Regulation of Folate Transport at the Blood-Brain Barrier

Regulation of Folate Transport at the Blood-Brain Barrier: A Novel Strategy for the Treatment of Childhood Neurological Disorders Associated with Cerebral Folate Deficiency

Camille Alam

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Pharmaceutical Sciences University of Toronto

Abstract

Regulation of Folate Transport at the Blood-Brain Barrier: A Novel Strategy for the Treatment of Childhood Neurological Disorders Associated with Cerebral Folate Deficiency

Camille Alam

Doctor of Philosophy

Graduate Department of Pharmaceutical Sciences University of Toronto

2020

Folates are crucial for the maintenance of central nervous system homeostasis. Folate transport is mediated by three major pathways, i.e., folate receptor alpha (FRα), proton-coupled folate transporter (PCFT) and reduced folate carrier (RFC), known to be regulated by transcription factors. Brain folate delivery occurs predominantly at the choroid plexus through concerted actions of FRα and PCFT; inactivation of these transport systems causes cerebral folate deficiency resulting in early childhood neurodegeneration. Increasing brain folate permeability or finding alternative routes for brain folate transport could lead to therapeutic benefits. The overall goal of this PhD thesis was to characterize folate transport at the blood-brain barrier (BBB) by examining the role and regulation of folate transporters (i.e., RFC) in brain folate uptake. The objectives of this thesis were to: i) determine RFC functional expression in several in vitro (primary or immortalized cultures of human/rodent brain microvessel endothelial cells) and ex vivo (isolated mouse brain capillaries) BBB models, ii) investigate the role of transcription factors, particularly vitamin D receptor (VDR) and nuclear respiratory factor 1 (NRF-1), in the regulation of RFC in

ii various BBB model systems, and iii) examine in vivo, using Folr1 (FRα) knockout mice, the relative contribution of RFC in overall brain folate uptake. We initially demonstrated that RFC is functionally expressed in in vitro systems representative of human (hCMEC/D3 cells) and rodent

(mouse brain capillaries) BBB, and that activation of VDR by its ligand, 1,25-dihydroxyvitamin

D3 or calcitriol, significantly increased RFC mRNA and protein expression as well as function.

We further showed in vivo, using Folr1 knockout mice, that loss of FRα substantially decreased folate delivery to the brain, but transport was restored through calcitriol administration.

Additionally, we provided in vitro and in vivo evidence that RFC expression and transport activity is inducible by another transcription factor, NRF-1. These findings demonstrate that augmenting

RFC functional expression through interaction with specific transcription factors could constitute a novel strategy for enhancing brain folate delivery. Modulating folate uptake at the BBB may have clinical significance due to the lack of established optimal therapy for neurometabolic disorders caused by loss of FRα or PCFT function.

iii

Acknowledgments

The studies presented in this thesis were conducted at the Department of Pharmaceutical Sciences, University of Toronto, with financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC) awarded to Dr. Reina Bendayan. I am also very grateful for the studentship support from the Ontario Graduate Scholarship (OGS), Centre for Pharmaceutical Oncology (CPO), Pfizer Canada Graduate Fellowship, Mary Gertrude L’Anson Scholarship, Benjamin Cohen Bursary Fund, Shaping Student Life and Learning Fund, and the Department of Pharmaceutical Sciences.

I would like to acknowledge everyone who contributed to this thesis, with special thanks to the following individuals:

Dr. Reina Bendayan, my thesis supervisor, for her mentorship, guidance, and motivation throughout my graduate studies. Thank you for giving me the opportunity to pursue doctoral studies under your supervision and for entrusting me with this research project. I am grateful for your trust and support.

Dr. Deborah L. O’Connor, co-author and member of my advisory committee, for allowing me to work in her laboratory and providing training and resources to complete the in vivo portion of this thesis.

Dr. Carolyn Cummins, a member of my advisory committee, for sharing her expertise in transporter regulation by transcription factors and providing helpful advice in the design of in vitro and in vivo experiments.

Dr. Shinya Ito, a member of my advisory committee, for sharing his knowledge on cellular drug transport and providing helpful feedback on in vitro functional assays.

Dr. I. David Goldman, a co-author and collaborator at the Departments of Medicine and Molecular Pharmacology, Albert Einstein College of Medicine, for his tremendous support and guidance during my research thesis. Thank you for sharing your scientific and clinical knowledge of folate transporters and for providing antibodies and inhibitors for the in vitro experiments.

Dr. Richard H. Finnell, a co-author and collaborator at the Departments of Medicine and Molecular and Cellular Biology, Baylor College of Medicine, for all his input on this project and for generously providing breeding pairs of Folr1 knockout mice.

Dr. Robert Steinfeld, a collaborator at the Department of Pediatric Neurology, University Children’s Hospital Zürich, for his initial insights on this work.

Dr. Susanne Aufreiter, a co-author and collaborator at the Translational Medicine Program of the Hospital for Sick Children, for sharing her expertise in folate quantification by LC-MS/MS. Thank you for helping develop the protocol for folate extraction and quantification from mouse tissues.

Dr. Tozammel Hoque, a co-author and research associate in the laboratory of Dr. R. Bendayan, for providing initial laboratory training and for assisting with immunoblotting and

iv immunocytochemical analysis of transporter expression in human and rodent in vitro blood-brain barrier models.

Dr. Bogdan Wlodarczyk at the Baylor College of Medicine, for providing excellent advice on establishing the breeding colonies of Folr1 knockout mice.

All the past and present members of the Bendayan Lab for creating a positive and friendly work environment: Amy Kao, Billy Huang, Julian Gilmore, Wanying Dai, Steven Choi, Nareg Kara- Yacoubian, Olanrewaju Kayode, Olena Kis, and Tamima Ashraf. I would like to specifically thank Amila Omeragic and Sana-Kay Whyte-Allman for all the laughs and words of encouragement; the past five years would not have been the same without you both. I would also like to acknowledge the amazing graduate and undergraduate students that made significant contributions to the fulfillment of this research project: Adrian Turner, Constantine J. Georgiou, Marc Li, and Misaki Kondo.

My sincerest gratitude to my family and friends for their unwavering love and support. To my parents, Apolinar and Dinah Alam, and my sister Diane, thank you for inspiring me to choose my own path and pursue my goals. I am forever grateful for your trust, understanding, and guidance.

Lastly, I would like to thank my significant other, Richy Seto, for being my constant source of happiness and strength. You have taught me to be resilient and patient during challenging situations. This would not have been possible without your unconditional love and dedication.

Table of Contents

Abstract ...... ii

Acknowledgments ...... iv

Table of Contents ...... vi

List of Tables ...... xi

List of Figures ...... xii

List of Appendices ...... xiv

List of Abbreviations ...... xv

Chapter 1 ...... 1

Introduction ...... 1

1.1 Folate Family of Compounds ...... 3

1.1.1 Folate Derivatives ...... 3

1.1.2 Physiological Importance of Folates ...... 5

1.1.3 Global Folate Status ...... 6

1.2 Folate Transport Pathways ...... 10

1.2.1 Folate Receptors (FRs; FOLR) ...... 12

1.2.2 Reduced Folate Carrier (RFC; SLC19A1) ...... 14

1.2.3 Proton-Coupled Folate Transporter (PCFT; SLC46A1) ...... 16

1.2.4 Other Folate Transporters ...... 18

1.3 Regulation of Major Folate Transport Pathways ...... 22

1.3.1 Dietary Folate Levels ...... 22

1.3.2 Alcohol or Ethanol ...... 23

1.3.3 Transcription Factors ...... 24

1.3.4 Nuclear Receptors ...... 26

1.4 Importance of Folates in Central Nervous System (CNS) Homeostasis ...... 29

1.5 Neurological Consequences of Folate Deficiency ...... 32 vi

1.5.1 Neural Tube Defects (NTDs) ...... 32

1.5.2 Hereditary Folate Malabsorption ...... 33

1.5.3 Cerebral Folate Deficiency ...... 34

1.5.4 Kearns-Sayre Syndrome ...... 35

1.5.5 Autism Spectrum Disorders ...... 36

Rationale ...... 39

Goal ...... 40

Hypothesis ...... 40

Specific Objectives ...... 40

Chapter 2 ...... 41

Regulation of Reduced Folate Carrier (RFC) by Vitamin D Receptor at the Blood- Brain Barrier ...... 41

6.1 Abstract ...... 42

6.2 Introduction ...... 42

6.3 Materials and Methods ...... 45

6.3.1 Materials ...... 45

6.3.2 Cell Culture ...... 45

6.3.3 Rodent Brain Capillary Isolation ...... 46

6.3.4 Gene Expression Analysis ...... 47

6.3.5 Protein Expression Analysis ...... 47

6.3.6 Transport Assays ...... 48

6.3.7 Calcitriol Treatment ...... 49

6.3.8 siRNA Treatment ...... 49

6.3.9 Data Analysis ...... 49

6.4 Results ...... 50

6.4.1 Expression of Folate Receptor/Transporters at the BBB ...... 50

vii

6.4.2 Methotrexate Uptake by hCMEC/D3 Cells ...... 52

6.4.3 Effect of Calcitriol Treatment on RFC Expression in hCMEC/D3 Cells ...... 53

6.4.4 Effect of Calcitriol Treatment on RFC Function in hCMEC/D3 Cells ...... 55

6.4.5 Downregulation of RFC Expression by VDR siRNA in hCMEC/D3 Cells ...... 56

6.4.6 Effect of Calcitriol Treatment on RFC Expression in Isolated Mouse Brain Capillaries ...... 57

6.5 Discussion ...... 57

6.6 Acknowledgements ...... 62

Chapter 3 ...... 63

Upregulation of Reduced Folate Carrier by Vitamin D Enhances Brain Folate Uptake in Mice Lacking Folate Receptor Alpha ...... 63

7.1 Abstract ...... 64

7.2 Significance ...... 64

7.3 Introduction ...... 65

7.4 Materials and Methods ...... 67

7.4.1 Materials ...... 67

7.4.2 Cell Culture ...... 67

7.4.3 Mouse Brain Capillary Isolation ...... 68

7.4.4 Immunocytochemical Analysis ...... 68

7.4.5 Animal Model ...... 69

7.4.6 Calcitriol (1,25(OH)2D3) Treatment in Mice ...... 69

7.4.7 Quantification of Calcium, Phosphorus, and Calcitriol in Mouse Plasma ...... 70

7.4.8 Gene Expression Analysis ...... 70

13 7.4.9 Preparation of Intravenous [ C5]-5-formylTHF and Standard LC-MS/MS Solutions ...... 71

13 7.4.10 Distribution of [ C5]-5-formylTHF in Mouse Plasma and Brain Tissue ...... 71

7.4.11 Tissue Preparation for LC-MS/MS ...... 71

viii

13 13 7.4.12 Quantification of Plasma and Tissue [ C5]-5-formylTHF or [ C5]-5- methylTHF Concentrations by LC-MS/MS...... 72

7.4.13 Pharmacokinetic Analysis ...... 73

7.4.14 Data Analysis ...... 74

7.5 Results ...... 74

7.5.1 Localization of Folate Receptor/Transporters in Mouse BBB ...... 74

7.5.2 Expression of Folate Receptor/Transporters in WT and Folr1 KO Mice ...... 74

7.5.3 Effect of Calcitriol on RFC Expression in WT and Folr1 KO Mice ...... 76

7.5.4 Systemic Effects of Calcitriol Treatment ...... 78

7.5.5 Assessment of Folate Levels (5-formylTHF and 5-methylTHF) in WT and Folr1 KO Mice ...... 80

13 7.5.6 Plasma and Tissue Distribution of [ C5]-5-formylTHF in WT Mice ...... 81

13 7.5.7 Effect of Calcitriol Treatment on Plasma and Tissue Distribution of [ C5]-5- formylTHF in Folr1 KO Mice ...... 82

7.6 Discussion ...... 83

7.7 Acknowledgements ...... 89

Chapter 4 ...... 90

Nuclear Respiratory Factor 1 (NRF-1) Upregulates the Expression and Function of Reduced Folate Carrier (RFC) at the Blood-Brain Barrier ...... 90

8.1 Abstract ...... 91

8.2 Introduction ...... 91

8.3 Materials and Methods ...... 93

8.3.1 Materials ...... 93

8.3.2 Cell Culture ...... 94

8.3.3 Mouse Brain Capillary Isolation ...... 94

8.3.4 Gene Expression Analysis ...... 95

8.3.5 Protein Expression Analysis ...... 96

8.3.6 Transport Assays ...... 96 ix

8.3.7 Pyrroloquinoline Quinone (PQQ) Treatments ...... 97

8.3.8 siRNA Transfection ...... 98

8.3.9 Data Analysis ...... 98

8.4 Results ...... 98

8.4.1 Expression of NRF-1 and PGC-1α at the BBB ...... 98

8.4.2 Effect of PQQ Treatment on NRF-1 and PGC-1α Expression ...... 99

8.4.3 Effect of PQQ Treatment on RFC Functional Expression ...... 102

8.4.4 Effect of PQQ Treatment on Efflux Transporters ...... 104

8.4.5 Downregulation of RFC Expression by NRF-1 and PGC-1α siRNA ...... 105

8.4.6 Effect of PQQ Treatment on RFC Expression in Mice ...... 107

8.5 Discussion ...... 110

8.6 Acknowledgements ...... 114

Chapter 5 ...... 115

Overall Discussion ...... 115

Limitations ...... 121

Future Directions ...... 124

11.1 Transcriptional Regulation of RFC by VDR or NRF-1 ...... 124

11.2 Effect of PQQ-mediated NRF-1 Activation on RFC Functional Expression In Vivo .....125

11.3 Clinical Significance of NRF-1 Activation for the Treatment of Childhood Neurological Disorders ...... 126

Conclusion ...... 128

References ...... 129

Appendices ...... 162

List of Relevant Publications ...... 175

List of Tables

Table 1-1. Regulation of Major Folate Transport Pathways by Transcription Factors ...... 28

List of Figures

Figure 1-1. Chemical structures of oxidized and reduced folates ...... 4

Figure 1-2. Folate-mediated one-carbon metabolism ...... 6

Figure 1-3. Cellular expression of human and rodent folate transport pathways in different tissues ...... 12

Figure 1-4. Protein structure of FRα bound to folic acid ...... 13

Figure 1-5. Membrane topology of human RFC ...... 16

Figure 1-6. Membrane topology of human PCFT ...... 17

Figure 1-7. Membrane topology of ABC transporters ...... 20

Figure 1-8. Membrane topology of SLC transporters ...... 22

Figure 6-1. Relative expression of major folate transport systems in various in vitro and ex vivo models of the BBB ...... 51

Figure 6-2. Methotrexate uptake by hCMEC/D3 cells ...... 52

Figure 6-3. Effect of calcitriol treatment on RFC expression in hCMEC/D3 cells ...... 54

Figure 6-4. Effect of calcitriol (500nM) or ethanol (vehicle control) treatment for 24 h on methotrexate uptake ...... 55

Figure 6-5. Effect of VDR downregulation on RFC expression in hCMEC/D3 cells ...... 56

Figure 6-6. Effect of calcitriol treatment on RFC expression in isolated mouse brain capillaries ...... 57

Figure 7-1. Cellular localization of major folate transport pathways in primary mouse brain microvascular endothelial cells representative of the BBB ...... 75

Figure 7-2. Relative expression of major folate transport pathways in various tissues of WT and Folr1 KO mice ...... 75

Figure 7-3. Effect of calcitriol treatment on RFC and P-gp expression in WT mice ...... 77

Figure 7-4. Effect of calcitriol treatment on RFC and P-gp expression in Folr1 KO mice ...... 77

Figure 7-5. Effect of calcitriol treatment on body weight, plasma calcitriol, calcium, and phosphorus in WT and Folr1 KO mice ...... 79

Figure 7-6. Basal folate levels in WT versus Folr1 KO mice...... 80 xii

13 13 Figure 7-7. Plasma and brain distribution of [ C5]-5-formylTHF and its metabolite [ C5]-5- methylTHF in WT mice ...... 82

13 Figure 7-8. Effect of calcitriol treatment on plasma and brain distribution of [ C5]-5-formylTHF in Folr1 KO mice ...... 83

Figure 8-1. Relative expression of NRF-1 and PGC-1α in human (hCMEC/D3 cell line) and rodent (isolated mouse brain capillaries) in vitro models of the BBB ...... 99

Figure 8-2. Effect of PQQ treatment on NRF-1 and PGC-1α expression in hCMEC/D3 cells . 100

Figure 8-3. Effect of PQQ treatment on the expression of NRF-1 target genes in hCMEC/D3 cells ...... 101

Figure 8-4. Effect of PQQ treatment on RFC functional expression in hCMEC/D3 cells ...... 103

Figure 8-5. Effect of PQQ treatment on the expression of ABC membrane transporters in hCMEC/D3 cells ...... 104

Figure 8-6. Effect of NRF-1 and PGC-1α knockdown on RFC functional expression in hCMEC/D3 cells ...... 106

Figure 8-7. Effect of in vivo PQQ treatment in wild type mice ...... 108

Figure 8-8. Effect of in vivo PQQ treatment on expression of NRF-1 target genes in wild type mice ...... 109

Figure 11-1. Putative VDR response elements within the human SLC19A1 (RFC) promoter .. 125

xiii

List of Appendices

Appendix A: Supplemental Data for Chapter 2...………….…………….………...……….162

Appendix B: Supplemental Data for Chapter 3……...…….……………..………...... …….166

Appendix C: Supplemental Data for Chapter 4……...…….…………….………...... ….173

xiv

List of Abbreviations

1,25(OH)2D3, 1,25-dihydroxyvitamin D3 or calcitriol

ABC, ATP-binding cassette

AhR, aryl hydrocarbon receptor

AP, activating protein

ATF-1, activating transcription factor 1

ATPase, adenosinetriphosphatase

AUC, area under the concentration-time curve

BBB, blood-brain barrier

BCRP, breast cancer resistance protein

BCSFB, blood-cerebrospinal fluid barrier

CDX2, caudal-type homeobox transcription factor 2

C/EBP, CCAAT/enhancer-binding protein

CFD, cerebral folate deficiency

CL, clearance

CNS, central nervous system

CREB, cAMP responsive element binding protein

CSF, cerebrospinal fluid

ER, estrogen receptor

FR, folate receptor

GR, glucocorticoid receptor hBMEC, primary cultures of human brain-derived microvascular endothelial cells hCMEC/D3, immortalized cultures of human cerebral microvessel endothelial cells

HFM, hereditary folate malabsorption

HNF4α, hepatocyte nuclear factor 4α

KLF4, Krüppel-like factor 4 xv

KO, knockout

LC-MS/MS, liquid chromatography-tandem mass spectrometry

MeCP2, methyl-CpG-binding protein 2

MRP, multidrug resistance-associated protein

MTHFR, methylenetetrahydrofolate reductase

NRF, nuclear respiratory factor

NTD, neural tube defect

OAT, organic anion transporter

OATP, organic anion-transporting polypeptide

PCFT, proton-coupled folate transporter

PGC-1α, peroxisome proliferator-activated receptor-γ coactivator-1α

P-gp, P-glycoprotein

PMX, pemetrexed

PQQ, pyrroloquinoline quinone qPCR, quantitative polymerase chain reaction

RAR, retinoic acid receptor

RBE4, immortalized cultures of rat brain microvessel endothelial cells

RFC, reduced folate carrier

SAM, S-adenosylmethionine siRNA, small interfering RNA

SLC, solute carrier

Sp1, specificity protein 1 t1/2b, plasma terminal elimination half-life

TEER, transendothelial electrical resistance

Tfam, mitochondrial transcription factor A

TFB1M, mitochondrial transcription factor B type 1 xvi

TFB2M, mitochondrial transcription factor B type 2

THF, tetrahydrofolate

USF, upstream stimulatory factor

V, volume of distribution

VDR, vitamin D receptor

WT, wild type

YY1, yin-yang 1

xvii

Chapter 1 Introduction

This chapter contains excerpts from published as well as unpublished review articles indicated below.

