some groups have found success with delivering MBs retro-orbitally [52], placement of a tail vein catheter provides an easy route by which MBs can be introduced with relatively little stress, minimizing the risk of destroying the MBs before they reach the systemic circulation. Tail vein catheters also allow contrast agents and therapeutic agents to be delivered as a bolus or infusion during sonication, which maximizes the delivery across the BBB. Typically, 22 and 27 gauge catheters are used in rats and mice, respectively. Lastly, the ability to propagate ultrasound efficiently into the brain is paramount to producing consistent increases in BBB per- meability. In terms of animal preparation steps that can be employed to achieve this goal, removal of hair along the path of ultrasound propagation prevents air bubbles from being trapped in the follicles, which will scatter ultrasound. If using an acoustic control algorithm, trapped air bubbles may also contribute to acoustic emissions in the frequency range being monitored (trig- gering a premature drop in PNP). Thus, the complete removal of hair from the top of the head with depilatory cream is necessary to produce consistent FUS treatments. The skull can attenuate a significant proportion of ultrasound and distort the focus in larger experimental animals (e.g., rabbits, porcine, and nonhuman primates) and human patients, necessitat- ing either skull thinning, cranial windows, or the use of large phased arrays with skull corrections. However, in small rodents the skull is thin enough that ultrasound can be focused in the brain without significant skull heating or prohibitive distortion of the focus, provided the skull is normal to the direction of ultra- sound propagation and the frequency is relatively low. 274 Dallan McMahon et al.

5.2 Degassed, Of considerable importance is the use of deionized, degassed water Deionized Water as a medium to propagate ultrasound. Dissolved gases and micro- scopic bubbles can impede the efficiency of ultrasound propaga- tion, which creates a source of variance in the delivery of acoustic energy and the detection of acoustic emissions. Additionally, the increased conductivity of ionized water can provide a conduction path for electromagnetic interference, increasing the noise received by the hydrophone.

5.3 Microbubbles The addition of intravenously administered MBs to the FUS proce- dure has been integral in developing a reliable way to increase the permeability of the BBB, while minimizing the risk of tissue dam- age. There are several options available for MB type (Definity, Optison, SonoVue, in-house made, etc.), all with specific advan- tages and disadvantages. Effort should be made to understand how different MB types respond to FUS, as well as their pharmacokinet- ics, in order to design robust experiments. Due to variations in size distribution, shell properties, concentration, etc., it is difficult to compare doses of differing MB types. While gas volume may have value in approximating equivalent [45] doses, in vivo pilot studies are required whenever developing new protocols or changing exist- ing ones. Definity MBs are used in our lab due to their Health Canada and FDA approval, the extensive literature available on their response to FUS [87, 88], and their commercial availability. Definity MBs must be allowed to acclimatize to room temper- ature prior to activation. Once activated, an 18 gauge blunt tip needle is used to gently draw up the MBs into a 1 mL syringe. When handling MBs, effort should be made to reduce impacts and pressure changes which may destroy significant proportions of formed bubbles. To reduce the change in pressure caused by draw- ing MBs from the vial, a second needle can be used as a vent. Additionally, in our experience, the use of large gauge, blunt tipped needles for MB draws are best to maximize the chance of successful FUS treatments.

5.4 Anesthesia Like any procedure that requires the induction and maintenance of anesthesia in rodents, there are a variety of options; however, it may be of value for FUS procedures to consider the effects of the anesthetic agent on the activity and integrity of circulating MBs. To this end, McDannold et al. have investigated the effects of anesthetics on the degree of MRI contrast enhancement following FUS. They found that isoflurane, with oxygen as a carrier gas, induces significantly less gadolinium contrast enhancement follow- ing FUS when compared to intraperitoneally administering a keta- mine/xylazine cocktail [89]. These effects may be attributed to differences in local vessel diameter and blood flow caused by the different anesthetics, which could potentially alter the local bubble concentration and the interaction of the MBs with vessel walls. FUS-Mediated Increase in BBB Permeability 275

The same group has also shown that the choice of carrier gas is likely a large contributor to differences in BBB effects following FUS. Isoflurane, with medical air as a carrier gas, results in a significantly higher level of increased BBB permeability when com- pared to oxygen, as well as a higher magnitude of harmonic emis- sions [77]. These effects can likely be attributed to oxygen having a negative impact on MB circulation time [90, 91].

6 MRI-Guided FUS Experiments in Rodents

While localized increases in BBB permeability can be achieved with FUS in the absence of MRI, the use of this technology can increase the precision and accuracy of targeting and enable posttreatment assessment without sacrificing the animal. These benefits can add to the scientific rigor of a chronic treatment study by enabling a comparison of the degree and location of increased BBB perme- ability across time and between animals. However, in the absence of access to MRI, stereotaxically targeted FUS has been shown to have a low margin of error in spatial accuracy [92].

6.1 Co-Registrations For simplicity, MRI-guided FUS will be described for the use of of FUS System with MR commercially available equipment, specifically the RK100 System Scanner (FUS Instruments Inc., Toronto, CAN) and BioSpec 7T MRI (Bruker Scientific Instruments, Billerica, Massachusetts, USA). Co-registration of transducer positioning system and MRI spatial coordinates is achieved by sonicating an MRI-compatible sled with an open water reservoir that can be coupled to the water tank housing the transducer. While sonicating at a PNP capable of producing a small, controlled fountain in water (e.g., 0.5 MPa at 1.68 MHz; this PNP will depend of the frequency the transducer is being driven at), the transducer is moved in the vertical direction until the fountain is centered at the surface of the water in the open reservoir. A 3D-printed plastic fiducial marker is physically centered at the fountain and secured, demarking the location of the ultra- sound focus. A 3D FLASH tripilot scan of the sled with open water reservoir and plastic insert is performed. From these three imaging planes, the coordinates of the focus are defined within the RK100 software by locating the level of the water (vertical coordinate of the focus) and the center of the plastic insert (two horizontal coordi- nates of the focus). This co-registration of the transducer coordi- nates with the MRI coordinates enables FUS targets to be chosen directly from MR images.

6.2 Animal The anesthetized rodent, with the hair removed from the top of the Positioning in the FUS head and tail vein catheter placed, is first secured in a supine System position on an MRI-compatible (Fig. 2a). The sled provided with the RK100 system includes a height adjustable bite bar and nose 276 Dallan McMahon et al.

Fig. 2 MRI-guided FUS. (a) The rodent is positioned supine on an MRI-compatible sled with the top of the skull coupled to a polyimide membrane. The bottom of the membrane is coupled to a tank filled with degassed, deionized water, housing the transducer/hydrophone assembly. (b) FUS is targeted from T2-weighted images (targets indicated by red circle with cross). (c) The effect of FUS on BBB permeability is assessed by gadolinium-contrast enhancement using T1-weighted images

cone assembly that allows the user to secure the head of the rodent, provide inhaled anesthetic, and scavenge exhalation products. A polyimide window in the sled, filled with degassed/deionized water, provides a route to propagate ultrasound from the water tank (which houses the transducer) to the brain of the rodent. Before positioning the animal on the sled, ultrasound gel is placed on the polyimide window in the location that the top of the skull will rest. It is important that there is enough gel to couple the top of the head to the polyimide window, but not too much as to make it difficult for the rodent to breathe (especially relevant for mice, due to their small size). When positioning the animal on the sled the incisors are hooked into the bite bar and the nose cone is advanced to fit snugly around the snout of the animal. Special care must be taken to ensure there are no air bubbles trapped between the top of the head and the polyimide membrane, as this will inhibit ultrasound propagation and interfere with the acoustic emissions used for acoustic feedback control. Tape can be used to secure the body to the sled, reducing the potential for the head to change position while moving the sled to and from the MRI, which would reduce the accuracy of targeting. It may be desirable to shift the precise position of the head depending on the region of the brain being targeted (the skull is ideally normal to the direction of ultrasound propagation), but in general, the horizontal plane of the brain should be parallel to the polyimide membrane. FUS-Mediated Increase in BBB Permeability 277

The body temperature of the animal should be maintained while under anesthetic. This can be accomplished with warm water bags, MRI-compatible heating pads, forced-air warming sys- tem, etc. Given the size of mice, temperature maintenance is espe- cially important for reducing the rate of death while under anesthetic.

6.3 Targeting FUS The imaging sequences required for basic targeting of FUS in from MR Images rodents are quite simple. In general, T2-weighted sequences are sufficient to visualize brain anatomy with a level of detail that allows either direct targeting of specific areas (Fig. 2b) or targets to be estimated based on their proximity to visible structures (brain atlases can be useful for this). For targeting areas in the brain that are not readily visible on T2-weighted images and are not in pre- dictable locations, such as certain types of implanted tumors, gado- linium contrast enhanced T1-weighted imaging can be useful. While these sequences are the most commonly used in our lab, any MRI sequence that provides sufficient spatial information about the intended FUS targets can be used for targeting. The animal is imaged while secured on the MRI-compatible sled. It is important to be gentle while moving the sled from the bore of the MRI to the FUS system to avoid head movement, which will introduce error into MRI-based targeting (while the RK100 system is MRI-compatible, it does not fit into the bore of the Bruker 7T, thus the sled and animal is moved between the MRI and RK100). T2-weighted images are imported into the FUS Instruments software and targets are placed in the regions of inter- est. Depending on the pulse repetition frequency, duty cycle, dis- tance between targets, and speed of the motors moving the transducer, there is a limit to the number of targets that can be sonicated with one injection of MBs. When using a pulse repetition frequency of 1 Hz, 10 ms burst length, and 1 mm spacing between targets, the RK100 can sonicate four targets per dose of MBs.

6.4 Sonication Once target planning is complete, the sled is coupled to the water Procedure for MRI- tank of the FUS system, and sonication parameters have been Guided FUS chosen, Definity MBs are diluted in saline. Dilution ratios of 1:49 and 1:9 are used for mice and rats, respectively, to minimize the volume of fluid administered intravenously. The MBs and saline are mixed gently in the syringe by inverting repeatedly until the solu- tion appears homogeneous. If a control algorithm is used to cali- brate PNP, and a baseline measurement of acoustic emissions is used in this algorithm, sonications should be started before MB administration. Otherwise, immediately prior to the start of soni- cation, a dose of 20 μL/kg of Definity is administered via the tail vein catheter. This can either be done by a slow bolus injection or with the use of an infusion pump [93]. 278 Dallan McMahon et al.

The ultrasound parameters used to induce increased BBB per- meability are perhaps the most varied feature of the FUS procedure between labs. This may be due to the number of factors that can influence the effectiveness of treatment (MB handling, animal type, animal age, transducer geometry, anesthetic protocol, etc.). For this reason, it is imperative that thorough pilot studies are per- formed that include histological evaluation of targeted brain areas and, ideally, biochemical and behavioral evaluation; it is not advis- able to simply follow the parameters used in published articles without in-house validation studies. As a starting point, a constant PNP may be determined based on values found in the literature. McDannold et al. have found that the threshold PNP for increased BBB permeability is approximately 0.5 MPa at 1 MHz and scales with the inverse of the square root of frequency in MHz [86]. This pressure needs to be estimated in the brain, accounting for attenu- ation along the path of ultrasound propagation. Higher PNPs result in increased BBB permeability and eventually tissue damage. The FUS parameters commonly used in our lab for rodents include 10 ms burst length, 1 Hz pulse repetition frequency, 120 s treatment duration, and 0.551–1.68 MHz transducer frequency. For experiments using an acoustic controller, PNP starts at 0.128–0.250 MPa (depending on the driving frequency) and is increased 0.008–0.025 MPa every burst until acoustic emissions at the sub- or first ultraharmonic frequencies reach 3.5 times the mean of the baseline signal (baseline is collected during the first 10 bursts, without MBs in circulation). Once triggered, the PNP is reduced by half and remains until the end of the sonication. This strategy has been shown to produce more consistent FUS treatments [85]. Once the sonication is completed, the treatment can be assessed by MRI (permeability, edema, hemorrhage, blood- oxygen-level dependent response, etc.). Depending on the design and purpose of the study, the animal may then be recovered, main- tained under anesthesia, sonicated a second time (at least 15 min between MB injections is recommended to allow Definity concen- trations to be reduced), imaged, etc.

6.5 BBB Permeability There are a variety of approaches that can be used to confirm the Quantification increase in BBB permeability following FUS and/or quantify this effect, as well as several points to consider when designing this part of an experiment. First, it is important to remember that BBB permeability is not binary (open vs. closed) and that substances of differing physical properties (mass, charge, etc.) can traverse the BBB at different rates [94]. It is also critical to understand that the FUS-mediated changes in BBB permeability will begin to decay toward baseline immediately following sonication; this rate of decay will also depend on the physical properties of the substance travers- ing the BBB [94]. Thus, when designing or interpreting FUS-Mediated Increase in BBB Permeability 279

experiments using FUS to increase BBB permeability, it is impor- tant to carefully consider the methods used. There are three common approaches to confirm or quantify the increase in BBB permeability following sonication: (1) administra- tion of an MRI contrast agent and subsequent imaging in vivo, (2) administration of a tracer and subsequent ex vivo evaluation, and (3) immunohistochemistry of endogenous substances that permeate the BBB following FUS.

6.5.1 MRI Contrast enhanced MRI is a simple way to assess FUS treatment and enables a quantification of the affected brain volume without necessitating the removal of the brain. This is beneficial for studies that investigate the effects of repeated treatments [51] or long-term survival [56] following FUS. Typically, a contrast agent is adminis- tered intravenously during sonication and imaging, specific to that agent, is subsequently performed (e.g., T1-weighted scan with a gadolinium-based contrast agent). The volume of brain displaying increased BBB permeability can be calculated by determining the number of voxels with intensities at least 2.5 standard deviations above the mean intensity of non-sonicated regions of the brain [95] (Fig. 3). This method can also be used to track the return of BBB

Fig. 3 Gadolinium contrast enhancement following sonication. (a) Horizontal plane of T1-weighted image following FUS-mediated increase in BBB perme- ability. Gadolinium contrast enhanced areas (three seen in this image) indicate areas of the brain for which FUS has had an impact of BBB permeability. (b) The volume of enhancement can be determined by first calculating the mean and standard deviation of voxel intensity in the non-sonicated contralateral hemi- sphere (blue box). Then the number of voxels in the sonicated hemisphere (red box) which are at least 2.5 standard deviations above the mean intensity of the non-sonicated hemisphere (purple pixels) are counted. Knowing the volume of each voxel, the volume of enhancement can be calculated 280 Dallan McMahon et al.

permeability to baseline by administering contrast agent and imag- ing at regular time intervals following sonication; however, it is important to remember that the use of different contrast agents will result in different half closure times (time required for BBB perme- ability to be reduced by half). This effect has been demonstrated by Marty et al. who showed that half closure times for MRI contrast agents are inversely related to their size [94]. To obtain more quantitative assessments of BBB permeability following sonication, dynamic contrast enhanced (DCE) MRI can be performed. This method uses T1 mapping, the change in con- trast enhancement over time, and an arterial input function to model the rate of gadolinium movement across the BBB (or your contrast agent of choice). While DCE-MRI has been used to assess BBB permeability following FUS [40, 96–99], this method can be more technically challenging.

6.5.2 Evans Blue Dye In situations where MRI is not available or not feasible, the intra- venous administration of dyes that do not penetrate the naive BBB in high concentrations can be used to evaluate FUS treatments ex vivo. The most commonly used dye for this purpose is Evans blue (EB), which has a very high affinity for serum albumin. Given this affinity, EB acts as an approximately 68 kDa dye, only permeat- ing the BBB in high concentrations where FUS has had an effect. Typically, EB is administered during, or shortly after sonication, and allowed to circulate for 1 h or more; however, the dye will not remain in the brain parenchyma indefinitely, thus this technique is not appropriate for long-term survival studies. During this time, rodents should be kept anesthetized due to the potential for dis- comfort. After an hour, rodents are transcardially perfused with ice-cold phosphate buffer (0.1 M) until the fluid exiting the right atrium runs clear to remove blood and EB from circulation. This tissue can be examined in several ways. Sonicated brain areas can be dissected, homogenized in trichloroacetic acid, and absorbance measured at 620 nm in the supernatant [100]. When normalized to non-sonicated control tissue, this absorbance measurement can be used to compare relative effects of FUS on BBB permeability [101]. Alternatively, following phosphate buffer, rodents can be perfused with 4% paraformaldehyde (in 0.1 M phosphate buffer), sectioned, and fluorescently imaged (Ex 540 nm/ Em 680 nm) (Fig. 4) or qualitatively assessed [102]. EB can be used for acute studies and can provide investigators with information regarding the success of FUS treatment. While EB may provide a semi-quantitative indication of the degree to which BBB permeability has been increased following FUS, the complex dynamic between influx and efflux across the BBB that can occur during the extended EB circulation time, combined with the fact that EB can only provide information about a single time FUS-Mediated Increase in BBB Permeability 281

Fig. 4 Evans blue dye fluorescence following sonication. EB can be administered i.v. following a FUS-mediated increase in BBB permeability. The dye is allowed to circulate and extravasate in the regions of the brain that have been affected by FUS. Animals are transcardially perfused with saline, followed by a fixative (e.g., 4% paraformaldehyde in 0.1 M phosphate buffer). After 24 h of postfixation, brains can be vibratome sectioned at 100 μm(a) and fluorescently imaged (b). An IVIS 100 imaging system with an excitation filter of 615–665 nm and an emission filter of 695–770 nm can be used to identify regions of EB extravasation and quantify the effect

point, may add uncertainty to quantitative analyses. In addition to EB, other dyes, such as sodium fluorescein, may be useful in asses- sing the success of FUS treatment [100].

6.5.3 Leakage The extravasation of endogenous compounds from the systemic of Endogenous Substances circulation into the brain can also be used as a measure to assess the effect of FUS on BBB permeability. Generally, a substance that does not cross the naive BBB at a high rate can be stained and used to demark the location of increased BBB permeability ex vivo. The benefit of this technique is that it does not require the administra- tion of an exogenous dye; however, it may be more difficult to compare treatments between animals, given the concentration of any substance of interest in circulation may differ significantly between animals, altering the amount available to extravasate fol- lowing FUS. Despite this limitation, immunostaining of IgM and IgG has been used previously to indicate the location of increased BBB permeability and has been shown to positively correlate to post-FUS gadolinium contrast enhancement [50]. As with the administration of dye, animals must be sacrificed (transcardially perfused) within a certain time frame following sonication to avoid substantial removal (efflux, phagocytosis, degradation, etc.) of the substance from brain tissue. The duration of this delay will depend on the substance being evaluated; in a mouse model of Alzheimer’s disease (TgCRND8), immunodetection of IgM and 282 Dallan McMahon et al.

IgG in sonicated brain areas remained significantly elevated for at least 4 days post FUS [50]. The availability of a variety of methods to assess FUS treatment allows great flexibility in experimental design, but requires thoughtful planning to ensure that the methods are suited to answer the questions being asked. While some of the more com- mon approaches have been discussed here, there exist others. Methods of evaluating the magnitude of acoustic emissions at the sub- and ultraharmonic frequencies detected during sonication have shown predictive value in determining the success of FUS treatment [85, 103], as well as tissue damage [72]. Additionally, decreases in the uptake of 18F-2-fluoro-2-deoxy-d-glucose, detected with positron emission tomography, has been shown to correlate to regions of increased BBB permeability following FUS [73]. It is also important to consider that while it is relatively common for long-term survival or repeated treatment studies to conduct pilot studies to determine optimal sonication parameters and use these for subsequent experiments without evaluating each FUS treatment, this approach is prone to wide variations in FUS treatments that aren’t considered in the interpretation of results. This increase in variance may mask significant effects or lead to conclusions that are not applicable to optimized, properly assessed, FUS treatments.

7 Two-Photon Fluorescence Microscopy FUS Experiments

Two-photon fluorescence microscopy is a powerful tool for visua- lizing biological responses as it enables in vivo imaging on a micro- scopic scale with decreased risk of photobleaching [104]. In the context of FUS, it allows the visualization of FUS-mediated BBB effects and subsequent vascular and cellular responses in real time. Compared to single-photon confocal microscopy, two-photon microscopy has superior deep tissue imaging capabilities but slightly inferior spatial resolution [105]. An important distinction between confocal microscopy and two-photon microscopy is the method by which high spatial reso- lution is achieved. Confocal microscopy uses pinholes, such that only the photons from the focal plane are transmitted to the detec- tor, and all other photons are excluded. However, undetected photons just outside of the focal plane can still bleach the tissue. In addition, the scattering of photons is high in thick samples, and thus phototoxicity becomes more probable as users are forced to compensate for this loss by increasing fluorescence excitation [106]. Conversely, two-photon microscopy only excites photons at the focal plane; thus, phototoxicity is much reduced. In addition, the near-infrared wavelengths used in two-photon microscopy are FUS-Mediated Increase in BBB Permeability 283

absorbed to a lesser degree in biological tissue than are the shorter wavelengths used in single-photon confocal microscopy [107]. One limitation to two-photon microscopy is that the spatial resolution is lower than that of confocal microscopy. Since the resolution of microscopy systems is inversely proportional to the wavelength, the resolution of a two-photon microscope is thus twice lower than the same fluorophore excited by a single photon [107]. FUS-mediated increases in BBB permeability is detected in two-photon microscopy by the leakage of fluorescent dextrans from blood vessels into the extravascular space. Image analysis can then be performed to investigate factors such as the kinetics of different leakage types [108, 109], and differences in BBB perme- ability between disease models and wild-type animals [110].

7.1 Preparation Single element ring transducers are used for two-photon micros- of Transducers for FUS copy experiments such that the objective lens can be positioned in the center of the transducer. The geometric focus of the transducer should be aligned with the focus of the objective lens. Specifications on the fabrication and characterization of ring transducers used in our experiments can be found in Dr. Tam Nhan’s thesis [109]. Briefly, the transducers are composed of PZT-4 and have a diameter of 10 mm, thickness of 1.5 mm, height of 1.1 mm, and are driven at 1.1–1.2 MHz. The focus of the transducer is within 1 mm under the coverslip, which corresponds to the maximum imaging depths of current two-photon microscopy systems [109, 111]. To ensure that there are no air bubbles between the transducer and the brain, and to facilitate application of the transducer onto the brain, it is advisable to attach the ring transducer to a glass coverslip. This reduces torsion of the wires attached and stabilizes the transducer. Typically, ring transducers are adhered to a 12 mm cover glass using cyanoacrylate glue. Ideally, there should be no air bubbles between the transducer and cover glass in the glue. The glue should be allowed to cure for 24 h. The transducer can be coupled to the brain surface by using a drop of saline or 1% agarose. Both compounds reduce the presence of air bubbles, thereby facilitating effective ultrasound propagation to the brain. Typically used FUS parameters for our ring transducers are: 1.1–1.2 MHz driving frequency, 120 s sonication duration, 10 ms pulse length, 1 Hz pulse repetition frequency, estimated in situ PNP range of 0.4–0.8 MPa, and 20 μL/kg Definity MBs.

7.2 Animal In vivo two-photon microscopy of the brain requires gaining opti- Preparation: cal access to the brain by either removing a small piece of skull Implantation of Cranial (cranial window), or by thinning the skull. Cranial window pre- Windows parations are preferred in longitudinal studies that require animals to be reimaged over weeks or months, whereas skull thinning 284 Dallan McMahon et al.

experiments are limited by the time required for skull regrowth [112]. Although rethinning the skull is possible, the skull thinning procedure is technically difficult and requires the skull to be a precise thickness, balancing good optical clarity with minimal trauma to the brain [113]. In addition, optical properties of the skull deteriorates with each rethinning procedure [112]. Notably, cranial windows have been observed to affect dendritic spine dynamics [114] and gliosis [115], whereas skull thinning has been associated with microglial activation and dendritic blebbing [112]. However, the use of immunomodulators, such as dexameth- asone, and allowing animals to recover for 1–3 weeks after sterile cranial window procedures, can reduce inflammation before imag- ing begins [116, 117]. Finally, the skull can attenuate ultrasound propagation, such that in situ pressures may not reflect the pressure profiles of transducers measured in water. Thus, all the experiments conducted in our lab use cranial windows. Two-photon microscopy experiments for FUS studies can be acute or chronic. In acute experiments, the sonication and imaging occur on the same day that the cranial window is prepared. In chronic experiments, the cranial window is made 3 weeks in advance of sonication and imaging. This allows any neuroinflam- mation resulting from the cranial window preparation to reside, and therefore the results obtained will be closer to the naive state of the animal [116–118]. Chronic preparations also allow animals to be imaged repeatedly over the course of weeks. The following section details the methodology of both types of experiments. More detailed methods for similar preparations of chronic cranial win- dows can be found in other publications [116–118]. A list of equipment and materials to be autoclaved can be found in Supple- mentary Materials. Animals are first anesthetized in an induction chamber using a mix of medical air, oxygen, and isoflurane. A ketamine/xylazine cocktail can be used instead of isoflurane. In addition to the effects of different anesthetics and carrier gases on FUS studies mentioned in Sect. 5.4, some groups have reported that use of isoflurane is associated with a higher susceptibility to dural bleeding [116, 117]. Once a sufficient plane of anesthesia has been achieved, animals should be weighed; this is necessary to determine the dose of drugs and dyes used during the experiment. Various instru- ments, such as a rectal probe or a pulse oximeter sensor (e.g., MouseOx products from Starr Life Sciences Corp), can be used to monitor the animal’s physiological health during the experi- ment. For acute experiments, access to the systemic circulation must be established for the injection of dextrans and MBs, typically by catheterizing a tail vein. For chronic cranial windows, carprofen, a non-steroidal anti-inflammatory drug, and dexamethasone, a steroidal drug that counteracts edema, can be injected subcutane- ously for pain management and reduction of inflammation. FUS-Mediated Increase in BBB Permeability 285

One of the most common artifacts in two-photon microscopy studies results from breathing motions. To prevent this, it is impor- tant to ensure that the animal can breathe with ease. This can be achieved by ensuring that the head of the animal is level with the rest of the body: First, the incisors of the animal are hooked into the bite bar, then the angle and height of the nose cone and height of the platform that carries the body, can be adjusted. Lubricant is then applied to the eyes to prevent damage to the cornea from drying out or from irritants (e.g., depilatory cream). Although it is possible to cut and resuture the scalp over the cranial window, chronic imaging will require the skin to be reo- pened every experiment. This will cause discomfort to the animal and may also initiate an immune response which could confound results. Removing the scalp may also be necessary depending on the size of the transducer used, as skin folds under the transducer can hide the presence of unwanted air bubbles. Although this may seem invasive, proper application of a topical antibiotic cream on the edges of the removed skin post-surgery will aid the healing process. Prior to removing the scalp, it is necessary to remove the fur on the head of the animal. This will prevent fur from falling into the cranial window during surgery, which can lead to infection, and is also a necessary step for ultrasound propagation. To remove the fur, use small clippers, and then apply a depilatory cream to the head of the animal using a cotton swab. Leaving the cream on for too long can burn the epidermis, which will impair recovery. When the fur is sufficiently removed, the scalp should be cleaned with water and soap, ensuring that there is no dander or depilatory cream residue left on the skin of the animal. A local anesthetic such as lidocaine can be injected subcutaneously on the head of the animal prior to removing the scalp. For chronic cranial window surgeries, special care must be taken to prepare the surgery site in a sterile manner; detailed pro- tocols for chronic brain imaging have been published [116, 117]. To clean the scalp, wash with alternating wipes of betadine and 70% EtOH. To remove the scalp, lift up the skin between the eyes and the ears of the animal with a pair of forceps, and cut with surgical scissors longitudinally along the animal’s head such that the parietal bones are exposed. The periosteum must be removed. During the craniotomy procedure, saline can be used to cool the skull. To ensure that saline and bone dust do not accumu- late on the brain surface, a piece of autoclaved Kimwipe can be adhered to the edge of the skull with a drop of saline. Gelfoam is often used during craniotomy procedures to stop bleeding and to keep the brain surface moist. Gelfoam should be pre-soaked in saline; this will soften the Gelfoam and increase its absorptive capacity. The cranial windows used for FUS experiments are typically 3–4 mm in diameter, located between lambda and bregma, on one 286 Dallan McMahon et al.

of the parietal bones. If lambda and bregma cannot be located visually, gently pushing down on the parietal bones with a cotton swab may aid in locating these landmarks. The craniotomy proce- dure consists of an initial outline of the cranial window by lightly sanding the skull with the drill burr, then deepening the grooves, and finally removing the resulting bone island using a pair of fine surgical forceps. Throughout this process, it is important to cool the skull by intermittently dispensing drops of saline. If bleeding occurs, it must be stopped by applying pieces of wet Gelfoam before continuing the drilling. Blood will obscure the different layers of bone, which will hamper the surgeon’s ability to gauge the amount of bone removed. Depending on the genotype and strain of mouse, the skull may be heterogeneous in density and thickness. As the skull is thinned, pial vessels will become visible on the brain surface. Usually, small cracks in the thinnest areas of the skull are a good indication that drilling is almost complete. When the bone island depresses with gentle pressure from forceps, it is ready to be removed. Attempting to remove the bone island prematurely can cause parts of the bone island to dig into the brain, damaging the dura and causing inflammation and bleeding. Once the bone island has been removed, the brain must be kept moist. This can be done by applying drops of saline or wet Gelfoam. If there is slight bleeding, apply pieces of wet Gelfoam until the bleeding has stopped. If bleeding is excessive, it may be necessary to sacrifice the animal, especially for acute experiments, as a compro- mised BBB will allow fluorescent dextrans to leak into the paren- chyma and obscure the field-of-view. The coverslip can be adhered to the skull using cyanoacrylate glue. A sterile surgical spatula can be used to spread the glue evenly around the perimeter of the coverslip.