Contents of following sections were taken and updated from: Alam C, Kondo M, Goldman ID, O’Connor DL, and Bendayan R. (2019). Clinical implications of folate transport in the central nervous system. Review manuscript in preparation; submitted to Trends in Pharmacological Sciences (TIPS).

1.1 Folate Family of Compounds; 1.1.1 Folate Derivatives; 1.1.2 Physiological Importance of Folates; 1.1.3 Global Folate Status; 1.1.3.1 Folate Deficiency; 1.1.3.2 Folate Supplementation

1.2 Folate Transport Pathways; 1.2.1 Folate Receptors (FRs); 1.2.2 Reduced Folate Carrier (RFC); 1.2.3 Proton-Coupled Folate Transporter (PCFT); 1.2.4 Other Folate Transporters; 1.2.4.1 ATP- Binding Cassette (ABC) Transporters; 1.2.4.2 Solute Carrier (SLC) Transporters

1.3 Regulation of Major Folate Transport Pathways; 1.3.1 Dietary Folate Levels; 1.3.2 Alcohol or Ethanol; 1.3.3 Transcription Factors; 1.3.4 Nuclear Receptors

1.4 Importance of Folates in Central Nervous System (CNS) Homeostasis

1.5 Neurological Consequences of Folate Deficiency; 1.5.1 Neural Tube Defects (NTDs); 1.5.2 Hereditary Folate Malabsorption; 1.5.3 Cerebral Folate Deficiency; 1.5.4 Kearns-Sayre Syndrome; 1.5.5 Autism Spectrum Disorders

Author Contributions: C Alam (literature review; initial draft of sections on folate transport pathways, folate absorption and distribution, and neurological consequences of folate deficiency; Figures 2 and 3), M Kondo (literature review; initial draft of sections on folate status in North America and regulation of major folate transport pathways; Figure 3), ID Goldman (editorial review of several manuscript drafts), DL O’Connor and R Bendayan (overall conceptual and editorial review of several manuscript drafts)

In addition, some contents of the following sections were taken and reproduced with permission from: Alam C, Whyte-Allman SK, Omeragic A, Bendayan R. (2016). Role and modulation of drug transporters in HIV-1 therapy. Advanced Drug Delivery Reviews, 103:121-143. Copyright 2016 Elsevier.

1.2.4.1 ATP-Binding Cassette Transporters

1.2.4.2 Solute Carrier Transporters

Author Contributions: C Alam (literature review; initial draft of sections on blood-intestinal mucosa barrier and regulation of drug transporters by HIV-1 pharmacotherapy; initial submission; responses to reviewers’ comments; resubmission; Figure 2 and Table 5), SK Whyte-Allman (literature review; initial draft of sections on HIV-1 infection, effect of drug transporters on antiretroviral drug disposition, and male and female genital systems; responses to reviewers’ comments; Figure 2 and Table 2), A Omeragic (literature review; initial draft of sections on HIV- 1 pharmacotherapy, brain barrier sites, and regulation of drug transporters by HIV-1 infection; responses to reviewers’ comments; Figure 2 and Tables 1, 3, 4), and R Bendayan (overall conceptual and editorial review of several manuscript drafts and responses to reviewers’ comments).

1.1 Folate Family of Compounds

Folates are a family of water-soluble vitamins that are essential for key biosynthetic processes in mammalian cells. They are best known for their role as one-carbon donors in the synthesis and repair of DNA, production of amino acids, and regulation of gene expression (Tibbetts & Appling, 2010). Mammals obtain folates from dietary sources as they lack the metabolic capacity for folate biosynthesis. Traditionally, folates were acquired from food such as liver, green leafy vegetables, beans, and other legumes; however, folic acid fortification of white wheat flour and select grain products in North America and other developed countries now represent a predominant source of dietary folate. Emerging evidence also suggests that microbial folate biosynthesis in the colon likely plays a role in meeting folate requirements of mammals (Aufreiter et al., 2009; T. H. Kim, Yang, Darling, & O’Connor, 2004; Lakoff et al., 2014). Historically, nutritional folate deficiency is one of the most common vitamin deficiencies worldwide causing a range of health issues including megaloblastic anemia, birth defects in developing embryos, and neurodevelopmental disorders in infants (R. L. Bailey, West, & Black, 2015). Recently it has been proposed that elevated consumption of high levels of synthetic folate (folic acid) may also pose negative health outcomes such as masking of vitamin B12 deficiency (Smith, Kim, & Refsum, 2008). Alterations in folate status have major biological ramifications especially in the developing central nervous system (CNS).

1.1.1 Folate Derivatives

‘Folate’ is a generic term assigned to a group of structurally and functionally related compounds belonging to the B9 family of vitamins. The folate molecule consists of a p- aminobenzoic acid core linked to a pteridine ring at one end and to glutamic acid at the other (Figure 1-1) (Lucock, 2000; R Zhao, Matherly, & Goldman, 2009). The parent structure is folic acid, an oxidized synthetic form often used in dietary supplements and food fortification. Folic acid itself is not biologically active, but it is readily reduced in the intestine or liver and converted into various forms of active tetrahydrofolate (THF) molecules such as 5-methylTHF, 5,10- methyleneTHF, 5,10-methenylTHF, and 10-formylTHF. THF derivatives associate with different one-carbon moieties (i.e., methyl, methylene, methenyl, or formyl) at the N5 or N10 positions enabling them to serve as one-carbon donors in cellular biosynthetic processes (Lucock, 2000; R Zhao, Matherly, et al., 2009). They also undergo glutamate conjugation reactions catalyzed by

folylpolyglutamate synthetase, in order to produce THF polyglutamates with enhanced cellular retention and are preferred substrates of folate-dependent enzymes. Folate derivatives differ in terms of the oxidation state of the pteridine ring, the type of one-carbon substituent, and the extent of glutamate conjugation (Figure 1-1).

Folate status of Canadian and Korean women

Fig. 1. The chemical structure of folic acid or pterolylmonoglutamic acid. Fig. 2. Simplified diagram of intracellular folate metabolism involving DNA ٙ٘ Figure 1-1. Chemical structures of oxidized and reduced folates. The folic acid molecule consistsbiosynthesis of a and methylation. pteridine ring, p-aminobenzoic acid core, and glutamic acid. Folate derivates are denoted by the “R” substitutions listedlactation. in the box The below. rationale They for differ and from history each of other recommending based on the oxidation folic state of the pteridine ring, typeacid-containing of one-carbon substituent supplements at N5 during and N10 the position periconceptionals, and the length period of the glutamic acid chain. Figure reproduced with permission from (Lindzon & O’Connor, 2007). and pregnancy in Canada is described as is the folic acid fortification policy. The impact of folic acid fortification is discussed and unresolved and immerging issues associated with Due to their role in cell growth and replication, folates were further exploited for their this policy are described. While the incidence of NTDs pre-folic likewise metabolized via folate-dependent reactions. applications in anticancer therapy. Structural analogs of folates, known as antifolates, have been acid fortification of the food supply in Canada were higher than Folate in the form of 5-methyltetrahydrofolate, is involved in developed and usedthat ofclinically South Korea to target today, and bloodinhibit folate folate levels-dependent of Korean enzymes women involved inremethylation crucial of homocysteine to methionine. The latter is a metabolic pathwaysare strikingly such as DNA similar. synthesis. We will Clinically briefly relevantexplore theseantifolates parallels include in methotrexate,precursor for S-adenosylmethionine (SAM), the principle methyl an attempt to understand whether folic acid fortification of the pemetrexed, raltitrexed, and pralatrexate (Gonen & Assaraf, 2012; Matherly, Wilson, group & Hou, donor in the body. SAM is involved in methylating food supply in South Korea might be worth consideration. cytosine in DNA and is thought to play a key role in post- 2014). These compounds utilize the same transport pathways as physiological folates to enter transcription regulation of gene expression. Myelin maintenance tumor cells and inhibit folate-dependent metabolic reactions. and neural function are likewise dependent on the methylation Function reactions involving SAM.

As described above, folates function in many coenzyme forms in acceptance, redox processing and transfer of one-carbon units. Recommended Dietary Intake? In order to carry out these functions, folates in nature are typically 4

reduced to either di-hydro or tetrahydrofolate forms with Folates are synthesized in plants and certain bacteria but not hydrogens at the 5, 6, 7 and 8 positions (Fig. 1). Further, various in mammalian cells, and thus, they must be obtained in the diet. one-carbon units can be carried at the N5 and N10 positions The recommended dietary allowance (RDA=dietary intake or bridging the same. Finally, in nature, a significant proportion sufficient to meet the requirements of 97-98% of healthy of folates are polyglutamylated, meaning they have several individuals) of folate for adults, is 400g/day of dietary folate glutamates linked together to create what is commonly referred equivalents (DFEs) (Institute of Medicine, 1998). The concept to as a polyglutamate tail. of DFEs for folate was introduced in North America in 2000 Embryonic, fetal and infant growth occurs more rapidly than to account for the differences in bioavailability of synthetic folic at any other stage of the life cycle. The anabolic activity that acid and naturally occurring folates (Institute of Medicine, 1998). must occur during pregnancy and lactation to support this growth, Compared to folic acid consumed alone (a relative availability and requisite DNA, RNA and amino acid biosynthesis dictates of 100%), folic acid ingested with food or used as a fortificant an elevated dietary requirement for folate. Folate, in the form is thought to be only ~85% available. In contrast, naturally of 5,10-methylene tetrahydrofolate, during DNA synthesis acts occurring food folates are thought to be only ~50% available as a methyl donor for the enzyme thymidylate synthase which (Institute of Medicine, 1998; Pfeiffer et al., 1997). Thus, folic converts deoxyuridine monophosphate to thymidine mono- acid is calculated to be 1.7 (85 divided by 50) more available phosphate (Fig. 2). Folate in the form of 10-formyl tetrahy- than naturally occurring food folates. Hence, in order to convert drofolate is necessary for the synthesis of the purines adenine all forms of dietary folate into DFEs, the following calculation and guanine, the nucleic acid building blocks of DNA and RNA. was developed: g of DFEs provided = g food folate + (g folic The amino acids methionine, serine, glycine and histidine are acid × 1.7). The Estimated Average Requirement (EAR = dietary

1.1.2 Physiological Importance of Folates

Folate plays a central role in the regulation of cell function in all tissues. It exists in multiple coenzyme forms to facilitate the acceptance, redox processing, and transfer of one-carbon units during key biosynthetic processes. In particular, folate-mediated one-carbon metabolism is involved in at least three major interrelated metabolic cycles required for: i) de novo synthesis of thymidine and purine building blocks of DNA and/or RNA, ii) production of amino acids, and iii) biosynthesis of the universal methylating agent, S-adenosylmethionine (SAM) (Lucock, 2000; Tibbetts & Appling, 2010). As shown in Figure 1-2, folate metabolism begins with the two-step reduction of folic acid to an unsubstituted THF molecule by dihydrofolate reductase to form 5,10- methyleneTHF. Folate in the form of 5,10-methyleneTHF is involved in nucleotide synthesis by serving as a methyl donor in the conversion of deoxyuridine monophosphate to deoxythymidine monophosphate (thymidylate). Oxidation of 5,10-methyleneTHF to 10-formylTHF also enables the production of purine bases, adenine and guanine. Additionally, 5,10-methyleneTHF can be further reduced to 5-methylTHF by methylenetetrahydrofolate reductase (MTHFR) in the presence of riboflavin (vitamin B2). 5-methylTHF is the major circulating form of folate in the body and requires active transport to vital tissues. 5-methylTHF is also an important coenzyme in the conversion of homocysteine to methionine in a cobalamin (vitamin B12)-dependent reaction. Methionine is a precursor for SAM, which is the universal methyl donor for numerous methylation reactions in the body including the methylation of DNA, proteins, phospholipids, and neurotransmitters. SAM-mediated methylation is also a key mechanism in epigenetic programming early in life and in the biosynthesis of hormones such as epinephrine. Myelin maintenance and neural function are likewise dependent on methylation processes involving SAM. Furthermore, folates are involved in amino acid metabolism. The conversion of SAM to homocysteine enables production of either cysteine or methionine. The formation of 5,10- methyleneTHF from THF also acts as an important branch point in the generation of glycine and serine. Given the numerous roles of folate coenzymes in the body, it is evident that folate serves as a crucial component of cell growth and survival. Hence, changes in folate status can consequently disrupt normal cellular processes.

FOLIC ACID CYCLE DNA CYCLE Folic Acid dTMP DHFR

DNA DHF METHYLATION CYCLE B6 Synthesis DHFR TYMS SHMT MAT dUMP Serine THF Methionine SAM R Glycine MTHFD AICART Purines 10-FormylTHF GART Methylation MTHFD 5-FormylTHF MS MT Reactions 5,10-MethenylTHF B12 MTHFD

5,10-MethyleneTHF 5-MethylTHF Homocysteine SAH R-CH3

MTHFR B2 AHCY

Figure 1-2. Folate-mediated one-carbon metabolism. Folate metabolism (green) is at the branch of two major metabolic cycles – DNA cycle (blue) and methylation cycle (pink). This set of reactions, collectively known as folate-mediated one-carbon metabolism, are essential for the production of: i) thymidine and purine precursors of nucleic acids, ii) several amino acids required for protein synthesis, and iii) the universal methylation agent, S-adenosylmethionine (SAM).

Abbreviations: AHCY, S-adenosylhomocysteine hydrolase; AICART, 5-amino-4-imidazolecarboxamide ribonucleotide transformylase; DHF, dihydrofolate; DHFR, dihydrofolate reductase; dUMP, deoxyuridine monophosphate; dTMP, deoxythymidine monophosphate; GART, β-glycinamide ribonucleotide transformylase; MAT, methionine adenosyl transferase; MS, methionine synthase; MT, methionine transferase; MTHFD, methylenetetrahydrofolate dehydrogenase; MTHFR, methylenetetrahydrofolate reductase; SAH, S-adenosylhomocysteine; SAM, S-adenosylmethionine; SHMT, serine hydroxymethyltransferase; THF, tetrahydrofolate; TYMS, thymidilate synthase.

1.1.3 Global Folate Status

Folate requirements in mammals are largely met through dietary sources since they lack the enzymatic capacity for folate biosynthesis. Excellent naturally occurring sources of folate include green leafy vegetables, liver, beans, and legumes. However, at present, white wheat flour and other grain products fortified with folic acid have become the predominant source of dietary folate in North America and other developed countries. Folate requirements vary depending on age, sex, or physiological state (Institute of Medicine; Food and Nutrition Board, 1998). According to the Food and Nutrition Board at the Institute of Medicine, the recommended daily allowance of folate for adults is 400 μg/day. For children, it varies between 150-400 μg/day depending on age (i.e., 150 for 1-3 years, 200 for 4-8 years, 300 for 9-13 years, and 400 for 14-18 years). An adequate 6

intake dose of 65 and 80 μg/day has also been established for infants 0-6 months and 7-12 months old, respectively. For pregnant and lactating women, the recommended daily allowance is 600 and 500 μg/day, respectively. Furthermore, Health Canada recommends women of childbearing age (15-45 years) to maintain a folate-rich diet and take a daily multivitamin supplement containing 400 μg of folic acid to prepare for a healthy pregnancy (Health Canada, 2007).

1.1.3.1 Folate Deficiency

Historically, nutritional folate deficiency is one of the most common vitamin deficiencies worldwide, and continues to be so in countries that have not mandated folic acid fortification of their food supply. Children, pregnant women, and individuals with malabsorptive disorders are at the greatest risk for folate deficiency. To assess low folate status, the World Health Organization has defined cut-off values of <3 ng/ml (<6.8 nmol/L) for serum/plasma and <100 ng/ml (<226.5 nmol/L) for red blood cell folate (World Health Organization; Department of Nutrition for Health and Development, 2012). Serum folate is reflective of recent dietary folate intake, whereas red blood cell folate is an indicator of long-term folate status since erythrocytes have a turnover rate of approximately 120 days. The first sign of suboptimal folate intake begins with a decrease in serum folate, followed by a reduction in red blood cell folate and an increase in serum homocysteine levels. Given the physiological roles of folate, deficiency often results in disorders associated with defective cell division, accumulation of toxic metabolites, or impairment of methylation reactions. The primary clinical manifestation of folate deficiency is the onset of megaloblastic anemia, which is characterized by the appearance of large and abnormally nucleated erythrocytes resulting from ineffective cell proliferation (R. L. Bailey et al., 2015). Symptoms also include fatigue, irritability, headache, lack of concentration, heart palpipations, and shortness of breath (Institute of Medicine; Food and Nutrition Board, 1998). Folate requirement and risk of deficiency is also highest during anabolic periods of life such as in pregnancy (Lindzon & O’Connor, 2007; Tamura & Picciano, 2006). Folic acid supplementation is highly recommended during the periconceptional period to reduce the risk of neural tube defects (NTDs) (i.e., spina bifida and anencephaly) and other congenital malformations (i.e., congenital heart defects, oral cleft lip, and palette) in the developing embryo (Badovinac, Werler, Williams, Kelsey, & Hayes, 2007; Czeizel, Tímár, & Sárközi, 1999; MRC Vitamin Study Research Group, 1991). However, it has been estimated that only 30% of women globally take folic acid supplements prior to conception (Ray, Singh, & Burrows, 2004), and to date there remains a lack of reliable global

estimates for folate deficiency in the world’s most vulnerable populations (i.e., women of reproductive age, pregnant women, and young children residing in low to middle-income countries) (McLean, de Benoist, & Allen, 2008; Rogers et al., 2018). Additionally, folate status has been shown to have an inverse association with the development of malignancies including colorectal, prostate, and breast cancer (Y. Kim, 1999, 2003). This correlation can be explained by the crucial role of folate in DNA synthesis and repair. In fact, in vitro and in vivo evidence suggest that folate deficiency may induce DNA strand breaks, chromosomal and genomic instability, uracil misincorporation, and increased incidence of mutations (Y. Kim, 2003). Other disorders implicated in suboptimal folate status may also be a direct result of elevated serum homocysteine concentration, which relies on folate for its clearance from the body. Hyperhomocysteinemia has been positively associated with stroke (Huang et al., 2012; Huo et al., 2015) and cognitive decline in Alzheimer’s disease and dementia (Ravaglia et al., 2005; Seshadri et al., 2002).