7.2.1 Acute Cranial In acute experiments, there are two options for what is directly Window Preparations contacting the brain surface: (1) A small cover glass (5 or 8 mm) can be placed directly on top of the brain, or (2) the glass coverslip that is glued onto the transducer can be placed directly on top of the brain. The choice will depend on the size of the transducer, and the size of the cranial window. If the surface area of the transducer is much larger than the cranial window, it is advisable to cover the brain surface with a 5 mm coverslip first, and then to couple that to the transducer’s coverslip using saline or 1% agarose, and cyanoac- rylate glue (Fig. 5). The two main concerns are whether the trans- ducer is sufficiently coupled to the brain (no air bubbles between the transducer and brain), and whether the geometric focus of the transducer is within the imaging window. FUS-Mediated Increase in BBB Permeability 287

Fig. 5 Coupling of transducer with cranial window. The ring transducer has been coupled to the cranial window with a drop of saline. The connection is sealed with 1% agarose, which holds the two together as it solidifies. Cyanoacrylate glue may be applied on the transducer’s cover glass and agarose in order to stabilize the transducer. Note the extent of fur removed from the scruff of the mouse

Fig. 6 Chronic cranial window preparations. In chronic preparations, the cranial window may become “cloudy” in the days following surgery (a). This may indicate a compromise in sterility during the surgery, leading to infection, or inflammation. To ensure clarity of the window for optimal imaging, accurate neurophysiology, and good health of the animal, these animals should be excluded from the study. (b) A 5 mm cover glass covers the cranial window; pial vessels are clearly observed. Cyanoacrylate glue covers the perimeter of the cover glass and extends liberally over the skull. Small bleeds can be seen, particularly around the top right. If blood does not clear in the days following surgery, the animal should be excluded from the study

7.2.2 Chronic Cranial Once the cyanoacrylate glue is completely dry, the surgery is essen- Window Preparations tially done (Fig. 6). Postsurgical care consists of subcutaneous injection of prophylactic antibiotics, ketoprofen and carprofen for management of pain and inflammation, saline if the animal is dehy- drated, and a generous application of topical antibiotic cream (e.g., neomycin) on the edges of the cut scalp. Animals should be 288 Dallan McMahon et al.

recovered in their home cage with a heat lamp. Allow the animal to recover for 2–3 weeks to allow inflammation to subside [117, 118]. Soft foods may help encourage animals to eat.

7.3 Fluorescent To visualize FUS effects on the BBB, large-molecular weight dex- Dextrans, Imaging trans can be used. Common examples include Texas Red 70 kDa Parameters, (Invitrogen, Burlington, ON, Canada), and FITC 70 kDa. Note and Sonication that we have observed dextrans of lower molecular weight to leak Workflow out of the blood vessels, free of FUS exposure, during long experi- ments. Typically, dextrans are purchased in powder form, and can be solubilized in PBS. Solubilized dextrans can be stored in the freezer according to the producer’s specifications. Care should be taken to vortex the thawed dextran fully dissolve large pieces of dye, which may block blood vessels. A volume appropriate to the ani- mal’s weight is injected into the systemic circulation, typically via a tail vein catheter or retro-orbital injections. Tail vein injections are preferable as the catheter can be made long enough to extend outside of the microscope’s enclosure. This allows users to injection dextrans and MBs while the microscope is acquiring images. A XYZT depth stack is set up by finding a region-of-interest that includes the vessels-of-interest (penetrating arterioles, diving venules, capillaries, etc.), and then setting up the appropriate HV levels at different depths according to the chosen step-size. Typical imaging parameters for BBB experiments are 300–800 um in depth, 5–10 μm step-size, and 10–20 T-stacks (Fig. 7); these

Fig. 7 Two-photon imaging of cerebral vasculature. This maximum projection image shows a depth stack 330 μm in volume, and 507.9 μm  507.9 μm in the XY plane. Capillaries and surface blood vessels can be seen (Objective lens: XLPLN 25X W NA 1.05, dye: Texas Red 70 kDa, excitation wavelength: 890 nm, Z step-size: 5 μm) FUS-Mediated Increase in BBB Permeability 289

parameters can be adjusted according to the goal of the experiment. Care should be taken to avoid photobleaching. Since the effects of FUS on BBB permeability will be visualized as dye leaking into the extravascular space, saturation of signal may occur, so it is advisable to set the HV to a level that will allow signal intensity changes to be fully captured. The objective should be chosen such that it offers sufficient magnification and resolution, and fits within the inner diameter of the transducer. Typically, 25–40 objectives are used. Excitation parameters should be determined according to the dex- tran used (e.g., ~810–900 nm excitation wavelength for Texas Red). Once one T-stack has been completed, MBs (Definity, Lantheus Medical Imaging, Billerica, MA, USA), diluted 1:10 v/ v in saline, are slowly injected through the tail vein (with the door of the microscope closed) using a 1 mL syringe (0.02 mL/kg) [111]. This is then followed by a saline flush that will push all the MBs out of the tail vein catheter and into the mouse. Injecting slowly will help protect the integrity of the MBs. Sonication can begin during the saline flush. The forward and reflected RF-power values are recorded. This process occurs while the microscope is acquiring images.

7.4 Image Analysis Like all microscopy, image processing for two-photon microscopy depends on the focus of the study. To evaluate kinetics of dye leakage as a representative model for drug delivery, the signal intensity between the intra- and extravascular spaces can be evalu- ated [111]. Olympus Fluoview is often used in two-photon imag- ing to collect data, and can also be used for basic image processing, such as creating maximum projection images, measuring the length of objects (e.g., cells), and measuring distance between fluorescent objects (e.g., β-amyloid plaques to blood vessels). ImageJ/FIJI is another image-processing tool to use due to its open-source nature and compatibility with MATLAB. The BioFor- mats toolbox has several useful plugins for image processing of volume stacks. For batch processing, a simple MATLAB script can be written through the MIJI interface in order to automate image processing. Images may need to be prescreened to ensure that breathing artifacts do not affect measurements. Deconvolution software (e.g., AutoQuant) may also be helpful.

8 Conclusion

The BBB is the single largest impediment to the delivery of thera- peutic agents to the brain [1]. Significant investments and advances in our understanding of various neuropathologies have by and large failed to result in effective treatments. While a variety of factors have surely contributed to this slow progress, it is clear that there is a 290 Dallan McMahon et al.

need for new strategies. In providing an avenue for the noninvasive, targeted delivery of therapeutic agents to the brain, the use of FUS to increase BBB permeability presents the opportunity to rethink our approach to treating neuropathologies. The use of FUS to increase BBB permeability has great poten- tial to significantly impact strategies for treating a variety of diseases of the CNS and the transition to clinical testing is already under- way; however, there remains considerable preclinical work to be done to determine the driving mechanisms behind the effects of FUS on BBB permeability, characterizing any long-term effects of repeated treatment, fully assessing the biological events which fol- low FUS, among other. The combination of in vivo two-photon imaging and FUS continues to provide valuable information regarding the effects of oscillating MBs on the vasculature of living animals. MRI-guided FUS is currently the most flexible platform for assessing the success of sonication in long-term survival and repeated treatment studies. It is our hope that the procedures and considerations discussed here provide a starting point to begin designing experiments or refining existing protocols. The continued advancement of preclinical FUS research is important for future clinical translation and has great potential for discovery and development of novel applications.

Acknowledgements

We gratefully acknowledge our funding sources: Canadian Insti- tutes of Health Research (MOP 119312) and the National Insti- tutes of Health (R01 EB003268). The authors would also like to thank Marc Santos for his MATLAB script used to analyze gadolin- ium contrast enhancement, Shawna Rideout-Gros for her expertise with the cranial windows, and Marcelline Ramcharan for her administrative assistance.

Supplementary Materials

Equipment:

l Surgical scissors* l Surgical forceps* l Fine surgical forceps* l Cotton swabs* l Fibreless swabs* l Gelfoam l Stereotax for rodents FUS-Mediated Increase in BBB Permeability 291

– Bite bar – Ear bars l Warming pad l Dental drill l Micro drill burr* (0.5 mm, Fine Science Tools) l Circular cover glass (5, 8, or 12 mm diameter, #1 thickness, Warner Instruments) l Drapes* l Kimpwipes* l Small glass beaker or Petri dish* l Spatula* l Dissection microscope l 1 mL syringe l 5 mL syringe * Autoclave (or use new) if doing sterile surgeries for a chronic cranial window. Chemicals:

l Depilatory cream (sensitive) l Baby shampoo l Warm water l Agarose Drugs:

l Isoflurane l Ketoprofen (5 mg/kg) l Carprofen (5 mg/kg) l Dexamethasone (0.2 mg/kg) l 1% lidocaine l Lactated Ringer’s solution, or saline l Eye lubrication l Topical antibiotic cream l Prophylactic antibiotics

References

1. Pardridge WM (2005) The blood-brain bar- 3. Nagy Z, Pappius HM, Mathieson G, Hu¨tt- rier: bottleneck in brain drug development. ner I (1979) Opening of tight junctions in NeuroRx 2:3–14 cerebral endothelium. I. Effect of hyperos- 2. Kristensson K, Olsson Y (1971) Uptake of molar mannitol infused through the inter- exogenous proteins in mouse olfactory cells. nal carotid artery. J Comp Neurol 185: Acta Neuropathol 19:145–154 569–578 292 Dallan McMahon et al.

4. Tabatabaei SN, Girouard H, Carret A-S, Mar- 15. Mei J, Cheng Y, Song Y, Yang Y, Wang F, tel S (2015) Remote control of the permeabil- Liu Y, Wang Z (2009) Experimental study ity of the blood-brain barrier by magnetic on targeted methotrexate delivery to the rab- heating of nanoparticles: a proof of concept bit brain via magnetic resonance imaging- for brain drug delivery. J Control Release guided focused ultrasound. J Ultrasound 206:49–57 Med 28:871–880 5. Kiyatkin EA, Sharma HS (2009) Permeability 16. Fan C-H, Ting C-Y, Liu H-L, Huang C-Y, of the blood-brain barrier depends on brain Hsieh H-Y, Yen T-C, Wei K-C, Yeh C-K temperature. Neuroscience 161:926–939 (2013) Antiangiogenic-targeting drug- 6. Man S-M, Ma Y-R, Shang D-S, Zhao W-D, loaded microbubbles combined with focused Li B, Guo D-W, Fang W-G, Zhu L, Chen Y-H ultrasound for glioma treatment. Biomaterials (2007/4) Peripheral T cells overexpress 34:2142–2155 MIP-1α to enhance its transendothelial 17. The´venot E, Jorda˜o JF, O’Reilly MA, migration in Alzheimer’s disease. Neurobiol Markham K, Weng Y-Q, Foust KD, Kaspar Aging 28:485–496 BK, Hynynen K, Aubert I (2012) Targeted 7. Nakamuta S, Endo H, Higashi Y, Kousaka A, delivery of self-complementary adeno-asso- Yamada H, Yano M, Kido H (2008) Human ciated virus serotype 9 to the brain, using immunodeficiency virus type 1 gp120- magnetic resonance imaging-guided focused mediated disruption of tight junction proteins ultrasound. Hum Gene Ther 23:1144–1155 by induction of proteasome-mediated degra- 18. Hsu P-H, Wei K-C, Huang C-Y, Wen C-J, dation of zonula occludens-1 and -2 in human Yen T-C, Liu C-L, Lin Y-T, Chen J-C, Shen brain microvascular endothelial cells. J Neu- C-R, Liu H-L (2013) Noninvasive and tar- rovirol 14:186–195 geted gene delivery into the brain using 8. Karyekar CS, Fasano A, Raje S, Lu R, Dowl- microbubble-facilitated focused ultrasound. ing TC, Eddington ND (2003) Zonula occlu- PLoS One 8:e57682 dens toxin increases the permeability of 19. Burgess A, Huang Y, Querbes W, Sah DW, molecular weight markers and chemothera- Hynynen K (2012) Focused ultrasound for peutic agents across the bovine brain micro- targeted delivery of siRNA and efficient vessel endothelial cells. J Pharm Sci knockdown of Htt expression. J Control 92:414–423 Release 163:125–129 9. Schilling L, Wahl M (1994) Opening of the 20. Wang S, Olumolade OO, Sun T, Samiotaki G, blood-brain barrier during cortical superfu- Konofagou EE (2015) Noninvasive, neuron- sion with histamine. Brain Res 653:289–296 specific gene therapy can be facilitated by 10. Tuladhar A, Morshead CM, Shoichet MS focused ultrasound and recombinant adeno- (2015) Circumventing the blood–brain bar- associated virus. Gene Ther 22:104–110 rier: local delivery of cyclosporin A stimulates 21. Jorda˜o JF, Ayala-Grosso CA, Markham K, stem cells in stroke-injured rat brain. J Con- Huang Y, Chopra R, McLaurin J, trol Release 215:1–11 Hynynen K, Aubert I (2010) Antibodies tar- 11. Hynynen K, McDannold N, Vykhodtseva N, geted to the brain with image-guided focused Jolesz FA (2001) Noninvasive MR imaging- ultrasound reduces amyloid-beta plaque load guided focal opening of the blood-brain bar- in the TgCRND8 mouse model of Alzhei- rier in rabbits. Radiology 220:640–646 mer’s disease. PLoS One 5:e10549 12. Treat LH, Nathan M, Natalia V, Yongzhi Z, 22. Kinoshita M, McDannold N, Jolesz FA, Karen T, Kullervo H (2007) Targeted delivery Hynynen K (2006) Targeted delivery of anti- of doxorubicin to the rat brain at therapeutic bodies through the blood-brain barrier by levels using MRI-guided focused ultrasound. MRI-guided focused ultrasound. Biochem Int J Cancer 121:901–907 Biophys Res Commun 340:1085–1090 13. Treat LH, McDannold N, Zhang Y, 23. Kinoshita M, McDannold N, Jolesz FA, Vykhodtseva N, Hynynen K (2012) Improved Hynynen K (2006) Noninvasive localized anti-tumor effect of liposomal doxorubicin delivery of Herceptin to the mouse brain by after targeted blood-brain barrier disruption MRI-guided focused ultrasound-induced by MRI-guided focused ultrasound in rat gli- blood-brain barrier disruption. Proc Natl oma. Ultrasound Med Biol 38:1716–1725 Acad Sci U S A 103:11719–11723 14. Kovacs Z, Werner B, Rassi A, Sass JO, Martin- 24. Lipsman N, Schwartz ML, Huang Y, Lee L, Fiori E, Bernasconi M (2014) Prolonged sur- Sankar T, Chapman M, Hynynen K, Lozano vival upon ultrasound-enhanced doxorubicin AM (2013) MR-guided focused ultrasound delivery in two syngenic glioblastoma mouse thalamotomy for essential tremor: a proof- models. J Control Release 187:74–82 of-concept study. Lancet Neurol 12:462–468 FUS-Mediated Increase in BBB Permeability 293

25. Elias WJ, Lipsman N, Ondo WG, brain barrier with focused ultrasound: associ- Ghanouni P, Kim YG, Lee W, Schwartz M, ation with cavitation activity. Phys Med Biol Hynynen K, Lozano AM, Shah BB, Huss D, 51:793 Dallapiazza RF, Gwinn R, Witt J, Ro S, Eisen- 36. Sheikov N, McDannold N, Vykhodtseva N, berg HM, Fishman PS, Gandhi D, Halpern Jolesz F, Hynynen K (2004) Cellular mechan- CH, Chuang R, Butts Pauly K, Tierney TS, isms of the blood-brain barrier opening Hayes MT, Cosgrove GR, Yamaguchi T, induced by ultrasound in presence of micro- Abe K, Taira T, Chang JW (2016) A rando- bubbles. Ultrasound Med Biol 30:979–989 mized trial of focused ultrasound thalamot- 37. Sheikov N, McDannold N, Sharma S, Hyny- omy for essential tremor. N Engl J Med nen K (2008) Effect of focused ultrasound 375:730–739 applied with an ultrasound contrast agent on 26. Lipsman N, Meng Y, Bethune AJ, Huang Y, the tight junctional integrity of the brain Lam B, Masellis M, Herrmann N, Heyn C, microvascular endothelium. Ultrasound Med Aubert I, Boutet A, Smith GS, Hynynen K, Biol 34:1093–1104 Black SE (2018) Blood–brain barrier opening 38. Deng J, Jinmu D, Qin H, Feng W, in Alzheimer’s disease using MR-guided Yingjiang L, Zhibiao W, Zhigang W, focused ultrasound. Nat Commun 9:2336 Qingtao Z, Bo L, Yuan C (2011) The role of 27. Carpentier A, Canney M, Vignot A, Reina V, Caveolin-1 in blood–brain barrier disruption Beccaria K, Horodyckid C, Karachi C, induced by focused ultrasound combined Leclercq D, Lafon C, Chapelon J-Y, with microbubbles. J Mol Neurosci Capelle L, Cornu P, Sanson M, Hoang- 46:677–687 Xuan K, Delattre J-Y, Idbaih A (2016) Clini- 39. Sun T, Zhang Y, Power C, Alexander PM, cal trial of blood-brain barrier disruption by Sutton JT, Aryal M, Vykhodtseva N, Miller pulsed ultrasound. Sci Transl Med 8:343re2 EL, McDannold NJ (2017) Closed-loop con- 28. Vykhodtseva NI, Hynynen K, Damianou C trol of targeted ultrasound drug delivery (1995) Histologic effects of high intensity across the blood-brain/tumor barriers in a pulsed ultrasound exposure with subharmo- rat glioma model. Proc Natl Acad Sci U S A nic emission in rabbit brain in vivo. Ultra- 114:E10281–E10290 sound Med Biol 21:969–979 40. Park J, Aryal M, Vykhodtseva N, Zhang Y-Z, 29. Hosseinkhah N, Nazanin H, Goertz DE, Kul- McDannold N (2017) Evaluation of perme- lervo H (2015) Microbubbles and blood–- ability, doxorubicin delivery, and drug reten- brain barrier opening: a numerical study on tion in a rat brain tumor model after acoustic emissions and wall stress predictions. ultrasound-induced blood-tumor barrier dis- IEEE Trans Biomed Eng 62:1293–1304 ruption. J Control Release 250:77–85 30. Lewin PA, Bjo L (1982) Acoustically induced 41. Liu H-L, Hua M-Y, Chen P-Y, Chu P-C, Pan shear stresses in the vicinity of microbubbles C-H, Yang H-W, Huang C-Y, Wang J-J, Yen in tissue. J Acoust Soc Am 71:728–734 T-C, Wei K-C (2010) Blood-brain barrier dis- 31. Wu J (2002) Theoretical study on shear stress ruption with focused ultrasound enhances generated by microstreaming surrounding delivery of chemotherapeutic drugs for glio- contrast agents attached to living cells. Ultra- blastoma treatment. Radiology 255:415–425 sound Med Biol 28:125–129 42. Yang F-Y, Wang H-E, Liu R-S, Teng M-C, Li 32. Dayton P, Klibanov A, Brandenburger G, Fer- J-J, Lu M, Wei M-C, Wong T-T (2012) Phar- rara K (1999) Acoustic radiation force in vivo: macokinetic analysis of 111In-labeled liposo- a mechanism to assist targeting of microbub- mal doxorubicin in murine glioblastoma after bles. Ultrasound Med Biol 25:1195–1201 blood-brain barrier disruption by focused 33. Apfel RE, Holland CK (1991) Gauging the ultrasound. PLoS One 7:e45468 likelihood of cavitation from short-pulse, 43. Szablowski JO, Lee-Gosselin A, Lue B, low-duty cycle diagnostic ultrasound. Ultra- Malounda D, Shapiro MG (2018) Acousti- sound Med Biol 17:179–185 cally targeted chemogenetics for the 34. Hynynen K, McDannold N, Vykhodtseva N, non-invasive control of neural circuits. Nat Raymond S, Weissleder R, Jolesz FA, Sheikov Biomed Eng 2:475–484 N (2006) Focal disruption of the blood-brain 44. Stavarache MA, Petersen N, Jurgens EM, Mil- barrier due to 260-kHz ultrasound bursts: a stein ER, Rosenfeld ZB, Ballon DJ, Kaplitt method for molecular imaging and targeted MG (2018) Safe and stable noninvasive focal drug delivery. J Neurosurg 105:445–454 gene delivery to the mammalian brain follow- 35. McDannold N, Vykhodtseva N, Hynynen K ing focused ultrasound. J Neurosurg:1–10. (2006) Targeted disruption of the blood – https://doi.org/10.3171/2017.8. JNS17790 294 Dallan McMahon et al.

45. Song K-H, Fan AC, Hinkle JJ, Newman J, 55. O’Reilly MA, Jones RM, Barrett E, Schwab A, Borden MA, Harvey BK (2017) Microbubble Head E Investigation of the safety of focused gas volume: a unifying dose parameter in ultrasound-induced blood-brain barrier blood-brain barrier opening by focused ultra- opening in a natural canine model of aging. sound. Theranostics 7:144–152 Theranostics. https://doi.org/10.7150/ 46. Alkins R, Burgess A, Ganguly M, Francia G, thno.20621 Kerbel R, Wels WS, Hynynen K (2013) 56. Scarcelli T, Jorda˜o JF, O’Reilly MA, Ellens N, Focused ultrasound delivers targeted immune Hynynen K, Aubert I (2014) Stimulation of cells to metastatic brain tumors. Cancer Res hippocampal neurogenesis by transcranial 73:1892–1899 focused ultrasound and microbubbles in 47. Burgess A, Alison B, Ayala-Grosso CA, adult mice. Brain Stimul 7:304–307 Milan G, Jorda˜o JF, Isabelle A, Kullervo H 57. Mooney SJ, Shah K, Yeung S, Burgess A, (2011) Targeted delivery of neural stem cells Aubert I, Hynynen K (2016) Focused to the brain using MRI-guided focused ultra- ultrasound-induced neurogenesis requires an sound to disrupt the blood-brain barrier. increase in blood-brain barrier permeability. PLoS One 6:e27877 PLoS One 11:e0159892 48. Ting C-Y, Fan C-H, Liu H-L, Huang C-Y, 58. Bremner JD, Narayan M, Anderson ER, Staib Hsieh H-Y, Yen T-C, Wei K-C, Yeh C-K LH, Miller HL, Charney DS (2000) Hippo- (2012) Concurrent blood-brain barrier open- campal volume reduction in major depression. ing and local drug delivery using drug- Am J Psychiatry 157:115–118 carrying microbubbles and focused ultra- 59. MacQueen GM, Campbell S, McEwen BS, sound for brain glioma treatment. Biomater- Macdonald K, Amano S, Joffe RT, ials 33:704–712 Nahmias C, Young LT (2003) Course of ill- 49. Baseri B, Choi JJ, Deffieux T, Samiotaki G, ness, hippocampal function, and hippocampal Tung Y-S, Olumolade O, Small SA, volume in major depression. Proc Natl Acad Morrison B, Konofagou EE (2012) Activa- Sci U S A 100:1387–1392 tion of signaling pathways following localized 60. Neumeister A, Wood S, Bonne O, Nugent delivery of systemically administered neuro- AC, Luckenbaugh DA, Young T, Bain EE, trophic factors across the blood-brain barrier Charney DS, Drevets WC (2005) Reduced using focused ultrasound and microbubbles. hippocampal volume in unmedicated, remit- Phys Med Biol 57:N65–N81 ted patients with major depression versus con- 50. Jorda˜o JF, The´venot E, Markham-Coultes K, trol subjects. Biol Psychiatry 57:935–937 Scarcelli T, Weng Y-Q, Xhima K, O’Reilly M, 61. Sheline YI, Sanghavi M, Mintun MA, Gado Huang Y, McLaurin J, Hynynen K, Aubert I MH (1999) Depression duration but not age (2013) Amyloid-β plaque reduction, endoge- predicts hippocampal volume loss in medically nous antibody delivery and glial activation by healthy women with recurrent major depres- brain-targeted, transcranial focused ultra- sion. J Neurosci 19:5034–5043 sound. Exp Neurol 248:16–29 62. Kovacs ZI, Kim S, Jikaria N, Qureshi F, 51. Burgess A, Dubey S, Yeung S, Hough O, Milo B, Lewis BK, Bresler M, Burks SR, Eterman N, Aubert I, Hynynen K (2014) Frank JA (2017) Disrupting the blood-brain Alzheimer disease in a mouse model: MR barrier by focused ultrasound induces sterile imaging-guided focused ultrasound targeted inflammation. Proc Natl Acad Sci U S A 114: to the hippocampus opens the blood-brain E75–E84 barrier and improves pathologic abnormalities 63. Silburt J, Lipsman N, Aubert I (2017) Dis- and behavior. Radiology 273:736–745 rupting the blood-brain barrier with focused 52. Leinenga G, Go¨tz J (2015) Scanning ultra- ultrasound: perspectives on inflammation and sound removes amyloid-β and restores mem- regeneration. Proc Natl Acad Sci U S ory in an Alzheimer’s disease mouse model. A. https://doi.org/10.1073/pnas. Sci Transl Med 7:278ra33 1710761114 53. Leinenga G, Go¨tz J (2018) Safety and efficacy 64. Kovacs ZI, Burks SR, Frank JA (2017) Reply of scanning ultrasound treatment of aged to Silburt et al.: concerning sterile inflamma- APP23 mice. Front Neurosci 12:55 tion following focused ultrasound and micro- 54. Nisbet RM, Van der Jeugd A, Leinenga G, bubbles in the brain. Proc Natl Acad Sci U S Evans HT, Janowicz PW, Go¨tz J (2017) A. https://doi.org/10.1073/pnas. Combined effects of scanning ultrasound 1711544114 and a tau-specific single chain antibody in a 65. McMahon D, Hynynen K (2017) Acute tau transgenic mouse model. Brain inflammatory response following increased 140:1220–1230 FUS-Mediated Increase in BBB Permeability 295

blood-brain barrier permeability induced by 75. McDannold N, Arvanitis CD, focused ultrasound is dependent on micro- Vykhodtseva N, Livingstone MS (2012) Tem- bubble dose. Theranostics 7:3989–4000 porary disruption of the blood-brain barrier 66. Kovacs ZI, Burks SR, Frank JA (2018) by use of ultrasound and microbubbles: safety Focused ultrasound with microbubbles and efficacy evaluation in rhesus macaques. induces sterile inflammatory response propor- Cancer Res 72:3652–3663 tional to the blood brain barrier opening: 76. Mooney SJ, Nobrega JN, Levitt AJ, Hynynen attention to experimental conditions. Thera- K (2018) Antidepressant effects of focused nostics 8:2245–2248 ultrasound induced blood-brain-barrier 67. McMahon D, Hynynen K (2018) Reply to opening. Behav Brain Res 342:57–61 Kovacs et al.: concerning acute inflammatory 77. McDannold N, Zhang Y, Vykhodtseva N response following focused ultrasound and (2017) The effects of oxygen on ultrasound- microbubbles in the brain. Theranostics induced blood-brain barrier disruption in 8:2249–2250 mice. Ultrasound Med Biol 43:469–475 68. McMahon D, Bendayan R, Hynynen K 78. Baseri B, Choi JJ, Tung Y-S, Konofagou EE (2017) Acute effects of focused ultrasound- (2010) Multi-modality safety assessment of induced increases in blood-brain barrier per- blood-brain barrier opening using focused meability on rat microvascular transcriptome. ultrasound and definity microbubbles: a Sci Rep 7:45657 short-term study. Ultrasound Med Biol 69. O’Reilly MA, Hough O, Hynynen K (2017) 36:1445–1459 Blood-brain barrier closure time after con- 79. McDannold N, Vykhodtseva N, Raymond S, trolled ultrasound-induced opening is inde- Jolesz FA, Hynynen K (2005) MRI-guided pendent of opening volume. J Ultrasound targeted blood-brain barrier disruption with Med 36:475–483 focused ultrasound: histological findings in 70. Downs ME, Buch A, Sierra C, Karakatsani rabbits. Ultrasound Med Biol 31: ME, Chen S, Konofagou EE, Ferrera VP 1527–1537 (2015) Long-term safety of repeated blood- 80. Hynynen K, McDannold N, Sheikov NA, brain barrier opening via focused ultrasound Jolesz FA, Vykhodtseva N (2005) Local and with microbubbles in non-human primates reversible blood-brain barrier disruption by performing a cognitive task. PLoS One 10: noninvasive focused ultrasound at frequencies e0125911 suitable for trans-skull sonications. Neuro- 71. Kobus T, Vykhodtseva N, Pilatou M, Image 24:12–20 Zhang Y, McDannold N (2016) Safety valida- 81. Hynynen K, Jolesz FA (1998) Demonstration tion of repeated blood-brain barrier disrup- of potential noninvasive ultrasound brain tion using focused ultrasound. Ultrasound therapy through an intact skull. Ultrasound Med Biol 42:481–492 Med Biol 24:275–283 72. Jones RM, Deng L, Leung K, McMahon D, 82. Sun J, Hynynen K (1999) The potential of O’Reilly MA, Hynynen K (2018) Three- transskull ultrasound therapy and surgery dimensional transcranial microbubble imag- using the maximum available skull surface ing for guiding volumetric ultrasound- area. J Acoust Soc Am 105:2519–2527 mediated blood-brain barrier opening. Ther- 83. Pajek D, Hynynen K (2013) The application anostics 8:2909–2926 of sparse arrays in high frequency transcranial 73. Yang F-Y, Chang W-Y, Chen J-C, Lee L-C, focused ultrasound therapy: a simulation Hung Y-S (2014) Quantitative assessment of study. Med Phys 40:122901 cerebral glucose metabolic rates after blood- 84. Hwang JH, Tu J, Brayman AA, Matula TJ, brain barrier disruption induced by focused Crum LA (2006) Correlation between inertial ultrasound using FDG-MicroPET. Neuro- cavitation dose and endothelial cell damage Image 90:93–98 in vivo. Ultrasound Med Biol 32:1611–1619 74. Horodyckid C, Canney M, Vignot A, 85. O’Reilly MA, Hynynen K (2012) Blood-brain Boisgard R, Drier A, Huberfeld G, barrier: real-time feedback-controlled focused Franc¸ois C, Prigent A, Santin MD, Adam C, ultrasound disruption by using an acoustic Willer J-C, Lafon C, Chapelon J-Y, Carpen- emissions-based controller. Radiology tier A (2016) Safe long-term repeated disrup- 263:96–106 tion of the blood-brain barrier using an 86. McDannold N, Vykhodtseva N, Hynynen K implantable ultrasound device: a multipara- (2008) Blood-brain barrier disruption metric study in a primate model. J induced by focused ultrasound and circulating Neurosurg:1–11. https://doi.org/10.3171/ preformed microbubbles appears to be 2016.3.JNS151635 296 Dallan McMahon et al.

characterized by the mechanical index. Ultra- 97. Wu S-K, Chu P-C, Chai W-Y, Kang S-T, Tsai sound Med Biol 34:834–840 C-H, Fan C-H, Yeh C-K, Liu H-L (2017) 87. Goertz DE, Wright C, Hynynen K (2010) Characterization of different microbubbles in Contrast agent kinetics in the rabbit brain assisting focused ultrasound-induced blood- during exposure to therapeutic ultrasound. brain barrier opening. Sci Rep 7:46689 Ultrasound Med Biol 36:916–924 98. Vlachos F, Tung Y-S, Konofagou E (2011) 88. Stapleton S, Goodman H, Zhou Y-Q, Permeability dependence study of the focused Cherin E, Henkelman RM, Burns PN, Foster ultrasound-induced blood–brain barrier FS (2009) Acoustic and kinetic behaviour of opening at distinct pressures and microbubble definity in mice exposed to high frequency diameters using DCE-MRI. Magn Reson ultrasound. Ultrasound Med Biol Med 66:821–830 35:296–307 99. Park J, Zhang Y, Vykhodtseva N, Jolesz FA, 89. McDannold N, Zhang Y, Vykhodtseva N McDannold NJ (2012) The kinetics of blood (2011) Blood-brain barrier disruption and brain barrier permeability and targeted doxo- vascular damage induced by ultrasound bursts rubicin delivery into brain induced by focused combined with microbubbles can be influ- ultrasound. J Control Release 162:134–142 enced by choice of anesthesia protocol. Ultra- 100. Kaya M, Ahishali B (2011) Assessment of sound Med Biol 37:1259–1270 permeability in barrier type of endothelium 90. Itani M, Mattrey RF (2012) The effect of in brain using tracers: Evans blue, sodium inhaled gases on ultrasound contrast agent fluorescein, and horseradish peroxidase. longevity in vivo. Mol Imaging Biol 14: Methods Mol Biol 763:369–382 40–46 101. Yang F-Y, Fu W-M, Yang R-S, Liou H-C, 91. Mullin L, Gessner R, Kwan J, Kaya M, Borden Kang K-H, Lin W-L (2007) Quantitative MA, Dayton PA (2011) Effect of anesthesia evaluation of focused ultrasound with a con- carrier gas on in vivo circulation times of ultra- trast agent on blood-brain barrier disruption. sound microbubble contrast agents in rats. Ultrasound Med Biol 33:1421–1427 Contrast Media Mol Imaging 6:126–131 102. Raymond SB, Treat LH, Dewey JD, McDan- 92. Bing C, Ladouceur-Wodzak M, Wanner CR, nold NJ, Hynynen K, Bacskai BJ (2008) Shelton JM, Richardson JA, Chopra R (2014) Ultrasound enhanced delivery of molecular Trans-cranial opening of the blood-brain bar- imaging and therapeutic agents in Alzhei- rier in targeted regions using astereotaxic mer’s disease mouse models. PLoS One 3: brain atlas and focused ultrasound energy. J e2175 Ther Ultrasound 2:13 103. Arvanitis CD, Livingstone MS, 93. O’Reilly MA, Waspe AC, Ganguly M, Hyny- Vykhodtseva N, McDannold N (2012) Con- nen K (2011/4) Focused-ultrasound disrup- trolled ultrasound-induced blood-brain bar- tion of the blood-brain barrier using closely- rier disruption using passive acoustic timed short pulses: influence of sonication emissions monitoring. PLoS One 7:e45783 parameters and injection rate. Ultrasound 104. Squirrell JM, Wokosin DL, White JG, Bavister Med Biol 37:587–594 BD (1999) Long-term two-photon fluores- 94. Marty B, Larrat B, Van Landeghem M, cence imaging of mammalian embryos with- Robic C, Robert P, Port M, Le Bihan D, out compromising viability. Nat Biotechnol Pernot M, Tanter M, Lethimonnier F, Me´r- 17:763–767 iaux S (2012) Dynamic study of blood – brain 105. Helmchen F, Denk W (2005) Deep tissue barrier closure after its disruption using ultra- two-photon microscopy. Nat Methods sound: a quantitative analysis. J Cereb Blood 2:932–940 Flow Metab 32:1948–1958 106. Svoboda K, Yasuda R (2006) Principles of 95. Choi JJ, Pernot M, Small SA, Konofagou EE two-photon excitation microscopy and its appli- (2007) Noninvasive, transcranial and loca- cations to neuroscience. Neuron 50:823–839 lized opening of the blood-brain barrier 107. So PTC (2001) Two-photon fluorescence using focused ultrasound in mice. Ultrasound light microscopy. In: eLS. John Wiley & Med Biol 33:95–104 Sons, Ltd, Hoboken 96. Chai W-Y, Chu P-C, Tsai C-H, Lin C-Y, Yang 108. Cho EE, Drazic J, Ganguly M, Stefanovic B, H-W, Lai H-Y, Liu H-L (2018) Image- Hynynen K (2011) Two-photon fluorescence guided focused-ultrasound CNS molecular microscopy study of cerebrovascular dynamics delivery: an implementation via dynamic in ultrasound-induced blood-brain barrier contrast-enhanced magnetic-resonance imag- opening. J Cereb Blood Flow Metab ing. Sci Rep 8:4151 31:1852–1862 FUS-Mediated Increase in BBB Permeability 297

109. Nhan TQ (2015) Drug delivery to the brain 114. Xu H-T, Pan F, Yang G, Gan W-B (2007) by focused ultrasound and microbubble Choice of cranial window type for in vivo mediated blood-brain barrier disruption: imaging affects dendritic spine turnover in vascular-level investigation using two-photon the cortex. Nat Neurosci 10:549–551 fluorescent microscopy. University of 115. Yan P, Bero AW, Cirrito JR, Xiao Q, Hu X, Toronto Wang Y, Gonzales E, Holtzman DM, Lee J-M 110. Burgess A, Nhan T, Moffatt C, Klibanov AL, (2009) Characterizing the appearance and Hynynen K (2014) Analysis of focused growth of amyloid plaques in APP/PS1 ultrasound-induced blood-brain barrier per- mice. J Neurosci 29:10706–10714 meability in a mouse model of Alzheimer’s 116. Goldey GJ, Roumis DK, Glickfeld LL, Kerlin disease using two-photon microscopy. J Con- AM, Reid RC, Bonin V, Schafer DP, Ander- trol Release 192:243–248 mann ML (2014) Removable cranial windows 111. Nhan T, Burgess A, Cho EE, Stefanovic B, for long-term imaging in awake mice. Nat Lilge L, Hynynen K (2013) Drug delivery to Protoc 9:2515–2538 the brain by focused ultrasound induced 117. Holtmaat A, Bonhoeffer T, Chow DK, blood-brain barrier disruption: quantitative Chuckowree J, De Paola V, Hofer SB, evaluation of enhanced permeability of cere- Hu¨bener M, Keck T, Knott G, Lee W-CA, bral vasculature using two-photon micros- Mostany R, Mrsic-Flogel TD, Nedivi E, copy. J Control Release 172:274–280 Portera-Cailliau C, Svoboda K, Trachtenberg 112. Yang G, Pan F, Parkhurst CN, Grutzendler J, JT, Wilbrecht L (2009) Long-term, high-res- Gan W-B (2010) Thinned-skull cranial win- olution imaging in the mouse neocortex dow technique for long-term imaging of the through a chronic cranial window. Nat Protoc cortex in live mice. Nat Protoc 5:201–208 4:1128–1144 113. Dorand RD, Barkauskas DS, Evans TA, 118. Cao VY, Ye Y, Mastwal SS, Lovinger DM, Petrosiute A, Huang AY (2014) Comparison Costa RM, Wang KH (2013) In vivo of intravital thinned skull and cranial window two-photon imaging of experience- approaches to study CNS immunobiology in dependent molecular changes in cortical neu- the mouse cortex. Intravital 3. https://doi. rons. J Vis Exp. https://doi.org/10.3791/ org/10.4161/intv.29728 50148 Chapter 17

Evaluation of Blood–Brain Barrier Permeability and Integrity in Juvenile Rodents: Dynamic Contrast-Enhanced (DCE), Magnetic Resonance Imaging (MRI), and Evans Blue Extravasation

Trish Domi, Faraz Honarvar, and Andrea Kassner

Abstract

Focal ischemic stroke is the result of a blockage in an artery that leads to decreased blood flow to the neuronal cells in the brain. The middle cerebral artery (MCA) is the most common artery that is occluded in adult and pediatric stroke patients. The pathophysiology is challenging to study in either of these popula- tions because of the highly variable clinical state in humans. Many of these variables can be eliminated when using in vivo models of stroke in rodents. Here, we describe a technique called the transient MCA occlusion (tMCAo) model in a juvenile rat model of stroke. This technique utilizes a filament that is advanced to block the origin of the MCA to induce focal ischemia. The filament is then retracted 60–90 min later allowing for secondary reperfusion. By incorporating reperfusion, this model mimics embolic strokes in humans and provides the opportunity to uncover injury associated with reflow through ischemic tissue. We are particu- larly interested in the reperfusion-induced injury to the blood–brain barrier (BBB) that follows after blood flow to the ischemic brain is restored. Our goal is to provide the reader with guidelines on how to execute the tMCAo surgical procedure, with notes highlighting the advantages and limitations of the method. We also include directions on how to conduct the techniques used to evaluate the permeability of the blood–brain barrier including Evans blue extravasation, a histological procedure, and dynamic contrast-enhanced (DCE) magnetic resonance imag- ing (MRI), a technique used to evaluate blood–brain barrier permeability that can be applied to study stroke in a rodent model.

Key words BBB: blood–brain barrier, MCAo: middle cerebral artery occlusion, tMCAo: transient middle cerebral artery occlusion, stroke, MRI: magnetic resonance imaging, Evans Blue

Abbreviations

BBB Blood–brain barrier CCA Common carotid artery DCE Dynamic contrast-enhanced imaging EB Evans Blue ECA External carotid artery ICA Internal carotid artery

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_17, © Springer Science+Business Media, LLC, part of Springer Nature 2019 299 300 Trish Domi et al.

MCA Middle cerebral artery MCAo Middle cerebral artery occlusion RPM Revolutions per minute tMCAo Transient middle cerebral artery occlusion TTC Triphenyl tetrazolium chloride

1 Introduction

Ischemic stroke, which accounts for the majority of stroke occur- rences, is the result of an obstruction within a cerebral blood vessel, leading to a decrease in blood flow to regions of the brain. In pediatric and adult patients, occlusions of the middle cerebral artery (MCA) are among the most common causes of ischemic stroke [1, 2]. Traditionally focal ischemia models involved permanent occlu- sion of a major cerebral artery such as the MCA. However, since vessel occlusion is seldom permanent in human stroke, more recent developments have incorporated reperfusion (following ischemia) into the design of the animal model. The most widely used of these is the transient middle cerebral artery occlusion (tMCAo) method used to mimic the reperfusion injury that can occur following recanalization in human stroke patients [3]. The tMCAo model allows for manipulation of the length of ischemia and reproducible reperfusion with relative ease [4, 5]. The tMCAo model using an intraluminal suture is the most common technique used to occlude the MCA. In this model, a filament (also referred to as suture) is inserted into the MCA, interrupting the blood flow in this artery. This arrests blood flow nearby, including the lenticulo-striate arteries that supply the basal ganglia. By employing a silicone-coated filament in this procedure, subarachnoid hemorrhage and premature reperfusion are reduced [6], and a denser ischemia can be produced [7]. Filament occlusion of the MCA results in reproducible lesions in the cortex and stria- tum [1]. In addition, ischemic intervals can be varied in this model depending on the time point chosen for reperfusion and resulting in ischemic lesions with varying degrees of severity [1]. Further- more, the ability to study reperfusion provides the opportunity to assess secondary injury that can occur following ischemic stroke such as damage to the blood–brain barrier. Under normal physio- logical conditions, the blood–brain barrier plays an essential role in protecting the brain from pathogens. However, after ischemic stroke, the barrier is disrupted and this leads to further injury with the infiltration of leukocytes. In this chapter, we describe a procedure of modeling ischemia/ reperfusion injury in a juvenile rat model of stroke. This is followed by a description of the procedures used to assess/quantify the status Evaluation of Blood-Brain Barrier Permeability and Integrity in Juvenile... 301

of the blood–brain barrier after ischemic injury. These include in vivo dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) and histological techniques using Evans Blue extravasation.

2 Materials

2.1 Presurgical l Subject(s): Juvenile male Sprague-Dawley rats (Charles-River Procedure Laboratories, Sherbrooke, Canada) weighing approximately 150–200 g at time of surgery. l Isoflurane (avoid inhaling of gas, as isoflurane can be toxic). l Isoflurane chamber. l Heating blanket (Gaymar Medsearch, USA). l Surgical pads. l Surgical tape. l Nose cone (for constant isoflurane and oxygen supply). l Ophthalmic liquid gel. l Clippers (or hair removal gel). l 70% isopropyl. l Betadine (10% povidone-iodine).

2.2 Surgical Surgical tape Procedure l Magnifying surgical loupes (Roboz, magnification 4.0Â-340). l 0.9% saline. l Sterile surgical kit containing: – Scalpel blade (Swann Morton, USA). – 3 mm cutting edge Vannas scissors (Cedarlane, USA). – Ultrafine tweezers (Electron Microscopy Sciences, USA). – Microvascular clip (Harvard apparatus, USA). l 2 Â 5 mL syringes. l 7 Â 1 mL syringes. l 3 Â 30G needles. l 3 Â 3-0 silk sutures (Doccol, Sharon, MA, USA). l 4-0 vicryl suture (Doccol, Sharon, MA, USA). l Filament (silicon-coated suture) (Doccol, Sharon, MA, USA). l 0.03 mg/kg injections of Buprenorphine (0.33 mL diluted in 10 mL saline). 302 Trish Domi et al.

2.3 Behavioral Garcia test: a test of neurological outcome, assesses fore and hind Testing limb strength as well as sensory neglect (refer to Sect. 3.3 for a detailed explanation).

2.4 MRI Procedure l 3.0T clinical MRI system (Philips, Netherlands). l 8-channel wrist coil. l Isoflurane (avoid inhaling of gas, as isoflurane is toxic). l Isoflurane chamber. l Nose cone (for constant isoflurane and oxygen supply to the animal). l Oxygen supply. l Surgical pads (to be placed underneath the rat during scanning to protect scanner hardware). l Surgical tape. l Gadolinium-DTPA (Bayer Healthcare Pharmaceuticals Inc.) (do not ingest, as substance can be toxic in large amounts): used as a contrast dye for dynamic contrast-enhanced imaging. l Line (approximately 30 cm in length with a diameter equivalent to the thickness of a 25G needle). l Butterfly needle (27G). l Hamilton syringe (0.1 mL). l 0.9% saline. l 1 mL syringes. l 25G needles.

2.5 Sacrificing and l Guillotine (Kent Scientific, USA). Brain Extraction l Dissecting scissors 5 ½” Straight (Cedarlane, USA). Procedure l Bone cutting forceps (Roboz Surgical, USA). l Surgical spatula (The Medical Supply Group, USA). l Weighing paper sheets.

2.6 Histological l TTC stock powder (Sigma-Aldrich, USA). Procedures l Weighing paper sheets. 2.6.1 Triphenyl l Micro weighing scale (Scales Galore, USA). Tetrazolium Chloride (TTC) l Single edge razor blade (e.g., Titan tools, USA). Staining l 0.9% saline. l 50 mL tube. l 10 mL syringe. l Aluminum foil. Evaluation of Blood-Brain Barrier Permeability and Integrity in Juvenile... 303

l Brain matrix (Roboz surgical, USA). l Ice.

2.6.2 Evans Blue l EB stock powder (Sigma-Aldrich, USA) (Use caution when (EB) Staining handling. Be careful about spills). l P200 Pipette. l P1000 Pipette. l Pipette tips. l 1.5 mL microtube. l Single edge razor blade (e.g., Titan tools, USA). l Weighing paper sheets. l Micro weighing scale (Scales Galore, USA). l Brain matrix (Roboz surgical, USA). l Formaldehyde, 36.5–37% (do not smell or ingest, as substance is highly carcinogenic). l Fine point waterproof marker (Sharpie). l 50 mL centrifuge tube. l Microtube holder. l Microtube freezer box. l Heating block. l Aluminum foil. l Ultrasonic cell disruptor (Misonix sonicator 3000, Cole- Parmer, USA). l Task wipers (Kimberley-Klark Kimtech Kimwipes, USA). l Distilled water. l 70% Ethanol. l Spectrophotometer. l 96-well plate.

3 Methods

3.1 Presurgical Sprague Dawley male rats (Charles-River Laboratories, Sher- Procedure brooke, Canada) weighing approximately between 150 and 200 g were used in this study. Buprenorphine (0.03 mg/kg, stock con- centration of 0.3 mg/mL is diluted into 0.01 mg/mL with 0.9% saline) dosage needed toward to end of surgery is calculated based on weight: mass (kg) Â (0.03/0.0099). A heating pad is used to maintain body temperature between 36.5 and 38.0 C throughout the surgery. A surgical pad is placed onto the heating pad and secured by surgical tape. The rat is then placed in an isoflurane chamber and anesthetized using isoflurane (5% of isoflurane for 304 Trish Domi et al.

induction mixed with a flow rate of 1–2 L/minute of oxygen from Oxygen supply at 1.0 Bar 21 C). Flow of the gas mixture is then redirected to the nose cone and the rat is transferred to the cone, while the anesthesia is maintained with isoflurane at 1.5–2.5%. Ophthalmic liquid gel is applied to the eyes (to prevent drying during the procedure). Positioning the rat in a supine position, the neck is shaved using hair clippers or hair removal gel and the skin is prepared for surgery by disinfecting with 10% betadine (povidone-iodine) and then 70% isopropyl.

3.2 Surgical Focal ischemia is induced using the tMCAo model as described by Procedure Longa et al. [8], but with some modifications. A ventral midline incision is made using a scalpel blade, and the superficial fascia is 3.2.1 Middle Cerebral dissected using 3 mm cutting edge Vannas scissors and ultrafine Artery Occlusion Surgery tweezers. The sternohyoid, digastric, and sternomastid muscles are identified by blunt dissection and displaced to reveal the carotid artery. Further blunt dissection is used to identify the common carotid artery (CCA), internal carotid artery (ICA), external carotid artery (ECA), and vagus nerve. The common carotid artery (CCA) is isolated from the vagus nerve and ligated temporarily using a microvascular clip (Notes 1, 2, and 3 in Sect. 4.2)(magnifying surgical loupes can be used from this stage for identification of smaller vessels). The first bifurcation of the CCA, which leads to the external carotid artery (ECA) and internal carotid artery (ICA), is identified as shown in Fig. 1a. The soft tissue around the ECA and

ab ACA

MCA Ant. cerebral a.

Mid. cerebral a. AchA HTA Post. comm. a. Post. cerebral a.

Sup. cerebellar. a Int. carotid a. PCOM

Mastoid bulla Basilar a Lingual a. Pterygopal. a ICA Ext. max a. PPA Suture Sup. thyr. a.

Sternomastoid mus. Occip. a. ECA Ext. carotid a. Com carotid a.

Fig. 1 (a) Vascular brain anatomy of the rat. (b) The filament (suture) enters from the ECA, into the CCA, and then advanced until it reaches the MCA. Figures adapted from [8, 14] Evaluation of Blood-Brain Barrier Permeability and Integrity in Juvenile... 305

ICA is blunt-dissected without harming the arteries, and one knot is tied at the distal end of the ECA using a 3-0 silk suture. Next, the ICA and CCA are also temporarily ligated by positioning two other 3-0 silk sutures under the vessels (one suture per vessel) and lifting the sutures, thereby occluding the vessels. Microvascular clips can be used to keep the sutures in place during the lifting. The ECA is then cut beside the knot, leaving a vessel stump at the bifurcation. Next, a small opening is made at the stump section of the ECA using a 30G needle to insert the filament—a silicon- coated suture (Doccol, Sharon, MA, USA). The temporarily ligated section of the ICA is loosened to allow the suture to enter the ICA through the ECA stump (see Fig. 1b). The diameter of the filament is based on the data provided by the filament manufacturer (Doccol, Sharon, MA, USA), and adjusted according to the weight of the rat. For juvenile rats (weight 150–200 g), a 0.33-mm diam- eter filament with a silicone coating of 3–4 mm is usually a good choice. Once the tip of the filament reaches the ICA, a 3-0 silk suture is used to form a knot placed below the filament insertion in the ECA stump to ensure the filament stays in position. In order to carefully advance the filament (approximately by 17–20 mm) into the MCA, the ECA stump can be gently moved around to change the angle of the filament for less resistance (see Note 4 for further details). Once the filament is in place (Fig. 2), the 3-0 suture around the ECA stump is tightened, and the incision is closed with a 4-0 vicryl suture. After the surgical procedure, the filament remains in place for 60 to 90 min. During this period, the rat is transferred back to its cage and allowed to recover. After this period, the rat is anesthetized again to establish reperfusion. For this procedure, the previous incision is reopened. Before retracting the filament, the temporary ligation on the CCA and ICA is rein- stated, and the knot placed in the ECA stump below the insertion is loosened. The filament is then withdrawn carefully until the tip is near the insertion opening (Note 5). Next, the knot at the ECA stump is tightened immediately after the removal of the filament. After reperfusion is confirmed by removal of all ligation around CCA and ICA, the incision is sewn back up using a 4-0 vicryl suture. Immediately following the reperfusion procedure (depend- ing on the amount of blood loss), rats are given 3–7 mL of 0.9% saline using a 5 mL syringe subcutaneously to prevent dehydration. During the rat’s recovery period prior to reperfusion (but after completion of filament insertion), the rat’s neurological function is assessed using the Garcia test (described below in Sect. 3.3)[9]. 306 Trish Domi et al.

Fig. 2 Diagram of the Circle of Willis with placement of intraluminal occluding filament. 1 represents placement of filament for occlusion of the MCA, and retraction to 2 allows recirculation through the Circle of Willis. ACA anterior cerebral artery, ACOA anterior communicating artery, BA basilar artery, CCA common carotid artery, ECA external carotid artery, MCA middle cerebral artery, PCA posterior cerebral artery, PCOA posterior communicating artery, VA verte- bral artery. Figure adapted from [15]

3.2.2 Pain Relief Postoperative pain relief is provided with Buprenorphine Procedure (0.03 mg/kg, stock concentration of 0.3 mg/mL is diluted into 0.01 mg/mL with 0.9% saline) administered intraperitoneally using a 1 mL syringe. Buprenorphine is administered every 8 h for the first 24 h, and every 12 h thereafter for a total duration of 72 h. Upon completion of the surgery, the rats are disconnected from isoflurane, and returned to their cages to allow recovery.

3.3 Behavioral To test for neurological function, The Garcia test is administered Testing upon arousal. This test assesses fore and hind limb strength as well as sensory neglect [9], and is based on five factors, each being scored out of 3 with a maximum total score of 15 (minimum score: 0, indicating no activity; maximum score: 3, indicating max- imum activity):

l Spontaneous activity. l Symmetry in the movement of four limbs. l Forepaw outstretching. l Body proprioception. l Response to Vibrissae touch. Evaluation of Blood-Brain Barrier Permeability and Integrity in Juvenile... 307

Fig. 3 Rat (as indicated by the annotation) positioned in the wrist coil of an MR scanner

3.4 MRI Scan The MRI scans in our institution are performed on a 3.0T clinical Procedure MRI system (Philips, Netherlands) (see Note 3) equipped with an 8-channel wrist coil for clinical use. The 8-channel wrist coil is 3.4.1 Pre-scan positioned on the MRI table and the nose cone attached to the Preparation Isoflurane tank inside the MRI scan room is secured inside the coil, slightly in front of the center of localization (where the rat’s head would be positioned). The rat is then positioned in the Isoflurane chamber, and induction begins at 5% Isoflurane with 1–2 L/min flow of Oxygen from the Oxygen supply. Once the rat is fully anesthetized, airflow is redirected to the nose cone (Fig. 3), and the Isoflurane is reduced and maintained to 2% at a flow rate of 1 L/ min. For dynamic contrast-enhanced (DCE) MRI, a line is attached to a butterfly needle and is then filled with saline from a syringe (with a 25G needle) inserted into its other end. The butterfly needle is then inserted into the tail vein of the anesthetized rat, and secured into position along with the tube by surgical tape. The syringe containing saline is then removed, and a Hamilton syringe, filled with 0.1 mL of Gadolinium-DTPA (the contrast dye), is attached to the end of the line and secured with surgical tape. Immediately prior to the start of the DCE scan, the same quantity of Gadolinium is injected into the tube from the Hamilton syringe, thereby pushing the saline into the tail vein and replacing it. It is important that no Gadolinium is injected into the vein at this point, as it would enter the blood circulation prior to the DCE scan and render the scan results inaccurate. Once the DCE scan is started, baseline data will be acquired for approximately 25 s, followed by injection of the Gadolinium inside the Hamilton syringe (as well as the contrast inside the line) at a steady flow rate for 10–15 s. 308 Trish Domi et al.

3.4.2 MRI Acquisition The MRI sequences used are standard anatomical scans for tissue localization, diffusion weighted imaging (DWI) to identify acute ischemia, and DCE MRI to assess BBB permeability. The para- meters for the DWI and DCE sequences are as follows: DWI: 2D Turbo spin echo (TSE) acquisition; b ¼ 0, 1000; TR/TE ¼ 1000/33 ms; number of slices: 6; Flip angle: 90; Slice gap: 2 mm; slice thickness: 1 mm; FOV: 100 Â 85 mm2; pixel size: 0.28 Â 0.28 mm. DCE: T1-weighted dynamic 3D gradient echo acquisition; TR/TE ¼ 6.3/2.2, FOV ¼ 100 Â 85 mm2; matrix ¼ 168 Â 142; slices ¼ 12; slice thickness ¼ 1 mm; volumes ¼ 36; temporal resolution 6.5 s; acquisition time ¼ 4:20 min; Gadolinium-DTPA is injected through the tail vein at a rate of 60 μL over 10–15 s beginning approximately 25 s after the start of the DCE acquisition.

3.4.3 MR Image Analysis To quantify the degree of water diffusion from DWI data, apparent diffusion coefficient (ADC) maps are created using in-house soft- Diffusion Weighted Imaging ware in MATLAB v.7.11 (Mathworks, Natick, MA, USA). ADCs (DWI) are calculated by fitting the normalized logarithmic signal-intensity decay as a function of the b-value. Ischemic areas are identified as regions where there is reduced diffusion on the ADC maps, relative to normal cortex [10]. This was used as the basis for delineating the lesion region-of-interest (ROI). For representing healthy brain tissue, a second ROI within the homologous location in the con- tralateral hemisphere is defined. ADCs from the ROIs are recorded for all slices where ischemia is present.

Dynamic Contrast- As described above, DCE MRI consists of repeated T1-weighted Enhanced Imaging images to track the pathway of a gadolinium contrast agent injected into the tail vein [10]. The extent of contrast accumulation can be measured as a function of time and a uni-directional two-compart- ment model [11] model can be used to determine the degree of BBB disruption as a result of stroke. In this model, parametric maps of permeability-surface area product (KPS) are calculated on a pixel-by-pixel basis (units: mL/100 g/min). The calculation involves fitting the time-varying MR signals from the extravascular (tissue) region to an intravascular (input) function and performing a linear regression to estimate the rate of contrast leakage in a given region. The chosen input function is often the sagittal sinus, which is used as a surrogate of the arterial input function [12]. The KPS values from infarcted areas are then normalized against the homol- ogous normal tissue in the contralateral hemisphere to generate the ratio of infarcted KPS to contralateral KPS (see Fig. 4). Please note that the ROIs are copied over from the ADC maps to ensure that measurements are taken from the same areas. Evaluation of Blood-Brain Barrier Permeability and Integrity in Juvenile... 309

Fig. 4 Representative MR images. (a) Representative T2W image of a coronal slice from an infarcted rat brain. The bright area inside the green demarcated region is where infarct has occurred. (b) Representative KPS map of a coronal slice from an infarcted rat brain. The gradient colors indicate the extent of BBB permeability from dark blue (low permeability) to dark red (high permeability). Figure adapted from [13]

3.5 Sacrificing and After completion of MRI, the rat is scarified via decapitation using a Brain Extraction guillotine. The brain is then extracted (within a 10-min period) by Procedure carefully cutting through the skull from the posterior side with surgical dissecting scissors. This is followed by removing the skull with bone cutting forceps, and subsequent removal of the brain with a surgical spatula. After the extraction, the brain is placed on a sheet of weighing paper and placed in a À20  C freezer for 20 min.

3.6 Histological During this 20 min of freezing time, the TTC solution is prepared Procedures by weighing 0.3 g of TTC powder using a micro weighing scale. The powder is then poured into a 50 mL centrifuge tube contain- 3.6.1 Triphenyl ing 15 mL of 0.9% saline (saline added by a 10 mL syringe). The Tetrazolium Chloride (TTC) tube is shaken until the powder dissolves completely, and then Staining wrapped in aluminum foil and placed in ice until it is used (This step is strongly encouraged because TTC is light sensitive). Next, the brain is taken out of the freezer and placed in a brain matrix where it is cut into 2 mm slices using a single edge razor blade (usually multiple blades are required). The slices are then removed and placed into a petri dish, which has been prepared with the TTC solution (described above). The slices are submerged into the TTC solution. This will change the color of the tissue. After approxi- mately 15 min, the contrast between ischemic tissue (white area) and healthy tissue (pink area) is maximized and photographs of the slices will be taken for record keeping (see Fig. 5). The slices are also flipped around to make note of any differences occurring on the other side of the slices.

3.6.2 Evans Blue For successful EB staining, an EB solution needs to be prepared and (EB) Staining injected prior to sacrifice. A 4% EB solution [13] is prepared by adding 400 mg of EB powder (weighed using a micro weighing 310 Trish Domi et al.