Nutritional folate deficiency has multiple contributing factors. In most cases, poor or restricted diet leads to inadequate folate intake. This can be further exacerbated by malabsorption conditions resulting from chronic/heavy alcohol consumption (Cravo et al., 1996; De La Vega et al., 2001), diseases of the small intestine (i.e., celiac disease, tropical sprue), and co-administration of drugs that interfere with folate metabolism (i.e., methotrexate, sulfasalazine) (L. H. Allen, 2008). Folate status is also influenced by inherited disorders of folate metabolism and transport. A common single nucleotide polymorphism (677C>T) on the MTHFR gene leads to the substitution of alanine to valine, which produces a thermolabile variant of the MTHFR enzyme with 50-70% decrease in function when folate intake is limited (Frosst et al., 1995; Goyette et al., 1994). MTHFR is responsible for the conversion of 5,10-methyleneTHF to 5-methylTHF, hence impaired MTHFR activity consequently impairs SAM production and cellular methylation reactions. During periods of folate insufficiency, individuals that are homozygous (TT) for the C677T allele have shown higher plasma homocysteine and lower plasma folate concentrations compared to people with wild type 677CC MTHFR gene (Frosst et al., 1995). The presence of this mutation has also been associated with an increased risk of cardiovascular disorders, NTDs, adverse pregnancy outcomes, genome hypomethylation, and several cancers (Ueland, Hustad, Schneede, Refsum, & Vollset, 2001). The prevalence of MTHFR 677C>T polymorphism contains regional and ethnic variations. The 677TT genotype is most common in individuals of Mexican (32%), Chinese (26%) or Southern Italian (20%) descent, and least common in African populations (0.3-0.8%) (Botto &

Yang, 2000; Wilcken et al., 2003). For Caucasian populations, approximately 8-14% in North America, 6-14% in Northern Europe, and 15-20% in Southern Europe have been documented to be homozygous (TT) for the MTHFR 677C>T mutation (Wilcken et al., 2003). A study in Northern China also reported a prevalence of 35.1% for the 677TT genotype (Crider et al., 2011), and approximately 12 to 18% anywhere else in Asia (Wilcken et al., 2003). A second common polymorphism on the MTHFR gene is the 1298A>C substitution, which causes a change from glutamate to alanine. This mutation also leads to decreased MTHFR enzyme activity but without marked effects on homocysteine or folate plasma levels in individuals who are homozygous (CC) for the A1298C allele (N. M. J. van der Put et al., 1998; Weisberg I, Tran P, Christensen B, Sibani S, 1998). Additionally, inactivating mutations on genes that encode for specific folate transport systems (i.e., FOLR1, SLC46A1) can further impair folate absorption and distribution (Qiu et al., 2006; Steinfeld et al., 2009). These mutations, although rare, have been demonstrated to be the underlying mechanism of several debilitating childhood disorders such as hereditary folate malabsorption and cerebral folate deficiency. These conditions are further discussed in Section 1.5.

1.1.3.2 Folate Supplementation

As part of a public health strategy to reduce folate-associated NTDs, in 1998 mandatory folic acid fortification of white wheat flour and select grain products was initiated in the United States and Canada. Several other countries such as Chile, Costa Rica, Australia, and South Africa have implemented similar policies. Since then, fortified “grains” represent the largest source of dietary folate. Serum folate concentrations in North American populations have at least doubled (normal cut-off of <6.8 nmol/L for deficiency) compared to pre-folic acid fortification, and red blood cell folate concentrations have increased by at least 60% (normal cut-off of <226.5 nmol/L for deficiency) (R. L. Bailey et al., 2012). Folate deficiency is virtually non-existent in American and Canadian populations. In the United States, post-fortification (1999-2010) levels of serum and red blood cell folate in the general population (4-80 years of age) were reported to be greater than 40 and 1000 nmol/L, respectively (C M Pfeiffer et al., 2012). The same study showed that there is an extremely low prevalence (<1%) of suboptimal serum and red blood cell folate concentrations in the United States post-fortification, regardless of age, gender, race/ethnicity, and dietary supplement use (C M Pfeiffer et al., 2012). In Canada, 40% of the population (6-45 years of age) has shown high levels of red blood cell folate (>1360 nmol/L) (Colapinto, O’Connor, & Tremblay,

2011). Approximately 78% of Canadian women of childbearing age (15-45 years) also showed folate concentrations that are already optimal for NTD risk-reduction (>906 nmol/L). Furthermore, unmetabolized folic acid is frequently detected in the circulation of healthy North Americans living in the post-fortification era (Christine M Pfeiffer et al., 2015; Mary R. Sweeney, McPartlin, Weir, Daly, & Scott, 2006). This may be explained by the limited activity of dihydrofolate reductase, the enzyme responsible for folic acid conversion to metabolically active THF derivative in human liver. High oral doses of folic acid (>260-280 μg) are known to saturate the hepatic metabolic capacity of dihydrofolate reductase, thereby resulting in the appearance of circulating unmetabolized folic acid which may pose a potential health risk (M R Sweeney, Mcpartlin, Weir, & Scott, 2003; Mary R. Sweeney et al., 2006). Elevated consumption of high levels of folic acid has also been linked to a number of negative health outcomes such as masking of vitamin B12 deficiency, which can lead to a missed diagnosis and high risk of cognitive impairment (Smith et al., 2008). Several reports have also suggested that folate supplementation can accelerate the progression of established neoplasms, particularly in rodent models of colorectal cancer (Y. Kim, 2003; Song, Sohn, et al., 2000; Song, Medline, Mason, Gallinger, & Kim, 2000). Available data from human studies, however, remain limited and controversial (Stevens et al., 2011; Wien et al., 2012). More recently, researchers have focused on investigating the epigenetic impacts of maternal folate status on offspring disease phenotype. Several human studies have found that high maternal folate status is associated with increased incidence and/or prevalence of insulin resistance (Yajnik et al., 2008), atopic dermatitis (Kiefte-de Jong et al., 2012), compromised respiratory health (Håberg, London, Stigum, Nafstad, & Nystad, 2009), and retinoblastoma (Orjuela et al., 2012) in the offspring. Additional research is required to confirm these initial findings and further understand the relationship between high folate intake and offspring health.

1.2 Folate Transport Pathways

Most naturally occurring dietary folates exist as polyglutamates that must be hydrolyzed into monoglutamate forms for transport across cell membranes or tissue barriers. In humans and pigs, this is achieved by the enzyme glutamate carboxypeptidase II, which is located at the brush- border membrane of the jejunum (Shafizadeh & Halsted, 2009). In rats, hydrolysis is facilitated by γ-glutamyl hydrolase located within the intestinal lumen (Shafizadeh & Halsted, 2007). Gastrointestinal absorption of folate monoglutamates occurs primarily across the apical brush- border of the proximal jejunum (R Zhao et al., 2008; R Zhao, Matherly, et al., 2009). There is also 10

evidence in the literature demonstrating folate absorption across the colon, where monoglutamylated folates are derived mainly from resident microflora (Aufreiter et al., 2009; T. H. Kim et al., 2004; Lakoff et al., 2014). Dietary folates are metabolized into active THF derivatives (i.e., 5-methylTHF) either during transit through the intestinal mucosa (in rodents) and/or the liver (in humans) (Patanwala et al., 2014). Folates in the liver are either converted to polyglutamate storage forms, secreted into the bile, or they can enter the systemic circulation for distribution to peripheral tissues. 5-methylTHF represents the major circulating form of folates in plasma and the predominant form entering human metabolism (Lucock, 2000; R Zhao, Matherly, et al., 2009). Circulating folates that are not bound to serum proteins are eventually delivered to the kidneys, where they can be filtered and reabsorbed. Since folates are anionic at physiological pH, they are unable to cross biological membranes through diffusion and must rely on specific transport systems for their permeability into cells and across epithelia. In mammalian tissues, at least three major folate transport pathways have been characterized – the folate receptors (FRs), reduced folate carrier (RFC), and the proton-coupled folate transporter (PCFT) (R Zhao, Diop- Bove, Visentin, & Goldman, 2011). The cellular localization of folate transport pathways in a select number of tissues is illustrated in Figure 1-3.

Figure 1-3. Cellular expression of human and rodent folate transport pathways in different tissues. Cellular localization of folate receptors and transporters are depicted in intestinal epithelia (A), liver hepatocytes (B), kidney proximal tubules (C), brain capillary endothelial cells (D), choroid plexus epithelia (E), and placental syncytiotrophoblasts (F). Depending on their localization in polarized epithelial/endothelial cells, folate receptors or transporters may display secretory or absorptive function. Influx transporters (green) primarily mediate folate uptake into cells, while efflux transporters (blue) mediate cellular export of folates. Figure created with BioRender.

1.2.1 Folate Receptors (FRs; FOLR)

At least four FR isoforms (α, β, γ, δ), encoded by the FOLR multigene family (FOLR1, FOLR2, FOLR3, and FOLR4, respectively), have been identified in mammalian tissues (Antony, 1996; Elnakat & Ratnam, 2004). FRα and FRβ are glycosylphosphatidylinositol-anchored cell surface glycoproteins involved in cellular uptake of folates (Antony, 1996; Elnakat & Ratnam, 2004; Kamen & Smith, 2004). FRγ is a secreted protein whose function is currently not well understood (Antony, 1996; Elnakat & Ratnam, 2004; Kamen & Smith, 2004). FRδ, also known as Juno, is not involved in folate transport, but serves as an important cell surface receptor that facilitates fertilization of mammalian eggs (Bianchi, Doe, Goulding, & Wright, 2014). Among all folate transport pathways, FRα and FRβ bind folic acid and its reduced derivatives with the highest 12

LETTER doi:10.1038/nature12327

Structural basis for molecular recognition of folic acid by folate receptors

Chen Chen1,2*,JiyuanKe1*,X.EdwardZhou1,WeiYi3, Joseph S. Brunzelle4,JunLi5,Eu-LeongYong5, H. Eric Xu1,3 &KarstenMelcher1

Folate receptors (FRa,FRb and FRc) are cysteine-rich cell-surface Folates (vitamin B9) are important one-carbon donors for the syn- glycoproteins that bind folate with high affinity to mediate cellular thesis of purines and thymidine—essential components of nucleic uptake of folate. Although expressed at very low levels in most tissues, acids—and indirectly, via S-adenosyl methionine, for methylation of folate receptors, especially FRa,areexpressedathighlevelsinnumer- DNA, proteins and lipids9. Folate deficiency is therefore associated ous cancers to meet the folate demand of rapidly dividing cells under with many diseases, including fetal neural tube defects, cardiovascular

low folate conditions1–3. The folate dependency of many tumours has disease and cancers10. In adult tissues, folate is mainly taken up by been therapeutically and diagnostically exploited by administration reduced folate carrier, a ubiquitously expressed anion channel that has 4,5 11 affinity of (K anti-FRm = 1-a10antibodies, nM) at high-affinityneutral pH antifolatesand facilitate,folate-basedimaging transport through recrelativelyeptor-mediated low folate-binding affinity (Km 5 1–10 mM) . By contrast, 6–8 12 agents and folate-conjugated drugs and toxins . To understand high-affinity uptake of the food supplement folic acid (Kd , 1nM) endocytosishow (Antony, folate binds 1996; its Chen receptors, Chen we et determinedal., 2013; Kamen the crystal & Smith, structure 2004) of (Figureand 1 the-4) physiologically. Folate prevalent folate N5-methyltetrahydrofolate (5- uptake ishuman initiated FR witha in complexa folate molecule with folic binding acid at 2.8to the A˚ resolution. FR on the FRcella surface,has a followedmTHF) requires by the the function of three subtypes of folate receptor (FRa, globular structure stabilized by eight disulphide bonds and contains FRb and FRc), which are cysteine-rich glycoproteins that mediate invagination of the plasma membrane forming a vesicle (endosome) that circulates in the a deep open folate-binding pocket comprised of residues that are folate uptake through endocytosis. Inside of the cell, the acidic envir- endosomalconserved compartment, in all receptor where subtypes.it is later Thacidifiedefolatepteroatemoietyisburied to a pH of approximately 6-onment6.5. The of folate the endosome promotes the release of folate from receptors, inside the receptor, whereas its glutamate moiety is solvent-exposed which is then transported into the cytoplasm by proton-coupled folate ligand is subsequently released from the receptor and exported into the cytoplasm by a process and sticks out of the pocket entrance, allowing it to be conjugated to transporter13. The expression of folate receptors is largely restricted that operatesdrugs optimally without adverselyat low pH. affecting FRa binding. The extensive inter- to cells important for embryonic development (for example, placenta actions between the receptor and ligand readily explain the high and neural tubes) and folate resorption (kidney). Among the three folate-binding affinity of folate receptors and provide a template FR isoforms, FRa is the most widely expressed, with very low levels for designing more specific drugs targeting the folate receptor system. in normal tissues, but high expression levels in many tumours14.As

N139 α3–α4 loop Figure 1 | Structure of FRa bound to folic acid. a, Two views of the a b α4 N139 α4 α4 α3 complex, with FRa in green, folic acid in grey, NAG in orange and the α3 α1–α2 loop NAG α5 disulphide bonds depicted as yellow α5 α5 α3 α2 sticks. The N and C termini are α2 NAG N47 4 labelled. b, Ribbon diagram of FRa, α1 90° α1 β N47 β4 N179 β3 with folic acid and NAG in green NAG stick presentations, overlaid with the β3 α6 2 β β2 semi-transparent receptor surface. β2 N179 β1 β1 C c, Charge distribution surface of FRa β1 N with a close-up view of the ligand- α6 α6 β1–β2 loop binding pocket entrance. Folic acid C N C carbon atoms are coloured grey, N nitrogen atoms blue, and oxygen atoms red. A colour-code bar c N139 (bottom) shows an electrostatic scale Figure 1-4. Protein structure of FRα bound to folic acid. FRα has a globular structure comprised of four from 23to13 eV. long α-helices (α1, α2, α3, α6), two short α-helices (α4, α5), four short β-strands (β1-4), and many loop regions. It also contains three predicted N-glycosylation sites at N47, N139, and N179. Figure reproducW140 ed with permission from (Chen Chen et al., 2013).

N179 K136 N179 180º N47

FRs are characterized by their restricted pattern of tissue expression. In humans,Y60 FRα is predominantly expressed in epithelial cells of the choroid plexus, proximal kidney tubules, uterus, W102 C ovary, fallopian tube, placenta, and retina (Kamen & Smith, 2004; N Parker et al., 2005). It is C typically localized on the apical surface of these polarized epithelial cells, except in the retina –3.000 3.000 where FRα is expressed at the basolateral membrane (Chancy et al., 2000), as well as in choroid 1Program for Structural Biology and Drug Discovery, Van Andel Research Institute, 333 Bostwick Avenue North East, Grand Rapids, Michigan 49503, USA. 2National University of Singapore Graduate plexus epitheliumSchool for Integrative where Science it has and been Engineering, detected National in both University apical of Singapore, and basolateral Singapore 117456, compartments Singapore. 3VARI/SIMM through Center, Center for Structure and Function of Drug Targets, CAS-Key Laboratory 4 immunohistochemicalof Receptor Research, analysis Shanghai Institute(Grapp of Materiaet al., Medica,2013) Chinese. A number Academy of of Sciences,studies Shanghaihave also 201203, reported China. robusLife Sciencest Collaborative Access Team, Synchrotron Research Center, Northwestern University, Argonne, Illinois 60439, USA. 5Department of Obstetrics & Gynecology, National University Hospital, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 119074, Singapore. expression*These of authorsFRα contributedin a variety equally of to non this work.-mucinous adenocarcinomas of the ovary, cervix, uterus,

brain, and to a lesser extent in breast, colon, and renal carcinomas (Parker et al., 2005; Toffoli et al., 1997; Weitman, Frazier, & Kamen, 1994; M. Wu, Gunning, & Ratnam, 1999). Furthermore, significant levels of FRβ are present in human placenta and hematopoietic cells (i.e., CD34+) and tissues (i.e., spleen and thymus) (Ratnam, Marquardt, Duhring, & Freisheim, 1989; Reddy et al., 1999; Ross et al., 1999), as well as in malignant cells of chronic and acute myelogenous leukemia (Ross et al., 1999; Hui Wang, Zheng, Behm, & Ratnam, 2000). On the other hand, FRγ is constitutively secreted at low levels from normal and leukemic hematopoietic tissues such as the spleen, thymus, and bone marrow (Shen, Wu, Ross, Miller, & Ratnam, 1995). High levels of FRδ expression has also been documented in mouse oocytes (Bianchi et al., 2014) and CD4+CD25+ regulatory T-cells (Yamaguchi et al., 2007).

Of the four FR isoforms, FRα has been the most characterized in terms of function and physiological importance. In the kidney, FRα located at the apical surface of proximal renal tubules constitutes an efficient reabsorption mechanism for filtered folates, which is essential for the conservation of the vitamin (Birn, Spiegelstein, Christensen, & Finnell, 2005; Selhub, Nakamura, & Carone, 1987). FRα localized to the basolateral membrane of choroid plexus epithelial cells also serves as the primary route for 5-methylTHF uptake into the CNS (Grapp et al., 2013). Inactivation of this receptor can lead to alterations in normal folate homeostasis causing increased renal clearance of folates (Birn et al., 2005) as well as cerebral folate deficiency during early childhood (Grapp et al., 2012, 2013; Vincent T Ramaekers et al., 2005; Steinfeld et al., 2009). In mice, it was further demonstrated that FRα-mediated transport is essential during embryonic and fetal development. Several groups have shown using Folr1 null (Folr1-/-) mice that functional inactivation of FRα produces embryos with severe growth retardation and multiple developmental abnormalities that altogether contribute to embryonic lethality (Piedrahita et al., 1999; Spiegelstein et al., 2004; Tang & Finnell, 2003; Tang, Santillano, Wlodarczyk, Miranda, & Finnell, 2005; H. Zhu, Cabrera, et al., 2007). However, supplementation of pregnant dams with high doses of dietary folate during gestation enables live birth of Folr1 null (Folr1-/-) animals with normal phenotype.

1.2.2 Reduced Folate Carrier (RFC; SLC19A1)

The RFC, a member of the solute carrier (SLC) superfamily of facilitative transporters and product of the SLC19A1 gene, facilitates folate delivery to mammalian cells and tissues (Matherly & Hou, 2008). It is a transmembrane protein consisting of 12 transmembrane domains with

intracellularly oriented amino and carboxyl termini (Cao & Matherly, 2004; Ferguson & Flintoff, 1999; X. Y. Liu & Matherly, 2002), and a single N-glycosylation site located in the loop connecting the first and second transmembrane domains (Wong, Zhang, Proefke, & Matherly, 1998) (Figure 1-5). RFC functions optimally at pH 7.4 as a high-capacity bidirectional antiporter exchanging folates with intracellular organic phosphates (Goldman, Lichtenstein, & Oliverio, 1968; Matherly & Hou, 2008; Rongbao Zhao & Goldman, 2013). Compared to FRα, RFC exhibits a relatively lower affinity for 5-methylTHF and other reduced folate derivatives (Km = 2-7 μM), and an even lower affinity for the oxidized folic acid (Km = ~200 μM) (Goldman et al., 1968; Rongbao Zhao & Goldman, 2013).

RFC is ubiquitously expressed in mammalian tissues and constitutes a very effective route for folate transport. High levels of human SLC19A1 transcripts have been detected in the liver, placenta, kidney, lung, bone marrow, intestine, and various areas of the CNS including brain microvessel endothelial cells (Araújo, Gonçalves, & Martel, 2010; Whetstine, Flatley, & Matherly, 2002). In mice, Slc19a1 transcript levels were highest in kidney and brain tissues, followed by the muscle, liver, uterus, small intestine, and lung (M. Liu et al., 2005). Localization of human or rodent RFC has also been documented at the apical brush-border membrane of the small intestine and colon, basolateral side of renal proximal tubules, apical and basolateral compartments of placental syncytiotrophoblast, hepatocyte plasma membranes, and the apical membranes of choroid plexus and retinal pigment epithelium (Chancy et al., 2000; Grapp et al., 2013; Solanky, Requena Jimenez, D’Souza, Sibley, & Glazier, 2010; Y Wang, Zhao, Russell, & Goldman, 2001; Yasuda et al., 2008). Specific cellular localization of RFC enables specialized tissue functions including folate absorption in the colon, renal secretion of antifolates (Kneuer, Honscha, & Honscha, 2005), transplacental folate transport (Yasuda et al., 2008), as well as a potential role in folate uptake at the choroid plexus and blood-brain barrier (BBB) (Alam et al., 2019; Alam, Hoque, Finnell, Goldman, & Bendayan, 2017; R Zhao, Matherly, et al., 2009). Due to these important physiological roles, changes in RFC functional expression can exacerbate the effects of folate deficiency and contribute to the development of fetal abnormalities, cardiovascular disease, and neurological disorders. In fact, it has been shown that inactivation of the Slc19a1 gene is embryonic lethal in mice due to failure of hematopoietic tissue function (R. Zhao et al., 2001).

184 Zhanjun Hou and Larry H. Matherly

Figure 4.2 Human RFC topology model. A topology model is shown for human RFC, with 12 TMDs, internal N- and C-termini, and a loop domain connecting TMD6 and 7. Figure 1-5. MembraneThe structurally topology and functionally of human important RFC. amino RFC acids, consists as described of 12 in transmembrane the text, are domains, intracellularly shownoriente asd amino red circles. and carboxyl A conserved termini, stretch and of a amino single acids N-glycosylation (Lys204–Arg214) site in (Asn58) the located in the loop connectingTMD6/TMD7 the loop first domain, and second which is transmembrane important for transport domains activity,. The is shown positions as yellow of structurally or functionally importantcircles. N-glycosylation amino acids, occurs as reported at Asn58, from which published is labeled mutagenesis as a red triangle. studies, are depicted as red circles. Conserved amino acids shown in yellow are important for RFC transport activity. Figure reproduced with permission from (Hou & Matherly, 2014).

The linker domain connecting TMD6 and TMD7 in RFC (Fig. 4.2)ispoorly 1.2.3 Protonconserved-Coupled with theFolate except Transporterion of a Lys204–Arg214 (PCFT; segment. SLC46A1 Deleting) segments (49 or 60 amino acids; positions 215–263 and 204–263, respectively) of the The PCFT,TMD6/TMD7 encoded by linker the fromSLC46A1 human gene, RFC offers abolished another transport transmembrane activity (Liu, route for folate Witt, & Matherly, 2003). Interestingly, replacingthedeletedsegmentswith uptake. PCFT73 is or an 84 integral amino acid membrane segments protein of nonhomologous comprised sequenceof 12 transmembrane from SLC19A2 domains with intracellularlyrestored directed transport, amino although and carboxyl there was terminian absolute (Rongbao requirement Zhao, for Unal, the human Shin, & Goldman, RFC 204–214 peptide sequence. Finally, when human RFC was expressed as 2010), and twoindividual N-glycosylation TMD1–6 and sites TMD7–12 located half in themolecules, loop connecting transport activity the first was and second transmembranerestored domains (Witt, (Unal, Stapels, Zhao, & Matherly, Qiu, & 2004 Goldman,). These studies 2008) establish(Figure that 1-6 neither). It functions as a the N- or C-termini nor the TMD6/TMD7 linker is critical for folate substrate unidirectionalbinding symporter and membrane that co-transports translocation. folates The primary along with role of protons the TMD6-TMD7 into the cell (Qiu et al., 2006). PCFT loopis functiona domainlly is todistinct provid fromeappropriatespacingbetweentheTMD1–6and RFC as it operates optimally at acidic pH 5.0-5.5 with TMD7–12 segments for optimal carrier function. Km of ~1 μM for folic acid and reduced folate derivatives (Qiu et al., 2006, 2007).

In humans and rodents, PCFT is abundantly expressed at the apical brush-border membrane of the duodenum and proximal jejunum, serving as a principal mechanism for the gastrointestinal absorption of dietary folates (Inoue et al., 2008; Qiu et al., 2006, 2007; Urquhart et al., 2010). High levels of PCFT expression were also reported in the apical brush-border membrane of the kidney, sinusoidal membrane of the liver, basolateral membranes of the placenta and retinal pigment 16

epithelium, and to a lesser extent in the colon, testis, lung, and brain tissues (Gnana-Prakasam et al., 2011; Qiu et al., 2007; Shayeghi et al., 2005; Urquhart et al., 2010; R Zhao et al., 2008; Rongbao Zhao & Goldman, 2013). PCFT has also been proposed to play a role in FRα-mediated uptake of 5-methylTHF at the choroid plexus (R Zhao, Min, et al., 2009). The critical role of PCFT-mediated transport is exemplified in hereditary folate malabsorption (Geller, Kronn, Jayabose, & Sandoval, 2002; Qiu et al., 2006). This autosomal recessive disorder is a product of loss-of-function mutations occurring within the SLC46A1 gene, and is clinically associated with severely low folate levels in the blood and CSF causing anemia, immune deficiency, and neurological deficits among infants (Aluri et al., 2018; Geller et al., 2002; Qiu et al., 2006; R Zhao, Aluri, & Goldman, 2017). These symptoms were further confirmed in a PCFT-deficient mouse model, where homozygous null (Slc46a1-/-) mice developed severe systemic folate deficiency, macrocytic anemia, and elevated levels of homocysteine in plasma and brain tissues compared to the wild type (Slc46a1+/+) and heterozygous (Slc46a1+/-) animals (Salojin et al., 2011).

Biology of Facilitative Folate Transporters 191

Figure 4.4 Schematic structure of human PCFT membrane topology. A topology model is shown for human PCFT, with 12 TMDs and internal N- and C-termini. Structurally or Figure 1-6. Membranefunctionally important topology amino of human acids, as determined PCFT. PCFT from publishedconsists of mutagenesis 12 transmembrane studies domains, intracellularlyand oriented in patients amino with and hereditary carboxyl folate termini, malabsorption and two (HFM),N-glycosylation are shown as site reds circles.(Asn58 and Asn68) located in the GXXXGloop connecting putative oligomerization the first and second motifs transmembrane are shown as yellow domains. circles The (Phe157 positions and of structurally or functionallyGly158 important in the amino G155XXXG acids,159 motifas reported are shown from as published red circles mutagenesis because they studies are also and struc- in patients with hereditary folateturally malabsorption and functionally, are important).depicted as N-glycosylation red circles. Amino occurs acids at Asn58 shown and in Asn68 yellow (shown represent putative oligomerizationas motifs. red triangles). Figure reproduced with permission from (Hou & Matherly, 2014).

A number of structurally and functionally important amino acids in human PCFT have been identified. This includes molecular characterization of human PCFT mutants that result in loss of function in HFM cases, and systematic mutagenesis of these HFM mutants. From considerations of 17 species homologies and amino acid charge characteristics in relation to their TMD localization, other residues have been identified as important to PCFT transport. Thus, residues have been implicated as critical to proton coupling (i.e., Glu185 (TMD5);Unal, Zhao, & Goldman, 2009) or substrate binding (i.e., His281 (TMD7); Unal, Zhao, et al., 2009), or both (i.e., Arg376 (TMD10);Mahadeo et al., 2010)(Fig. 4.4). His247, which is localized to the loop region separating TMD6 and TMD7, is predicted to lie in the cytoplasmic opening of the water-filled translocation pathway where it interacts with Ser172 to limit substrate access to the folate-binding pocket (thus determining substrate selectivity) (Unal, Zhao, et al., 2009). His281 faces the extracellular region in TMD7 (Fig. 4.4) and is believed to play an important role in PCFT protonation, which aug- ments substrate binding to the carrier (Unal, Zhao, et al., 2009).

1.2.4 Other Folate Transporters

Apart from the three highly specific folate transport pathways discussed above, other transport proteins have been identified to mediate folate transport across biological membranes, particularly members of the ATP-binding cassette (ABC) and SLC superfamilies. This section will briefly highlight the contribution of several membrane-associated ABC and SLC transporters to folate disposition in mammalian cells and tissues.

1.2.4.1 ATP-Binding Cassette (ABC) Transporters

The ABC superfamily is one of the largest and most ubiquitously expressed protein families known to date. Members of this transporter family are involved in the translocation of endogenous and exogenous substrates and metabolites against their concentration gradient using energy from ATP hydrolysis (Dean, 2002; Dean, Rzhetsky, & Allikmets, 2001; Hollenstein, Frei, & Locher, 2007). ABC transporters are characterized by the presence of a highly conserved nucleotide binding domain containing three distinct motifs designated as Walker A, Walker B, and an ABC signature or C motif, as well as a hydrophobic transmembrane domain consisting of membrane-spanning helices (Figure 1-7). ABC superfamily members are also classified as half or full transporters, with full transporters constituting at least two sets of nucleotide binding and transmembrane domains, while half transporters only consist of one of each domain (Ho & Kim, 2005). At present, 48 ABC genes have been confirmed in humans and are grouped into seven subfamilies (ABCA-G) (Dean, 2002).

Several transporters belonging to the ABC superfamily, particularly the multidrug resistance-associated proteins (MRPs; ABCC) and breast cancer resistance protein (BCRP; ABCG2), have been recognized as low affinity, high-capacity transporters of folates and antifolates

(Km = 0.2-2 mM) (Assaraf, 2006). Depending on their cellular localization, these transporters may acquire secretory or absorptive function in different tissues. For example, the expression of MRP2 and BCRP at the apical brush-border membrane of the small intestine may mediate folate export into the intestinal lumen and oppose PCFT function (Visentin, Diope-Bove, Zhao, & Goldman, 2014). In contrast, basolaterally localized MRP3 can play a role in the absorption of dietary folates across enterocytes, as demonstrated by the impaired gastrointestinal absorption and subsequent decrease in the bioavailability of folic acid and methotrexate in MRP3 null (Abcc3-/-) mice (Kitamura, Hirouchi, Kusuhara, Schuetz, & Sugiyama, 2008; Kitamura, Kusuhara, & Sugiyama,

2010). These findings, however, may only be limited to oxidized forms of folate since similar effects were not observed with 5-formylTHF and 5-methylTHF. MRP1 and MRP5 are also present on the basolateral side of the human jejunum, but their contribution to the vectorial transport of folates across enterocytes has not been fully examined (Hooijberg et al., 2003; Wielinga et al., 2005). In the liver, biliary secretion of folates is largely mediated by MRP2 and BCRP which are expressed in the apical/canalicular membranes of hepatocytes (Visentin et al., 2014). Using MRP2 deficient (Abcc2-/-) rats or mice, several groups have shown a marked decrease in biliary excretion of methotrexate and THF derivatives compared to wild type controls (Cuiping Chen et al., 2003; Kusuhara et al., 1998; M. Masuda et al., 1997; Vlaming et al., 2011). Similarly in BCRP null (Abcg2-/-) mice, inactivation of BCRP function resulted in increased methotrexate levels in the blood and liver tissues of knockout animals (Vlaming et al., 2011). There is also evidence of MRP4 expression at the basolateral membrane of human hepatocytes, suggesting a potential role for this transporter in facilitating folate transport towards the systemic circulation (Z. S. Chen et al., 2002). Several ABC transporters are also present in the kidney. MRP1, MRP3, and MRP5 are localized to the basolateral membrane of proximal renal tubules and are potentially involved in folate export into systemic circulation (Sissung et al., 2012). In contrast, MRP2 and MRP4 are expressed at the apical side of renal epithelial cells and could facilitate folate secretion into the urine (Ho & Kim, 2005). The presence of ABC transporters in human and rodent brain barrier sites, such as the blood- cerebrospinal fluid barrier (BCSFB) of the choroid plexus or in brain microvessel endothelial cells representative of the BBB, can also influence folate delivery to the CNS. BCRP is localized to the apical side of the choroid plexus epithelium (Tachikawa et al., 2005) and to the luminal membrane of brain microvessel endothelial cells (Cooray, Blackmore, Maskell, & Barrand, 2002). MRP1, MRP4, and MRP5 are expressed at the basolateral membrane of choroid plexus cells (Ashraf, Kao, & Bendayan, 2014; Borst, Evers, Kool, & Wijnholds, 2000). On the other hand, MRP1 and MRP4 are present in both luminal and abluminal membranes of the BBB endothelium, whereas MRP2 and MRP5 are mainly localized to the luminal membrane of the BBB (Ashraf et al., 2014; Borst et al., 2000).

TMD

Extracellular

Intracellular

NH2 Walker A C motif Walker B COOH

NBD

Figure 1-7. Membrane topology of ABC transporters. ABC transporters are characterized by the presence of a hydrophilic nucleotide binding domain (NBD) comprised of three distinct motifs (Walker A, Walker B, signature C motif) as well as a hydrophobic transmembane domain (TMD) containing membrane-spanning α-helices. Full transporters consist of at least two sets of nucleotide binding and transmembrane domains, while half transporters consist of only one of each domain. It is important to note that the number of transmembrane helices or the cellular orientation of amino and carboxyl termini may vary between different members of the ABC family. Figure adapted from (Dean, 2002; Dean et al., 2001; Hollenstein et al., 2007).

1.2.4.2 Solute Carrier (SLC) Transporters

The SLC superfamily is comprised of a group of secondary and tertiary active transporters that utilize pre-established ionic gradients to mediate the influx and/or efflux of a wide variety of substrates (Lin, Yee, Kim, & Giacomini, 2015). SLC transporters are important determinants of the absorption, distribution, and elimination of many organic anionic and cationic compounds (Roth, Obaidat, & Hagenbuch, 2012). To date, at least 360 members of the SLC family have been identified and are divided into 48 subfamilies (SLC1-48) (He, Vasiliou, & Nebert, 2009). There is an abundance of diversity in terms of the membrane topology of members of the SLC family, but these transporters are generally comprised of membrane-spanning α-helical domains with intracellularly oriented amino terminal and an intracellular/extracelullar carboxyl terminal (Figure 1-8).

Apart from RFC and PCFT, several other members of the SLC superfamily have been implicated in folate transport. Organic anion-transporting polypeptides (OATPs; SLCO/SLC21),

particularly OATP1B1 and OATP1B3, are expressed in the basolateral membrane of hepatocytes and work in concert with MRP2 and BCRP to facilitate folate uptake from the hepatic portal system for subsequent excretion into the bile. OATP1B1 and OATP1B3 have been shown to transport methotrexate (Kt = 25-40 µM), but not folic acid or 5-methylTHF (Abe et al., 2001). Using a transgenic mouse model with liver-specific functional expression of human OATP1B1, van de Steeg et al. observed a decrease in the area under the plasma concentration-time curve of intravenously administered methotrexate, as well as a 2-fold increase in liver methotrexate levels, demonstrating the importance of OATP1B1 function in hepatic folate clearance (van de Steeg et al., 2009). Clinical studies have also reported that polymorphisms in OATP1B1 are associated with alterations in the clearance of methotrexate in children with leukemia (Ramsey et al., 2013; Treviño et al., 2009). Furthermore, OATP2B1 is localized to the basolateral membrane of hepatocytes and the apical membrane of the small intestine. OATP2B1 has been demonstrated to be a low-affinity transporter of antifolates (i.e., raltitrexed; Ki = 70 µM > pemetrexed; Km = 300

µM > methotrexate, Ki = 600 µM) at pH 4.5-5.5 (Visentin, Chang, Romero, Zhao, & Goldman,

2012). In the kidney, OATP1A2 mediates methotrexate transport (Km = 450 µM) across the apical brush-border membrane of distal nephrons (Badagnani et al., 2006). Similarly, kidney-specific rat organic anion transporters, OAT-K1 and OAT-K2, were observed in the apical brush-border membrane of proximal renal tubules and may play an important role in the reabsorption of 5- methylTHF and methotrexate (Km = 2 μM) (S. Masuda, 2003). In contrast, organic anion transporters (OATs; SLC22) such as OAT1, OAT2, and OAT3 are expressed at the basolateral side of renal proximal tubules and are thought to favour methotrexate secretion in the kidneys (Km = 10-700 μM) (Cha et al., 2001; Takeda et al., 2002; Uwai et al., 2004). On the other hand, OAT4 is present at the apical brush-border membrane of the renal proximal tubule with an estimated Km of 18 μM for methotrexate (Takeda et al., 2002). Lastly, the presence of OATP1A2 and OAT3 at the luminal and abluminal membranes of the BBB, and the localization of OAT1 and OAT3 to the apical compartment of choroid plexus epithelial cells, could further modulate folate homeostasis in the brain (Ashraf et al., 2014).

Extracellular

Intracellular COOH NH 2

Figure 1-8. Membrane topology of SLC transporters. SLC transporters are generally comprised of membrane-spanning α-helical domains with intracellularly oriented amino and carboxyl termini. It is important to note that the number of transmembrane helices or the cellular orientation of amino and carboxyl termini may vary between different members of the SLC family. Figure adapted from (He et al., 2009; Lin et al., 2015).

1.3 Regulation of Major Folate Transport Pathways

Various molecular mechanisms exist to positively or negatively regulate the expression and/or function of the three highly specific folate transport pathways. This section will discuss four major regulatory mechanisms involving folate receptors and transporters, specifically dietary folate level, alcohol consumption, transcription factors, and nuclear receptors.