Fig. 5 Representative TTC staining of 2 mm coronal slices of stroke rats (a–e) and non-stroke rats (f). (a–c) Slices from 72-h survival rats, whereas (d) and (e) are slices from 24-h survival rats. (f) A slice from a 3-h survival non-stroke rat

scale)to10mLofsaline in a 50 mL centrifuge tube. The tube is then shaken and vortexed so that EB is completely dissolved. The EB solution is then divided into smaller aliquots by adding 0.7 mL of the solution to an empty 1.5 mL microtube (totaling 14 micro- tubes). The aliquots are stored in a À20  C freezer and thawed prior to usage. Two hours prior to sacrifice, the rat is weighed as the EB injection volume is calculated for each rat based on their weight: EB Injection VolumeðÞ¼ mL Rat WeightðÞÂ kg 2mL=kg: Depending on the rat’s weight, the appropriate number of EB aliquots is then thawed, and the EB is drawn into a syringe, with the remainder of the solution being disposed. Next, the rat is placed in the isoflurane chamber and isoflurane gas is released (same gas mixtures as before). Once the rat is fully anesthetized, it is moved to the nose cone and positioned for EB injection (Fig. 6). For that, a 24G catheter is injected into tail vein, and the needle is pulled back to verify its placement in the vein. If blood begins to pool in the needle tip, the catheter was inserted successfully. A 1 mL syringe containing EB solution is then attached to the catheter, and the solution is injected into the tail vein. After completion of the EB injection, the rat is allowed to recover in its cage. Two hours after EB injection, the rat is anesthetized again and sacrificed by decapitation using a guillotine. The brain is then extracted as described in Sect. 3.5. Next, the brain is placed in a À20  C freezer for 20–25 min. During this period, 14 empty 1.5 mL tubes are weighed and labeled with slice number and designated slice hemisphere (right or left). Once the brain is removed from the freezer it is placed onto the brain matrix for Evaluation of Blood-Brain Barrier Permeability and Integrity in Juvenile... 311

Fig. 6 Preparation of anesthetized rat for EB tail vein injection. (a) Nose cone, (b) syringe containing EB, (c) syringe containing saline, (d) 24G catheter

slicing. In order to secure the brain while slicing, one single edge razor blade is placed in front of the brain, and one in between the cerebrum and cerebellum. The first slice is cut by placing a razor blade 2 mm posterior to the first blade, and then a second slice is cut by placing a blade 2 mm posterior to the previous blade. The first and second blades are then removed to extract the first slice. After removing the slice, another razor blade is placed 2 mm poste- rior to the third razor blade and the third blade is then removed. The same procedure is repeated until a total of seven slices are cut per brain. Once all slices are extracted, each one is cut in half (left and right hemispheres), and each hemisphere is stored in the pre-labeled (using a fine point waterproof marker) 1.5 mL microtube. The tubes are placed on ice in a microtube freezer box. Each tube is then weighed again with tissue inside, and weight is recorded. The slices are stored at À80  C until further EB analysis is performed. For EB extravasation analysis, the samples are retrieved from the freezer. Under the fume hood, 500 μLofformaldehyde (3.5%) is added to each tube (Note 4). The samples are then incubated on a heating block at 50  C for 20 h. After incubation, the tissue is broken down by using the ultrasonic cell disruptor (Misonix sonicator 3000, Cole-Parmer, USA) for 3 s per sample. The cell disruptor tip is cleaned with task wipers and distilled water between every sample sonication. The samples are centri- fuged at 10,000 RPM, 20  C for 30 min. During this period, 312 Trish Domi et al.

4096 μg ¼ 4.096 mg of EB powder is weighed. The EB powder is then poured into a 1.5 mL microtube labeled “4096.” In order to prepare the standards by serial dilution, 13 other tubes are labeled as follows: 2048, 1024, 512, 256, 128, 64, 32, 16, 8, 4, 2, 1, 0. Before pipetting the centrifuged samples into a 96-well plate, the positions of the samples are marked on the plate lid. Wells A1 to B6 are left empty to be used for standards. Starting at B7, 150 μLof each sample is pipetted into the wells in duplicates (i.e., B7 and B8 are both sample 1). Once all samples are pipetted into the wells, the same procedure is repeated for the standards. To prepare the standards (starting from 4096 μgofEBto0μg of EB, with every tube containing half the amount of EB from the previous tube), 1000 μL of formaldehyde is added to the tube containing 4096 μg of EB. Serial dilution is created by first adding 500 μL of formaldehyde to the other labeled tubes, and 500 μLof solution from the tube labeled 4096 is added into the tube labeled 2048 and vortexed. The 500 μL of solution from the tube labeled 2048 is then added into the tube labeled 1024 and vortexed. This process is repeated for all of the labeled tubes. For the tube labeled 0, only 500 μL of formaldehyde is added. Standards are plated as: 128, 64, 32, 16, 8, 4, 2, and 0 (in duplicates). Once the standards are inside the wells, the plate is covered with aluminum foil and put on ice in a freezer box. The plate is then transported as fast as possible to the spectrophotometer. The plate template is set up according to the number of samples in the plate, and scanned at 620 nm. The quantified values for the samples are determined based on a standard curve created by the spectrophotometer. It is important to make sure that the standard curve has a high R value (between 0.9 and 1), or else the output values would be inaccurate. In addition, if the standard deviation for a sample is too high, the sample solution may not be homogeneous. In that case, the plate can be shaken and rescanned.

4 Notes

4.1 Surgical 1. In order to have an adequate view of the common carotid Procedure artery, the pericarotid fat is often excised. In addition, the vagus nerve is carefully separated along the entire pathway of the common carotid artery so that microsurgical clips can be placed. 2. It is important that the microsurgical clip is placed as proxi- mally as possible to the common carotid artery and as distally as possible to the internal carotid artery. This way, the subsequent insertion of the silicone-coated suture is much easier. Evaluation of Blood-Brain Barrier Permeability and Integrity in Juvenile... 313

3. If the microsurgical clips are larger than what is comfortable for blocking the ICA and CCA, sutures (which are smaller and more flexible) can be used as alternatives to tie the arteries. 4. Perhaps the most difficult step of the procedure is advancing the silicone-coated suture into the ICA. This is due to fact that the tip of the suture can get obstructed at the base of the skull or diverted into the pterygopalatine artery. In order to ease full advancement of the silicone-coated suture the head of the rat can be hyper-extended by approximately 10. In addition, the trajectory of the suture should be such that the force is directed slightly medially and rostrally. If a lot of resistance is felt while advancing the filament, there is a chance that the filament is too large for the vessel. In that case, it is recommended that another filament with a smaller size is used instead. 5. A common reason for lethal experimental failure during sur- gery (even by the most experienced surgical technicians) is subarachnoid bleeding due to the retraction of the silicone- coated suture. It is therefore important that this step is con- ducted as carefully and slowly as possible to minimize the risks of trauma to the circle of Willis, and avoid bleeding.

4.2 Preparations 1. Make sure the butterfly needle is replaced for every scan, as for MRI Scanning using it repeatedly for different scans would make the needle tip dull and reduce the rate of successful injection inside the tail vein. 2. In order to test if the butterfly needle is inside the vein, the syringe attached to the other end of the tube can be pulled back slightly; if the needle is in the vein, the displaced saline inside the tube is replaced with blood. As soon as blood is spotted inside the tube, the syringe containing saline can be used to push the blood back to the vein so that no blood clot is formed inside the tube. 3. The 3 T Philips clinical MRI system with dedicated hardware is specific to our setup. In general, clinical MRI scanners at 3.0 T are not well suited for small animal studies due to the reduced signal-to-noise. Dedicated animal MRI systems (e.g., Bruker) at higher fields, e.g., at 7 T are well equipped to handle these scans we described.

4.3 EB Procedure 1. In order to better spot the vein prior to EB injection, a con- tainer filled with warm water can be placed on the rat tail, which causes vein dilation and therefore allows for better visibility of the vessel. 2. Before injecting EB, it is important to ensure that the needle/ catheter is in the tail vein and that the catheter/needle hasn’t poked through the vein. As a precaution, injection of some 314 Trish Domi et al.

saline is recommended to check the needle/catheter is posi- tioned correctly in the vein. If a white bulge is forming in the area of injection, that is typically an indication that the needle is not in the vein. 3. Do not conduct EB analysis any later than 8 weeks of the sample being stored in a À80  C freezer, as the EB embedded within the slices may degrade. 4. It is important to ensure that the tissue is submerged in the formaldehyde solution, so that its contents are preserved.

References

1. Engel O, Kolodziej S, Dirnagl U, Prinz V 9. Garcia JH, Wagner S, Liu KF, Hu XJ (1995) (2011) Modeling stroke in mice - middle cere- Neurological deficit and extent of neuronal bral artery occlusion with the filament model. J necrosis attributable to middle cerebral artery Vis Exp 47:e2423. https://doi.org/10.3791/ occlusion in rats. Statistical validation. Stroke 2423 26(4):627–634; discussion 635. Retrieved 2. Tsze DS, Valente JH (2011) Pediatric stroke: a from http://www.ncbi.nlm.nih.gov/ review. Emerg Med Int 2011:1–10. https:// pubmed/7709410 doi.org/10.1155/2011/734506 10. Merali Z, Huang K, Mikulis D, Silver F, Kass- 3. Uluc¸ K, Miranpuri A, Kujoth GC, Aktu¨re E, ner A (2017) Evolution of blood-brain-barrier Bas¸kaya MK (2011) Focal cerebral ischemia permeability after acute ischemic stroke. PLoS model by endovascular suture occlusion of the One 12(2):e0171558. https://doi.org/10. middle cerebral artery in the rat. J Vis Exp (48). 1371/journal.pone.0171558 doi:https://doi.org/10.3791/1978 11. Patlak CS, Blasberg RG (1985) Graphical eval- 4. Kawamura S, Shirasawa M, Fukasawa H, Yasui uation of blood-to-brain transfer constants N (1991) Attenuated neuropathology by nilva- from multiple-time uptake data. Generaliza- dipine after middle cerebral artery occlusion in tions. J Cereb Blood Flow Metab 5 rats. Stroke 22(1):51–55. Retrieved from (4):584–590. https://doi.org/10.1038/ http://www.ncbi.nlm.nih.gov/pubmed/ jcbfm.1985.87 1987673 12. Ewing JR, Knight RA, Nagaraja TN, Yee JS, 5. Takano K, Tatlisumak T, Bergmann AG, Gib- Nagesh V, Whitton PA et al (2003) Patlak plots son DG, Fisher M (1997) Reproducibility and of Gd-DTPA MRI data yield blood-brain reliability of middle cerebral artery occlusion transfer constants concordant with those of using a silicone-coated suture (Koizumi) in 14C-sucrose in areas of blood-brain opening. rats. J Neurol Sci 153(1):8–11. Retrieved Magn Reson Med 50(2):283–292. https:// from http://www.ncbi.nlm.nih.gov/ doi.org/10.1002/mrm.10524 pubmed/9455971 13. Merali Z, Leung J, Mikulis D, Silver F, Kassner 6. Carmichael ST (2005) Rodent models of focal A (2015) Longitudinal assessment of imati- stroke: size, mechanism, and purpose. Neu- nib’s effect on the blood-brain barrier after roRx 2(3):396–409. https://doi.org/10. ischemia/reperfusion injury with permeability 1602/neurorx.2.3.396 MRI. Transl Stroke Res 6(1):39–49. https:// 7. Langheinrich AC, Yeniguen M, Ostendorf A, doi.org/10.1007/s12975-014-0358-6 Marhoffer S, Kampschulte M, Bachmann G 14. Canazza A, Minati L, Boffano C, Parati E, et al (2010) Evaluation of the middle cerebral Binks S (2014) Experimental models of brain artery occlusion techniques in the rat by ischemia: a review of techniques, magnetic res- in-vitro 3-dimensional micro- and nano com- onance imaging, and investigational cell-based puted tomography. BMC Neurol 10:36. therapies. Front Neurol 5:1–15. https://doi. https://doi.org/10.1186/1471-2377-10-36 org/10.3389/fneur.2014.00019 8. Longa EZ, Weinstein PR, Carlson S, Cummins 15. Mhairi Macrae I (1992) New models of focal R (1989) Reversible middle cerebral artery cerebral ischaemia. Br J Clin Pharmacol 34 occlusion without craniectomy in rats. Stroke (4):302–308. Retrieved from http://www. 20(1):84–91 Retrieved from http://www.ncbi. ncbi.nlm.nih.gov/pubmed/1457262 nlm.nih.gov/pubmed/2643202 Chapter 18

Recording Leukocyte Rolling and Adhesion on Meningeal Vessels by Intravital Microscopy

Aline Silva de Miranda, Thiago Macedo Cordeiro, Milene Alvarenga Rachid, and Antoˆnio Lu´ cio Teixeira

Abstract

Leukocyte infiltration in the central nervous system (CNS) has been implicated in several neuroinflamma- tory diseases, being an important step in the development of inflammatory response in the brain. The use of intravital microscopy technique allows direct in vivo assessment of leukocyte rolling and adhesion on cortical meningeal vessels. The current review highlights the use of intravital microscopy and its association with neuroinflammatory parameters in different models of experimental brain diseases such as HSV-1 encephalitis, dengue encephalitis, cerebral malaria, and hepatic encephalopathy. Herein, we also describe a detailed protocol of intravital microscopy of meningeal vessels, its advantages and major concerns.

Key words Brain, Pia mater vessels, Intravital microscopy, Leukocyte rolling, Leukocyte adhesion, Chemokines, Mice

1 Introduction: Overview of Leukocyte Recruitment

There are at least tree known routes for leukocytes entry into the central nervous system (CNS). The most extensively studied is the one from blood into the parenchymal perivascular space though the blood brain barrier (BBB). Other potential routes include the choroid plexus into the cerebrospinal fluid, and the post-capillary venules at the pial surface into subarachnoid and Virchow-Robin perivascular spaces [1–3]. A fourth route has also been suggested, and involves migration from subependymal vessels via the epen- dyma into the ventricles [4]. By migrating from blood to the CNS compartments, leukocytes can exert immune surveillance, being able to engage inflammatory and immune responses during host defense and neuroinflammatory diseases [5]. The recruitment of immune cells into the CNS follows the paradigm of leukocyte extravasation across other vascular beds, being characterized by a sequential and tightly controlled multistep process (Fig. 1). This process involves, among other factors, fluid

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_18, © Springer Science+Business Media, LLC, part of Springer Nature 2019 315 316 Aline Silva de Miranda et al.

Fig. 1 (a) Schematic diagram representing the highly regulated process of leukocyte migration through the blood-brain barrier, including six steps: rolling, activation, adhesion, crawling, protrusion, and transmigration, directed by the concentration gradient of inflammatory mediators, mainly chemokines (pink dots). (b) Emphasis on the molecules involved on the rolling step represented by selectins (light blue) and their respective glycosylated ligands on the surface of leukocytes (orange). (c) Emphasis on the molecules involved on the activation step represented by chemokines (pink) binding to G-protein-coupled receptor (GPCR) (green) that generates a signal to activate integrins (brown) on leukocytes’ surface. (d) Emphasis on the molecules involved on the adhesion step represented by concomitant binding of chemokines (pink) and their GPCR (green) and integrins (brown) binding to their endothelial counter receptors of the immunoglobulin superfamily, specially VCAM-1 and ICAM-1 (dark blue). (e) Emphasis on the molecules involved on the crawling step represented by interactions of the same molecules involved in the adhesion step, but with increased number of interactions. (f) Emphasis on the molecules involved on the activation step represented by transmigration of the leukocyte between the endothelial cells following a gradient of immune mediators (pink dots). (g) Emphasis on the end of the process, with the leukocyte inside the brain tissue

dynamics within the vasculature and molecular interactions between circulating leukocytes and the vascular endothelium [5, 6]. The first step is known as rolling and is characterized by a short and initial transient contact of the circulating cells with the vascular endothelium mediated by adhesion molecules of the selectin family (L-, E-, or P-selectin) expressed on endothelial surface and their Intravital Microscopy in Mouse Brain 317

respective glycosylated ligands [e.g., P-selectin glycoprotein ligand (PSGL-1)] on leukocytes. Very late antigen-4 (VLA-4) can also support rolling. These interactions are of low affinity and allow the leukocytes to roll along the vascular wall with gradually lower velocity, giving the traveling leukocytes the opportunity to scan endothelial surfaces for luminal immobilized chemotactic factors from the family of chemokines [7]. In a second sequential step called activation, chemokines by binding to G-protein-coupled receptors (GPCRs) expressed on leukocytes deliver an inside-out signal that activates integrins. Integrin activation induces both conformational changes and clus- tering, determining enhanced avidity and affinity of the leukocyte integrin for its endothelial counter-receptors of the immunoglobu- lin superfamily, specifically vascular cell adhesion molecule-1 (VCAM-1) and intercellular adhesion molecule-1 (ICAM-1) or alternative docking sites such as fibronectin connecting segment-1 (FN CS-1) [8]. The binding of leukocytes high affinity/avidity integrins (VLA-4, LFA-1, and Mac-1) to their endothelial ligands (ICAM-1 and VCAM-1) generates cytoplasmic signaling cascades in both leukocytes and endothelial cells. As a result, these binding interactions lead to integrin-mediated post-arrest immune cell adhesion, strengthening and polarization, followed by immune cell crawling on the endothelium [6]. During the crawling process, immune cells probe the endothe- lium with invadosome-like protrusions in search of optimal sites for transmigration [7, 9]. The last stage in this multistep process is transmigration or diapedesis that seems to occur through two distinct pathways: paracellular diapedesis through the endothelial junctions, and transcellular diapedesis by inducing the formation of pore-like structures in the endothelium [7, 10].

2 Using Intravital Microscopy to Study Models of Neuroinflammatory Diseases

The development of intravital microscopy technique (Fig. 2) allowed the study of leukocyte rolling and adhesion processes on cerebral cortical pial vessels, paving the way for the better under- standing of the pathogenesis of several CNS diseases. For instance, the role of the chemokines CCL2 and CCL5 was investigated in experimental autoimmune encephalomyelitis (EAE) by intravital microscopy of the pial microvasculature. A high expres- sion of CCL2 and CCL5, and increased adhesion and consequent migration of mononuclear cells into the brain were found in MOG35–55-induced EAE. Mononuclear cells recruitment to the brain was prevented by the administration of anti-CCL2 antibodies [11] or a modified chemokine that interferes with CCL2 function [12]. Similar findings were observed in the absence of the kinin B2 318 Aline Silva de Miranda et al.

Fig. 2 Schematic diagrams showing: (a) Administration of fluorescent dye rhodamine 6G in order to label leukocyte mitochondria, and craniotomy of the right parietal bone with exposure of the underlying brain pial vessels. (b) Placement of the mouse in the intravital microscope for leukocyte rolling and adhesion analysis. (c) Image capture of leukocytes rolling and adhesion on pial vessels of noninfected (controls) C57BL6 mice. (d) Image capture of leukocytes rolling and adhesion on pial vessels of C57BL6 mice following 6 days of infection with Plasmodium berghei ANKA (PbA), a well-known model of cerebral malaria. (e) Intravital microscopy quantification of leukocytes rolling (left) and adhesion (right) in the pial vessels of controls and PbA-infected mice on day 6 post-infection (6 dpi)

receptor, the major receptor for bradykinin, with decreased CCL2 production and CCL2-mediated leukocyte adhesion in EAE [11]. The absence of the signaling molecule phosphoinositide 3-kinase-γ (PI3Kγ), a molecule implicated in driving leucocyte migration, did not influence leukocyte rolling and adhesion in EAE, supporting the hypothesis that these processes were CCL2- Intravital Microscopy in Mouse Brain 319 dependent. However, the lack of PI3Kγ clearly improved the dis- ease symptoms and neuropathology in parallel with an increase of apoptotic events in the CNS, indicating that PI3Kγ might play a role in leukocytes survival once they migrate into the brain [13]. Similarly, the platelet-activating factor (PAF), an important mediator of immune responses, seems to play a role in the induc- tion and development of EAE without influencing rolling and adhesion of leukocytes [14]. PAF changed the profile of the inflam- matory infiltrate with a bias toward the recruitment of polymor- phonuclear leukocytes along with improvement of EAE clinical symptoms [14]. The analysis of brain microvasculature by intravital microscopy can contribute to the investigation of leukocyte recruitment in response to CNS infectious. The intracranial inoculation of a neu- rotropic Herpes simplex virus-1 (HSV-1) strain, a well-known murine model of HSV-1 infection, caused signs of encephalitis and death by day 6 of infection. Using intravital microscopy, our group found a significant increase of leukocyte rolling and adhesion in the brain microvasculature of infected mice associated with high expression of chemokines (CCL2, CCL3, CCL5, CXCL1, and CXCL9) and TNF at early (1-day post-infection) and late (5 days post-infection) points of the disease. Histological analyses con- firmed diffuse meningoencephalitis characterized mainly by mono- nuclear cell infiltrates [15]. Further studies showed that the chemokine CCL5 (also known as RANTES, regulated upon activa- tion, normal T cell expressed and presumably secreted) is essential for leukocyte recruitment in this HSV-1 encephalitis model. Accordingly, treatment with anti-CCL5 or Met-RANTES, an antagonist of the CCL5 receptors CCR1 and CCR5, had no effect on viral titers but significantly decreased the number of leukocytes adherent to the pial microvasculature at days 1 and 3 after infection [16]. The intravital microscopy revealed that CCL5, possible through binding on CCR1 and CCR5 receptors, drives leukocyte adhesion to brain pial vessels, and subsequent migration in the context of HSV-1 infection. However, blocking of CCR1 and CCR5 did not affect virus replication, suggesting that other immune mechanisms are involved in the process of infection con- trol [16]. Genetic deletion of CCR5 increased the levels of chemo- kines (CCL2, CCL5, CXCL1, and CXCL9) in the brain of HSV-1 infected mice in parallel with enhanced leukocyte adhesion in brain microvasculature, predominantly of neutrophils, and reduction of viral load at an early course of the infection. In the early stage of HSV-1 infection, the lack of CCR5 might boost the immune response in an attempt to promote viral clearance, which may be detrimental to the host in late stages [17]. HSV-1 viral replication control might also be mediated by mechanisms related to TNF signaling since mice lacking the recep- tor 1 for TNF (TNFR1 or p55) presented higher HSV-1 viral 320 Aline Silva de Miranda et al.

replication and severe meningitis compared with infected wild-type mice. Accordingly, a decrease in the expression of the chemokines CCL3 and CCL5 in the brain in parallel with reduced leukocyte adhesion on the pial vasculature, as revealed by intravital micros- copy, corroborated the role of CCL5 in leukocyte recruitment in response to HSV-1 infection [18]. Another important immune mediator in HSV-1 encephalitis seems to be PAF since the pharma- cological blockage of PAF receptor or its absence by gene deletion promotes increased survival in association with less inflammation as evidenced by decreased leukocyte-endothelial cell interaction in meningeal vessels and less intense brain infiltration of mononuclear cells [19]. As HSV-1 infection, intracranial injection of dengue virus (DENV) serotype 3 caused encephalitis in C57BL/6 mice, leading to death around 8 days post-infection. Intravital microscopy analy- sis of pial microvasculature revealed increase of leukocyte rolling and adhesion at days 3 and 6 post-infection. Significant increase of the cytokines IFN-γ, TNF, and chemokines CCL2, CCL5, CXCL1, and CXCL2 levels was also found in the brain of DENV- 3-infected animals. Increased numbers of neutrophils, CD4+ and CD8+ T cells were detected notably at day 6 post-infection, preced- ing the development of motor signs of disease at day 7 post- infection [20]. In a cerebral malaria (CM) model induced by intraperitoneal administration of 106 Plasmodium berghei (strain ANKA) parasi- tized red blood cells in C57Bl/6 susceptible mice, intravital micros- copy also showed that the increase in leukocyte rolling and adhesion on pial microvasculature occurs at day 5 post infection, preceding the development of neurological signs around day 6 post-infection. High concentrations of the chemokine CXCL9 were observed in the brain and serum of infected mice at day 5 post-infection whereas the levels of the other chemokines (CCL2, CCL3, CCL5, and CXCL1) increased later on day 7 post-infection [21]. These findings suggest that CXCL9 (not CCL5 as observed in HSV-1 infection) plays a role in the recruitment of leukocytes to the CNS in response to malaria infection. The absence of PAF or the pharmacological blockage of its signaling delayed mortality, reduced vascular plugging and hemorrhage, decreased brain levels of inflammatory cytokines, and markedly reduced changes in vas- cular permeability. However, no effect was found in leukocyte roll- ing and adherence on pial vessels as assessed by intravital microscopy at day 6 post-infection [22]. The intravital microscopy was also used to evaluate leukocyte recruitment to the brain in response to Plasmodium berghei NK65 infection, a well-recognized model of systemic (not cerebral) malaria. Increased leukocyte rolling and adhesion on pial venules of infected mice compared with noninfected animals was revealed by intravital microscopy, which were associated with BBB Intravital Microscopy in Mouse Brain 321

permeability and chemokine production, showing that CNS were also affected by the Plasmodium berghei NK65 infection [23]. The pivotal steps (rolling and adhesion) of leukocytes recruit- ment to the brain have also been studied by intravital microscopy in sterile neuroinflammatory conditions. For example, increased brain levels of cytokines (TNF and IL-1β) and chemokines (CXCL1, CCL2, CCL3, and CCL5) were associated with behavioral impairment and morphological changes in astrocytes and brain endothelial cells in a model of encephalopathy secondary to thioacetamide-induced acute liver failure. The intravital microscopy showed increase in leukocyte rolling but not adhesion on pial microvasculature, providing functional evidence of endothelial cells activation in this model [24]. Altogether, these studies highlight the potential application of intravital microscopy to better understand leukocyte recruitment during CNS pathophysiological conditions as well as the role of inflammatory mediators like chemokines and their receptors in this specific context. Intravital microscopy might be a valuable tool to the search and ultimately the identification of promising therapeu- tic targets. Intravital microscopy presents some interesting features that contribute to the understanding of leukocyte traffic in brain micro- vasculature: direct in vivo visualization of immune cells rolling and adhesion, real-time evaluation of leukocyte migration steps, and the possibility to study specifically the recruitment process through administration of testing drugs just before its assessment. However, as the majority of techniques, intravital microscopy also present limitations that should be taken into consideration such as visuali- zation restricted to the pial vessels, no strict correlation between the intensity of rolling and adhesion with the number of leukocytes in brain parenchyma, no visualization of cells transmigration, and difficulty to define interacting cell types [25].

3 Materials

All materials including reagents and dyes used to perform intravital microscopy are described as follows:

l Male or female C57BL/6 mice or genetic modified mice in C57BL/6 background (20–25 g), aged 8–12-week-old. The animals were housed in groups of six mice per cage in a room- controlled temperature (25 C) with food and water ad libitum. l A sterile mixture of 10 mg/kg of body weight xylazine and 150 mg/kg ketamine all diluted in phosphate buffered saline (PBS 1Â) containing, in mmol/L: NaCl 137, KCl 2.7,  Na2HPO4 10, and KH2PO4 1.8, at 37 C, pH 7.35. 322 Aline Silva de Miranda et al.

l Fluorescent dye rhodamine 6G (0.5 mg/kg; Sigma, St. Louis, MO, USA). Rhodamine is a useful dye that labels leukocyte mitochondria, allowing cell visualization by epi-illumination. l A high- speed drill (Dremel, New York, USA) used to perform craniotomy in order to expose the underlying pial vasculature. l Heating pad (Fine Science Tools Inc., North Vancouver, Canada) was used throughout the experiment to maintain mouse rectal temperature at 36.8–37 C. l Artificial cerebrospinal fluid, an ionic composition containing, in mmol/L: NaCl 132, KCl 2.95, CaCl2 1.71, MgCl2 0.64,  NaHCO3 24.6, dextrose 3.71, and urea 6.7, at 37 C, pH 7.35. The superfusate was bubbled continuously with 10% O2,6%O2, and 84% N2, which maintains a gas tension and a pH comparable with those of normal CSF. l An Olympus (Center Valley, PA, USA) model B201 microscope (Â20 objective lens, corresponding a 100 μm of area) outfitted with a fluorescent light source (epi-illumination at 510–560 nm, using a 590 nm emission filter). l A silicon-intensified camera (Optronics Engineering DEI-470) mounted on the microscope projected the image onto a monitor (Olympus, Center Valley, PA, USA). l A videocassette recorder (VHS, Semp Toshiba, model x685) and/or Adobe Premier 4.0 software and SigmaScan Pro 4.0 software (SPSS Chicago, IL).

4 Methods

4.1 Mice Preparation At first, mice were anesthetized by intraperitoneal injection of a and Craniotomy sterile mixture of 10 mg/kg of body weight xylazine and 150 mg/ kg of body weight ketamine. After complete anesthesia, the tail vein was cannulated to venous administration of the fluorescent dye rhodamine 6G as well as for additional anesthetic if necessary. Mice rectal temperature were constantly monitored and kept at 36.8–37 C with a heating pad. The craniotomy (Fig. 3) was performed using a high-speed drill in the right parietal bone. Dura mater and arachnoid meninges were removed to expose the underlying brain pial vessels. Importantly, the craniotomy did not disrupt the vascular barrier of the pial microvasculature as fluores- cently labeled proteins remained within the vasculature. All experi- ments were performed during 1 h and the exposed brain was constantly superfused with the artificial CSF to maintain the prepa- ration stable. Intravital Microscopy in Mouse Brain 323

Fig. 3 Schematic diagram of the craniotomy performed for intravital microscopy. Sections a–d illustrate, respectively, the exposition of the animal’s scalp with a small scalpel, followed by the creation of an oval window laterally to the sagittal suture on the right parietal bone and the complete exposure of the pial vessels by removal of the oval piece of bone and underlying leptomeninges

4.2 Visualization In order to observe leukocyte/endothelium interactions, leuko- of Leukocyte Rolling cytes were fluorescently labeled by intravenous administration of and Adhesion in Brain rhodamine 6G (0.5 mg/kg body weight) and observed using a Pial Vessels microscope (Olympus B201, Â20 objective lens, corresponding a 100 μm of area) outfitted with a fluorescent light source (epi-illumination at 510–560 nm, using a 590 nm emission filter). A silicon-intensified camera (Optronics Engineering DEI-470) mounted on the microscope projected the image onto a monitor (Olympus). The number of rolling and adherent leukocytes was determined offline during video playback analysis using a videocas- sette recorder. Alternatively, the video images were digitalized with a personal computer using Adobe Premier 4.0 software, and x-y coordinate data for each cell image were obtained using SigmaScan Pro 4.0 software (SPSS Chicago, IL).

4.3 Quantification To quantify leukocyte rolling and adherence in brain pial microvas- of Leukocyte Rolling culature, three to four vessels were analyzed for animal. Leukocytes and Adhesion in Brain were considered adherent to the venular endothelium if they Pial Vessels remained stationary for a minimum of 30 s. Rolling leukocytes were defined as cells moving at a velocity lower than that of ery- throcytes within a given vessel. Pial vessels with diameters ranging from 50 to 120 μm were used, as most adhesion occurred in vessels 324 Aline Silva de Miranda et al.

of these sizes. Due to great variability in size of these vessels (compared with that of other tissues studied using intravital micros- copy), we expressed leukocyte adhesion as number of cells/ 100 μm. A significant amount of platelet deposition was also noted on endothelium. This was quantified as the percentage area of vessel covered by fluorescently labeled platelets.

5 Notes

Optimal cranial window surgeries are crucial for the successful of the technique, but require training to become familiar with the fine surgical procedures. The surgery should be performed under the most sterile condition as possible to avoid infection and with extremely care to avoid dura mater damage and bleeding. If the dura is manipulated excessively or punctured, the preparation should not be used for imaging. During the craniotomy, it is important not to apply pressure to the skull and brain while drilling, which might lead to cortical trauma and, as a consequence, inflam- mation. It is recommended to stop the drilling when the bone flap becomes loose. Gently lift the flap after applying a drop of artificial CSF to prevent dehydration and bleeding of the dura mater. In order to avoid biased assessment of leukocyte rolling and adhesion we first check whether the blood is flowing in all vessels. Then, selection of the vessels to be measured is made and it should include venules and arterioles of different diameters (in our mouse pial preparations, most vessels range between 50 and 120 μm) and cover different locations within the area exposed by the window. In our studies, measurements are made in 3–4 vessels for each animal. Precise location of each spot to be measured is annotated in the picture of the pial vasculature. We quantify rolling and adhesion in the same spots predefined to measure blood flow. Quantification of leukocyte adhesion is made by counting the number of leukocytes in a 100 μm-vessel length. Rolling is quantified by counting the number of leukocytes traveling at a velocity significantly slower than blood velocity in the same 100 μm length, during 30 s.

6 Conclusion

Intravital microscopy is a valuable tool that provides in vivo details of cellular recruitment dynamics upon interventions such as drug and antibodies administration, or under physiological and patho- logical conditions. It allows measuring alterations in vessel dia- meters, blood flow, adherence and rolling of leukocytes, platelets and other blood elements, cellular proliferation, vascular leakage, among other potential applications, providing unique advances in our understanding of immune responses [26]. Intravital Microscopy in Mouse Brain 325

Over the past years, significant improvements in the intravital microscopy methods have been made, especially regarding the fluorescence-labeled markers and the experiment duration. For instance, specific antibodies like anti-CD45-TxR and Albumin- FITC allow the identification of the immune cell type as well as the assessment of perfusion and vascular leakage, respectively [27]. It is possible to perform cranial windows that allow visualiza- tion of cells in living brain over time. With appropriate prepara- tions, the same brain area could be analyzed longitudinally, from hours to several months [28–30]. Advances in microscopy methods and expansion of the existing biosensors and fluorophores have brought intravital imaging among the most powerful tools employed in biomedical research.

Acknowledgments

The authors would like to thank FAPEMIG (Fundac¸a˜o de Amparo a` Pesquisa do Estado de Minas Gerais, Brazil), CNPq (Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico, Brazil), and CAPES (Coordenac¸a˜o de Aperfeic¸oamento de Pessoal de Nı´vel Superior) for financial support. ASM is a 2016 NARSAD Young Investigator Grant Awardee from the Brain and Behavior Research Foundation.

References

1. Engelhardt B (2006) Regulation of immune doi.org/10.1111/j.1750-3639.2007.00067. cell entry into the central nervous system. x Results Probl Cell Differ 43:259–280 6. Greenwood J, Heasman SJ, Alvarez JI, Prat A, 2. Engelhardt B, Ransohoff RM (2005) The ins Lyck R, Engelhardt B (2011) Review: and outs of T-lymphocyte trafficking to the leucocyte-endothelial cell crosstalk at the CNS: anatomical sites and molecular mechan- blood-brain barrier: a prerequisite for success- isms. Trends Immunol 26(9):485–495. ful immune cell entry to the brain. Neuro- https://doi.org/10.1016/j.it.2005.07.004 pathol Appl Neurobiol 37(1):24–39. https:// 3. Ransohoff RM, Kivisakk P, Kidd G (2003) doi.org/10.1111/j.1365-2990.2010.01140. Three or more routes for leukocyte migration x into the central nervous system. Nat Rev 7. Muller WA (2011) Mechanisms of leukocyte Immunol 3(7):569–581. https://doi.org/10. transendothelial migration. Annu Rev Pathol 1038/nri1130 6:323–344. https://doi.org/10.1146/ 4. Alvarez JI, Teale JM (2007) Differential annurev-pathol-011110-130224 changes in junctional complex proteins suggest 8. Johnston B, Butcher EC (2002) Chemokines the ependymal lining as the main source of in rapid leukocyte adhesion triggering and leukocyte infiltration into ventricles in murine migration. Semin Immunol 14(2):83–92. neurocysticercosis. J Neuroimmunol 187 https://doi.org/10.1006/smim.2001.0345 (1–2):102–113. https://doi.org/10.1016/j. 9. Steiner O, Coisne C, Cecchelli R, Boscacci R, jneuroim.2007.05.005 Deutsch U, Engelhardt B, Lyck R (2010) Dif- 5. Man S, Ubogu EE, Ransohoff RM (2007) ferential roles for endothelial ICAM-1, ICAM- Inflammatory cell migration into the central 2, and VCAM-1 in shear-resistant T cell arrest, nervous system: a few new twists on an old polarization, and directed crawling on blood- tale. Brain Pathol 17(2):243–250. https:// brain barrier endothelium. J Immunol 185 326 Aline Silva de Miranda et al.

(8):4846–4855. https://doi.org/10.4049/ 17. Vilela MC, Lima GK, Rodrigues DH, Lacerda- jimmunol.0903732 Queiroz N, Pedroso VSP, Miranda AS, Rachid 10. Engelhardt B, Wolburg H (2004) Mini-review: MA, Kroon EG, Campos MA, Teixeira MM, transendothelial migration of leukocytes: Sellner J, Teixeira AL (2013) Absence of CCR5 through the front door or around the side of increases neutrophil recruitment in severe her- the house? Eur J Immunol 34(11):2955–2963. petic encephalitis. BMC Neurosci 14:19–19. https://doi.org/10.1002/eji.200425327 https://doi.org/10.1186/1471-2202-14-19 11. dos Santos AC, Barsante MM, Arantes RM, 18. Vilela MC, Lima GK, Rodrigues DH, Lacerda- Bernard CC, Teixeira MM, Carvalho-Tavares Queiroz N, Mansur DS, de Miranda AS, J (2005) CCL2 and CCL5 mediate leukocyte Rachid MA, Kroon EG, Vieira LQ, Campos adhesion in experimental autoimmune enceph- MA, Teixeira MM, Teixeira AL (2010) alomyelitis—an intravital microscopy study. J TNFR1 plays a critical role in the control of Neuroimmunol 162(1–2):122–129. https:// severe HSV-1 encephalitis. Neurosci Lett 479 doi.org/10.1016/j.jneuroim.2005.01.020 (1):58–62. https://doi.org/10.1016/j.neulet. 12. Handel TM, Johnson Z, Rodrigues DH, Dos 2010.05.028 Santos AC, Cirillo R, Muzio V, Riva S, 19. Vilela MC, Lima GK, Rodrigues DH, Lacerda- Mack M, Deruaz M, Borlat F, Vitte PA, Wells Queiroz N, Pedroso VS, de Miranda AS, TN, Teixeira MM, Proudfoot AE (2008) An Rachid MA, Kroon EG, Campos MA, Teixeira engineered monomer of CCL2 has anti- MM, Teixeira AL (2016) Platelet activating inflammatory properties emphasizing the factor (PAF) receptor deletion or antagonism importance of oligomerization for chemokine attenuates severe HSV-1 meningoencephalitis. activity in vivo. J Leukoc Biol 84 J Neuroimmune Pharmacol 11(4):613–621. (4):1101–1108. https://doi.org/10.1189/ https://doi.org/10.1007/s11481-016-9684- jlb.0108061 7 13. Rodrigues DH, Vilela MC, Barcelos LS, 20. Amaral DC, Rachid MA, Vilela MC, Campos Pinho V, Teixeira MM, Teixeira AL (2010) RD, Ferreira GP, Rodrigues DH, Lacerda- Absence of PI3Kγ leads to increased leukocyte Queiroz N, Miranda AS, Costa VV, Campos apoptosis and diminished severity of experi- MA, Kroon EG, Teixeira MM, Teixeira AL mental autoimmune encephalomyelitis. J Neu- (2011) Intracerebral infection with dengue-3 roimmunol 222(1):90–94. https://doi.org/ virus induces meningoencephalitis and behav- 10.1016/j.jneuroim.2010.02.016 ioral changes that precede lethality in mice. J 14. Rodrigues DH, Lacerda-Queiroz N, de Neuroinflammation 8:23. https://doi.org/10. Miranda AS, Fagundes CT, Campos RD, Ara- 1186/1742-2094-8-23 ntes RE, Vilela Mde C, Rachid MA, Teixeira 21. Lacerda-Queiroz N, Rodrigues DH, Vilela MM, Teixeira AL (2011) Absence of PAF MC, Miranda AS, Amaral DC, Camargos ER, receptor alters cellular infiltrate but not rolling Carvalho LJ, Howe CL, Teixeira MM, Teixeira and adhesion of leukocytes in experimental AL (2010) Inflammatory changes in the central autoimmune encephalomyelitis. Brain Res nervous system are associated with behavioral 1385:298–306. https://doi.org/10.1016/j. impairment in Plasmodium berghei (strain brainres.2011.02.036 ANKA)-infected mice. Exp Parasitol 125 15. Vilela MC, Mansur DS, Lacerda-Queiroz N, (3):271–278. https://doi.org/10.1016/j. Rodrigues DH, Arantes RME, Kroon EG, exppara.2010.02.002 Campos MA, Teixeira MM, Teixeira AL 22. Lacerda-Queiroz N, Rodrigues DH, Vilela (2008) Traffic of leukocytes in the central ner- MC, Rachid MA, Soriani FM, Sousa LP, vous system is associated with chemokine Campos RD, Quesniaux VF, Teixeira MM, up-regulation in a severe model of herpes sim- Teixeira AL (2012) Platelet-activating factor plex encephalitis: an intravital microscopy receptor is essential for the development of study. Neurosci Lett 445(1):18–22. https:// experimental cerebral malaria. Am J Pathol doi.org/10.1016/j.neulet.2008.08.072 180(1):246–255. https://doi.org/10.1016/j. 16. Vilela MC, Mansur DS, Lacerda-Queiroz N, ajpath.2011.09.038 Rodrigues DH, Lima GK, Arantes RM, 23. Lacerda-Queiroz N, Lima OC, Carneiro CM, Kroon EG, da Silva Campos MA, Teixeira Vilela MC, Teixeira AL, Teixeira-Carvalho A, MM, Teixeira AL (2009) The chemokine Araujo MS, Martins-Filho OA, Braga EM, CCL5 is essential for leukocyte recruitment in Carvalho-Tavares J (2011) Plasmodium ber- a model of severe Herpes simplex encephalitis. ghei NK65 induces cerebral leukocyte recruit- Ann N Y Acad Sci 1153:256–263. https://doi. ment in vivo: an intravital microscopic study. org/10.1111/j.1749-6632.2008.03959.x Acta Trop 120(1–2):31–39. https://doi.org/ 10.1016/j.actatropica.2011.04.020 Intravital Microscopy in Mouse Brain 327

24. Faleiros BE, Miranda AS, Campos AC, microcirculation using a closed cranial window. Gomides LF, Kangussu LM, Guatimosim C, J Vis Exp (45). https://doi.org/10.3791/ Camargos ERS, Menezes GB, Rachid MA, 2184 Teixeira AL (2014) Up-regulation of brain 28. Holtmaat A, Bonhoeffer T, Chow DK, cytokines and chemokines mediates neurotox- Chuckowree J, De Paola V, Hofer SB, icity in early acute liver failure by a mechanism Hubener M, Keck T, Knott G, Lee WC, independent of microglial activation. Brain Res Mostany R, Mrsic-Flogel TD, Nedivi E, 1578(Supplement C):49–59. https://doi.org/ Portera-Cailliau C, Svoboda K, Trachtenberg 10.1016/j.brainres.2014.07.001 JT, Wilbrecht L (2009) Long-term, high-reso- 25. Teixeira MM, Vilela MC, Soriani FM, Rodri- lution imaging in the mouse neocortex gues DH, Teixeira AL (2010) Using intravital through a chronic cranial window. Nat Protoc microscopy to study the role of chemokines 4(8):1128–1144. https://doi.org/10.1038/ during infection and inflammation in the cen- nprot.2009.89 tral nervous system. J Neuroimmunol 224 29. Mostany R, Portera-Cailliau C (2008) A crani- (1–2):62–65. https://doi.org/10.1016/j. otomy surgery procedure for chronic brain jneuroim.2010.05.018 imaging. J Vis Exp 12:680. https://doi.org/ 26. Prunier C, Chen N, Ritsma L, Vrisekoop N 10.3791/680 (2017) Procedures and applications of long- 30. Zuluaga-Ramirez V, Rom S, Persidsky Y term intravital microscopy. Methods (2015) Craniula: a cranial window technique 128:52–64. https://doi.org/10.1016/j. for prolonged imaging of brain surface vascula- ymeth.2017.06.029 ture with simultaneous adjacent intracerebral 27. Cabrales P, Carvalho LJ (2010) Intravital injection. Fluids Barriers CNS 12:24. https:// microscopy of the mouse brain doi.org/10.1186/s12987-015-0021-y Chapter 19

Molecular Imaging of Blood–Brain Barrier Permeability in Preclinical Models Using PET and SPECT

Vijayasree V. Giridharan, Tatiana Barichello, and Sudhakar Selvaraj

Abstract

The blood–brain barrier (BBB) with tightest junction separates the systemic circulation and brain micro- environment to protect the brain from insults, such as infections. The integrity of BBB is preserved by multi-structural and functional components. Increasing evidence indicates that BBB is used as an important marker measured in variety of pathological condition with large permeability leaks, such as brain tumors and multiple sclerosis, to more subtle disruption such as vascular diseases, cognitive decline, and dementia. Several imaging modalities are available to study disruption of the BBB. In this chapter, we described the protocols for nuclear imaging studies such as positron emission tomography (PET) and single photon emission computed tomography (SPECT) investigating BBB permeability in preclinical models.

Key words Blood–brain barrier, Positron emission tomography, Single photon emission computed tomography

1 Introduction

The blood–brain barrier (BBB) is the dynamic and complex barrier between the vascular-neural interface in central nervous system (CNS) that maintains brain homeostasis by preventing and regulat- ing the permeation of molecular and cellular components [1, 2]. The brain endothelial cell, the major component of BBB, is the part of larger neurovascular unit that contains pericytes, astrocytes, microglia, and neurons [3]. Normally, these cells are in resting state, but they go to the activated state when there are insults such as ischemia, infection, or an influx of albumin from blood to maintain the brain homeostasis [4]. Understanding the function and integrity of BBB is important due to several reasons. For example, the integrity of the BBB is compromised due to diseases or infection that could increase the BBB permeability by generating an immune or inflammatory response. On the other hand, when drugs are unable to pass through the BBB, it makes the treatment ineffective in neurological and psychiatric disorders

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_19, © Springer Science+Business Media, LLC, part of Springer Nature 2019 329 330 Vijayasree V. Giridharan et al.

[4, 5]. The three commonly used experimental approaches for the investigation of BBB permeability includes measurement of the cerebrospinal fluid (CSF)/blood albumin ratio, histologic assessment of the blood-derived proteins in the brain tissue, and brain molecular imaging, using magnetic resonance imaging (MRI) or positron emission tomography (PET) or single-photon emission computed tomography (SPECT) [6–8]. More than two decades ago, several research groups started to develop PET systems dedi- cated to animal studies. The growing interest in preclinical imaging studies, both in biological and medical basic research and in pharmaceutical industry, has recently induced the world-leading manufacturers of medical image equipment to invest in this field [9]. Small animal imaging represents a cutting-edge research method able to approach an enormous variety of physiological and pathological processes in which animal models of disease may be used to elucidate the mechanisms underlying the human condition and/or to allow a translational pharmacological (or other) evalua- tion of therapeutic tools. Molecular imaging avoids animal sacrifice and permits repetitive (i.e., longitudinal) studies on the same animal which becomes its own control. Small-animal molecular imaging has become an invaluable component of modern biomedical research that will gain probably an increasingly important role in the next few years [10]. The PET is a three-dimensional imaging technique based on nuclear medicine principles to study biological, pharmacological, and physiological function in vivo. The PET imaging involves at least three steps: first, the preparation of radiotracer by chemically incorporating a radionuclide into a molecule targeting either a specific site of action (receptor or enzyme) or a normal biochemical process (e.g., glucose consumption); second, the administration of radiotracer to the subject and subsequent imaging of the radio- tracer activity by the PET scanner; and third, the quantification of PET data into a clinically useful outcome measure by using mathe- matical modeling and computation. On the other hand, the prin- ciples of SPECT involve the emission of gamma-ray photons from the internally distributed radiopharmaceutical that penetrate through the subject body and are detected by a single or a set of collimated radiation detectors. Most of the detectors used in cur- rent SPECT systems are based on a single or multiple thallium doped sodium iodide (NaI(TI)) scintillation detectors. In SPECT, projection data are acquired from different views around the sub- ject. In this chapter, we described the protocols for the molecular imaging of BBB permeability in preclinical models using PET and SPECT. Molecular Imaging of Blood–Brain Barrier Permeability in Preclinical... 331

2 Materials

2.1 Blood–Brain We described the most common BBB disruption by lipopolysac- Barrier Disruption charide (LPS) and BBB opening by focused ultrasound (FUS). Models

2.2 BBB Disruption Rats were anesthetized using isoflurane anesthesia (1.5–2%) and with LPS placed in stereotaxic frame, the body temperature maintained at 37 C with a controlled heating pad system. Then they were intras- triatally injected with LPS at the dose of 50 μgin1μL PBS using the stereotaxic coordinates (A: 0.2 mm, L: 3.0 mm D: 5.0 mm). The needles were not removed before 3 min in order to prevent back diffusion of the injected solutions [11, 12].

2.3 BBB Opening Using a function generator (WF1974; NF Corporation, Yoko- by FUS hama, Japan) with a power amplifier (HSA4101; NF Corporation) a single sine wave was supplied as a burst pulse (frequency: 1 MHz, burst length: 50 ms, repetition frequency: 1 Hz, duty ratio: 5%, total sonication duration: 60 s). The calibrated needle-type hydro- phone (MH-28-10; Eastek Corporation, Tokyo, Japan) that was comparable with 0.49 MPa at the focus was used to measure the acoustic pressure of FUS. On the skull surface cones filled with degassed water and ultrasound gel were mounted (A: À0.2 or À5.0 mm, L: Æ3.0 mm from the bregma) as described elsewhere [22, 23]. A solution of microbubbles (GTS-MB; 100 μL/kg; Nepa Gene Co. Ltd., Ichikawa, Japan) in saline was injected intravenously 5 min after injection of 2% Evans blue (1.5 mL in saline, intrave- nous; Sigma-Aldrich), followed by sonication 15 s later [12–14].

2.4 Anesthesia In order to keep the animal still during imaging almost all PET During Imaging studies utilizes the anesthesia along with maintaining body temper- ature. Thus, anesthetic agent used is considered as confounding variable, particularly in brain-related studies. To minimize this variability it is necessary to characterize the effects of different anesthetics on biological systems and careful selection of anes- thetics required. On the other hand, tracers that are irreversibly trapped can be used to enable the distribution and uptake of tracer while the animal is conscious followed by scanning of the anesthe- tized animal after uptake is complete [15]. Alternatively, completely restrained animal in a tube or body cast or paralyzing drugs that prevent motion but do not interfere with brain function can be used. However, these procedures cause unwanted stress to animals and also lead to highly stressed conditions that can be a confound factor that affects the experimental results. A number of groups are in the early stages of exploring the use of motion detection systems or mounting a ring of counterbalanced detectors directly onto a rat’s head to permit PET studies in freely moving animals [16]. 332 Vijayasree V. Giridharan et al.

2.5 PET Probe There are several effective radiotracers available for studying the and Tracer brain receptors and other protein targets with PET. The potassium analog rubidium-82 (half-life 75 s) that crosses the BBB used as a PET tracer [17]. The [68Ga]EDTA (half-life 68 min) has been employed as a PET tracer to assess BBB permeability in multiple sclerosis [18]. PET probes 18 F-2 and 11C-3 are used in BBB permeability of monkey PET imaging [19]. Similarly, for single photon emission tomography technetium-DTPA or gallium- DTPA ([99mTc/67Ga] DTPA) is used [20]. However, the positron emitter, 68Ga conjugated with ethylenediaminetetraacetic acid with long half-life (67.7 min) is not suitable for repeated measurement of BBB permeability [21, 22]. Therefore, PET probe labeled with short half-live positron emitter is considered more suitable for quantitative imaging of BBB permeability. The suitable features of the PET tracer candidate are: (1) able to be labeled with short-lived positron emitter such as 11C (20.4 min); (2) highly metabolically stable; (3) able to pass BBB at slow rate; and (4) be able to quickly transport unidirectional from blood to the brain. Emerging evi- dence suggests that 14C-labeled 2-aminoisobutyric acid (AIB) has preferable kinetic properties for monitoring BBB permeability [23–27]. Okada et al. recently reported the efficient method for preparing 11C-labeled AIB ([3-11C] AIB) and they also confirmed in vivo stability in rat model [28].

2.6 Radiosynthesis The radiosynthesis of AIB contains the following steps. Base- 11 and Kinetics of 2- promoted α-[ C] methylation of methyl N-(diphenylmethylene)-D, 11 Amino-[3-11C] L-alaniate using iodo[ C] methane following hydrolysis was used to 11 11 Isobutyric Acid (AIB) synthesize [3- C]AIB. At the time of injection to the animal [3- C] AIB specific activity was maintained at 88.0 GBq/μmol and radio- chemical purity was more than 99.7%. In order to confirm whether the [3-11C]AIB accumulation was in accordance with BBB disruption the same brain slices were subjected to ex vivo autoradiography and Evans blue staining. To perform ex vivo autoradiography, rats were intrave- nously injected with 2%, 1.5 mL of Evans blue, 1 day after LPS/PBS administration. Following this, [3-11C] AIB (141–164 MBq) was injected intravenously. At 20 min after [3-11C]AIB administration blood was removed by cardiac perfusion with cold saline (heparin- ized), then the entire brain was snap-frozen and sectioned with 20 μm thickness using cryostat. Then it was exposed to imaging plate for 1 h, then the Evans blue fluorescence radioactivity were detected using bio imaging analyzer and the fluorescence image analyzer. Interested region was selected manually and relative concentration of radioactiv- ity and fluorescence per unit area were measured [28]. The kinetics and in vivo stability evaluation of [3-11C] AIB in the arterial blood demonstrated no radioactive metabolites in the arterial plasma during the experimental period. This in turn depicts Molecular Imaging of Blood–Brain Barrier Permeability in Preclinical... 333

the high in vivo stability of [3-11C] AIB. There were no statistical difference in the kinetics of [3-11C]AIB in the whole blood and plasma [32].

3 Methods

3.1 PET Protocol for According to Okada et al., the procedure for initiation of PET LPS and FUS Induced acquisition involves following steps. Firstly, heating water pump BBB Disruption system was used to maintain body temperature during entire study period. For the PET study of LPS model, rats were injected with [3-11C]AIB (34.2–76.3 MBq) at 24 h after the intrastriatal injection of LPS/PBS. For the PET study of FUS-model, [3-11C] AIB (37.51–46.65 MBq; n ¼ 4/group) was injected at 5 min and 1, 2, 5, and 24 h after exposure to FUS. In LPS model, the Evans blue was injected soon after the PET scanning, and in FUS model, Evans blue was injected before FUS insult; in both models the fluorescent images were taken using Inveon small-animal PET scanner (0.259 mm/pixel, 0.796 mm slice thickness; Siemens Medical Solutions USA, Knoxville, Tennessee, USA) in list mode. The list mode data were histogrammed in 21 frames: 60 s  5, 120 s  5, 180 s  5, and 300 s  6. A transmissions can with a 57Co point source was performed to correct attenuation and the image was reconstructed by filtered back projection. Reconstructed PET images were superimposed on T2-weighted spin echo mag- netic resonance (MR) (TR: 8000 ms, TE: 15 ms) template images of a normal male Sprague–Dawley rat, acquired on a 7.0-T scanner [magnet: Kobelco and Jastec (Kobe, Japan); console: Bruker Avance-I console (Bruker BioSpin, Ettlingen, Germany)], to place the volumes of interest on a transverse view. The radioactivity of 11C was expressed as the mean Æ SD of the standardized uptake value (SUV) after correcting for the physical decay of 11C to the time of injection [28].

3.2 BBB Permeability Figure 1 shows the typical summed PET image from 40 to 60 min Measurement Using and the corresponding Evans blue image along with the time- PET in LPS Model radioactivity curves. As shown in Fig. 1a, b the uptake of [3-11C] AIB was high in LPS-injected side where Evans blue had accumu- lated. But it was not seen in the PBS injected side (control). The time curve Fig. 1c shows that soon after the injection the uptake of [3-11C]AIB shoot up continuously for about 10 min. At the same time in the PBS injected side the uptake was low and constant throughout the experiment. There was significant difference between the PBS and LPS injected side in the uptake values after 2 min (P < 0.02; Fig. 1). The results demonstrate that [3-11C]AIB PET can be used noninvasively to evaluate the BBB permeability status in LPS-injected rats. 334 Vijayasree V. Giridharan et al.

Fig. 1 (a) A typical summed PET image of [3-11C]AIB (36.0 MBq, for 40–60 min); (b) The corresponding Evans blue image; (c) The time radioactivity curves of [3-11C]AIB (34.2–76.3 MBq) in the striatum treated with lipopolysaccharide (LPS) and PBS under isoflurane anesthesia (2%; n ¼ 4). Results expressed as means ÆSDs. [3-11C]AIB significantly accumulated in the LPS side as compared with the PBS side (P < 0.01, repeated-measures analysis of variance). Reprinted from “In-vivo imaging of blood–brain barrier permeability using positron emission tomography with 2-amino-[3-11C] isobutyric acid,” by Okada et al., 2015, Nuclear Medicine Communications, volume 36, p. 1239–48. Reprinted with permission

3.3 BBB Permeability The FUS model is used to evaluate the specificity of [3-11C]AIB in Measurement Using transient BBB leakage. Similar to LPS model the summed SUV PET in FUS Model images (40–60 min) of [3-11C]AIB showed increased accumula- tion of radioactivity in the FUS-sonication, as compared with the contralateral side, at 5 min, 1 2, 5, and 24 h time points after sonication (Fig. 2). At 1 h after [3-11C]AIB injection, the uptake values in the FUS-sonicated region were 0.79, 0.66, 0.56, 0.36, and 0.26 at 5 min and 1, 2, 5, and 24 h elapsed from FUS, respectively (Fig. 2a, b). The ratios of the areas under the time–- radioactivity curves (0–60 min) in the FUS-region to that in the contralateral region were 3.11, 2.62, 2.47, 1.79, and 1.57 at 5 min and 1, 2, 5, and 24 h after FUS exposure, respectively. These results indicated that the level of [3-11C]AIB uptake in the sonicated region decreased as the elapsed time increased. The results Molecular Imaging of Blood–Brain Barrier Permeability in Preclinical... 335

Fig. 2 (a) Typical PET images of [3-11C]AIB (67.8–74.9 MBq, for 40–60 min) at 5 min and 1, 2, 5, and 24 h after focused ultrasound (FUS) sonication. (b) Time–radioactivity curves of [3-11C]AIB [37.51–46.65 MBq under isoflurane anesthesia (2%)] in rats at various elapsed time points after FUS sonication (n ¼ 3–4). Results expressed as means Æ SD. [3-11C]AIB accumulated significantly in the FUS-sonicated side, compared with the contralateral side, at all elapsed time points (P < 0.05, repeated-measures analysis of variance). Reprinted from “In-vivo imaging of blood–brain barrier permeability using positron emission tomography with 2-amino-[3-11C] isobutyric acid,” by Okada et al., 2015, Nuclear Medicine Communications, volume 36, p. 1239–48. Reprinted with permission

demonstrate that the BBB-disrupted region following FUS and the accumulation of [3-11C]AIB was comparable with that of Evans blue, a standard indicator of BBB permeability. In dynamic PET studies, high levels of [3-11C]AIB accumulation were observed in the BBB-opened sides, but not in the control. In FUS model, [3-11C]AIB PET elucidated temporal changes in the BBB status after FUS sonication. These findings suggest that [3-11C]AIB PET could be used to quantitatively determine the BBB permeability status at a high level of sensitivity and might facilitate the optimiza- tion of BBB-opening protocols and the monitoring of temporal changes in BBB permeability [28]. Taken together, this chapter narrated the protocol to evaluate the BBB permeability using [3-11C]AIB as a PET probe in LPS-injected and FUS-sonicated, BBB-disrupted rat models.

3.4 Animal Male rats of Sprague–Dawley strain were used to measure the BBB Preparation for SPECT disruption. To reduce the distortion of the ultrasonic beam, crani- otomy was performed in anesthetized rats. To prevent dehydration prior to the ultrasound saline soaked gauze was used [29]. Briefly, a focused-ultrasound transducer (Imasonics, Besancon, France; diameter ¼ 60 mm, radius of curvature ¼ 80 mm, fre- quency ¼ 1.5 MHz, electric-to-acoustic efficiency ¼ 70%) was used to generate concentrated ultrasound energy. An arbitrary- function generator (33120A, Agilent, Palo Alto, CA, USA) was used to generate the driving signal fed to a radio frequency power 336 Vijayasree V. Giridharan et al.

amplifier (150A100B, Amplifier Research, Souderton, PA, USA) operating in burst mode. A polyvinylidene difluoride type hydro- phone was used to measure pressure distribution along the trans- ducer axis and in the radial direction (Onda, Sunnyvale, CA, USA; calibration range: 50 kHz–20 MHz). Pressure was measured by hydrophone at low output amplitudes and then extrapolated to higher outputs (negative peak pressure amplitude higher than 1.1 MPa) [30]. A second acrylic water tank was used for animal experiments. This tank had a bottom window of 4 Â 4cm2 to allow the entry of ultrasound energy. Animals were placed directly under the water tank with heads tightly attached to the window. The animal brain sonication were conducted with the presence of an ultrasound contrast agent, SonoVue (Bracco, Milan, Italy; SF6 coated with mean diameter ¼ 2.0–5.0 μm), which was injected intravenously before sonication. Each bolus injection contained 0.025 mL/kg of microbubble. Left striatal regions underwent burst-mode sonication for 30 s, with a burst length of 10 ms and a pulse repetition frequency of 1 Hz. Electric powers of 0.5, 1, 3, 5, and 10 W were used with the equivalent negative peak pressure amplitudes of 0.78, 1.1, 1.9, 2.45, and 3.48 MPa, respectively. After sonication, Evans blue was injected intravenously (3 g/mL of 0.9% saline) as a bolus immediately after sonication to observe the BBB disruption from gross brain sections. Evans blue dye has been able to conjugate to plasma albumin, and a number of studies have shown that a direct relationship between the extravasations of dye and albumin into animal cerebral tissues during inflammation [31], focused ultrasound-induced [29, 32] BBB disruption. Peak pressure amplitudes for SPECT/CT imaging and quantitative autoradiography (QAR) analysis ranged from 0.78 to 2.45 MPa.