1.3.1 Dietary Folate Levels

Dietary levels of folate have shown to modulate folate transporter expression in vitro and in vivo. Traditionally, it was observed that low levels of extracellular folate can upregulate folate receptor or transporter expression. An earlier study by McHugh and Cheng reported much higher uptake of folate in human nasopharyngeal carcinoma (KB) cells grown in folate deficient medium compared to cells cultured in folate sufficient conditions (McHugh & Cheng, 1979). Supporting this finding, Kane et al. observed significant increases in FR and RFC protein expression in KB cells as well as in human and mouse mammary tumor cells grown in folate deficient medium (Kane et al., 1988). Several groups later demonstrated that the upregulation of FRα expression was due to an increase in FRα stability resulting from increased receptor half-life or rate of receptor synthesis (Hsueh & Dolnick, 1993; W. Zhu, Alliegro, & Melera, 2001). In rats, dietary folate deficiency was shown to upregulate RFC mRNA and protein levels at the intestinal brush-border membrane (Said et al., 2000). Similarly, in mice, expression of RFC in the small intestine and FRα expression in the small intestine and kidney were significantly increased when the animals were 22

fed a folate deficient diet (M. Liu et al., 2005). Similar trends were observed for PCFT, where mice placed on a folate deficient diet showed a 13-fold increase in Slc46a1 mRNA in the proximal small intestine (Qiu et al., 2007). Alternatively, in rats, over-supplementation of folate led to decreased RFC and PCFT protein expression in intestinal cells (Dev, Ahmad Wani, & Kaur, 2011).

1.3.2 Alcohol or Ethanol

Alcohol or ethanol consumption was also documented to regulate the functional expression of folate transport systems. In micropigs, Villanueva et al. observed significant reduction in Slc19a1 mRNA (10-fold) in the jejunal brush-border membrane of animals supplemented with diet containing 40% ethanol for one year (Villanueva, Devlin, & Halsted, 2001). The study also showed that transport of [3H]-labeled folic acid across the jejunal mucosa was significantly impaired in ethanol-fed micropigs resulting in decreased Km and Vmax values (Villanueva et al., 2001). Supplementation of rats with 20% ethanol diet for 3 months also resulted in the downregulation of RFC protein expression and corresponding reduction (~55%) of [3H]-folic acid transport across the intestinal basolateral membrane (Hamid, Kiran, Rana, & Kaur, 2009). Similarly, Wani and Kaur studied folate transport across the colonic apical membranes of rats fed with 20% ethanol or control diet for 3 months, and reported that ethanol-fed animals showed significant decrease in RFC and PCFT protein levels in the colon compared to control (N. A. Wani & Kaur, 2011). These studies collectively show that consumption of ethanol negatively regulates folate transport across the intestine by decreasing the functional expression of folate transporters such as RFC and PCFT. It is also important to note that in the studies by Villanueva et al. and Hamid et al., the observed decrease in folic acid transport with chronic ethanol consumption may be primarily due to reduced PCFT expression as opposed to RFC, since it is now known that RFC has a relatively lower affinity for folic acid (Km = ~200 μM) compared to reduced folates (Km = 2-7 µM) (Goldman et al., 1968; Rongbao Zhao & Goldman, 2013).

Another potential mechanism of ethanol-induced folate deficiency may involve depletion of folate stores, specifically in the liver which is the major site of folate storage in the body. Tamura et al. have previously performed studies in monkeys supplemented with diet containing 50% ethanol for two years, and demonstrated that hepatic levels of 5-methylTHF was greatly reduced in ethanol-fed animals compared to control, showing that folate storage in the liver is severely decreased with high consumption of alcohol (Tamura, Romero, Watson, Gong, & Halsted, 1981).

Wani and colleagues further reported that supplementation of rats with 20% ethanol diet significantly decreased the expression of folylpolyglutamate synthetase, an enzyme essential for the addition of glutamate groups to 5-methylTHF so it can be retained and stored in the liver (N. Wani, Hamid, & Kaur, 2012). The same study observed a reduction in the levels of polyglutamylated 5-methylTHF in various tissues including the liver, suggesting that ethanol can decrease hepatic folate stores by downregulating the expression of folylpolyglutamate synthetase (N. Wani et al., 2012).

Several groups have also indicated that ethanol can regulate systemic folate levels by increasing urinary or fecal excretion of folate. McMartin et al. showed that supplementation of rats with ethanol (~30%) for 12 weeks resulted in increased urinary folate excretion compared to animals fed with control diet (McMartin et al., 1989). Similarly, monkeys that were fed with diet containing 50% ethanol for four years contained much higher levels of folic acid in their urine and feces compared to control (Tamura & Halsted, 1983). A human study by Russell and colleagues involving chronic alcoholic patients also observed a 20-40% increase in urinary folate excretion following oral consumption of 95% ethanol (administered at 2 h intervals) (Russell et al., 1983). Eichner and Hillman further showed that serum folate of healthy as well chronic alcoholic individuals was significantly decreased following an oral or intravenous dose of 10-95% ethanol (Eichner & Hillman, 1973).

1.3.3 Transcription Factors

Numerous reports in the literature suggest a role for transcription factors in modulating folate receptor or transporter expression (Table 1-1). Transcription factors are proteins that bind to specific sites on DNA and control the fate of transcription of the targeted DNA sequence. They function to either activate or repress the expression of target genes and consequently alter cellular processes (Latchman, 1997). Specificity protein 1 (Sp1) is a transcription factor that binds to GC- rich sequences that are present within promoter regions of genes encoding for different FR isoforms (Fernandez-Zapico et al., 2011). In human KB cells and cervical carcinoma (HeLa) cells, Sp1 was observed to mediate the upregulation of FOLR1 by binding to its promoter region and inducing full promoter activity for the transcription of FRa (Saikawa, Price, Hance, Chen, & Elwood, 1995). The FOLR2 gene, encoding for the FRβ isoform, was also reported to be positively regulated by concerted actions of Sp1and another transcription factor known as nuclear respiratory

factor 2 (NRF-2) (Bruni, Polosa, Gadaleta, Cantatore, & Roberti, 2010; Sadasivan, Cedeno, & Rothenberg, 1994). In mouse embryonic fibroblast cells, Sp1 was demonstrated to regulate the transcriptional activity of the FOLR3 gene, which encodes for the FRγ isoform (Huiquan Wang, Ross, & Ratnam, 1998). In contrast, the transcription factor, activating protein 1 (AP1), has been shown to repress FOLR2 promoter activity, as the downregulation of FRβ expression was associated with increased AP1 promoter binding in human acute myelogenous leukemia (KG-1) cells (Hao, Qi, & Ratnam, 2003).

The SLC19A1 gene, which encodes for RFC, is regulated by a host of transcription factors through various complex mechanisms involving 6 promoter regions (A1/A2, A, B, C, D, and E). Similar to the FOLR genes, Sp1 remains an important regulatory factor for SLC19A1 transcription. Whetstine et al. showed that Sp1 works in concert with cAMP responsive element binding protein (CREB-1) to transactivate the promoter A region of SLC19A1 (Whetstine, Witt, & Matherly, 2002). The same group demonstrated that in the presence of a corepressor, c-Jun, the transactivating effects of Sp1 was abolished (Whetstine, Witt, et al., 2002). Furthermore, transfection of human fibrosarcoma cells with the AP2 transcription factor resulted in over 3-fold increase in SLC19A1 promoter activity, indicating a role for AP2 in upregulating RFC expression (Whetstine, Witt, et al., 2002). Transcription factors belonging to the b-Zip family, such as CREB- 1, c-Jun, and activating transcription factor 1 (ATF-1), have also been found to play an important regulatory role in the transcription of RFC via promoter region B (Whetstine & Matherly, 2001). Additionally, Sp1-mediated transactivation of promoter region B of SLC19A1 was observed to synergistically occur with upstream stimulatory factor proteins, USF1 and USF2a (M. Liu et al., 2004). On the other hand, co-transfection of drosophila melanogaster cells with another transcription factor, Ikaros 2, showed a downregulation of promoter B activity (M. Liu et al., 2004). Transcription factors, USF1 and GATA-binding factor 1 (GATA1), were further shown to synergistically increase SLC19A1 promoter activity at the A1/A2 regions of human HepG2 cells (Payton, Liu, Ge, & Matherly, 2005). The promoter region C of SLC19A1 can also be transactivated by Sp1 and the beta isoform of CCAAT/enhancer-binding protein (C/EBP b) in drosophila and HepG2 cells (Payton, Whetstine, Ge, & Matherly, 2005). A potential upregulation of RFC by the E2F-1 transcription factor has also been reported in human osteosarcoma cells, however the specific promoter region involved was not fully characterized (Sowers et al., 2003). More recently, Gonen and Assaraf investigated the role of NRF-1 in the modulation of RFC

expression in HeLa cells and observed a significant decrease in SLC19A1 mRNA expression following NRF-1 knockdown (Gonen & Assaraf, 2010). In vivo silencing of NRF-1 also showed significant reduction of RFC protein levels in mouse hepatocytes (Sid, Siow, Shang, Woo, & Karmin, 2018), further suggesting a role for NRF-1 in regulating RFC expression. In contrast, the aryl hydrocarbon receptor (AhR) has been reported to repress RFC functional expression by binding to distinct sites on the SLC19A1 promoter (Halwachs et al., 2010). Halwachs et. al. have extensively demonstrated in rats that activation of AhR through treatment with the AhR agonist, 2,3,7,8-tetrachlordibenzo-p-dioxin (TCDD), resulted in a significant reduction of RFC expression and functional activity (Halwachs et al., 2010).

Unlike SLC19A1, the SLC46A1 gene encoding for PCFT has one known promoter region with several regulatory sites for transcription factors. Two trans-activating transcription factors, AP1 and yin-yang 1 (YY1), were initially discovered in human cervical carcinoma cells after reportedly observing a ~50% decrease in SLC46A1 promoter activity following inactivation of AP1 and YY1 (Kalenik, Chen, Bradley, Chen, & Lee, 1997; Stark, Gonen, & Assaraf, 2009). In human embryonic kidney cells, Krüppel-like factor 4 (KLF4) and hepatocyte nuclear factor 4α (HNF4α) were also shown to synergistically transactivate SLC46A1 (Furumiya, Inoue, Ohta, Hayashi, & Yuasa, 2013). However, in the presence of transcription factors, C/EBPα and caudal- type homeobox transcription factor 2 (CDX2), the transactivating effects of KLF4 on the SLC46A1 gene is antagonized (Furumiya et al., 2013). Furthermore, NRF-1 was shown to bind to the promoter region of SLC46A1 and increase its transcriptional activity (Gonen & Assaraf, 2010).

1.3.4 Nuclear Receptors

Another family of transcription factors, known as nuclear receptors, have been reported to modulate cellular or tissue expression of folate transport pathways (Table 1-1). Nuclear receptors regulate the expression of their target genes in response to specific ligand activators, thereby affecting downstream processes of reproduction, cell proliferation, metabolism, and development (Benoit et al., 2006). The hormone nuclear receptor, retinoic acid receptor (RAR), was demonstrated to upregulate FRβ expression in human myeloid leukemia cells following treatment with the RAR agonist, all-trans retinoic acid (Hui Wang et al., 2000). Using HeLa cells in a tumor xenograft model, Zhang et al. showed that activation of the glucocorticoid receptor (GR) by dexamethasone resulted in increased promoter activity, transcription, and mRNA expression of

FRα (Tran, Shatnawi, Zheng, Kelley, & Ratnam, 2005). In the same study, the promoter activity of the FOLR1 gene was inhibited by the GR antagonist, mifepristone (RU486), further confirming the role of GR as a regulator of FRα expression (Tran et al., 2005). Similarly, co-transfection experiments using an in vitro model of the visceral endoderm (F9 embryo carcinoma cells) showed that HNF4α binds strongly to the promoter region of FOLR1 and increases promoter activity (Salbaum, Finnell, & Kappen, 2009; Z. Zhang et al., 2004). In contrast, the estrogen receptor (ER) was found to negatively regulate the transcription of FOLR1 in human cervical carcinoma cells (Kelley, Rowan, & Ratnam, 2003). PCFT is also positively regulated by the nuclear receptors HNF4α and the vitamin D receptor (VDR) (J J Eloranta et al., 2009; Furumiya et al., 2013). As mentioned previously, HNF4α has been shown to upregulate PCFT expression in human embryonic kidney cells by transactivating the SLC46A1 gene synergistically with the KLF4 transcription factor (Furumiya et al., 2013). Eloranta et al. also demonstrated that treatment of human epithelial colorectal adenocarcinoma (Caco-2) cells and rat duodenal biopsies with the

VDR activating ligand, 1,25-dihydroxyvitamin D3 or calcitriol, significantly increased PCFT mRNA expression and subsequent uptake of [3H]-labeled folic acid (J J Eloranta et al., 2009). At present, there is limited knowledge on the role of nuclear receptors in modulating the functional expression of RFC; hence, additional research is needed.

Table 1-1. Regulation of Major Folate Transport Pathways by Transcription Factors

Target Gene/ Transcription Factor Effect on Target Gene References Gene Product Expression or Function

FOLR1/ FRα GR, HNF4α, Sp1 é (Saikawa et al., 1995; Salbaum et al., 2009; Tran et al., 2005; Z. Zhang et al., 2004)

ER ê (Kelley et al., 2003)

FOLR1/ FRβ NRF-2, RAR, Sp1 é (Bruni et al., 2010; Sadasivan et al., 1994; Hui Wang et al., 2000)

AP1 ê (Hao et al., 2003)

SLC46A1/PCFT AP1, HNF4α, KLF4, NRF-1, é (J J Eloranta et al., 2009; Furumiya et al., VDR, YY1 2013; Gonen & Assaraf, 2010; Kalenik et al., 1997; Stark et al., 2009)

C/EBPα, CDX2 No effect (Furumiya et al., 2013)

SLC19A1/RFC AP2, C/EBPb, CREB-1, E2F-1, é (Gonen & Assaraf, 2010; M. Liu et al., 2004; GATA1, NRF-1, Sp1, USF1, Payton, Liu, et al., 2005; Payton, Whetstine, et al., 2005; Sid et al., 2018; Sowers et al., 2003; Whetstine & Matherly, 2001; Whetstine, Witt, et al., 2002)

AhR, c-Jun, Ikaros2 ê (Halwachs et al., 2010; M. Liu et al., 2004; Whetstine, Witt, et al., 2002)

Note: Transcription factors shown in bold belong to the nuclear receptor superfamily.

1.4 Importance of Folates in Central Nervous System (CNS) Homeostasis

The role of folates in one-carbon metabolism is crucial for the maintenance of CNS homeostasis. There are two potential routes for folate delivery into the brain – the BCSFB at the choroid plexus and the BBB at the cerebral vasculature endothelium. To date, mechanisms of brain folate transport have been largely investigated at the BCSFB. Under physiological conditions, folate concentration in the CSF is 2- to 3-fold greater than in blood, indicating active transport across the choroid plexus. As shown in Figure 1-3, folate transport pathways have distinct polarized cellular distribution along the choroid plexus epithelium. In humans and rodents, FRα is predominantly expressed at the apical brush-border membrane of the CSF interface, and to a lesser extent at the basolateral membrane facing toward the systemic circulation (Grapp et al., 2013; Patrick, Kranz, van Dyke, & Roy, 1997; Selhub & Franklin, 1984). RFC is also abundantly expressed at the apical brush-border membrane of choroid plexus epithelial cells (Grapp et al., 2013; Y Wang et al., 2001), whereas PCFT has been detected mainly along the basolateral and intracellular compartments (Grapp et al., 2013; R Zhao, Min, et al., 2009).

Among the three major folate transport pathways, FRα constitutes a major transcytosis route for 5-methylTHF uptake across the BCSFB (Grapp et al., 2013). This is especially true under physiological plasma concentrations of 5-methylTHF (4-20 nM), which correspond to the high binding affinity of FRα for folates (Km = 1-10 nM) (Antony, 1996; Grapp et al., 2013; Kamen & Smith, 2004). Using polarized rat Z310 choroid plexus epithelial cells transfected with human FRα, Grapp et al. have demonstrated a unidirectional vesicular transport of folic acid and 5- methylTHF from the basolateral to the apical compartment, followed by the release of FRα-bound substrates as exosomal cargo into the CSF (Grapp et al., 2013). The same study also revealed in vivo that secreted FRα-positive exosomes can penetrate into brain parenchymal cells (i.e., astrocytes and neurons), thereby demonstrating the role of FRα in mediating transcellular transport of folates across the BCSFB and their subsequent shuttling to the brain parenchyma (Grapp et al., 2013). Work by Wollack et al. further reported that 5-methylTHF transport into primary rat choroid plexus epithelial cells can also occur through two different processes depending on extracellular folate status (Wollack et al., 2008). One mechanism described a high affinity FRα- dependent uptake of 5-methylTHF at low extracellular folate concentrations, whereas the other

mechanism appears to be independent of FRα and accounts for a lower affinity uptake of 5- methylTHF at higher extracellular concentrations (Wollack et al., 2008). These findings suggest that under conditions of abundant folate supply (i.e., folate supplementation), 5-methylTHF may follow other routes of internalization and transport across the choroid plexus, which could be mediated by the folate transporters RFC and PCFT. Furthermore, the cellular localization of RFC at the apical brush-border membrane of the choroid plexus suggests that it could function to extract folates from the CSF. However, it should be noted that this transporter can also operate bidirectionally and may potentially mediate 5-methylTHF uptake at this site (R Zhao et al., 2001). Moreover, PCFT has been proposed to contribute to FRα-mediated transport by providing a route for folate export from acidified endosomal compartments and into the cell cytoplasm (Bozard et al., 2010; R Zhao, Min, et al., 2009). This process is essential for maintaining normal folate homeostasis and function within the choroid plexus epithelium (Grapp et al., 2013).

At least two types of genetic disorders provide compelling evidence of the role of FRα and PCFT in folate transport across the choroid plexus. Please refer to Section 1.5 of this chapter for a more comprehensive discussion of these disorders. Briefly, cerebral folate deficiency and hereditary folate malabsorption are characterized by very low 5-methylTHF concentrations in the CSF (<5 nmol/L) caused by the presence of FRα autoantibodies (Vincent T Ramaekers et al., 2005) or through inactivating mutations that occur within the FOLR1 and SLC46A1 genes, respectively (Grapp et al., 2012; Qiu et al., 2006; Steinfeld et al., 2009; R Zhao et al., 2017). Patients with either of these disorders exhibit severe neurological deficits, but the symptoms can be alleviated or even prevented if CSF folate levels are normalized. Several lines of evidence suggest that folate supplementation, in the form of 5-formylTHF (folinic acid), can somewhat reduce the occurrence of these neurological disorders; however, current therapeutic approaches are not sufficiently effective (Kronn & Goldman, 2017; Torres et al., 2015; R Zhao et al., 2017). While cerebral folate deficiency and hereditary folate malabsorption share similar neurological symptoms, these disorders have distinct clinical presentation. Hereditary folate malabsorption occurs within a few months of birth and is characterized by defective intestinal folate absorption causing systemic folate deficiency, as well as impaired brain folate transport resulting in low CSF folate (Qiu et al., 2006; R Zhao et al., 2017). On the other hand, patients with cerebral folate deficiency exhibit normal blood folate levels and only present with neurological deficits after 2-3 years of age (Grapp et al., 2012; Steinfeld et al., 2009). These findings are indicative of the presence of another

transport pathway that enables cerebral folate uptake during embryonic/fetal development and early life.