3.5 BBB Before starting the static animal SPECT imaging the dynamic Measurement Using SPECT protocol was optimized to define the optimal scanning SPECT in FUS Model time. To perform this part, the rats were subjected to focused ultrasound pressure (1.9 MPa). The temperature in scanner room was 24 C and the animals were maintained 37 C using temperature-controlled bed. In the anesthetized (isoflurane; 2%) rats the radiolabeled agent 99mTc-DTPA (257.9 Æ 27.6 MBq/ 0.2 cc) was injected via the carotid artery with a catheter immedi- ately after brain sonication. During the dynamic SPECT protocol, the rats were placed in prone position with their skull opened at the center field of view and consisted of eight frames at a speed of 15/frame. To cover the whole brain, the acquisition protocol con- sists of 24 projections, 128 Â 128 matrix size/projection, 75 s/ projection with a scan range of 3 cm. After the SPECT imaging the reference CT scan was also made using standard-resolution setup integrated (tube voltage of 55 kVp, exposure time of 500 ms, and 180 projections) in the system. To cover the volume of interest Molecular Imaging of Blood–Brain Barrier Permeability in Preclinical... 337

(VOI) the helical pitch was set at 1.0 for helical scanning, using 3 cm axial scanning range. To use in static SPECT imaging the disruption index (DI) and optimal scanning time were measured. The DI is defined as the ratio of the mean counts per voxel in the cerebral lesion area compared to the non-lesion area. The optical scanning time is the time point when DI reached its peak value. Using four different ultrasound powers the animals underwent brain sonication. After the brain sonication, the blue dye was co-injected with the radiolabeled agent 241.2 Æ 25.5 MBq/ 0.2 cc of 99mTc-DTPA via the carotid artery. Results from the pilot dynamic study were used for static brain images. Accordingly, the static brain images were obtained 1.5 h after injection. The temperature of the animal and room were maintained as described in the pilot study. In order to improve the image quality the total scanning time is altered to 30 min; remaining procedures were similar to pilot study protocol. Increased uptake of radioactivity was observed over scalp, bone marrow, and salivary glands from the animal SPECT images. Nota- bly, the uptake was limited in the brain. In BBB-disrupted left striatum notable extravasation of radioactivity was observed. The distribution of 99mTc-DTPA at 15 min post injection was 0.379 Æ 0.018 and 0.243 Æ 0.004% ID (% injected dose)/mL in areas with and without BBB disruption, respectively. The amount of radioactivity accumulation within the resulting pixel size for SPECT and CT were 0.4 and 0.2 mm, respectively. The value of SPECT radioactivity counts represents the decay corrected injected time. The SPECT images along with the co-registered CT images were analyzed using PMOD image analysis software (PMOD Tech- nologies, Zurich, Switzerland). Using the fused SPECT/CT images the whole brain of VOI was manually drawn. The contour of the BBB opening area was determined by an automatic threshold-based algorithm using 20% of the maximum counts within the whole brain VOI. The undisturbed brain area was deter- mined in a similar fashion by an automatic threshold-based algo- rithm using 15% of the maximal counts within the whole brain VOI (Fig. 3). The total image counts within the VOIs were converted to absolute radioactivity by using an efficiency factor determined from the system performance study. The mean radio activities within the VOIs were converted to %ID/mL as normalized to the total injected dose. The radio activities within the BBB leakage lesions at different ultrasound power were determined and compared to the results obtained from the QAR [20].

3.6 Quantitative In PET neuro-receptor imaging, radiotracer activity in the arterial Autoradiography (QAR) or venous blood (input function) and in brain tissue (output response function) is sequentially measured over time after the administration of the selective radiotracer into the blood stream. If the arterial blood sampling is not available, the input function can 338 Vijayasree V. Giridharan et al.

Fig. 3 SPECT images at 1.5 h post injection were co-registered with the corresponding CT images for volume of interest (VOI) delineation. The entire brain VOI (dark blue) was delineated from CT images in the range of skull opening (white arrows). Disrupted (light blue) and non-disrupted VOI (green) were delineated from SPECT images using an automatic threshold-based algo- rithm. The color scale indicates the quantity of 99mTc-DTPA bio distribution. Reprinted from “Quantitative micro-SPECT/CT for detecting focused ultrasound- induced blood–brain barrier opening in the rat,” by Lin et al., 2009, Nuclear Medicine and Biology, volume 36, p. 853–861. Reprinted with permission

also be indirectly inferred from a reference region without target receptors [33]. In a simple region of interest model, a target brain region is identified and a time-activity of tracer for that region is estimated from the images. Regional tracer concentration provides the quantification esti- mates of receptor density. Therefore, the PET data represent tracer concentration (Bq/mL) per unit tissue. In PET studies, the radio- tracer is administered at tracer doses and thereby occupy a negligi- ble (<5–10%) percentage of target sites. The use of tracer doses will give specific binding of target sites, without significantly affecting the total number of available receptors [34]. To validate the SPECT results QAR need to be performed. Autoradiography is a sensitive and simple method of recording spatial distribution of radioisotope-labeled substances within a specimen material. This is a technique in which radioactive mole- cules make their location known by exposing photographic films or Molecular Imaging of Blood–Brain Barrier Permeability in Preclinical... 339

Fig. 4 (a) Brain pictures, (b) Brain sections, (c) SPECT/CT images, (d) QAR results at various pressure amplitudes ranged from 0.78 to 2.45 MPa. Reprinted from “Quantitative micro-SPECT/CT for detecting focused ultrasound-induced blood–brain barrier opening in the rat,” by Lin et al., 2009, Nuclear Medicine and Biology, volume 36, p. 853–861. Reprinted with permission

emulsions [35]. The Fig. 4 clearly demonstrates the blue dye staining of the left brain region in all animals following CT scans. The striatal region was serially sectioned with a thickness of 10 μm in coronal direction as in SPECT imaging and used for QAR. The dried brain sections and the standards were attached to phosphor image plate for 1 day exposure. The exposed plates were scanned by FLA5100 (Fujifilm, Tokyo, Japan) with a 25-μm resolution and 340 Vijayasree V. Giridharan et al.

analyzed using a Multi Gauge version 3.0 software (Fujifilm). The regions of interest were placed on the striatal area in order to measure the regional brain volume (mm2 Â 0.01 mm), either with or without radioactivity extravasation, and the activity was subsequently converted to Bq/mL. The uptake of tracer was then assessed as the percentage of injected 99mTc-DTPA dose/mL of tissue. The cell apoptosis was detected using DNA labeling with digoxigenin-deoxyuridine 50-triphosphate (dUTP) and terminal transferase, followed by immunocytochemical staining with peroxidase-coupled antidigoxigenin antibody and diaminobenzi- dine, was carried out with the reagents supplied in the Apoptag kit (ApopTag kit, Intergen, Purchase, NY, USA) (Fig. 4).

3.7 Limitations Among the few practical issues that confront PET as a technology when applied to small animal imaging, the first one is the cost of performing PET studies. The second major practical issue relates to access to PET tracers. In addition, PET is not available universally due to the need for a cyclotron as well as considerable technical expertise in the field. Further, it depends on the half-life of tracer repeated studies, and accurate quantitative assessment of BBB dis- ruption is not possible using this technique.

4 Conclusion

Positron Emission Tomography (PET) is a noninvasive imaging technique that offers the highest spatial and temporal resolution and has the added advantage that it can allow for true quantification of tracer concentrations in tissues. It is obvious that small animal PET has the capacity to make a valid and unique contribution in the measurement of BBB in vivo. The noninvasive, quantitative, and clinically translatable results from small-animal imaging make it an invaluable component of modern biomedical research. Thus, molecular imaging using PET and SPECT has now been proven and will doubtless form an essential component of applications in BBB permeability in small animal PET measurement.

Acknowledgments

The Translational Psychiatry Program is funded by the Department of Psychiatry and Behavioral Sciences, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth). National Institute for Molecular Medicine (INCT- MM) and Center of Excellence in Applied Neurosciences of Santa Catarina (NENASC). Molecular Imaging of Blood–Brain Barrier Permeability in Preclinical... 341

References

1. Abbott NJ, Ronnback L, Hansson E (2006) (2):370–380. https://doi.org/10.1038/ Astrocyte-endothelial interactions at the blood- jcbfm.2009.223 brain barrier. Nat Rev Neurosci 7(1):41–53. 12. Paxinos PWC (2004) The rat brain in stereo- https://doi.org/10.1038/nrn1824 taxic coordinates, 5th edn. Elsevier Academic 2. Jamieson JJ, Searson PC, Gerecht S (2017) Press, San Diego Engineering the human blood-brain barrier 13. Yang FY, Lin GL, Horng SC, Chang TK, Wu in vitro. J Biol Eng 11:37. https://doi.org/ SY, Wong TT, Wang HE (2011) Pulsed high- 10.1186/s13036-017-0076-1 intensity focused ultrasound enhances the rela- 3. Cardoso FL, Brites D, Brito MA (2010) Look- tive permeability of the blood-tumor barrier in ing at the blood-brain barrier: molecular anat- a glioma-bearing rat model. IEEE Trans Ultra- omy and possible investigation approaches. son Ferroelectr Freq Control 58(5):964–970. Brain Res Rev 64(2):328–363. https://doi. https://doi.org/10.1109/tuffc.2011.1897 org/10.1016/j.brainresrev.2010.05.003 14. Yang FY, Lin YS, Kang KH, Chao TK (2011) 4. Kuhnline Sloan CD, Nandi P, Linz TH, Reversible blood-brain barrier disruption by Aldrich JV, Audus KL, Lunte SM (2012) Ana- repeated transcranial focused ultrasound allows lytical and biological methods for probing the enhanced extravasation. J Control Release 150 blood-brain barrier. Annu Rev Anal Chem (1):111–116. https://doi.org/10.1016/j. (Palo Alto Calif) 5:505–531. https://doi.org/ jconrel.2010.10.038 10.1146/annurev-anchem-062011-143002 15. Kornblum HI, Araujo DM, Annala AJ, Tatsu- 5. Bentivoglio M, Mariotti R, Bertini G (2011) kawa KJ, Phelps ME, Cherry SR (2000) In vivo Neuroinflammation and brain infections: his- imaging of neuronal activation and plasticity in torical context and current perspectives. Brain the rat brain by high resolution positron emis- Res Rev 66(1–2):152–173. https://doi.org/ sion tomography (microPET). Nat Biotechnol 10.1016/j.brainresrev.2010.09.008 18(6):655–660. https://doi.org/10.1038/ 6. Skoog I, Wallin A, Fredman P, Hesse C, 76509 Aevarsson O, Karlsson I, Gottfries CG, Blen- 16. Lasbennes F, Lestage P, Bobillier P, Seylaz J now K (1998) A population study on blood- (1986) Stress and local cerebral blood flow: brain barrier function in 85-year-olds: relation studies on restrained and unrestrained rats. to Alzheimer’s disease and vascular dementia. Exp Brain Res 63(1):163–168 Neurology 50(4):966–971 17. Zunkeler B, Carson RE, Olson J, Blasberg RG, 7. Taheri S, Gasparovic C, Huisa BN, Adair JC, Girton M, Bacher J, Herscovitch P, Oldfield Edmonds E, Prestopnik J, Grossetete M, Shah EH (1996) Hyperosmolar blood-brain barrier NJ, Wills J, Qualls C, Rosenberg GA (2011) disruption in baboons: an in vivo study using Blood-brain barrier permeability abnormalities positron emission tomography and rubidium- in vascular cognitive impairment. Stroke 42 82. J Neurosurg 84(3):494–502. https://doi. (8):2158–2163. https://doi.org/10.1161/ org/10.3171/jns.1996.84.3.0494 strokeaha.110.611731 18. Velikyan I (2013) Prospective of (6)(8)Ga- 8. Wallin A, Blennow K, Fredman P, Gottfries radiopharmaceutical development. Theranos- CG, Karlsson I, Svennerholm L (1990) Blood tics 4(1):47–80. https://doi.org/10.7150/ brain barrier function in vascular dementia. thno.7447 Acta Neurol Scand 81(4):318–322 19. Doi H, Sato K, Shindou H, Sumi K, 9. Larobina MBA, Salvatore M (2006) Small ani- Koyama H, Hosoya T, Watanabe Y, Ishii S, mal PET: a review of commercially available Tsukada H, Nakanishi K, Suzuki M (2016) imaging systems. Curr Med Imaging Rev 2 Blood-brain barrier permeability of ginkgolide: (2):187–192 comparison of the behavior of PET probes 10. Cuccurullo V, Di Stasio GD, Schilliro` ML, 7alpha-[(18)F]fluoro- and 10-O-p-[(11)C] Mansi L (2016) Small-animal molecular imag- methylbenzyl ginkgolide B in monkey and rat ing for preclinical cancer research: .μPET and μ. brains. Bioorg Med Chem 24(21):5148–5157. SPECT. Curr Radiopharm 9(2):102–113 https://doi.org/10.1016/j.bmc.2016.08.032 11. Aid S, Silva AC, Candelario-Jalil E, Choi SH, 20. Lin KJ, Liu HL, Hsu PH, Chung YH, Huang Rosenberg GA, Bosetti F (2010) WC, Chen JC, Wey SP, Yen TC, Hsiao IT Cyclooxygenase-1 and -2 differentially modu- (2009) Quantitative micro-SPECT/CT for late lipopolysaccharide-induced blood-brain detecting focused ultrasound-induced blood- barrier disruption through matrix metallopro- brain barrier opening in the rat. Nucl Med teinase activity. J Cereb Blood Flow Metab 30 342 Vijayasree V. Giridharan et al.

Biol 36(7):853–861. https://doi.org/10. emission tomography with 2-amino-[3-11C] 1016/j.nucmedbio.2009.04.011 isobutyric acid. Nucl Med Commun 36 21. Kessler RM, Goble JC, Bird JH, Girton ME, (12):1239–1248. https://doi.org/10.1097/ Doppman JL, Rapoport SI, Barranger JA mnm.0000000000000385 (1984) Measurement of blood-brain barrier 29. Liu HL, Wai YY, Chen WS, Chen JC, Hsu PH, permeability with positron emission tomogra- Wu XY, Huang WC, Yen TC, Wang JJ (2008) phy and [68Ga]EDTA. J Cereb Blood Flow Hemorrhage detection during focused- Metab 4(3):323–328. https://doi.org/10. ultrasound induced blood-brain-barrier open- 1038/jcbfm.1984.48 ing by using susceptibility-weighted magnetic 22. Schlageter NL, Carson RE, Rapoport SI (1987) resonance imaging. Ultrasound Med Biol 34 Examination of blood-brain barrier permeabil- (4):598–606. https://doi.org/10.1016/j. ity in dementia of the Alzheimer type with ultrasmedbio.2008.01.011 [68Ga]EDTA and positron emission tomogra- 30. Hynynen K (1991) The threshold for ther- phy. J Cereb Blood Flow Metab 7(1):1–8. mally significant cavitation in dog’s thigh mus- https://doi.org/10.1038/jcbfm.1987.1 cle in vivo. Ultrasound Med Biol 17 23. Washburn LC, Blair LD, Byrd BL, Sun TT (2):157–169 (1985) Comparison of 68Ga-EDTA, [1-11 31. Udaka K, Takeuchi Y, Movat HZ (1970) Sim- C]alpha-aminoisobutyric acid, and [99mTc] ple method for quantitation of enhanced vas- sodium pertechnetate in an experimental cular permeability. Proc Soc Exp Biol Med 133 blood-brain barrier lesion. Int J Nucl Med (4):1384–1387 Biol 12(4):267–269 32. Hynynen K, McDannold N, Vykhodtseva N, 24. Knight RA, Nagaraja TN, Ewing JR, Nagesh V, Jolesz FA (2001) Noninvasive MR imaging- Whitton PA, Bershad E, Fagan SC, Fensterma- guided focal opening of the blood-brain barrier cher JD (2005) Quantitation and localization in rabbits. Radiology 220(3):640–646. https:// of blood-to-brain influx by magnetic resonance doi.org/10.1148/radiol.2202001804 imaging and quantitative autoradiography in a 33. Laruelle M, Slifstein M, Huang Y (2002) Posi- model of transient focal ischemia. Magn Reson tron emission tomography: imaging and quan- Med 54(4):813–821. https://doi.org/10. tification of neurotransporter availability. 1002/mrm.20629 Methods 27(3):287–299 25. Blasberg RG, Fenstermacher JD, Patlak CS 34. Innis RB, Cunningham VJ, Delforge J, (1983) Transport of alpha-aminoisobutyric Fujita M, Gjedde A, Gunn RN, Holden J, acid across brain capillary and cellular mem- Houle S, Huang SC, Ichise M, Iida H, Ito H, branes. J Cereb Blood Flow Metab 3(1):8–32. Kimura Y, Koeppe RA, Knudsen GM, Knuuti J, https://doi.org/10.1038/jcbfm.1983.2 Lammertsma AA, Laruelle M, Logan J, 26. Weyerbrock A, Walbridge S, Saavedra JE, Kee- Maguire RP, Mintun MA, Morris ED, fer LK, Oldfield EH (2011) Differential effects Parsey R, Price JC, Slifstein M, Sossi V, of nitric oxide on blood-brain barrier integrity Suhara T, Votaw JR, Wong DF, Carson RE and cerebral blood flow in intracerebral C6 (2007) Consensus nomenclature for in vivo gliomas. Neuro-Oncology 13(2):203–211. imaging of reversibly binding radioligands. J https://doi.org/10.1093/neuonc/noq161 Cereb Blood Flow Metab 27(9):1533–1539. 27. Leroy C, Roch C, Koning E, Namer IJ, Nehlig https://doi.org/10.1038/sj.jcbfm.9600493 A (2003) In the lithium-pilocarpine model of 35. Ogasawara K, Ito H, Sasoh M, Okuguchi T, epilepsy, brain lesions are not linked to changes Kobayashi M, Yukawa H, Terasaki K, Ogawa A in blood-brain barrier permeability: an autora- (2003) Quantitative measurement of regional diographic study in adult and developing rats. cerebrovascular reactivity to acetazolamide Exp Neurol 182(2):361–372 using 123I-N-isopropyl-p-iodoamphetamine 28. Okada M, Kikuchi T, Okamura T, Ikoma Y, autoradiography with SPECT: validation Tsuji AB, Wakizaka H, Kamakura T, Aoki I, study using H2 15O with PET. J Nucl Med Zhang MR, Kato K (2015) In-vivo imaging of 44(4):520–525 blood-brain barrier permeability using positron Chapter 20

Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier Injury and Damage Measurement

Pavani Sayana, Jean Pierre Oses, Tatiana Barichello, and Vijayasree V. Giridharan

Abstract

Blood–brain barrier (BBB) is a highly selective semipermeable membrane with tight junctions formed from closely wedged epithelial cells and is the principal regulator of exchange of materials between the blood and the brain, simultaneously protecting it from the systemic insults. Any impairment in the functional coordi- nated effort or formation of the components of the BBB neurovascular bundle may lead to neurological disturbances from entry of extraneous elements. The BBB disruption causes an imbalance in tight junction proteins, adherens junction proteins, matrix metalloproteinases, cell adhesion molecules, and a myriad of neuroinflammatory proteins. Therefore, the modalities, which can enable the accurate detection of BBB breach in advance via protein biomarker assay to underpin targeted medicines, are of growing priority. In this chapter, we will summarize the BBB pathology measurement methods used in the detection of microvascular protein biomarkers such as Western blot, immunohistochemistry (IHC), enzyme-linked immunosorbent assay (ELISA), and real-time polymerase chain reaction (RT-PCR). In addition, we also emphasize on the BBB genomics and the method used to detect functionally related BBB genes in preclinical model.

Key words Blood–brain barrier, Brain microvasculature, Tight junction proteins, Western blot, Immunohistochemistry, Elisa and RT-PCR

1 Introduction

The blood–brain barrier (BBB) is formed by continuous endothe- lial cells of brain microvascular network and it is one of the tightest barriers in higher organisms that maintains the cerebral homeosta- sis. Anatomically, BBB is composed of microvascular endothelial cells, pericytes, and the foot-processes of astrocytes together they are responsible for the transport between the systemic circulation and brain microenvironment [1]. The key component of the BBB integrity is the formation of tight junctions. Tight junctions (TJ) are composed of transmem- brane and cytoplasmic proteins linked to an actin-based

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1_20, © Springer Science+Business Media, LLC, part of Springer Nature 2019 343 344 Pavani Sayana et al.

cytoskeleton that allows them to form a seal and rapid modulation and regulation centrally. The TJ which creates the intracellular barrier conceptualized into compartments with the transmembrane barrier proteins such as claudins, occludin, junction adhesion mol- ecule (JAM), peripheral scaffolding proteins such as zonula occlu- dens (ZO) protein, and afadin [1]. The integral membrane protein claudins form the primary seal of the TJ by binding homo-typically on adjacent endothelial cells. The regulatory protein, occludin correlated with increased electrical resistance across the barrier and decreased para-cellular permeability [2]. The JAM, member of immunoglobulin superfamily is also localized at the TJ which can function in association with platelet endothelial cellular adhesion molecule 1 (PECAM-1) to regulate leukocyte migration [3]. For the structural support TJ are also made up of several other proteins such as ZO-1 to 3 and cingulin. The ZO-1 to 3 [4] belong to a family of proteins known as membrane-associated guanylate kinase- like proteins [5]. Several studies suggest that levels of different types of TJ proteins were altered during BBB injury [6–8]. Apart from the TJ proteins, the other markers altered during BBB injury include endothelial inflammatory markers such as adhe- sion molecules and chemokines. Upon injury in BBB, the cerebral endothelial cells express adhesion molecules and chemokines and their potential to act as antigen presenting cells and participate in death receptor signaling are intricately involved in inflammatory processes. The clinical and preclinical evidences demonstrated increased expression of intercellular adhesion molecule 1 (ICAM)-1, vascular cell adhesion protein (VCAM)-1, PECAM- 1, and E-selectin levels on brain microvascular endothelial cells after BBB injury [9, 10]. The subgroups of small cytokines otherwise called chemokines that showed increased expression during in vitro BBB injury includes monocyte chemoattractant protein-1 (MCP-1), regulated on activation, normal T cell expressed and secreted (RANTES), and macrophage inflammatory protein (MIP) 1β. Further, pericytes, a minor population that act as a macrophage, contribute numerous immune functions and get altered during BBB injury. The levels of major histocompatibil- ity complex (MHC) II antigens, B7 (peripheral membrane protein) and cluster of differentiation (CD) 40, and Fc receptor are increased on perivascular macrophages on central nervous system (CNS) inflammation [11, 12]. Additionally, the glial markers CD11B, glial fibrillary acidic protein (GFAP) were found to be increased after BBB breakdown [13]. Recent studies over the past decades developed gene microar- ray technologies to the brain microvasculature unit which is called as BBB genomics. The BBB genomics reveals the BBB-enriched genes that are expressed at very high levels at the BBB that can be detected in a whole-brain gene microarray [14]. Additionally, stud- ies have also demonstrated the genes encoding proteins shown to be involved in BBB function that can be categorized under Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 345

junctional proteins, transporters, and metabolizing enzymes [15]. All the above given markers are detected by means of different detecting modules. The main detecting modules include Western blot, immunohistochemistry (IHC), ELISA, and RT-PCR. Thus, in this chapter, we emphasized on the different detection system used to identify the protein markers altered during BBB injury.

2 Detection of Brain Microvascular Proteins by Western Blot

2.1 Materials Tissue culture flask or dish is used on ice to hold the tissue intact. The platform is set with a centrifuge and micro centrifuge tubes to 2.1.1 Tissue Lysis centrifuge the cells, then lysed with lysis buffer or radioimmuno- and Protein Extraction precipitation assay buffer (RIPA buffer) (50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 1% nonidet P-40 (NP-40) or 0.1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 1 mM sodium orthovanadate, 1 mM NaF) along with protease inhibitor tablet (Roche). Protein concentration is measured with bicinchoninic acid (BCA) assay reagents.

2.1.2 Protein Separation The stage is set with 0.1% SDS, 10% polyacrylamide gel electropho- by SDS-PAGE Gel resis (SDS-PAGE) unit along with a mini-protean 3 electrophoresis Electrophoresis cell (BioRad), both supplied with a constant voltage power supply, power pac 300 (BioRad). Three buffers, running buffer or sample buffer (25 mM Tris, 190 mM glycine, 0.1% SDS), loading buffer or mercaptoethanol solution (10% 2-mercaptoethanol, 80% sample buffer, 0.004% bromophenol blue), and electrophoresis buffer (15.0 g Tris base, 72.0 g glycine, 5.0 g SDS, distilled H2O to 1000 mL) along with polyacrylamide gel are used in electrophoresis.

2.1.3 Protein Transfer The protein gel is transferred onto the polyvinylidene fluoride from the Gel membrane (PVDF) or to nitrocellulose membrane with the use of to the Membrane Whatman no. 2 filter papers, a transfer electrophoresis unit with coolant (chamber, sandwich holder, Scotch-Brite-like pads) and a mini trans-blot electrophoretic transfer cell (BioRad). During the transfer process, the transfer buffer components are wet with the transfer buffer (25 mM Tris base, 190 mM glycine, 20% methanol).

2.1.4 Immunoblotting The setting was with a container for membrane incubation and and Antibody Incubation washing, and a bench top shaker/rotating platform. Polyclonal antibodies as primary ones and horseradish peroxidase (HRP)- conjugated, species-specific antibodies as secondary antibodies and three buffers, Ponceau S staining buffer (0.2% (w/v) Ponceau S, 5% glacial acetic acid), tris-buffered saline with tween 20 (TBS-T) buffer (1 M Tris–HCL, pH 7.5, 1.5 mM NaCl, 0.1% Tween 20), and blocking buffer (3% bovine serum albumin (BSA) in TBST), are used during the process. 346 Pavani Sayana et al.

Western blot analysis of TJ proteins was done by using the primary antibodies, anti-occludin (1:1000) [6–8], anti-claudin-1 (1:1000), anti-claudin-3 (1:1000), and anti-claudin-5 (1:500) [8, 16, 17]. In a similar manner, the primary antibodies used for ZO-1, JAM-1, connexin 43, and epithelial membrane protein-1 (EMP-1) were anti-ZO-1 (1:500) [6, 18] anti-JAM-1 [8], anti- connexin-43 (1:1000) [7], and anti-EMP1 (1:500) [19]. The matrix metalloproteinases (MMP) were measured with primary antibodies such as anti-MMP-2 (1:300) and anti-MMP-9 (1:1000) [18]. Adherens junction proteins (AdJ) were measured with Anti-β-catenin antibody (1:1000), anti-phosphorylated β-catenin antibody (1:1000), anti-VE-cadherin (1 μg/mL), and anti-E-cadherin antibody (1:1000) [17]. Later on all the proteins are treated with species-appropriate HRP-conjugated secondary antibody (1:5000). Western blot assay of TJ proteins like claudin- 5, occludin, ZO-1, and AdJ proteins such as p120-catenin and β-catenin is shown in Fig. 1a, where GAPDH was used as the loading control and for band density normalization. On usage of exogenous basic fibroblast growth factor (bFGF) in traumatic brain injury (TBI) mouse with BBB damage, the level of protein expres- sion noted as optical density on Western blot for TJ proteins (claudin-5, occludin, and -ZO-1) increased considerably when compared to sham and TBI groups ( p < 0.05) (Fig. 1b)[20].

Fig. 1 (a) Protein expression of claudin-5, occludin, zonula occludens-1, p120-catenin, and β-catenin for the sham, TBI, and bFGF treatment groups. GAPDH was used as the loading control and for band density normalization. (b) The optical density analysis of claudin-5, occludin, and zonula occludens-1, *P < 0.05 versus the sham group. #P < 0.05 versus the TBI group. Data are the mean values Æ SEM, n ¼ 6. Reprinted from “bFGF Protects against Blood-Brain Barrier Damage through Junction Protein Regulation via PI3K-Akt- Rac1 Pathway Following Traumatic Brain Injury”, Wang et al., 2016, Molecular Neurobiology, volume 53. p 7298–311. Reprinted with permission. TBI traumatic brain injury, bFGF basic fibroblast growth factor, GAPDH glyceraldehyde 3-phosphate dehydrogenase, ZO-1 zonula occludens-1 Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 347

2.1.5 Imaging and Data For chemiluminescence (ECL), peroxidase substrate enhanced Analysis (ECL) kit, and to capture the chemiluminescent signals, a charged-coupled device (CCD) camera-based imager are used. Data analysis and interpretation of results is performed with Image J software.

2.2 Methods Wash tissue by adding cold phosphate buffered saline (PBS) and rocking gently by placing the tissue culture flask on ice. Dislodge 2.2.1 Tissue Lysis cells with a cell scraper and pipette into micro centrifuge tubes to and Protein Extraction centrifuge at 1500 rpm for 5 min. Discard the supernatant and add 180 μL lysis buffer to each well supplemented with 20 μL protease inhibitor cocktail. Incubate for 15 min at 4 C under agitation, scrape wells with a pipette tip and collect the homogenate in 1.5 mL tubes and centrifuge the tubes at 11,600 rpm for 10 min at 4 C. Store the supernatant (protein mix) in clean tubes at À80 C until processing. Measure the protein concentration by the bicinchoninic acid assay following the standard protocol or by using a spectrophotometer [21].

2.2.2 Protein Separation Resolve the 30 μg of total protein lysate for each sample by the by SDS-PAGE Gel standard technique of SDS-PAGE. Run samples with 10 mM Electrophoresis β-mercaptoethanol (reducing conditions) or without (non-reducing conditions) depending on the antibody used to detect the protein of interest. Load equal amounts of protein (20 μg) into wells of SDS-PAGE gel, along with weight markers (6 μg). Run the gel for 5 min at 50 V for separating gel and increase the voltage to 100–150 V for stacking the gel to finish the run in about 1 h (Pierce BCA Assay kit).