The BBB represents another potential route for folate delivery into the brain. Studies from a few groups have reported the presence of RFC and PCFT at the human and rodent BBB, but FRα has not been detected at this site (Araújo et al., 2010; Steinfeld et al., 2009; X. Wang, Cabrera, Li, Miller, & Finnell, 2013; Y Wang et al., 2001). Evidence in the literature and the clinic further support the role of the BBB in brain folate transport. High rates of 5-methylTHF delivery to the brain parenchyma appears to be consistent with the function of the vascular BBB (R Zhao, Matherly, et al., 2009). An earlier study conducted by Wu and Pardridge also reported that 5- methylTHF transport into isolated human brain capillaries was saturable and sensitive to inhibition by low levels of 5-methylTHF or folic acid (D. Wu & Pardridge, 1999). Similarly, Araújo et al. observed a saturable, time-dependent uptake of folic acid and 5-methylTHF by immortalized cultures of rat brain microvessel endothelial cells (RBE4 cells) representative of the BBB (Araújo et al., 2010). Clinical observations have also reported that some patients with hereditary folate malabsorption can present with mild neurological deficits despite suboptimal CSF folate levels, suggesting that in the absence of choroid plexus function, brain folate uptake can still occur via the BBB, however the efficiency of transport varies among patients. Furthermore, the earlier onset of hereditary folate malabsorption compared to cerebral folate deficiency could be due to the presence of severe systemic folate deficiency, which may limit the protective effect of BBB folate transport (R Zhao et al., 2017). In contrast, despite normal blood folate levels in patients with cerebral folate deficiency and presumed normal folate uptake at the BBB, the loss of FRa activity and consequent low CSF folate levels ultimately lead to neurological deficits. These clinical observations indicate that there may be insufficient folate transport across the BBB to adequately supply neural cells situated in close proximity to the cerebral ventricles with folates that are normally supplied via the CSF (Lehtinen et al., 2011; Lehtinen & Walsh, 2011). Thus, modulating folate transport at the level of the BBB could significantly contribute to folate uptake into the CNS.

1.5 Neurological Consequences of Folate Deficiency

Cerebral folate transport is crucial for proper brain development and function, hence nutritional folate deficiency is often associated with various types of neurological and neurodevelopmental complications. This section will discuss five neurological disorders that have been linked to folate deficiency.

1.5.1 Neural Tube Defects (NTDs)

The association between maternal folate status and NTDs exemplifies the importance of folates during early brain development. NTDs are common severe birth defects that arise early in embryogenesis due to the failure of neural tube closure leaving parts of the CNS exposed (Pitkin, 2007). Depending on the location of the lesion, NTDs may progress into either spina bifida affecting the spine, or into anencephaly that is characterized by significantly reduced or completely missing cerebral hemispheres (Lindzon & O’Connor, 2007; Pitkin, 2007). The prevalence of spina bifida and anencephaly in the United States is approximately 5.5 to 6.5 per 10,000 births (Williams et al., 2015). Poor folate status in women of childbearing age has been linked to pregnancies affected by NTDs. There is also clear evidence demonstrating that periconceptional folic acid supplementation can prevent the occurrence of these birth defects (Williams et al., 2015). A randomized double blind prevention trial (n = 1,817) conducted by the UK Medical Research Council in 1991 revealed that folic acid supplementation (4 mg/day) during the periconceptional period had a 72% protective effect in women who previously experienced NTD-affected pregnancies (MRC Vitamin Study Research Group, 1991). Another study by Czeizel et al. demonstrated that in women without previous NTD-affected pregnancies and plan to become pregnant again, the occurrence of NTDs was fully abolished with the use of folate-containing (0.8 mg folic acid) multivitamin-mineral supplements (n = 2,052), while six cases of NTDs were found in the placebo control group (n = 2,104) (Czeizel, 1993).

To date, the exact mechanisms underlying the protective effect of folic acid against NTDs remain unclear, but it has been postulated to involve folate’s role in DNA methylation (Blom, Shaw, Den Heijer, & Finnell, 2006). This hypothesis describes how MTHFR serves as a crucial junction in folate-mediated one-carbon metabolism by directing one-carbon groups towards either the DNA cycle or the methylation cycle. Therefore, genetic defects on this enzyme (i.e., MTHFR 677C>T or 1298A>C) may lead folate in the direction of DNA/RNA biosynthesis instead of the 32

production of SAM. It has been reported that women homozygous for the 677C>T polymorphism have a 60% increased risk of NTD-affected pregnancy (Blom et al., 2006). Work by van der Put et al. also indicated that MTHFR polymorphisms serve as important genetic risk factor for the development of folate-related NTDs (N. M. van der Put et al., 1995; N. M. J. van der Put et al., 1998). Due to the body of evidence supporting NTD prevention through folate supplementation, mandatory folic acid fortification of white wheat flour and select grain products was implemented in the United States and Canada since 1998. Post-fortification monitoring indicates that this strategy has been effective in decreasing the occurrence of NTDs (Arth et al., 2015; Williams et al., 2015; W. Yang, Carmichael, & Shaw, 2016). To date, women planning a pregnancy are also advised to take prenatal supplements with high levels of folic acid (400-5000 μg/day) for at least 3 months prior to conception and throughout the first trimester of pregnancy, in addition to the consumption of folate rich foods (Wilson, Genetics Committee, & Motherisk, 2007).

1.5.2 Hereditary Folate Malabsorption

Folate deficiency in the CNS, denoted by low 5-methylTHF concentrations in the CSF, has been implicated in several disorders involving defects in folate transport systems. Hereditary folate malabsorption is a condition attributed to impaired folate transport across the intestine and choroid plexus epithelia resulting from mutations in the SLC46A1 gene which encodes for PCFT (Aluri et al., 2018; Qiu et al., 2006; D S Shin et al., 2011; Daniel Sanghoon Shin, Zhao, Yap, Fiser, & Goldman, 2012; Rongbao Zhao et al., 2007). PCFT primarily mediates gastrointestinal absorption of dietary folates (Inoue et al., 2008; Qiu et al., 2006, 2007; Urquhart et al., 2010) and has also been reported to work in concert with FRα at the choroid plexus to facilitate folate export from acidified endosomes (R Zhao, Min, et al., 2009). Hereditary folate malabsorption is a rare disorder that has only been observed in approximately 37 patients (R Zhao et al., 2017). It typically manifests a few months after birth causing newborns to develop severe anemia, gastrointestinal symptoms such as diarrhea and stomatitis, recurrent infections due to immune deficiency, and neurologic manifestations including developmental delays, peripheral neuropathies, motor impairment, ataxia, and mental retardation (Geller et al., 2002; Kronn & Goldman, 2017; R Zhao et al., 2017). The severity of neurological symptoms may vary from patient to patient but if left untreated, patients further develop severe recurrent seizures.

To date, the treatment for hereditary folate malabsorption involves parenteral administration of 5-formylTHF or folinic acid. While this treatment approach is sufficient to correct for the anemia, immune dysfunction, and gastrointestinal problems, the accompanying neurological defects are not easily addressed since optimal levels of CSF folate are difficult to achieve. It is also important to note that folate concentration in the CSF tends to vary with age and has been reported to be at its highest during infancy and early childhood at ∼100−150 nM, then decreasing to ∼50−90 nM by the age of six, and >60 nM during puberty (Ormazábal et al., 2011; Perez-Duenas et al., 2011). These levels are difficult to achieve in patients with hereditary folate malabsorption, even with the administration of high doses of parenteral folate (Torres et al., 2015). Therefore, increasing folate permeability into the brain or finding alternative routes for brain folate transport can have significant therapeutic benefits.

1.5.3 Cerebral Folate Deficiency

Cerebral folate deficiency is a childhood neurological syndrome caused by: i) inactivating mutations in the FOLR1 gene encoding for FRα (Grapp et al., 2012; Steinfeld et al., 2009) or ii) the generation of autoantibodies against this receptor (Vincent T Ramaekers et al., 2005). This is a very rare disorder with only 13 confirmed cases to date (Al-Baradie & Chaudhary, 2014; Cario, Bode, Debatin, Opladen, & Schwarz, 2009; Grapp et al., 2012; Perez-Duenas et al., 2011; Steinfeld et al., 2009). The pathophysiology of cerebral folate deficiency is attributed to impaired folate transport across the choroid plexus; therefore, patients have normal concentrations of blood folate but very low levels of 5-methylTHF in the CSF (<5 nM). Patients with this disorder appear normal at birth and throughout infancy, and only exhibit neurological symptoms (i.e., abnormal brain myelination, psychomotor regression, epilepsy, ataxia) at 2-3 years of age. As previously indicated, the late onset of cerebral folate deficiency relative to hereditary malabsorption could be due in part to the protective effect of BBB folate transport; however, folate uptake across the BBB may be insufficient to adequately supply neural cells with folate to sustain normal brain development (Lehtinen et al., 2011; Lehtinen & Walsh, 2011).

Currently, the standard approach of treatment for cerebral folate deficiency is through oral supplementation of 5-formylTHF. Treatment with high doses of 5-formylTHF has been promising for some patients leading to dramatic improvements in neurological outcomes including increased CSF 5-methylTHF concentrations, normalization of appearance and speech, improved motor

skills, and reduced frequency of epileptic seizures (V T Ramaekers & Blau, 2004). However, in other cases, treatment response is either minimal or requires a considerable amount of time to take effect (i.e., 3 years to reverse cortical white matter changes). Due to the variability in patient response to 5-formylTHF therapy, there is a clear need for a more standardized optimal therapy for cerebral folate deficiency.

1.5.4 Kearns-Sayre Syndrome

Folate deficiency in the CNS has also been observed in Kearns-Sayre syndrome, a rare neuromuscular disorder caused by deletions in mitochondrial DNA (Shemesh & Margolin, 2018). Individuals with Kearns-Sayre syndrome are characterized by chronic progressive external ophthalmoplegia, where drooping and paralysis of the eyelids are observed, and pigmentary retinopathy that can lead to mild impaired vision. As the disease progresses, other associated symptoms can develop including cardiac conduction disorders, muscle weakness, neurologic dysfunction, endocrine disorders and finally, cerebral folate deficiency (Shemesh & Margolin, 2018).

Allen et al. were the first to report the association between Kearns-Sayre syndrome and cerebral folate deficiency, upon observing that a female patient diagnosed with Kearns-Sayre syndrome exhibited low plasma and CSF folate levels (R. J. Allen, DiMauro, Coulter, Papadimitriou, & Rothenberg, 1983). In a similar report by Dougados et al., another female patient with Kearns-Sayre syndrome was found to have reduced CSF folate (Dougados, Zittoun, Laplane, & Castaigne, 1983). More recent studies have aimed to elucidate the connection between Kearns- Sayre syndrome disease pathogenesis to folate deficiency in the CSF, based on these early findings. In a study by Garcia-Carzola et al. where CSF 5-methylTHF levels were measured in patients with various mitochondrial diseases, a particularly strong association was observed between Kearns-Sayre syndrome patients and reduced 5-methylTHF concentration (Garcia- Cazorla et al., 2008). Serrano et al. also showed that individuals with Kearns-Sayre syndrome were not only deficient in CSF 5-methylTHF, in fact they also exhibited very low levels of total folate in the CSF, indicating an impaired transport of different folate derivatives across the choroid plexus (Serrano et al., 2010). Interestingly, it was documented that choroid plexus epithelial cells collected from patients with Kearns-Sayre syndrome are transformed into a more enlarged and granular phenotype (Tanji, Schon, Dimauro, & Bonilla, 2000), which may potentially cause the

failure of folate transport into the CSF leading to cerebral folate deficiency (Spector & Johanson, 2010).

1.5.5 Autism Spectrum Disorders

In recent years, folate deficiency has been associated with autism spectrum disorders. Several lines of evidence suggest that periconceptional supplementation of folic acid might reduce the risk of autism spectrum disorders in the offspring (Devilbiss, Gardner, Newschaffer, & Lee, 2015). A prospective Norwegian Mother and Child Cohort Study found that women who took up to 400 μg of folic acid per day, starting from 4 weeks before and 8 weeks after the start of pregnancy, were 39% less likely to experience an autism-affected pregnancy compared to those who did not take supplements (Surén et al., 2013). Similarly, a population-based, case control study in the United States involving 837 children reported that subjects who were born to mothers that consumed ~600 μg of folic acid per day during the first month of pregnancy had 38% lower risk of developing autism spectrum disorders (Schmidt et al., 2012). Another case-control cohort study with Israeli children (n = 45, 300) showed that use of folic acid or multivitamin supplements before and/or during pregnancy can significantly decrease the risk of autism spectrum disorders (Levine et al., 2018). In contrast, a longitudinal population-based study in pregnant Danish women and their children did not find a link between folic acid or multivitamin supplementation and autism (Virk et al., 2016). While these studies present a potential link between maternal folate status and autism spectrum disorders, the exact mechanisms pertaining to the protective role of folate supplementation have yet to be determined.

Interestingly, low CSF folate concentrations were documented in at least two types of autism spectrum disorders, particularly Rett syndrome and infantile low-functioning autism. Rett syndrome is primarily caused by a mutation in the methyl-CpG-binding protein 2 (MECP2) gene, which encodes for the MeCP2 protein that binds to methylated CpG islands and 5- hydroxymethylcytosine sites on DNA, enabling chromatin organization and transcriptional regulation (Galvão & Thomas, 2005; Mellén, Ayata, Dewell, Kriaucionis, & Heintz, 2012). As MeCP2 is most abundantly expressed in postmitotic neurons, it is considered to play a significant role in synapse maintenance and development (Shahbazian, Antalffy, Armstrong, & Zoghbi, 2002). Patients with Rett syndrome suffer from retardation of brain development starting from 6 to 18 months after birth, impairment in expressive language, expression of stereotypical hand

movements, gait and truncal ataxia, seizures, scoliosis, breathing dysfunction, and reduced growth (Hagberg, Goutières, Hanefeld, Rett, & Wilson, 1985; Neul et al., 2010). On the other hand, infantile autism is a neurodevelopmental disorder observed in children below 3 years old with symptoms ranging from impaired communication and social interaction skills as well as restrictive and repetitive patterns of behavior, interests, and activities (American Psychiatric Association, 2013). Individuals with low-functioning infantile autism express the most severe symptoms according to the Diagnostic and Statistical Manual of Mental Disorders, where severe impairment in communication and social interaction, and stereotypical, repetitive behavior are observed frequently (Harrington & Allen, 2014).

An initial study by Ramaekers et al. reported that patients with autism spectrum disorders, irrespective of having the MECP2 genetic mutation, are associated with low 5-methylTHF in the CSF (Vincent T. Ramaekers et al., 2003). Further studies from the same group observed the same results in patients diagnosed with infantile low-functioning autism (V. T. Ramaekers, Blau, Sequeira, Nassogne, & Quadros, 2007) and Rett syndrome (V T Ramaekers et al., 2007). Most patients that were found to have low CSF 5-methylTHF levels also presented with serum FRa autoantibodies, which function to block folate transport into the CSF by either blocking folate binding to FRa or by binding to an epitope distant from the folate-binding site and disrupting receptor function (Vincent T Ramaekers et al., 2005). These findings were further supported by Frye and colleagues who showed that out of 93 patients with autism spectrum disorders, approximately 75% were found with high prevalence of circulating FRa autoantibodies in serum (R. E. Frye, Sequeira, Quadros, James, & Rossignol, 2013). In the same study, the concentration of FRα autoantibodies were also found to correlate with low CSF folate (R. E. Frye et al., 2013). More recent work by the research groups of Ramaekers and Frye further report a high prevalence (over 50%) of FRα autoantibodies in autistic patients compared to control (Richard E. Frye et al., 2016; V. T. Ramaekers, Quadros, & Sequeira, 2013).

Additionally, it has been demonstrated in rats that supplementation with 5-formylTHF or folinic acid could prevent autism related cognitive and behavioural deficits caused by in utero exposure to FRα autoantibodies (Desai, Sequeira, & Quadros, 2017). Human trials involving oral dosing of 5-formylTHF to children with autism spectrum disorders have shown similar results, demonstrating significant improvements in verbal communication and stereotypical behaviour. In

particular, Ramaekers and colleagues reported partial recovery of neurological and social impairments in patients with infantile low-functioning autism after oral supplementation with 5- formylTHF (V. T. Ramaekers et al., 2007). Frye et al. also observed marked improvements in attention, language, and behaviour of over 60% of patients treated with high oral doses of 5- formylTHF compared to non-treated controls (R. E. Frye et al., 2013). In a recent randomized double-blind placebo-controlled trial conducted by the same group, a 12-week, high-dose oral 5- formylTHF supplementation significantly improved verbal communication in patients expressing FRα autoantibodies (R. E. Frye et al., 2018). However, despite the promising effects of 5- formylTHF intervention in autism spectrum disorders, there is still a need to establish more effective therapies. Frye and colleagues have reported that prolonged intake of 5-formylTHF at such high doses could result in adverse events including insomnia, gastroesophageal reflux, and worsening aggression (R. E. Frye et al., 2013). Since autism is a lifelong disorder, finding an optimal therapeutic approach with limited adverse effects is vital.

Rationale

Brain folate transport is essential for the maintenance of CNS homeostasis; hence, nutritional folate deficiency has been implicated in a number of neurodevelopmental conditions particularly NTDs. Folate uptake at the choroid plexus, which occurs through concerted actions of FRa and PCFT, is crucial to cerebral folate delivery (Grapp et al., 2013; R Zhao, Min, et al., 2009). Inactivation of FRa or PCFT through loss-of-function mutations or presence of autoantibodies can lead to severe folate deficiency in the CNS (i.e., <5 nmol/L of 5-methylTHF in CSF), causing the development of early childhood neurodegenerative disorders, cerebral folate deficiency and hereditary folate malabsorption, respectively (Grapp et al., 2012; Qiu et al., 2006; Vincent T Ramaekers et al., 2005; Steinfeld et al., 2009; R Zhao et al., 2017). These disorders differ in terms of time of onset and overall clinical presentation, but both present with neurological defects characterized by developmental delays, abnormal brain myelination, psychomotor regression, ataxia, and recurrent seizures. Additionally, low CSF folate has been observed in other childhood neurological disorders such as Kearns-Sayre syndrome (Garcia-Cazorla et al., 2008; Serrano et al., 2010) and autism spectrum disorders (V. T. Ramaekers et al., 2007, 2013). Emerging evidence suggests that normalization of CSF folate levels through administration of parenteral or oral doses of 5-formylTHF (folinic acid) can alleviate some of the neurological symptoms associated with folate deficiency; however, treatment response tend to vary among patients (Kronn & Goldman, 2017; Torres et al., 2015; R Zhao et al., 2017). Furthermore, neurological complications, particularly seizures, are difficult to control since it remains a challenge to achieve near normal levels of CSF folate even with high doses of 5-formylTHF (Torres et al., 2015). Thus, increasing folate permeability into the CNS or finding alternative routes for brain folate transport can have significant therapeutic benefits.

The cerebral vasculature or BBB could represent an alternative route for brain folate transport. It has been proposed that 5-methylTHF uptake into the brain can potentially occur through the BBB (Araújo et al., 2010; D. Wu & Pardridge, 1999). There is also evidence confirming the presence of RFC and PCFT at the human and rodent BBB (Araújo et al., 2010; X. Wang et al., 2013; Y Wang et al., 2001), although their contribution to overall brain folate uptake has yet to be determined. Additionally, several groups have demonstrated that folate transport pathways are regulated by a wide range of transcription factors or nuclear receptors (J J Eloranta et al., 2009; Gonen & Assaraf, 2010). In particular, activation of the VDR nuclear receptor and 39

NRF-1 transcription factor were shown to increase RFC and/or PCFT functional expression in various in vitro and ex vivo systems (J J Eloranta et al., 2009; Gonen & Assaraf, 2010). To date, the mechanisms of brain folate transport have been primarily characterized at the choroid plexus through FRα or PCFT. However, there is limited knowledge on the role and regulation of folate transport pathways, particularly RFC, in brain microvesel endothelial cells which constitute the BBB. Modulating folate uptake at the BBB could potentially represent a novel strategy for enhancing brain folate delivery for the treatment of neurometabolic disorders caused by folate deficiency.

Goal

The goals of this thesis are i) to examine the contribution of folate transporters, specifically RFC, in brain folate transport and ii) identify the role of specific transcription factors in regulating RFC functional expression at the BBB.

Hypothesis

In the context of FRα or PCFT mutations, which render the major mechanism of brain folate transport ineffective, the folate transporters (i.e., RFC) expressed at the BBB could play a significant role in brain folate uptake. Furthermore, induction of RFC via interactions with transcription factors may provide alternative routes for effective folate permeability into the CNS.

Specific Objectives 1. To determine the functional expression of RFC in several in vitro (primary and immortalized cultures of human and rodent brain microvessel endothelial cells) and ex vivo (isolated mouse brain capillaries) models of the BBB, as well as examine the role of transcription factors (i.e., VDR) in the regulation of RFC.

2. To investigate the relative contribution of RFC in brain folate uptake and assess the role of VDR in the regulation of RFC in vivo, using wild type and Folr1 (FRα) knockout mice.

3. To examine the role of other transcription factors, such as NRF-1, in the regulation of RFC functional expression at the BBB, using in vitro cell culture systems (immortalized cultures of human brain microvessel endothelial cells) and in vivo in mice. 40

Chapter 2 Regulation of Reduced Folate Carrier (RFC) by Vitamin D Receptor at the Blood-Brain Barrier

This work is published and reproduced in this thesis with permission from: Alam C, Hoque MT, Finnell RH, Goldman ID, and Bendayan R. (2017). Regulation of reduced folate carrier (RFC) by vitamin D receptor at the blood-brain barrier. Molecular Pharmaceutics, 14(11):3848-3858. Copyright 2017 American Chemical Society.

Author Contributions:

Research design: C Alam (first author), R Bendayan (principal investigator)

Experiments and data analysis: C Alam (all figures), MT Hoque (Figures 6-1 and 6-5), ID Goldman (Figures 6-2, 6-4; contributed reagents), RH Finnell (developed and provided animals for ex vivo work), R Bendayan (all figures)

Writing of the manuscript: C Alam (manuscript preparation, revisions, submissions, and responses to reviewers’ comments); RH Finnell and MT Hoque (provided comments on initial submission); ID Goldman (editorial review of several manuscript drafts); R Bendayan (overall conceptual and editorial review of several manuscript drafts, responses to reviewers’ comments)

6.1 Abstract

Folates are essential for brain development and function. Folate transport in mammalian tissues is mediated by three major folate transport systems, i.e., reduced folate carrier (RFC), proton-coupled folate transporter (PCFT), and folate receptor alpha (FRα), known to be regulated by ligand-activated nuclear receptors such as vitamin D receptor (VDR). Folate uptake at the choroid plexus, which requires the actions of both FRα and PCFT, is critical to cerebral folate delivery. Inactivating FRα or PCFT mutations cause severe cerebral folate deficiency resulting in early childhood neurodegeneration. The objective of this study was to investigate the role of RFC in folate uptake at the level of the blood-brain barrier (BBB) and its potential regulation by VDR. We detected robust expression of RFC in different in vitro BBB model systems, particularly in immortalized cultures of human cerebral microvascular endothelial cells (hCMEC/D3) and isolated mouse brain capillaries. [3H]-methotrexate uptake by hCMEC/D3 cells at pH 7.4 was inhibited by PT523 and pemetrexed, antifolates with high affinity for RFC. We also showed that activation of VDR through calcitriol (1,25-dihydroxyvitamin D3) exposure upregulates RFC mRNA and protein expression as well as function in hCMEC/D3 cells and isolated mouse brain capillaries. We further demonstrated that RFC expression could be downregulated by VDR- targeting siRNA, further confirming the role of VDR in the direct regulation of this folate transporter. Together, these data suggest that augmenting RFC functional expression could constitute a novel strategy for enhancing brain folate delivery for the treatment of neurometabolic disorders caused by loss of FRa or PCFT function.

6.2 Introduction

Folates (vitamin B9) are water-soluble vitamins that serve as one-carbon donors in nucleotide synthesis, regulation of gene expression, and production of amino acids and neurotransmitters (Tibbetts & Appling, 2010). Folate requirements in mammals are predominantly met through dietary sources since they lack the enzymatic capacity for folate biosynthesis. Historically, nutritional folate deficiency is one of the most common vitamin deficiencies worldwide, and continues to be so in countries that have not mandated folic acid fortification of their food supply (R. L. Bailey et al., 2015). Folate deficiency is associated with a variety of negative health consequences including megaloblastic anemia, failure of neural tube closure, and congenital malformations in developing embryos, as well as developmental and neurological

defects in infants (Surtees, 2001). It has also been proposed to be a risk factor for cancer, heart disease, and stroke (L. B. Bailey, Rampersaud, & Kauwell, 2003). Maintaining sufficient folate levels requires adequate dietary intake as well as effective gastrointestinal absorption and distribution of folates to systemic tissues by specific transport systems.

Folate transport in mammalian tissues is mediated by three major transport systems: i) the folate receptor alpha (FRα; FOLR1), a glycosylphosphatidylinositol-anchored receptor that binds folates with high affinity (Km = 1-10 nM) at pH 7.4 and facilitates transport through receptor- mediated endocytosis (Elnakat & Ratnam, 2004; Kamen & Smith, 2004), ii) the proton-coupled folate transporter (PCFT; SLC46A1), which exhibits a relatively lower affinity for folates (Km = 1 μM) compared to FRα and functions as a proton cotransporter with optimal activity at pH 5.5 (Qiu et al., 2006; Rongbao Zhao & Goldman, 2013), and iii) the reduced folate carrier (RFC; SLC19A1), with a comparable influx Km of 2-7 μM for reduced folate uptake at physiological pH by operating as an antiporter exchanging folates with intracellular organic phosphates (Matherly & Hou, 2008; Rongbao Zhao & Goldman, 2013). Functional expression of the three folate transport systems has been reported to be modulated by nuclear receptors. These receptors belong to a large superfamily of DNA-binding transcriptional factors that regulate tissue expression of target genes in response to specific ligand activators (Benoit et al., 2006). The vitamin D receptor (VDR), in particular, was shown to regulate PCFT expression and function (J J Eloranta et al., 2009). Treatment with the natural VDR activating ligand, 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) or calcitriol, showed significant induction of SLC46A1/Slc46a1 (PCFT) mRNA in intestinal Caco-2 cells and rat duodenal biopsies, thereby resulting in enhanced uptake of folic acid (J J Eloranta et al., 2009). In the same study, the effect of VDR activation on SLC19A1/Slc19a1 (RFC) gene expression was variable and warranted further investigation. Several other nuclear receptors, notably the estrogen receptor (Kelley et al., 2003), glucocorticoid receptor (GR) (Tran et al., 2005), hepatocyte nuclear factor 4 alpha (HNF4α) (Salbaum et al., 2009), and nuclear respiratory factor-1 (NRF-1) (Gonen & Assaraf, 2010), are also thought to regulate mRNA and/or protein expression of major folate transporters, but their physiological roles have yet to be determined.

Folate transport in the brain is critical for the maintenance of central nervous system (CNS) homeostasis. FRα constitutes a major transcytosis pathway for folates across the choroid plexus epithelium (Grapp et al., 2013). Folate delivery into the brain occurs through the release of FRα- containing exosomes into the cerebrospinal fluid (CSF), which then cross the ependymal layer of 43

the ventricles and reach the brain parenchyma through passive diffusion. PCFT has been proposed to contribute to this FRα-mediated transport by facilitating folate export from FRα-containing exosomes into the CSF, as well as from acidified endosomes within choroid plexus epithelial cells (Grapp et al., 2013; R Zhao, Min, et al., 2009). Inactivation of FRα or PCFT through loss-of- function mutations can impair folate uptake at the choroid plexus resulting in extremely low CSF folate levels causing the early childhood neurodegeneration disorders, cerebral folate deficiency (CFD) (Grapp et al., 2012; Steinfeld et al., 2009) and hereditary folate malabsorption (HFM) (Qiu et al., 2006; R Zhao et al., 2017), respectively. These disorders differ in their time of onset and presentation. HFM presents within a few months after birth with failure to thrive and anemia, followed by neurological defects characterized by developmental delays, abnormal brain myelination, psychomotor regression, ataxia, and recurrent seizures. CFD, on the other hand, presents with neurological signs, alone, several years after birth since intestinal absorption and blood levels of folate are normal in this disorder. Improvement of CSF folate levels by administration of parenteral folinic acid in HFM or oral dosing in CFD improves these neurological deficits (Hyland, Shoffner, & Heales, 2010; R Zhao et al., 2017).

Apart from the choroid plexus, the cerebral vascular endothelium or blood-brain barrier (BBB) represents a major route of folate delivery into the brain. Despite limited knowledge of the mechanisms of folate uptake at this site, high rates of folate delivery to the brain parenchyma appear to be consistent with the function of the vascular BBB (Araújo et al., 2010; Pardridge, 2012; D. Wu & Pardridge, 1999). Earlier studies conducted by Wu and Pardridge reported that transport of the major circulating form of folates, 5-methyltetrahydrofolate (5-methylTHF), into isolated human brain capillaries was saturable and sensitive to inhibition by low levels of 5-methylTHF or folic acid (D. Wu & Pardridge, 1999). Similarly, Araújo et al. observed a saturable, time- dependent uptake of folic acid and 5-methylTHF by immortalized cultures of rat brain microvessel endothelial (RBE4) cells representative of the BBB (Araújo et al., 2010). Interestingly, there is also evidence suggesting the presence of RFC and PCFT at the human (Araújo et al., 2010) and rodent BBB (X. Wang et al., 2013; Y Wang et al., 2001), although their contribution to the overall brain uptake of folates has yet to be elucidated.

To date, the mechanisms of folate transport in the brain have been primarily characterized at the choroid plexus through the roles of FRα and PCFT. However, little is known about the functional expression of folate transporters, such as RFC and PCFT, and their regulation in brain 44

microvessel endothelial cells which constitute the BBB. The aim of the present study was to characterize RFC-mediated folate uptake using in vitro and ex vivo models of the BBB, as well as its potential regulation by the VDR nuclear receptor. Modulating folate uptake at the BBB through RFC could potentially represent a novel strategy for enhancing brain folate delivery for the treatment of neurometabolic disorders due to a failure of transport of folate across the choroid plexus into the CSF.

6.3 Materials and Methods

6.3.1 Materials

All cell culture reagents were obtained from Invitrogen (Carlsbad, CA, USA), unless indicated otherwise. Real-time quantitative polymerase chain reaction (qPCR) reagents, such as reverse transcription cDNA kits and qPCR primers, were purchased from Applied Biosystems (Foster City, CA, USA) and Life Technologies (Carlsbad, CA, USA), respectively. Primary rabbit polyclonal anti-SLC19A1 (RFC; AV44167) and anti-SLC46A1 (PCFT; SAB2108339) antibodies were obtained from Sigma-Aldrich (Oakville, ON, Canada). Mouse monoclonal anti-beta actin (sc-47778) and anti-VDR (sc-13133) antibody was supplied by Santa Cruz Biotechnology (Dallas, TX, USA). Anti-rabbit or anti-mouse horseradish peroxidase-conjugated secondary antibodies and calcitriol (1,25(OH)2D3) were purchased from Cedarlane Laboratories (Burlington, ON, Canada). Pre-designed and validated small interfering RNA (siRNA) against VDR and scrambled non- silencing control siRNA were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Tritium-labeled [3H]-methotrexate (23.4 Ci/mmol) was purchased from Moravek Biochemicals (Brea, CA, USA). Unlabeled methotrexate was obtained from Sigma-Aldrich (Oakville, ON, Canada). Pemetrexed and PT523 were generously provided by Dr. I. D. Goldman (Albert Einstein College of Medicine, New York, USA). All buffers and Triton X-100 were purchased from Sigma- Aldrich.

6.3.2 Cell Culture

Whole cell pellets of primary cultures of human brain-derived microvascular endothelial cells (hBMEC) were kindly provided by Dr. A. Prat (University of Montreal, QC, Canada); immortalized human cerebral microvessel endothelial cell line (hCMEC/D3) was generously provided by P.O. Couraud (Institut Cochin, Departement Biologie Cellulaire and INSERM, Paris,

France); rat brain microvessel endothelial cell line (RBE4) was provided by Dr. F. Roux (Hôpital Fernand Widal, Paris, France). hCMEC/D3 cells (passage 27-39) were cultured in Endothelial Cell Basal Medium-2 (Lonza, Walkersville, MD, USA), supplemented with vascular endothelial growth factor, insulin-like growth factor 1, epidermal growth factor, fibroblast growth factors, hydrocortisone, ascorbate, GA-1000, heparin and 2.5% fetal bovine serum (FBS), and grown on rat tail collagen type I-coated flasks and plates. RBE4 cells (passage 40-55) were cultured in a 50:50 mixture of minimal essential medium-alpha and HAM-F10 media supplemented with 1% penicillin/streptomycin and 10% FBS (US origin), and grown on rat tail collagen type I-coated flasks. All cell lines were maintained in a humidified incubator at 37°C, 5% CO2, and 95% air atmosphere with fresh medium replaced every 2 to 3 days. Cells were subcultured with 0.25% trypsin-EDTA upon reaching 95% confluence. For functional studies, cells were seeded into rat tail collagen-coated 24-well plates at a density of 6.3 x 104 cells/cm2 and subsequently used for experiments upon reaching 100% confluence (5 days). Cells used for calcitriol treatment were seeded into rat tail collagen-coated 6-well plates at an initial density of 3.3 x 104 cells/cm2 and grown to confluence for 5 days. For siRNA transfection studies, cells were plated in rat tail collagen-coated 6-well plates at an initial density of 7.8 x 104 cells/cm2 and grown to 80% confluence for 1 day before being subjected to transfection. The culture medium of plated cells was replaced every 2 days, and 24 h before the experiment.

6.3.3 Rodent Brain Capillary Isolation

Brain capillaries were isolated from adult male Wistar rats (250-300 g) and male or female C57BL/129/LM/Bc (10-12 weeks old) mice kindly provided by Dr. R. Finnell, as described previously (Chan & Cannon, 2017). Briefly, animals were anesthetized through isoflurane inhalation and decapitated once a deep anesthetic surgical plane was achieved. Brains were removed immediately and cortical gray matter was isolated and homogenized in ice-cold isolation buffer (phosphate-buffered saline or PBS containing calcium, magnesium, and supplemented with 5 mM glucose and 1 mM sodium pyruvate). Ficoll solution (30% final concentration) was added to the brain homogenates and the mixture was centrifuged at 5,800 g for 20 min at 4°C. The resulting pellet of capillaries was resuspended in isolation buffer supplemented with 1% bovine serum albumin (BSA) and filtered through a 300 μm nylon mesh. The filtrate containing the capillaries was passed through a 30 μm pluriStrainer and washed with 50 ml isolation buffer containing 1% BSA. Capillaries were harvested with 50 ml ice-cold isolation buffer and 46

centrifuged at 1,600 g for 5 min. The resulting pellet containing the capillaries was snap frozen in liquid nitrogen until further analysis.

6.3.4 Gene Expression Analysis

mRNA expression of specific genes of interest was quantified using qPCR. Total RNA was isolated from cell samples (hBMEC, hCMEC/D3, RBE4) and isolated rodent brain capillaries using TRIzol reagent (Invitrogen) and treated with DNase I to remove contaminating genomic DNA. RNA concentration (absorbance at 260 nm) and purity (absorbance ratio 260/280) was assessed using a Beckman Coulter DU Series 700 Scanning UV/vis Spectrophotometer. The RNA (2 μg) was then reverse transcribed to cDNA using a high-capacity reverse transcription cDNA kit (Applied Biosystems) according to the manufacturer’s instructions. Specific human/rat/mouse primer pairs for SLC19A1/Slc19a1 (RFC; Hs00953344_m1, Rn00446220_m1, Mm00446220_m1), SLC46A1/Slc46a1 (PCFT; Hs00560565_m1, Rn01471182_m1, Mm00546630_m1), FOLR1/Folr1 (FRa; Hs01124179_91, Rn00591759_m1, Mm00433355_m1), and NRI1I/Nri1i (VDR; Hs01045843_m1, Mm00437297_m1) were designed and validated by Life Technologies for use with TaqMan qPCR chemistry. All assays were done in triplicates with the housekeeping gene for human/rat/mouse cyclophilin B (Hs00168719_m1, Rn03302274_m1, Mm00478295_m1) as internal control. For each gene of interest, the critical threshold cycle (CT) was normalized to cyclophilin B using the comparative CT method. The difference in CT values (ΔCT) between the target gene and cyclophilin B was then normalized to -ΔΔC the corresponding ΔCT of the vehicle control (ΔΔCT) and expressed as fold expression (2 T) to assess the relative difference in mRNA expression for each gene.

6.3.5 Protein Expression Analysis

Western blotting was performed according to our published protocol with minor modifications (Hoque, Shah, More, Miller, & Bendayan, 2015). Briefly, whole cell or brain capillary lysates were obtained after lysing cells and rodent brain capillaries with a modified RIPA buffer [50 mM Tris pH 7.5, 150 mM NaCl, 1 mM EGTA, 1 mM sodium o-vanadate, 0.25% (v/v) sodium deoxycholic acid, 0.1% (v/v) sodium dodecyl sulfate (SDS), 1% (v/v) NP-40, 200 μM PMSF, 0.1% (v/v) protease inhibitor]. Protein concentration of the lysates was quantified using Bradford’s protein assay (Bio-Rad Laboratories) with BSA as the standard. For each sample, total protein (50 μg) was mixed in Laemmli buffer and 10% β-mercaptoethanol, separated on 10% SDS- 47

polyacrylamide gel, and electrotransferred onto a polyvinylidene fluoride membrane overnight. The blots were blocked for 2 h at room temperature in 5% skim milk Tris-buffered saline solution containing 0.1% Tween 20 and incubated overnight at 4°C with primary rabbit polyclonal anti- SLC19A1 (RFC) antibody (1:250), rabbit polyclonal anti-SLC46A1 (PCFT) antibody (1:250), murine monoclonal anti-VDR antibody (1:250), and murine monoclonal anti-beta actin antibody (1:1000). The blots were incubated for 1.5 h with corresponding horseradish peroxidase- conjugated anti-rabbit (1:5000) or anti-mouse (1:5000) secondary antibody. Protein bands were detected using enhanced chemiluminescence SuperSignal West Pico System (Thermo Fisher Scientific) and autoradiographed onto X-ray film.