2.2.3 Protein Transfer On separation of solubilized proteins, immobilize them on stable from the Gel surface like a membrane before antibody application. Cut four filter to the Membrane sheets and one PVDF membrane/Nitrocellulose membrane the same dimensions as the gel. Wet the components with the transfer buffer and place the gel into the PVDF membrane in 1Â transfer buffer for 10–15 min. Assemble the transfer sandwich by layering Scotch-Brite-like pad, two sheets of Whatman filter paper, acrylam- ide, gel, nitrocellulose membrane, two sheets Whatman filter paper, and a second Scotch-Brite-like pad and make sure no air bubbles are trapped in the sandwich. Place the sandwich in the transfer chamber with the PVDF membrane between the positive electrode and the gel. Fill the chamber with cold transfer buffer at 4 C and connect to power source and initiate the run at 20 mA for 90 min [22, 23].

2.2.4 Immunoblotting Briefly rinse the blot in water and stain it with Ponceau S solution to and Antibody Incubation check the transfer quality. Rinse off the Ponceau S satin with three washes using TBS-T. Block in 5% BSA in TBST at room tempera- ture for 1 h. Incubate overnight with primary antibody at 4 C. 348 Pavani Sayana et al.

Then rinse the blot 3–5 times for 5 min with TBST. Incubate the membrane under agitation (30 min at room temperature) with 10 mL of HRP-conjugated secondary antibody diluted using TBST with 5% skim milk for 2 h at room temperature. Rinse the blot 3–5 times for 5 min with TBST [21].

2.2.5 Imaging and Data Incubate the membrane for 1 min with 0.1 mL/cm2 of the peroxi- Analysis dase substrate enhanced ECL. Apply the ECL system to the blot and incubate the membrane for 1–2 min. Capture the chemilumi- nescent signals using a CCD camera-based imager or visualize the result in a dark room using a cassette to expose the membrane. Open the sample in the Image J software and select the band of interest. Represent the results with respect to loading protein con- trol (β-actin, tubulin, etc.).

2.3 Application The BBB disruption causes an imbalance in the TJ proteins such as occludin-1, claudin-1, claudin-3, claudin-5, ZO-1, JAM-1, gap junction alpha-1 (GJA-1) or connexin 43 and EMP-1 [8]. It also causes a disturbance in MMPs that are involved in the breakdown of extracellular matrix such as MMP-2 and MMP-9 [18] and cell adhesion molecules such as VCAM, ICAM, E-selectins, and L-selectins and AdJ proteins such as cadherins and catenins [8, 17, 24]. All the above proteins can be measured with Western or immunoblot assay.

3 Detection of Brain Microvascular Proteins by IHC

3.1 Materials Brain tissue is fixed in 4% paraformaldehyde solution, and if frozen tissue is desired, the sections are placed in 30% sucrose in 0.1 M 3.1.1 Tissue Fixation phosphate buffer.

3.1.2 Tissue Processing The platform is set with an automatic processor that made use of by Dehydration solvents such as increasing concentrations of ethanol (80, 95, and 100%), detergent like xylene to wash the tissue and paraplast (wax) plus tissue embedding medium to embed the tissue.

3.1.3 Slide Coating, The Fisher color frost slides or other glass slides coated with 2% of Paraffin Sectioning, 3-aminopropyltriethoxysilane in acetone and sta-on tissue section and Deparaffinization adhesive to stick tissue to slides, and solvents such as xylene and decreasing concentrations of ethanol (90, 70, and 50%) are used for slide coating and deparaffinization.

3.1.4 Second Slide About 0.005% poly-L-lysine hydrobromide in 0.1 M Tris–HCl, pH Coating and Frozen 8.0 is used for second coating and the slides are placed in a desicca- Sections tor overnight to protect from humidity before frozen sections are made. Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 349

3.1.5 Antigen Retrieval There are two methods of antigen travel followed: the heat-induced epitope retrieval (HIER) or enzyme-induced epitope retrieval (EIER). In EIER, protease is the most common enzyme used, hence also called protease-induced epitope retrieval (PIER) [25].

Heat-Induced Epitope The stage is arranged with a container with 500 mL of 0.01 M Retrieval sodium citrate buffer, pH 6.0, covered with aluminum foil and a hot plate to boil the container and 0.1 M PBS used to cool it.

Protease-Induced Epitope PIER technique uses 0.5% pepsin, 0.01 M HCL heated in a micro-  Retrieval wave and a water bath heated to 37 C.

3.1.6 Reducing About 0.3% methanolic peroxide and PBS washes are used to block Nonspecific Immune endogenous peroxidase, while other commonly used blocking buf- Staining fers include normal serum, BSA, or gelatin.

3.1.7 Antigen-Antibody The equipment required are a humid chamber, glass staining dishes Reaction that hold 10–50 slides or plastic racks that hold up to 20 slides, pipettes shaker to wash slides, PBS adjusted to pH 7.4 and 1 L volume, 0.05 M Tris–HCl, pH 7.6, antibody diluting buffer, nor- mal serum diluted in PBS, obtained from the same species as the secondary biotinylated link antibody, streptavidin-conjugated per- oxidase antibody, DAB as chromogen, and 1% H2O2. The IHC analysis of TJ proteins is conducted by using primary antibodies such as anti-occludin (1:500, Invitrogen) (1:100, Zymed) [8, 18, 19], anti-claudin-1 (Alexa Fluor 488), anti-clau- din-3 and anti-claudin-5 (1:250, Abcam) (1:50, Invitrogen) (1:500, Invitrogen) [8, 13, 16], anti-ZO-1 (1:200, Invitrogen) (1:100, Invitrogen) (Alexa Fluor 594) [18, 26], anti-JAM-A [27], and anti-EMP1 (1:300) [19]. The MMPs were measured with primary antibodies anti-MMP-2 (1:300, Millipore) and anti- MMP-9 (1:1000, Millipore) [18]. The AdJ proteins were measured with anti-β-catenin antibody, anti-VE-cadherin [17] and PECAM with anti-PECAM antibody (1:250, Millipore) [13]. Some of the inflammatory markers from BBB damage were analyzed using pri- mary antibodies such as anti-CD11B (1:250) [13], anti-CD31 (1:100) [4], anti-CD144 (1:100), anti-nerve growth factor (anti- NGF) (1:100, Chemicorn), anti-platelet derived growth factor (anti-PDGF) (1:1000, Chemicorn), anti- ciliary neurotrophic fac- tor (anti-CNTF) (1:2000, Santa), anti-tumor growth factor(TGF)- β1 (1:20,000, Chemicorn), anti-fibroblast growth factor (anti- FGF) (1:100, Santa), anti-growth associated protein (anti-GAP) (1:10,000, Santa) [28], anti-GFAP (1:250, Life Technologies) (1:400, Sigma-Aldrich) [13, 18], and anti-oligodendrocyte specific protein (anti-OSP) (1:100, BD Biosciences) [13]. Dual-label immunofluorescence staining results of endothelial cell marker, 350 Pavani Sayana et al.

Fig. 2 Dual-label immunofluorescence staining results of endothelial cell marker (a) CD31 (red) and different TJ proteins (b) Claudin-5 (c) Occludin, and (d) ZO-1 in the mouse brain 1 day after TBI. The nuclei are labeled by Hoechst. Scale bar ¼ 10 μm. Magnification was Â40. Reprinted from “bFGF Protects against Blood-Brain Barrier Damage through Junction Protein Regulation via PI3K-Akt-Rac1 Pathway Following Traumatic Brain Injury”, Wang et al., 2016, Molecular Neurobiology, volume 53. p 7298–311. Reprinted with permission. TBI traumatic brain injury, bFGF basic fibroblast growth factor, GAPDH glyceraldehyde 3-phosphate dehydroge- nase, ZO-1 zonula occludens-1

CD3 and TJ proteins such as claudin-5, occludin, and ZO-1 are presented in Fig. 2 from the TBI mouse brain 1 day after injury [20].

3.1.8 Counterstaining The stage is set with Hematoxylin (Harris formula) stain, acid and Mounting alcohol solution as a dip and distilled water, ethanol and xylene for washes. The stained tissue is placed on a slide and covered with coverslip after applying permount solution.

3.1.9 Amplification This uses tyramide signal amplification kit that contains Streptavi- Techniques din HRP, blocking agent, amplification diluent, and biotinyl tyramide.

3.1.10 Dual Labeling Dual labeling requires AEC Chromogen Kit (2.5 M acetate buffer, Using Indirect pH 5.0, 3-amino-9 ethylcarbazole (AEC) in N, N-dimethyl-form- Immunoperoxidase amide and 3% H2O2 in deionized water) in addition to DAB Technique chromogen kit, Hematoxylin stain, and crystal mount for slides. Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 351

3.1.11 Dual Labeling This method makes use of fluorochrome-conjugated biotinylated Using Different secondary antibody or fluorochrome-conjugated streptavidin and Fluorochromes mowiol mounting medium (add 2.4 g of mowiol to 6 g of glycerol, stir well and add 6 mL of water and leave for several hours at room temperature).

3.2 Methods Brain is immersed in 4% paraformaldehyde solution at room tem- perature and after 2 h is cut into 2 mm sections and placed in the 3.2.1 Tissue Fixation same fixative overnight at room temperature, until paraffin embed- ding [29]. If frozen sections are required, sections are placed in 30% sucrose in 0.1 M phosphate buffer at 4 C.

3.2.2 Tissue Processing The tissue is processed in an automatic processor using increasing by Dehydration concentrations of ethanol washes (80% ethanol for 30 min, two changes of 95% ethanol for 30 min and three changes of ethanol for 40 min) and then cleared with a detergent like xylene (two changes for 50 min) and finally embedded in wax (Paraplast) at 60 C for 30 min [30].

3.2.3 Slide Coating, All slides are coated with 2% of 3-aminopropyltriethoxysilane in Paraffin Sectioning, acetone, except Fisher color frost slides which are exempt from the and Deparaffinization coating need. Sta-On tissue section adhesive is used to ensure that the sections attach to the slides [31]. Then place the slides in an oven at 58–60 C for 30–60 min [29]. Paraffin sections are then deparaffinized with xylene two changes for 5 min, and decreasing concentrations of ethanol washes (absolute ethanol two changes for 3 min, 90% ethanol for 3 min, 70% ethanol for 2 min and 50% ethanol for 2 min) [31].

3.2.4 Second Slide To withstand washing off during staining, slides are coated with Coating and Frozen 0.005% poly-L-lysine hydrobromide in 0.1 M Tris–HCl, pH 8.0 for Sections 30 min. Slides are placed in a desiccator overnight and are cut into 15–18 μm frozen sections that are stored at À20 CorÀ70 C for several months [32].

3.2.5 Antigen Retrieval Slides rack is submerged in a container with 500 mL of 0.01 M sodium citrate buffer, pH 6.0, covered with aluminum foil and Heat-Induced Epitope boiled on a hot plate for 10 min and cooled on countertop for Retrieval 20 min and later in 0.1 M PBS [25, 33].

Protease-Induced Epitope Heat 0.01 M HCl in a microwave for 10 s, add 0.5% pepsin and  Retrieval place the container in a 37 C water bath for 30 min [25].

3.2.6 Reducing Nonspecific immune staining can be blocked by suppressing Nonspecific Immune endogenous biotin, peroxidase, and other enzymes. Endogenous Staining peroxidase is blocked with 0.3% methanolic peroxide for 20–30 min depending on the number of red cells and then washed twice with PBS for 3 min. Other common blocking buffers include normal serum, BSA, or gelatin [30]. 352 Pavani Sayana et al.

3.2.7 Antigen-Antibody Apply normal serum diluted in PBS for 15 min in a humid chamber Reaction after removing excess buffer. Shake off the normal serum and apply primary antibody diluted in diluting buffer for 2–4 h at room temperature. Rinse with PBS three times for 3 min and apply biotinylated link antibody for 30 min at room temperature after shaking off excess buffer. Then apply peroxidase-conjugated strep- tavidin, 1300 D for 30 min at room temperature and wash with two changes of 0.05 M Tris–HCl, pH 7.6 for 3 min and add slides for 5 min to 0.75 mL of 1% H2O2 mixed with 50 mL of DAB solution and wash with tap water for 3 min with intermittent shaking [34–37].

3.2.8 Counterstaining Stain sections with Hematoxylin (Harris formula) for 10–20 s and and Mounting wash with water. Dip slides in acid alcohol solution and wash with water if the nuclei are dark. Later wash three times with distilled water, two changes of ethanol, and two changes of xylene for 5, 2, and 2 min, respectively, and place a coverslip after applying a large drop of permount on the slide [31].

3.2.9 Amplification Tyramide amplification technique is used to increase the sensitivity Techniques of antigenic detection, when the antigen levels are low [38, 39].

3.2.10 Result When DAB is used as the substrate, antigen site is stained chocolate Interpretation brown color and the nuclei are stained light blue. Quantitative analysis of immunoreactivity is expressed manually as a score of 1+ to 5+, where 1+ is sparse, 2+ is mild, 3+ is moderate, 4+ is marked, and 5+ is maximal staining [39] or it can be done by computerized densitometry [31, 40].

3.2.11 Dual Labeling Dual labeling is done with two different chromogenic substrates Using Indirect such as 3,3’-diaminobenzidine (DAB) and 3-amino-9-ethylcarba- Immunoperoxidase zole (AEC). The AEC precipitates as an insoluble bright-red sub- Technique stance when exposed to peroxidase and hydrogen peroxide. After detecting the first antigen, to detect second antigen, use AEC substrate reagent for 8–10 min, wash with distilled water and place in water for 5 min, stain with Hematoxylin and wash with tap water, place a crystal mount and dry in the oven for 20 min at 60 C. The first antigen appears chocolate brown and the second one appears bright red in color [35–37].

3.2.12 Dual Labeling Fluorescence microscopy is a high-resolution method to detect two Using Different antigens in a single cell. For detecting first antigen, follow the Fluorochromes standard protocol, then shake off excess buffer and apply biotiny- lated secondary antibody for 30 min at room temperature. Wash slide three times with PBS for 5 min, remove excess buffer and apply Streptavidin linked to a fluorochrome and incubate at room temperature for 15–60 min and wash slides three times with PBS for 5 min. For detection of secondary antigen, shake off excess Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 353

buffer and apply normal serum for 15 min, remove normal serum and apply primary antibody for 2–4 h, both at room temperature, then wash with PBS, apply biotinylated secondary antibody, wash again and apply streptavidin linked to fluorochrome, wash again and mount using mowiol mounting medium and save the slides at 4 C. Slides are interpreted in a fluorescence microscope or a laser scanning confocal microscope, using optical filter sets [41].

3.3 Application The BBB disruption causes an impairment in the TJ proteins [7, 8, 19], MMP [18], cell adhesion molecules [17, 24] and inflamma- tory mediators [28] which can all be measured on IHC analysis.

4 Detection of Brain Vascular Proteins by ELISA

4.1 Materials For plate preparation, 96-well microplate that is coated with a capture antibody and a blocking buffer is required. Capture anti- 4.1.1 Plate Preparation body is made by diluting each vial with 0.5 mL of PBS (distilled by Coating with Capture water, 137 mM NaCl, 2.7 mM KCl, 8.1 mM Na HPO , 1.5 mM Antibody and Blocking 2 4 KH PO , pH 7.2–7.4, 0.2 μM filtered) without carrier protein to Buffer 2 4 the working concentration. After each addition, 0.05% Tween 20 in PBS, pH 7.2–7.4 is used as a wash buffer. Wash buffer is filled into the plate wells by using a squirt bottle, manifold dispenser or auto- washer.

4.1.2 Addition of Sample About 100 μL of sample with antigen or standards in reagent and Standards diluent are added per well. Recombinant human standards are prepared by reconstituting each vial with 0.5 mL of reagent diluent (1% BSA in PBS, pH 7.2–7.4, 0.2 μm filtered), a seven point standard curve using twofold serial dilutions in reagent diluent and then prepare 1000 μL of high standard per plate assayed at 2000 pg/mL.

4.1.3 Addition of Primary About 100 μL of the primary or detection antibody diluted in or Detection Antibody reagent diluent is added to each well. Biotin-labeled detection antibodies are reconstituted with 1.0 mL of reagent diluent and later diluted in PBS without carrier protein to the working concen- tration. ELISA analysis of adhesion molecules (ICAM-1, VCAM-1, E-selectin, and L-selectin) was made by use of ELISA kits from R&D systems Europe Ltd., Abington, UK, or R&D systems, Min- neapolis, MN, USA, whereas for metalloproteinases, commercially available ELISA kits from Biotrak, Arlington Heights, IL, USA, are also used. For cell adhesion molecules detection, recombinant human ICAM-1, VCAM-1, E-selectin, and L-selectin standards, each vial reconstituted with 0.5 mL of reagent diluent is used. Micro well strips is coated with the respective murine monoclonal capture antibody (mAb) against human VCAM-1, ICAM-1, E-selectin, and L-selectin, each vial reconstituted with 0.5 or 354 Pavani Sayana et al.

1.0 mL of PBS (for VCAM-1). In addition, biotinylated sheep monoclonal detection antibodies against human VCAM-1, ICAM-1, E-selectin, and L-selectin, each vial reconstituted with 1.0 mL of reagent diluent are used as primary antibodies. The micro well plates were then treated with the respective HRP-conjugated anti-adhesion molecule mAb diluted with the assay buffer. The detection limits of the assays were: sICAM-1, 0.1, 0.35, or 1.0 ng/mL; sVCAM-1, 0.5, 2, or 3 ng/mL; sE-selectin 0.1, 0.2, or 0.5 ng/mL; sL-selectin, 3.0 ng/mL; MMP-2, 0.40 ng/mL; and MMP-9, 0.60 ng/mL. The inter assay coefficient of variation was <8–10% in all cases [24, 42, 43].

4.1.4 Addition About 100 μL of the working dilution of Streptavidin-HRP is of Secondary Antibody added to each well.

4.1.5 Addition About 100 μL of substrate solution (TMB) is added to each well as of 3,30,5,50-Tetramethyl- chromogen. benzidine (TMB)

4.1.6 Addition of Stop About 50 μL of stop solution such as sulfuric acid (2 N H2SO4)is Solution added to each well.

4.2 Methods As a first step, dilute the capture antibody to the working concen- tration in PBS without carrier protein. Immediately coat a 96-well 4.2.1 Plate Preparation microplate with 100 μL diluted capture antibody. Capture antibody by Coating with Capture helps to capture the antigen from the sample. Seal the plate and Antibody and Blocking incubate overnight at room temperature. Aspirate each well and Buffer wash with wash buffer, repeating the process two times for three washes. Wash by filling each well with wash buffer (400 μL) using a squirt bottle, manifold dispenser, or auto-washer. Complete removal of liquid at each step is essential for good performance. After the last wash, remove any remaining wash buffer by aspirating or by inverting the plate and blotting it against clean paper towels. Block plates by adding 300 μL reagent diluent as blocking buffer to each well. Blocking buffer is added to block the remaining protein- binding sites on the plate. Incubate at room temperature for a minimum of 1 h. Repeat the aspiration and wash.

4.2.2 Addition of Sample About 100 μL of sample with antigen or standards in reagent and Standards diluent, or an appropriate diluent, is added to the well. Cover with an adhesive strip and incubate for 2 h at room temperature. Repeat the aspiration and wash of plate preparation.

4.2.3 Addition of Primary About 100 μL of the primary or detection antibody, diluted in or Detection Antibody reagent diluent, is added to each well. Cover with a new adhesive strip and incubate for 2 h at room temperature. Repeat the aspira- tion/wash of plate preparation. Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 355

4.2.4 Addition About 100 μL of the working dilution of Streptavidin-HRP is of Secondary Antibody added to each well. Cover the plate and incubate for 20 min at room temperature. Avoid placing the plate in direct light. Repeat the aspiration and wash of plate preparation.

4.2.5 Addition of TMB About 100 μL of substrate solution (TMB) is added to each well as chromogen. Incubate for 20 min at room temperature. Avoid placing the plate in direct light.

4.2.6 Addition of Stop About 50 μL of stop solution such as sulfuric acid is added to each Solution well. Gently tap the plate to ensure thorough mixing.

4.2.7 Measure Determine the optical density of each well immediately, using a the Optical Density microplate reader set to 450 nm. If wavelength correction is avail- able, set to 540 or 570 nm. If wavelength correction is not avail- able, subtract readings at 540 or 570 nm from the readings at 450 nm. This subtraction will correct for optical imperfections in the plate. Readings made directly at 450 nm without correction may be higher and less accurate.

4.3 Applications The adhesion molecules, such as ICAM-1, VCAM-1, L-selectin, E-selectin, and b-integrin, are known to play a key role in the development of neuroinflammatory diseases by facilitating leuko- cyte adhesion and transmigration across the BBB [44]. Their solu- ble forms, sL-selectin, sICAM-1, sICAM-3, and sVCAM-1, have been shown to be elevated in the serum and cerebrospinal fluid (CSF) from patients with inflammatory CNS diseases [43, 45]. In addition, many studies have shown an impairment in cell adhesion molecules such as VCAM, ICAM, E-selectins, L-selectins and AdJ proteins such as cadherins and catenins and MMPs such as MMP-2 and MMP-9 from BBB damage [17, 24, 43].

5 Detection of BBB Genes by Suppressive Subtraction Hybridization (SSH)

5.1 Materials The materials required for the isolation of poly (A)+ RNA includes micro centrifuges, polytron homogenizer, orbital shaker, ribonu- 5.1.1 Isolation of Poly clease (RNase)/Deoxy ribonuclease (DNase)-free and spin-X filter (A)+ RNA units, oligo d(T) cellulose, lysis buffer, binding buffer, low ionic strength binding buffer, elution buffer, and Proteinase K.

5.1.2 SSH The temperature cycler, 96-well dot-blot apparatus (BioRad), poly- merase chain reaction (PCR) select cDNA subtraction system (Clontech, Palo Alto, CA), α32P-dCTP (800 Ci/mmol) (Perkin Elmer, Boston, MA), first strand buffer, second strand buffer, Rsa I (Clontech), RsaI buffer, ligation buffer, hybridization buffer, E. coli supercompetent cell, glycogen and the chemicals include, phenol, chloroform, isoamylalcohol and ethanol. 356 Pavani Sayana et al.

5.1.3 Differential The hybridization incubator, QIAquick PCR purification and QIA- Hybridization prep spin kits (Qiagen,Valencia, CA), Whatman DE81 filter paper, differential screening blocking solution, Luria-Bertani (LB) culture broth and agar plates with ampicillin (LB/amp), α32P-dCTP and GenScreen Plus membranes (Perkin Elmer), DNA labeling kit (Amersham Megaprime), Sephadex G-25 spin columns (Roche), 20Â SSPE solution.

5.1.4 DNA Sequencing The DNA/RNA electrophoresis equipment (BioRad), oligodeox- and Northern Blot Analysis ynucleotides primers-M13 forward and reverse, agarose, 20 Â SSC solution and the chemicals includes formaldehyde, formamide, SDS (BioRad), tRNA (Invitrogen), EcoRI (New England Biolabs, Boston, MA), Denhardt’s solution, denatured salmon sperm DNA, and dextran sulfate (Sigma).

5.2 Methods It is necessary to conduct the isolation of brain capillary under RNase-free conditions. Poly(A)+ mRNA is isolated from brain 5.2.1 Isolation of Poly(A) capillaries or tissue samples using a single step method previously + RNA described by Boado et al. [46]. The brain capillaries isolated from rat brains or other rat tissues (kidney or liver) are added with 25 mL lysis buffer. The sample is then homogenized at full speed, follow- ing homogenization RNases and other cellular proteins are completely digested for 45 min at 45 C in an orbital shaker at approximately 600 rpm. The selection of Poly(A)+ RNA is per- formed in the same tube with the addition of 50 mg washed oligo d (T) cellulose. Using mixer the lysate-cellulose suspension is mixed for 30 min. The sample is then centrifuged for 2 min at 1500 G at and the supernatant is discarded. The Poly(A)+ RNA is washed off the resin twice with binding buffer containing 0.5 M NaCl. A final wash with low ionic strength binding buffer (0.1 M NaCl) is performed. The cellulose pellet is transferred to a 0.45-μm Spin-X filter unit. The Poly(A)+ RNA is eluted with two aliquots of 400 μL elution buffer. The Poly(A)+ RNA is collected in the micro tube of the filter unit. The yield of brain capillary Poly-(A)+ mRNA ranges from 2 to 5 μg from 20 rat brains. The quality of the isolated Poly (A)+ RNA can be determined by Northern blot analysis using actin cDNA probe.

5.2.2 SSH To identify tissue-specific gene expression at the BBB, the genomics methodology such as serial analysis of gene expression (SAGE) or subtractive suppressive hybridization (SSH) has been used. How- ever, the latter allows for the isolation of partial cDNAs corresponding to the BBB-enriched genes. The SSH is performed using the PCR-select cDNA subtraction system according to the manufacturer’s instructions (Clontech) [27]. The rat brain capillary Poly (A)+ RNA is used to produce tester cDNA, and the subtrac- tion procedure is completed using other rat tissue (liver or kidney) mRNA to generate driver cDNA [26, 47, 48]. The reaction is Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 357

followed with [32P]-dCTP. The RsaI used to digest tester or driver cDNA to obtain shorter, blunt end molecules, and two tester populations is created as either adaptor 1 or adaptor 2R and are independently ligated. The two populations of adaptor-ligated tes- ter cDNA are independently hybridized to the driver cDNA to enrich for differentially expressed sequences and hybridized a sec- ond time to generate a PCR template. A first-run PCR using PCR primer one amplifies differentially expressed sequences and is per- formed for 30 cycles (denaturation, 94 C for 30 s; annealing, 66 C for 30 s; extension, 72 C 1.5 min). A second-run PCR is performed for 15 cycles (denaturation, 94 C for 30 s; annealing, 68 C for 30 s; extension, 72 C 1.5 min), using nested PCR primers 1 and 2R. This second-run PCR further enriches for differ- entially expressed sequences and suppresses the background [27].

5.2.3 Differential The SSH-PCR products are cloned into the pCR2.1 vector and a Hybridization cDNA library is prepared in E. coli INVαF’ cells. Positive clones are identified by differential hybridization. The cDNA library is plated on Luria-Bertani medium/ ampicillin (LB/amp) plates, and white colonies are picked and individually grown in 200 L LB/amp medium in 96-well plates for 24 h at 37 C. Colonies are individu- ally blotted onto GeneScreen Plus membrane using a 96-well dot-- blot system (Bio-Rad, Richmond, CA, USA). Membranes are hybridized with [32P]-labeled subtracted or unsubtracted rat brain capillary tester cDNA and film autoradiography is performed with Biomax MS film (Kodak, Rochester, NY, USA) for 18 h at 23 C. The cDNA is purified with Qiaquick PCR purification kit (Qiagen, Santa Clarita, CA, USA) and labeled as previously described. Clones showing a strong hybridization signal with the subtracted probe compared with the unsubtracted one are selected for DNA sequencing and Northern blot analysis after release of the pCR2.1 insert with EcoRI. Northern blotting is performed as described previously. Clones showing a strong hybridization signal with the subtracted probe compared with the unsubtracted one were selected for DNA sequencing and Northern blot analysis [49].

5.2.4 DNA Sequence The isolated clones are subjected to DNA sequencing in both Analysis and Northern Blot directions using M13 forward and reverse primers. Using literature Analysis mining, similarities with other genes in GenBank were investigated using the BLAST program (NCBI, NIH). The QIAfilter maxi kit (Qiagen) was used to amplify and purify plasmid DNA from posi- tive clones [49]. The tissues from isolated poly (A)+ RNA, rat liver, rat kidney are used. The RNAs are resolved in agarose, formaldehyde gels then blotted on a GenScreen plus using capillary-blotting procedures with SSC solution overnight [50]. Following this, the membranes were rinsed with SSC, then placed in vacuum oven at 80 C for 2 h. Finally, hybridization is performed and exposed to X-ray film for autoradiography. 358 Pavani Sayana et al.

5.3 Application The standard genomic approaches based on gene microarrays using SSH revealed about 53 clones K1-K3; LK1-LK50 that are expressed at BBB and periphery. Genes expressed at BBB includes both known and novel. The clones K2 (novel), LK 48 (EST), the full sequence for BSAT1 encompassed the sequences of eight other clones found in the initial BBB library such as LK3 (novel), LK12 (novel), LK13 (novel), LK14 (novel), LK15 (novel), LK39(novel), LK45 (novel), and LK49 (novel). The clones hybridized to tran- scripts that are expressed in brain only at the BBB and the mRNA are also differentially expressed in some peripheral tissues such as LK1 (EST), LK17 (EST), and LK30(EST). Four additional clones that represented novel sequences not found in the EST or GenBank databases are LK20 (novel), LK35 (novel), LK38 (novel), and LK44 (novel). The known genes expressed at BBB are tissue plas- minogen activator (LK5); insulin-like growth factor (IGF)- 2 (LK6); the vascular endothelial growth factor receptor, flt-1 (LK 7, LK10, LK18, and LK43); the PC-3 gene product (LK16); the regulator of G protein signaling (Rgs)-5 (LK23); utrophin (LK24); IκB (LK27); connexin-45 (LK29); the transferrin receptor (LK34); the class I major histocompatibility complex (LK36); and organic anion transporting polypeptide type 2 or oatp2 (K1 and LK42).

6 Detection of BBB Genes by Quantitative Real-Time PCR (qRT-PCR)

6.1 Materials Thermocycler, TRIzol RNA isolation reagents, reaction buffer, dNTP mix, SuperScript III (Invitrogen) nuclease-free water, DTT, RNAse inhibitor, phenol, chloroform.

6.2 Methods The TRIzol®-solubilized tissue extract is used for the isolation of RNA from mice whole brain. To limit protein contamination and 6.2.1 RNA Isolation increase RNA yield the back extraction was performed after the phenol:chloroform extraction step by addition of extra TRIzol® and the organic and aqueous phases separated a second time. For the final precipitation of RNA from the aqueous phase, 0.15Â aqueous phase volume of 2 M sodium acetate pH 4.0 and an equal aqueous phase volume of isopropanol (Acros) is added, as carrier glycogen (Ambion).

6.2.2 DNAse Treatment The isolated RNA is treated with Turbo DNAse (Ambion) accord- and cDNA Synthesis ing to the protocol provided by the manufacturer. The SuperScript III (Invitrogen) is used to reverse transcript mRNA [51]. The resulting cDNA is stored at À20  C.

6.2.3 qRT-PCR The measurement of cDNA levels is performed by qRT-PCR using an ABI PRISM 7900 Sequence detection system version 2.3 [51]. For each gene the dissociation curve is analyzed to ensure Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 359

specific amplicon replication. All the reaction run in duplicates. Separate controls included no template control, no reverse tran- scriptase control, and positive control. To avoid the possibility of amplifying contaminating DNA, intron-spanning primer pairs are designed using primer express 3.0 software. The primers used are claudin 1 (forward 50-TCCTGCGTTTCGCAAAGC-30 and reverse 50-GGGTTGCCTGCAAAGTACTGTT-30), claudin 5 (forward 50-TGCCGCGAACAGTTCCTACT-30 and reverse 50-CCAGCT GCCCTTTCAGGTTA-30), occludin (forward 50-GGACTGGGT CAGGGAATATCC-30 and reverse 50-GCAGACCTGCATCAAAA TTTCTC-30), ZO-1 (forward 50-CTCGGAAAAATGAAGAATA TGGTC-30 and reverse 50-CACCATCTCTTGCTGCCAAA-30), ZO-2 (forward 50-GCAGCTTGAAGGACAGCATTC-30 and reverse 50-ATCATCCATCCCTTCCATCTTTC-30), ZO-3 (for- ward 50-GCCCGTATGGGTAGATGATCA-30 and reverse 50-TGTCAGCCTGTGTCTCCCATT-30), afadin (forward 50-TGAGCGAGGGATGGTGAAG-30 reverse and 50-CAGTA GCATCCTGAGTCCTGTTTTC-30), cingulin (forward 50-ATCCA GCCATGGTGCAGTTC-30 and reverse 50-GCCATAGA TGGTGGCCTTCA-30), cadherin (forward 50-ACGTGGCTGGA TGTGGAATC-30 and reverse 50-GGTCTGAATGAAGCTTGCC TACATA-30), beta-catenin (forward 50-ATGGAGCCGGAC AGAAAAGC-30 and reverse 50-CTTGCCACTCAGGGAAGGA-30), transferrin receptor (forward 50-GCACTCTTTGGACATGCTCATC -30 and reverse 50-ATGGCGGAAACTGAGTATGATTG-30), insulin receptor (forward 50-CCCAATGGCAACATCACACA-30 and reverse 50-GGAGCTTCAGCCCTTTGAGA-30), leptin receptor (forward 50-TGCTGAATTATACGTGATCGATGTC-30 and reverse 50-GGA TTGTGCTGGGTGACCAT-30), low-density lipoprotein receptor- related protein (forward 50-ATCCAAGGCATCCTGATCGA-30 and reverse 50-TCGTCTTCTGCTGCGTGTTG-30), norepinephrine transporter (forward 50-CTCAAGGAGGCCACGGTATG-30 and reverse 50-GCAAATGCAATCAAGACTCCAA-30), glucose trans- porter (forward 50-TCCTTATTGCCCAGGTGTTTG-30 and reverse 50-TAGCAGGGCTGGGACGAA-30), multidrug resistance (forward 50-GGCCGTGATGGAACTTGAAG-30 and reverse 50-CCTTTTTA CTCTTTTTGCCCATCT-30), breast cancer resistance protein (for- ward 50-GGAGGCAAGTCTTCGTTGCT-30 and reverse 50-GCA GGTTGAGGTGCTCCATT-30), gamma glutamyltransferase (for- ward 50-CCTGTGGGCAAGGGTTTG-30 and reverse 50-GGCA GAACACCTCGCACAA-30), DOPA decarboxylase (forward 50-AGGCAGCCTCCCCAGAGTT-30 and reverse 50-AGAGGAATG CGCCTGATCAG-30), junctional adhesion molecule 1 (forward 50-G ACAGCCTTTGATAGTGGTGAATACTA-30 and reverse 50-CCAC ATTCAGCTCCACAGCAT-30), protocadherin gamma, (forward 50-GGCACGAGCGGATCCC-30 and reverse 50-GTAGAGCTCCC ATCAGCAGCTT-30), connexin 43 (forward 50-TTTGACTTC 360 Pavani Sayana et al.

AGCCTCCAAGGA-30 and reverse 50-GTCTGGGCACCTCTC TTTCACT-30),connexin40(forward50-GAGGAGGAAAGGAAG CAGAAGG-30 and reverse 50-GGAAGCTCCAGTCACCCATCT- 30), connexin 37 (forward 50-CAAGCAGGCGAGAGAGGC-30 and reverse 50- ATGAGGATGCGGAAGATGAAGA-30), protein receptor tyrosine kinase (forward 50-GGCATTCCAGAACGTGAGAGA-30 and reverse 50-TGTGGGATCCGGATTGTTTT-30), erythropoietin receptor (forward 50-TACCAGCTCGAGGGTGAGTCA-30 and reverse 50-AAACTCGATGTGTCCGCTGTT-30), acetylcholinester- ase (forward 50-CCTGGATCCCTCGCTGAAC-30 and reverse 50-G ATTTGGAGTCTCGAGGGTCAT-30), butyrylcholinesterase (for- ward 50-TCCCCAAAGCATTGCAAGAT-30 and reverse 50-AGTCT GCATGTTGATTTCGGAAT-30),caveolin1(forward50-GTGGAC TCCGAGGGACATCTC-30 and reverse 50- GCCTTGTTGTT GGGCTTGTAG-30),basigin(forward50-GGTCTGCAAGTCCG ATGCAT-30 and reverse 50- GTGCTATTGGTGATTGCCTCTTC- 30), platelet endothelial cell adhesion marker (forward 50-TCCAGGTG TGCGAAATGCT-30 and reverse 50- TTTTCGGACTGGCAGCTG AT-30),andendoglin(forward50-GGGTGAGGTGACGTTTAC- CAC-30 and reverse 50-AGCATTCCGGGAAAATCCAGG-30).

6.3 Application There are about 87 genes known to play a role in BBB function and they were shown to be enriched in isolated brain micro-vessels [15]. The BBB-associated genes can be divided into (1) Genes encoding proteins involved in BBB function that includes claudin 1, claudin 5, occludin, ZO-1, ZO-2, ZO-3, afadin, cingulin, cad- herin, beta-catenin, transferrin receptor, insulin receptor, leptin receptor, low-density lipoprotein receptor-related protein, norepi- nephrine transporter, glucose transporter, multidrug resistance, breast cancer resistance protein, gamma glutamyltransferase, and DOPA decarboxylase and (2) genes encoding proteins possibly involved in BBB function that includes junctional adhesion mole- cule 1, protocadherin gamma, connexin 43, connexin 40, connexin 37, protein receptor tyrosine kinase, erythropoietin receptor, ace- tylcholinesterase, butyrylcholinesterase, caveolin 1, basigin, platelet endothelial cell adhesion marker, and endoglin.

7 Conclusion

A greater understanding of the BBB and brain microvasculature network is critical as the BBB and brain microvasculature play a key role in central drug delivery system. In order to identify the sub- stances that can be targeted for pharmacological manipulation and/or gene therapy continuous effort has been made on under- standing the molecular mechanisms of BBB during physiological and pathological conditions. In this chapter, we made an attempt to describe some of the protein markers that have been altered in brain Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 361

microvasculature unit during BBB injury. In addition to the protein markers the gene that expressed specifically at BBB and the gene encoding protein were also covered. The methods described in this chapter will aid researchers in selecting the protein markers and corresponding detection system after BBB injury.

Acknowledgments

The Translational Psychiatry Program is funded by the Department of Psychiatry and Behavioral Sciences, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth). National Institute for Molecular Medicine (INCT- MM) and Center of Excellence in Applied Neurosciences of Santa Catarina (NENASC).

References

1. Wilhelm I, Nyul-Toth A, Suciu M, down-regulating TLR4/MyD88 pathway. Int Hermenean A, Krizbai IA (2016) Heterogene- Immunopharmacol 56:29–35. https://doi. ity of the blood-brain barrier. Tissue barriers 4 org/10.1016/j.intimp.2018.01.002 (1):e1143544. https://doi.org/10.1080/ 8. Stamatovic SM, Sladojevic N, Keep RF, 21688370.2016.1143544 Andjelkovic AV (2015) PDCD10 (CCM3) 2. Hirase T, Staddon JM, Saitou M, Ando- regulates brain endothelial barrier integrity in Akatsuka Y, Itoh M, Furuse M, Fujimoto K, cerebral cavernous malformation type 3: role of Tsukita S, Rubin LL (1997) Occludin as a CCM3-ERK1/2-cortactin cross-talk. Acta possible determinant of tight junction perme- Neuropathol 130(5):731–750. https://doi. ability in endothelial cells. J Cell Sci 110 org/10.1007/s00401-015-1479-z (Pt 14):1603–1613 9. Lossinsky AS, Buttle KF, Pluta R, Mossakowski 3. Martin-Padura I, Lostaglio S, Schneemann M, MJ, Wisniewski HM (1999) Immunoultras- Williams L, Romano M, Fruscella P, Panzeri C, tructural expression of intercellular adhesion Stoppacciaro A, Ruco L, Villa A, Simmons D, molecule-1 in endothelial cell vesiculotubular Dejana E (1998) Junctional adhesion mole- structures and vesiculovacuolar organelles in cule, a novel member of the immunoglobulin blood-brain barrier development and injury. superfamily that distributes at intercellular Cell Tissue Res 295(1):77–88 junctions and modulates monocyte transmigra- 10. Williams KC, Zhao RW, Ueno K, Hickey WF tion. J Cell Biol 142(1):117–127 (1996) PECAM-1 (CD31) expression in the 4. Haskins J, Gu L, Wittchen ES, Hibbard J, Ste- central nervous system and its role in experi- venson BR (1998) ZO-3, a novel member of mental allergic encephalomyelitis in the rat. J the MAGUK protein family found at the tight Neurosci Res 45(6):747–757 junction, interacts with ZO-1 and occludin. J 11. Shukaliak JA, Dorovini-Zis K (2000) Expres- Cell Biol 141(1):199–208 sion of the beta-chemokines RANTES and 5. Anderson JM (1996) Cell signalling: MAGUK MIP-1 beta by human brain microvessel endo- magic. Curr Biol 6(4):382–384 thelial cells in primary culture. J Neuropathol 6. Liu J, Wang F, Liu S, Du J, Hu X, Xiong J, Exp Neurol 59(5):339–352 Fang R, Chen W, Sun J (2017) Sodium buty- 12. Zhang W, Smith C, Shapiro A, Monette R, rate exerts protective effect against Parkinson’s Hutchison J, Stanimirovic D (1999) Increased disease in mice via stimulation of glucagon like expression of bioactive chemokines in human peptide-1. J Neurol Sci 381:176–181. https:// cerebromicrovascular endothelial cells and doi.org/10.1016/j.jns.2017.08.3235 astrocytes subjected to simulated ischemia 7. Cheng X, Yang YL, Yang H, Wang YH, Du GH in vitro. J Neuroimmunol 101(2):148–160 (2018) Kaempferol alleviates LPS-induced 13. Roberts JM, Maniskas ME, Bix GJ (2018) neuroinflammation and BBB dysfunction in Bilateral carotid artery stenosis causes unex- mice via inhibiting HMGB1 release and pected early changes in brain extracellular 362 Pavani Sayana et al.

matrix and blood-brain barrier integrity in dodecyl sulfate--polyacrylamide gels to mice. PLoS One 13(4):e0195765. https:// unmodified nitrocellulose and radiographic doi.org/10.1371/journal.pone.0195765 detection with antibody and radioiodinated 14. Pardridge WM (2007) Blood-brain barrier protein A. Anal Biochem 112(2):195–203 genomics. Stroke 38(2 Suppl):686–690. 24. McDonnell GV, McMillan SA, Douglas JP, https://doi.org/10.1161/01.str. Droogan AG, Hawkins SA (1999) Serum solu- 0000247887.61831.74 ble adhesion molecules in multiple sclerosis: 15. Macdonald JA, Murugesan N, Pachter JS raised sVCAM-1, sICAM-1 and sE-selectin in (2010) Endothelial cell heterogeneity of primary progressive disease. J Neurol 246 blood-brain barrier gene expression along the (2):87–92 cerebral microvasculature. J Neurosci Res 88 25. Pileri SA, Roncador G, Ceccarelli C, Piccioli M, (7):1457–1474. https://doi.org/10.1002/ Briskomatis A, Sabattini E, Ascani S, Santini D, jnr.22316 Piccaluga PP, Leone O, Damiani S, Ercolessi C, 16. Lin R, Chen F, Wen S, Teng T, Pan Y, Huang Sandri F, Pieri F, Leoncini L, Falini B (1997) H (2018) Interleukin-10 attenuates Antigen retrieval techniques in immunohisto- impairment of the blood-brain barrier in a chemistry: comparison of different methods. J severe acute pancreatitis rat model. J Inflamm Pathol 183(1):116–123. https://doi.org/10. (Lond) 15:4. https://doi.org/10.1186/ 1002/(sici)1096-9896(199709)183:1<116:: s12950-018-0180-0 aid-path1087>3.0.co;2-2 17. Yang S, Gu C, Mandeville ET, Dong Y, 26. Shusta EV, Boado RJ, Mathern GW, Pardridge Esposito E, Zhang Y, Yang G, Shen Y, Fu X, WM (2002) Vascular genomics of the human Lo EH, Xie Z (2017) Anesthesia and surgery brain. J Cereb Blood Flow Metab 22 impair blood-brain barrier and cognitive func- (3):245–252. https://doi.org/10.1097/ tion in mice. Front Immunol 8:902. https:// 00004647-200203000-00001 doi.org/10.3389/fimmu.2017.00902 27. Diatchenko L, Lau YF, Campbell AP, 18. Yang Y, Yang LY, Orban L, Cuylear D, Chenchik A, Moqadam F, Huang B, Thompson J, Simon B, Yang Y (2018) Lukyanov S, Lukyanov K, Gurskaya N, Sver- Non-invasive vagus nerve stimulation reduces dlov ED, Siebert PD (1996) Suppression sub- blood-brain barrier disruption in a rat model of tractive hybridization: a method for generating ischemic stroke. Brain Stimul 11(4):689–698. differentially regulated or tissue-specific cDNA https://doi.org/10.1016/j.brs.2018.01.034 probes and libraries. Proc Natl Acad Sci U S A 19. Bangsow T, Baumann E, Bangsow C, Jaeger 93(12):6025–6030 MH, Pelzer B, Gruhn P, Wolf S, von 28. Wang X, Ju S, Chen S, Gao W, Ding J, Melchner H, Stanimirovic DB (2008) The epi- Wang G, Cao H, Tian H, Li X (2017) Effect thelial membrane protein 1 is a novel tight of electro-acupuncture on neuroplasticity of junction protein of the blood-brain barrier. J spinal cord-transected rats. Med Sci Monit Cereb Blood Flow Metab 28(6):1249–1260. 23:4241–4251 https://doi.org/10.1038/jcbfm.2008.19 29. Chen K, Wang N, Diao Y, Dong W, Sun Y, 20. Wang ZG, Cheng Y, Yu XC, Ye LB, Xia QH, Liu L, Wu X (2017) Hydrogen-rich saline Johnson NR, Wei X, Chen DQ, Cao G, Fu XB, attenuates brain injury induced by cardiopul- Li XK, Zhang HY, Xiao J (2016) bFGF protects monary bypass and inhibits microvascular against blood-brain barrier damage through endothelial cell apoptosis via the PI3K/Akt/ junction protein regulation via PI3K-Akt-Rac1 GSK3beta signaling pathway in rats. Cell pathway following traumatic brain injury. Mol Physiol Biochem 43(4):1634–1647. https:// Neurobiol 53(10):7298–7311. https://doi. doi.org/10.1159/000484024 org/10.1007/s12035-015-9583-6 30. Ramos-Vara JA, Miller MA (2014) When tissue 21. Mahmood T, Yang PC (2012) Western blot: antigens and antibodies get along: revisiting technique, theory, and trouble shooting. N Am the technical aspects of immunohistochemis- J Med Sci 4(9):429–434. https://doi.org/10. try--the red, brown, and blue technique. Vet 4103/1947-2714.100998 Pathol 51(1):42–87. https://doi.org/10. 22. Bittner M, Kupferer P, Morris CF (1980) Elec- 1177/0300985813505879 trophoretic transfer of proteins and nucleic 31. Nag S (1996) Immunohistochemical localiza- acids from slab gels to diazobenzyloxymethyl tion of extracellular matrix proteins in cerebral cellulose or nitrocellulose sheets. Anal Bio- vessels in chronic hypertension. J Neuropathol chem 102(2):459–471 Exp Neurol 55(3):381–388 23. Burnette WN (1981) "Western blotting": elec- 32. Simmons DA, Arriza JL, Swanson LW (1989) trophoretic transfer of proteins from sodium A complete protocol for in situ hybridization of Biomarkers for Microvascular Proteins Detection: Blood–Brain Barrier... 363

messenger RNAs in brain and other tissues 42. Droogan AG, McMillan SA, Douglas JP, Haw- with radio-labeled singlestranded RNA probes. kins SA (1996) Serum and cerebrospinal fluid J Histotechnol 12:169–180 levels of soluble adhesion molecules in multiple 33. Shi SR, Cote RJ, Taylor CR (1997) Antigen sclerosis: predominant intrathecal release of retrieval immunohistochemistry: past, present, vascular cell adhesion molecule-1. J Neuroim- and future. J Histochem Cytochem 45 munol 64(2):185–191 (3):327–343. https://doi.org/10.1177/ 43. Correale J, Bassani Molinas Mde L (2003) 002215549704500301 Temporal variations of adhesion molecules 34. Sternberger LA, Hardy PH Jr, Cuculis JJ, and matrix metalloproteinases in the course of Meyer HG (1970) The unlabeled antibody MS. J Neuroimmunol 140(1–2):198–209 enzyme method of immunohistochemistry: 44. Duan H, Luo Y, Hao H, Feng L, Zhang Y, preparation and properties of soluble antigen- Lu D, Xing S, Feng J, Yang D, Song L, Yan X antibody complex (horseradish peroxidase- (2013) Soluble CD146 in cerebrospinal fluid antihorseradish peroxidase) and its use in iden- of active multiple sclerosis. Neuroscience tification of spirochetes. J Histochem Cyto- 235:16–26. https://doi.org/10.1016/j.neuro chem 18(5):315–333. https://doi.org/10. science.2013.01.020 1177/18.5.315 45. Acar G, Idiman F, Kirkali GOS, Oktay G, 35. Hsu SM, Ree HJ (1980) Self-sandwich Cakmakci H, Idiman E (2005) Intrathecal method. An improved immunoperoxidase sICAM-1 production in multiple sclerosis- technic for the detection of small amounts of correlation with triple dose Gd-DTPA MRI antigens. Am J Clin Pathol 74(1):32–40 enhancement and IgG index. J Neurol 252 36. Hsu SM, Raine L, Fanger H (1981) A compar- (2):146–150. https://doi.org/10.1007/ ative study of the peroxidase-antiperoxidase s00415-005-0618-1 method and an avidin-biotin complex method 46. Boado RJ, Pardridge MM (1991) A one-step for studying polypeptide hormones with radio- procedure for isolation of poly(A)+ mRNA immunoassay antibodies. Am J Clin Pathol 75 from isolated brain capillaries and endothelial (5):734–738 cells in culture. J Neurochem 57(6):2136–2139 37. Elias JM, Margiotta M, Gaborc D (1989) Sen- 47. Li JY, Boado RJ, Pardridge WM (2001) Blood- sitivity and detection efficiency of the peroxi- brain barrier genomics. J Cereb Blood Flow dase antiperoxidase (PAP), avidin-biotin Metab 21(1):61–68. https://doi.org/10. peroxidase complex (ABC), and peroxidase- 1097/00004647-200101000-00008 labeled avidin-biotin (LAB) methods. Am J 48. Li JY, Boado RJ, Pardridge WM (2002) Rat Clin Pathol 92(1):62–67 blood-brain barrier genomics. II. J Cereb 38. Sabattini E, Bisgaard K, Ascani S, Poggi S, Blood Flow Metab 22(11):1319–1326. Piccioli M, Ceccarelli C, Pieri F, Fraternali- https://doi.org/10.1097/01.WCB. Orcioni G, Pileri SA (1998) The EnVision++ 0000040944.89393.0f system: a new immunohistochemical method 49. Boado RJ, Li JY, Nagaya M, Zhang C, Par- for diagnostics and research. Critical compari- dridge WM (1999) Selective expression of the son with the APAAP, ChemMate, CSA, LABC, large neutral amino acid transporter at the and SABC techniques. J Clin Pathol 51 blood-brain barrier. Proc Natl Acad Sci U S A (7):506–511 96(21):12079–12084 39. Adams JC (1981) Heavy metal intensification 50. Sambrook J, Russell DW (2001) Molecular of DAB-based HRP reaction product. J Histo- cloning. In: A laboratory manual. Cold Spring chem Cytochem 29(6):775 Harbor Laboratory Press, Cold Spring Harbor, 40. Adams JC (1992) Biotin amplification of biotin NY and horseradish peroxidase signals in histo- 51. Macdonald JA, Murugesan N, Pachter JS chemical stains. J Histochem Cytochem 40 (2008) Validation of immuno-laser capture (10):1457–1463. https://doi.org/10.1177/ microdissection coupled with quantitative 40.10.1527370 RT-PCR to probe blood-brain barrier gene 41. Nag S (2003) The blood-brain barrier. In: expression in situ. J Neurosci Methods 174 Biology and research protocols. Methods in (2):219–226. https://doi.org/10.1016/j. molecular medicine. Humana Press, New York jneumeth.2008.07.009 INDEX

A Cerebrospinal fluid (CSF) ...... 2, 18, 45, 137, 163, 164, 205–207, 224, 231–234, Albumin...... 21, 26, 27, 32, 236–239, 241–245, 315, 322, 324, 330, 355 166, 197–202, 329, 336 Chemokines...... 316, 317, Autocrine ...... 164 319–321, 344 B Choroid plexus (CP)...... 163–172, 205–227, 232, 233, 235, 236, 238–246, 315 Barrier integrity ...... 78, 99–101, Choroid plexus epithelial cell (CPEC) ...... 163, 109, 178, 194, 197–202 164, 166, 167, 169–171, 205–208, 213, 231, BBB-on-chip...... 100, 109, 110 233, 235, 236 Bioluminescence assays ...... 207, 209, 210 Cytokines ...... 10, 13, 26, 31, Biomarkers...... vii, 31, 343–361 81, 87, 88, 320, 321, 344 Blood-brain barrier (BBB)...... vii, 2–6, 9–36, 45–51, 55–69, 71–90, 99–110, 115–131, D 137–161, 175–184, 187–195, 197–202, 256, Danio rerio ...... vii, 6, 187–195 261, 267–290, 300–314, 329–339, 343–361 Drosophila ...... vii, 75, 175–184 Blood-cerebrospinal fluid barrier (BCSFB)...... 3, Drug delivery ...... vii, 15, 22, 46, 206, 207, 216, 231, 232, 236, 237, 239, 242, 57, 72, 116–121, 125, 126, 188, 195, 249, 251, 243, 245 268, 289, 359 Blood-retinal barrier (BRB) ...... 3, 5, 6, Drug transport ...... 19, 59, 78, 249, 250, 252, 253, 258, 261 108, 110, 188, 233, 258, 260–262 Brain Dynamic contrast-enhanced (DCE) delivery...... 11–22, 24–28, 36 imaging ...... 280, 300–314 diagnostics ...... 9–36 diseases...... 9, 10, E 19, 21, 26, 29–36, 188 microvascular endothelial cells (BMECs) ...... 45–51, Efflux...... vii, 10–12, 14, 59, 60, 64, 71, 74, 82, 84, 88, 105, 106, 108, 344 15, 27, 46, 57, 77, 78, 188, 194, 195, 206, 231, microvasculature...... 194, 319, 232, 236, 237, 239, 242, 243, 245, 250, 257, 321, 323, 344, 359 258, 260, 263, 265, 267, 280, 281 therapeutics ...... 9–36, Efflux transport ...... 231, 232, 64, 188, 268, 290 236, 237, 239, 242, 243, 245, 253, 257, 258, 263, 265 C Endothelial ...... 3, 10, 45, 55, 72, 100, 116, 138, 188, 197, 250, 267, 316, Calcium imaging ...... 210, 211, 329, 343 221–223 Endothelial junctions...... 77, 317 Cell culture filter inserts ...... 210, 216, Enzyme-linked immunosorbent assay (ELISA) ...... 345, 218–220, 225, 227 353–355 Cell dissociation ...... 176, 183, 212 Evans blue (EB) ...... 190, 192, Cell penetrating peptides (CPPs)...... 115–131 197–202, 280, 281, 303, 309, 311, 313, 314, Central nervous system (CNS)...... 2, 4, 331–336 6, 12, 18, 31, 46, 55, 57, 60, 71–73, 116–119, Ex vivo models...... vii, 73, 121, 122, 125–127, 131, 197, 207, 213, 242, 138, 159, 188, 233 243, 267, 290, 315, 317, 319–321, 329, 344, 355 Explant electrophysiology ...... 207

Tatiana Barichello (ed.), Blood-Brain Barrier, Neuromethods, vol. 142, https://doi.org/10.1007/978-1-4939-8946-1, © Springer Science+Business Media, LLC, part of Springer Nature 2019 365 BLOOD-BRAIN BARRIER 366 Index F Neurovascular unit (NVU)...... 4, 5, 45, 48, 55, 60, 68, 73, 75–80, 117, 197, 329 Focused ultrasound (FUS) ...... 267–290, 331, Nuclear magnetic resonance (NMR)...... 27, 335, 336, 338, 339 116, 122–126, 129–131 G O

ts Gal4/UAS/Gal80 ...... 176, 177 Organic anion transporters (OAT) ...... 232, G-protein-coupled receptor (GPCR) ...... 164, 244, 258, 261 167, 172, 224, 316, 317 Organic cation transporter (OCT) ...... 232, I 233, 244, 261 Organ-on-chip ...... 100 Immunohistochemistry ...... 18, 176, P 208, 209, 214, 215, 269, 279, 345 In vitro models...... 11–14, Paracellular ...... 5, 10, 12, 45–51, 64, 73, 76, 77, 105, 160, 207, 216 14, 24, 26–29, 31, 46, 48, 56, 59–61, 68, 72, 75, In vivo models ...... 12, 21, 100–102, 138, 145, 146, 149–151, 160, 188, 26, 73, 75, 195, 233 206, 250, 267, 269, 271, 317 Influx transports...... 254, 256, 261 Permeability...... vii, 5, 12, Insects ...... 175 35, 47–49, 57, 58, 60, 61, 68, 72–77, 100, 101, Integration plot...... 251, 106, 107, 109, 123, 127, 131, 138, 147, 254–257, 260, 261 187–190, 193, 194, 198, 201, 210, 216, 251, Intracerebroventricular administration ...... 233, 256, 260–262, 267–290, 300–314, 320, 236–238, 242 329–339, 344 Intravital microscopy ...... vii, 315–325 Pia mater vessels...... 163 L Positron emission tomography (PET)...... vii, 26, 30, 33, 34, 282, 329–339 Leukocyte adhesion ...... 315–325, 355 Primary cultures ...... 48, 59, 61, Leukocyte rolling ...... 315–325 68, 69, 74, 75, 77, 107, 164, 207, 208, 212, 213 M R

Magnetic resonance imaging (MRI)...... vii, 22, Rat...... 12, 15, 24, 27, 31, 26, 27, 29–33, 35, 268, 274–276, 278–280, 33, 34, 48–50, 59–61, 73–75, 105, 107, 199, 300–314, 330 200, 207, 208, 212, 213, 233, 235–246, 251, Mice ...... 4, 5, 18, 21, 24, 27, 252, 254, 256, 257, 259, 264, 300, 302–305, 29–31, 34, 35, 60, 120, 121, 126, 127, 140, 142, 307, 309, 310, 313, 332, 333, 335, 339, 356, 357 145, 146, 159, 164, 198–200, 207, 209, 211, Reactive oxygen species (ROS) ...... 81, 82, 87–89 215, 216, 224, 233, 270, 273, 276, 277, Real time-polymerase chain reaction 318–322, 358 (RT-PCR) ...... 345 Microbial translocation...... 137–161 Retinal uptake index (RUI)...... 251, 252, Microbubbles (MB) ...... 118, 268, 256, 257, 260, 261 270–272, 274, 275, 278, 331, 336 Rodents...... 6, 33, 45–51, Microdialysis...... 251–254, 164, 171, 194, 197–202, 207, 254, 256, 268, 257–259, 261, 262, 264 273–276, 278–280, 282, 290, 300–314 Microfluidic system ...... 55–69, 84 S Microinjections...... 118, 189, 191, 193 Microvascular endothelial cells...... 45, 64, 72, 343 Shear stress ...... 13, 59, 64, Middle cerebral artery occlusion (MCAO) ...... 31, 68, 71, 74, 82, 84, 88, 100, 108, 109, 269 304, 305 Single photon emission computed tomography N (SPECT) ...... 26, 30, 33, 34, 329–339 Neurodegenerative diseases...... 13, 22, siRNA transfection...... 223 73, 115–131, 224 Static platforms...... 55–69 BLOOD-BRAIN BARRIER Index 367 Stroke...... 26, 31, 33, 34, 106, 108, 138, 139, 160, 163–172, 182, 188, 73, 120, 121, 300, 308, 310 207, 213, 231, 232, 236, 237, 239, 242, 243, Subperineurial glial cells (SPG)...... 175, 245, 249, 250, 252, 253, 258, 261, 268, 271, 176, 178–181, 183 332, 343 Transporters ...... 12, 19, 46, T 47, 49, 50, 57, 61, 65, 72, 75, 77, 78, 118, 188, Theranostics ...... 115–131 194, 195, 216, 232, 233, 244, 250, 260, 261, Tight-junction proteins ...... 26, 82, 345, 359 Trojan horse ...... 15, 138, 89, 100, 102, 108, 189, 194 Transcellular ...... 11, 12, 14, 143, 146, 151–153 46, 57, 77, 102, 138, 143, 146, 150, 151, 160, Two-photon microscopy ...... 272, 282–285, 289 267, 269, 271, 317 W Transcytosis ...... 15, 18, 19, 22, 46, 164, 166, 169–172 Western blot ...... 164, 189, Transepithelial electrical resistance (TEER) ...... 12, 194, 213, 214, 269, 345–348 13, 22, 26, 48, 50, 60, 64, 75, 77, 99–110, 147, 171, 188, 216, 219, 220, 227 Z Transient middle cerebral artery occlusion Zebrafish...... vii, 6, (tMCAo) ...... 300, 304 22, 75, 187–195 Translocation...... 119, 127, 137, 149, 150, 152 Zonula occludens (ZO-1, ZO-2, and ZO-3) ...... 4, Transport ...... 4, 6, 11–15, 6, 46, 72, 81, 344 17–20, 22, 23, 46–49, 55, 59, 64, 77, 78, 100,