6.3.6 Transport Assays

Functional assays with tritium-labeled [3H]-methotrexate were performed following standard procedures from our laboratory with minor modifications (Kis, Zastre, Ramaswamy, & Bendayan, 2010). All experiments were performed using transport buffer consisting of Hanks’ balanced salt solution (HBSS) supplemented with 0.01% BSA and 25 mM HEPES (pH 7.4). Confluent hCMEC/D3 cell monolayers grown on 24-well plates were washed twice with transport buffer (pH 7.4) and then pre-incubated in the same buffer at 37°C for 20 min. Transport was initiated by adding 0.5 ml of transport buffer (pH 7.4) containing 50 nM [3H]-methotrexate at 37°C. To examine the inhibitory effects of folate analogs with high affinity for RFC, these agents were added to both pre-incubation and radioactive transport buffers. At the desired interval, the radioactive medium was aspirated and cells were washed twice with ice-cold PBS and solubilized in 1 ml of 1% Triton X-100 at 37°C for 40 min. The content of each well was collected and mixed with 3 ml of PicoFluor 40 scintillation fluid (PerkinElmer Life and Analytical Sciences), and total radioactivity was measured with a Beckman Coulter LS6500 Scintillation counter. For each experiment, correction for nonspecific binding and variable quench time was conducted by estimating the retention of radiolabeled compound in the cells after a minimum (zero) time of exposure. The “zero time” uptake (background) was determined by removing the radiolabeled solution immediately after its introduction into the well, followed by two washes of ice-cold PBS and collection of cells for liquid scintillation counting. Cellular uptake of the radiolabeled methotrexate was normalized to total protein content per well, which was measured by using the Bio-Rad DC Protein Assay kit with BSA as the standard.

6.3.7 Calcitriol Treatment

Confluent hCMEC/D3 cell monolayers grown on 6-well plates were treated with either ethanol (vehicle control) or calcitriol (50-500 nM) for a period of 6 or 24 h at 37°C. Freshly isolated mouse brain capillaries resuspended in 200 μl isolation buffer were also exposed to ethanol (vehicle control) or 100 nM calcitriol for 4 h at room temperature. At the desired time interval, treated cells or brain capillaries were harvested using TRIzol or RIPA lysis buffer and subsequently processed for gene and protein analyses, respectively. The effect of calcitriol on methotrexate uptake was further examined by treating hCMEC/D3 cell monolayers with 500 nM calcitriol for 24 h before conducting transport assays with [3H]-methotrexate at pH 7.4 and 37°C. To ensure that cells remained viable during treatment, all concentrations of calcitriol used in this study were tested with tetrazolium salts (MTT) assay. We confirmed that there was no significant reduction in cell viability in the calcitriol-treated groups compared to vehicle-treated or untreated controls (Appendix A, Figure A-1). The dose-dependent effect of calcitriol on SLC19A1 (RFC) mRNA expression was also assessed by exposing hCMEC/D3 cells to a wide range of calcitriol concentrations (0.1-500 nM) for 24 h. In all of the tested concentrations, RFC mRNA was induced by 40-50% relative to vehicle (Appendix A, Figure A-2).

6.3.8 siRNA Treatment

hCMEC/D3 cells plated in 6-well plates were subjected to siRNA transfection upon reaching 80% confluence (24 h). The transfection mix was prepared in Opti-MEM (Invitrogen) medium containing control or VDR siRNA and Lipofectamine RNAi MAX (Invitrogen) according to the manufacturer’s protocol. The final concentration of siRNA and Lipofectamine added to the cells were 100 nM and 2 μl/ml, respectively. Cells were cultured in the presence of transfection mixture for 24 h before replacing with fresh hCMEC/D3 cell medium the following day. Cells were grown for an additional 48 h before being harvested and used for Western blotting analysis to determine VDR and RFC protein expression.

6.3.9 Data Analysis

All experiments were repeated at least three times using cells from different passages or different rodent brain capillary preparations. Each data point from a single experiment represents triplicate measurements. Results are presented as mean ± S.E.M. All statistical analyses were performed using Prism 6 software (GraphPad Software Inc., San Diego, CA, USA). Statistical 49

significance between two groups was assessed by two-tailed Student’s t test for unpaired experimental values. Multiple group comparisons were performed using one-way analysis of variance (ANOVA) with Bonferroni’s post-hoc test. p < 0.05 was considered statistically significant.

6.4 Results

6.4.1 Expression of Folate Receptor/Transporters at the BBB

Relative mRNA and protein expression of RFC, PCFT, and FRα were documented in various in vitro and ex vivo models representative of the BBB (i.e., immortalized human (hCMEC/D3) and rat (RBE4) brain microvessel endothelial cell lines, primary cultures of human brain-derived microvascular endothelial cells (hBMEC), and isolated rat and mouse brain capillaries) by qPCR and Western blot analysis, respectively. Robust mRNA expression of SLC19A1/Slc19a1 (RFC) and SLC46A1/Slc46a1 (PCFT) was detected in all BBB model systems, particularly in the isolated rodent brain capillaries (Figure 6-1A). Interestingly, FOLR1/Folr1 (FRα) was not detected in human brain microvessel endothelial cells but was present in rodent brain capillaries. Corresponding Western blots also revealed multiple protein bands for RFC (58- 75 kDa) and PCFT (50-63 kDa) (Figure 6-1B). The multiple migratory protein bands could be attributed to post-translational modifications, such as differential glycosylation of N-linked glycosylation sites of these transmembrane proteins, as previously demonstrated by various groups (Unal et al., 2008; Wong et al., 1998) and our own laboratory through deglycosylation reactions.

hCMEC/D3 0.31 hBMEC RBE4 Rat Brain Capillary 0.21 Mouse Brain Capillary

0.11

0.01 0.003 mRNA expression 0.002 (compared to(compared cyclophilin B) 0.001 Relative SLC19A1/SLC46A1/FOLR1

0.000 SLC19A1 (RFC) SLC46A1 (PCFT) FOLR1 (FRα)

B RFC

hCMEC/D3 hBMEC RBE4 Rat BrainCapillary Mouse CapillaryBrain HEK HeLa

75 75 kDa 65 63 58 48 β-Actin (42 kDa)

PCFT

hCMEC/D3 hBMEC RBE4 Rat BrainCapillary Mouse CapillaryBrain HEK HeLa 63 63 kDa 55 50 48 β-Actin (42 kDa) Figure 6-1. Relative expression of major folate transport systems in various in vitro and ex vivo models of the BBB. (A) mRNA expression of human, rat, or mouse SLC19A1/Slc19a1 (RFC), SLC46A1/Slc46a1 (PCFT), and FOLR1/Folr1 (FRα) genes were determined in immortalized (hCMEC/D3) and primary (hBMEC) cultures of human brain microvessel endothelial cells, immortalized cultures of rat brain microvessel endothelial cells (RBE4), and rodent brain capillaries using TaqMan gene expression assay. Results are presented as mean relative mRNA expression ± S.E.M. normalized to the housekeeping human/rat/mouse cyclophilin B gene from n = 3 independent experiments. (B) Immunoblot analysis of RFC and PCFT protein expression was performed in the same BBB model systems. HEK293 and HeLa cells served as positive controls, while actin was used as a loading control. Multiple protein bands for RFC and PCFT are indicative of differential glycosylation of these transmembrane proteins. A representative blot is shown from n = 3 independent experiments.

6.4.2 Methotrexate Uptake by hCMEC/D3 Cells

The functional activity of folate transporters at the BBB was examined by performing transport assays with tritium-labeled [3H]-methotrexate using the well-characterized hCMEC/D3 cell line as an in vitro model of the human BBB. Transport assays were performed at extracellular pH 7.4, reflecting optimal activity for RFC and the presumed neutral pH of the BBB interface due to the high arterial blood flow at this site (Rongbao Zhao & Goldman, 2013). Methotrexate is an established RFC substrate (Km ~ 1-7 μM) and has been used in numerous studies to assess RFC- mediated transport (Matherly & Hou, 2008; Rongbao Zhao & Goldman, 2013). In hCMEC/D3 cells, uptake of [3H]-methotrexate was inhibited by PT523 (~50%) or pemetrexed (~70%), high- affinity substrates for this transporter, (i.e., Ki of 0.2 and 1 μM, respectively) (Yanhua Wang, Zhao, & Goldman, 2004). Influx was also inhibited by excess nonlabeled methotrexate (~70%) (Figure 6-2).

0.08 Control 10 µM PT523 0.06 10 µM Pemetrexed 10 µM Methotrexate 0.04 ** ** ** 0.02 (pmol/mg protein) H]-Methotrexate Uptake 3 [ 0.00

Figure 6-2. Methotrexate uptake by hCMEC/D3 cells. The inhibitory effects of PT523 (10 μM) or pemetrexed (10 μM), high-affinity substrates for RFC, or excess unlabeled methotrexate (10 μM) on [3H]- methotrexate uptake (50 nM) was measured over 5 s at pH 7.4 and 37°C. Results are presented as mean ± S.E.M. of n = 3-4 independent experiments. **, p < 0.01.

6.4.3 Effect of Calcitriol Treatment on RFC Expression in hCMEC/D3 Cells

The regulation of RFC by VDR nuclear receptor was further assessed using the hCMEC/D3 cell line. Initially, qPCR analysis was used to confirm the presence of VDR in two different systems representative of the human (hCMEC/D3 cells) and rodent (isolated mouse brain capillaries) BBB (Figure 6-3A). To examine the effect of VDR activation on RFC expression, hCMEC/D3 cells were exposed to increasing concentrations of the VDR ligand, 1,25(OH)2D3 or calcitriol, for 6 or 24 h. NRI1I (VDR) mRNA expression was unchanged following calcitriol treatment (Figure 6-3B). SLC19A1 (RFC) mRNA expression was induced by approximately 1.5- fold after exposure to 50-500 nM calcitriol for 6 and 24 h (Figure 6-3B). Similarly, RFC protein was increased by nearly 50% following 24 h treatment with 500 nM calcitriol, and this induction was consistent for the different glycosylated forms of RFC detected at 65 and 75 kDa (Figure 6- 3C). Significant increases in PCFT mRNA and protein expression was also observed following exposure to calcitriol, which is in agreement with what was previously observed in the intestine (Appendix A, Figure A-3) (J J Eloranta et al., 2009). However, it should be noted that PCFT might be a less relevant folate transporter at this site due to the presumed neutral pH at the BBB and the low pH optimum of PCFT.

A 0.03

0.02

0.01 (normalized to cyclophilin B) Relative NRI1I/Nri1i mRNA expression 0.00 hCMEC/D3 Mouse Brain Capillary

B RFC (mRNA) VDR (mRNA) 1,25(OH) D 1,25(OH) D 2.0 2 3 1.5 2 3 Vehicle Vehicle ** * ** 50 nM 50 nM 1.5 * * 100 nM 1.0 100 nM 500 nM 500 nM 1.0 0.5 0.5 NRI1I mRNA Expression (relative to (relative ethanol control) (relative to (relative ethanol control) SLC19A1 mRNA Expression Expression mRNA SLC19A1 0.0 0.0

6 h 6 h 24 h 24 h Treatment Duration Treatment Duration

C 1,25(OH)2D3 (nM) 0 50 100 500 Rat Kidney RFC 75 kDa 65 kDa

63 kDa

β-Actin 42 kDa

RFC Protein (63kDa) RFC Protein (65kDa) RFC Protein (75kDa)

150 1,25(OH) D 150 1,25(OH)2D3 200 1,25(OH) D 2 3 ** 2 3 Vehicle Vehicle * Vehicle 500 nM 500 nM 150 500 nM 100 100

100 50 50 50 RFC Protein Expression RFC Protein Expression RFC Protein Expression (relative to (relative ethanol control) (relative to (relative ethanol control) 0 to (relative ethanol control) 0 0

24 h 24 h 24 h Treatment Duration Treatment Duration Treatment Duration

Figure 6-3. Effect of calcitriol treatment on RFC expression in hCMEC/D3 cells. (A) mRNA expression of human NRI1I/rodent Nri1i gene was determined in immortalized cultures of human brain microvessel endothelial (hCMEC/D3) cells and isolated mouse brain capillaries using TaqMan gene expression assay. Results are presented as mean relative mRNA expression ± S.E.M. normalized to the housekeeping human cyclophilin B gene from n = 3 independent experiments. Significant increases in SLC19A1 mRNA (B) and RFC protein (C) expression were observed in hCMEC/D3 cells treated with calcitriol (50-500 nM) for 6 or 24 h compared to vehicle (ethanol) control. Multiple protein bands for RFC (63-75 kDa) are indicative of differential glycosylation of the transmembrane protein. Rat kidney lysates served as positive control, while actin was used as a loading control. Results are presented as mean ± S.E.M. for n = 3-4 independent experiments. *, p < 0.05; **, p < 0.01.

6.4.4 Effect of Calcitriol Treatment on RFC Function in hCMEC/D3 Cells

To verify whether the RFC induction following calcitriol exposure would also lead to an enhancement in its functional activity, transport assays were subsequently conducted in hCMEC/D3 cells treated with calcitriol (500 nM) or ethanol (vehicle control) for 24 h. Uptake of [3H]-methotrexate between 5 s to 1 min was increased by 30-40% in calcitriol-treated cells compared to vehicle control (Figure 6-4A). [3H]-methotrexate uptake by induced hCMEC/D3 cells, measured after 1 min, was also effectively blocked by pemetrexed suggesting that the transport was specifically mediated by RFC (Figure 6-4B).

** Control 0.10 500 nM 1,25(OH)2D3 * ** ** * **

0.05 (pmol/mg protein) H]-Methotrexate Uptake 3 [ 0.00 5 10 30 40 50 60

Time (sec)

B **** *** 0.10 Control

500 nM 1,25(OH)2D3

500 nM 1,25(OH)2D3 + 10 µM Pemetrexed

0.05 (pmol/mg protein) H]-Methotrexate Uptake 3 [ 0.00 Figure 6-4. Effect of calcitriol (500nM) or ethanol (vehicle control) treatment for 24 h on methotrexate uptake. (A) Cellular uptake of [3H]-methotrexate (50 nM) was measured over 1 min at pH 7.4 and 37°C. (B) Uptake of [3H]-methotrexate (50 nM) by calcitriol-treated hCMEC/D3 cells, measured after 1 min, was inhibited by pemetrexed (10 μM). Results are presented as mean ± S.E.M. for n = 4 independent experiments. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

6.4.5 Downregulation of RFC Expression by VDR siRNA in hCMEC/D3 Cells

VDR targeting siRNA was used to further confirm the direct involvement of VDR in the regulation of RFC in hCMEC/D3 cells. Transfection of cells with VDR siRNA resulted in more than 50% downregulation of VDR protein expression (Figure 6-5). A corresponding decrease in RFC expression by 30-40% was also observed in VDR siRNA-treated cells compared to scrambled siRNA-transfected controls, as depicted by protein bands between 63-75 kDa. Similar reduction in PCFT protein levels were also observed in hCMEC/D3 cells transfected with VDR siRNA relative to control (Appendix A, Figure A-4).

A siRNA

Control VDR HEK

RFC 75 kDa

65 kDa

63 kDa

VDR 48 kDa

β-Actin 42 kDa

B 150 Control siRNA VDR siRNA

100

* * ** (% (% Control) 50 RFC Protein Expression

VDR

RFC (63 kDa) RFC (65 kDa) RFC (75 kDa)

Figure 6-5. Effect of VDR downregulation on RFC expression in hCMEC/D3 cells. (A) Significant decreases in VDR and RFC protein expression were observed in cells transfected with VDR siRNA compared to scrambled siRNA-treated control. Multiple protein bands for RFC (63-75 kDa) are indicative of differential glycosylation of the transmembrane protein. The HEK293 cell line served as positive control, while actin was used as a loading control. (B) Relative levels of VDR and RFC expression were determined by densitometric analyses. Results are expressed as percentage change normalized to control siRNA and reported as mean ± S.E.M. for n = 3 independent experiments. *, p < 0.05; **, p < 0.01. 56

6.4.6 Effect of Calcitriol Treatment on RFC Expression in Isolated Mouse Brain Capillaries

The effect of calcitriol-mediated activation of VDR on RFC expression was additionally investigated in isolated mouse brain capillaries, which are considered to be a robust ex vivo model of the BBB. Treatment of mouse brain capillaries with 100 nM calcitriol for 4 h showed over 50% increase in Slc19a1 mRNA and a corresponding 50% induction in RFC protein, supporting our in vitro findings and further demonstrating the role of VDR in regulating RFC expression (Figure 6- 6).

A RFC (mRNA) VDR (mRNA)

2.0 2.0 * 1,25(OH)2D3 1,25(OH)2D3 Vehicle Vehicle 1.5 100 nM 1.5 100 nM

1.0 1.0

0.5 0.5

0.0 mRNAExpression Nri1i 0.0 Slc19a1 mRNA Expression Expression mRNA Slc19a1 (relative to ethanol control) (relative to ethanol control) 4 h 4 h Treatment Duration Treatment Duration

B 1,25(OH) D (nM) 2 3 RFC Protein (65 kDa) 0 100 Rat Kidney 200 1,25(OH)2D3 * Vehicle RFC 65 kDa 150 100 nM

β-Actin 42 kDa 100

50 RFC Protein Expression

(relative to (relative ethanol control) 0

4 h Treatment Duration Figure 6-6. Effect of calcitriol treatment on RFC expression in isolated mouse brain capillaries. Significant increases in Slc19a1 mRNA (A) and RFC protein (B) expression were observed in mouse brain capillaries treated with 100 nM calcitriol for 4 h compared to vehicle (ethanol) control. Results are presented as mean ± S.E.M. for n = 3 independent experiments, where each experiment contained pooled brain tissues from 3 to 4 animals per group. *, p < 0.05.

6.5 Discussion

Folates play a crucial role in the development and function of the CNS. Folate uptake at the choroid plexus, which requires the actions of FRa and PCFT, is critical to cerebral folate delivery (Grapp et al., 2013; R Zhao, Min, et al., 2009; R Zhao et al., 2017). As previously 57

indicated, folate uptake at the choroid plexus begins with FRα-mediated transcytosis, followed by the export of folates from FRα-containing exosomes directly into the CSF, presumably via PCFT (Grapp et al., 2013; R Zhao, Min, et al., 2009). PCFT-mediated folate export from acidified endosomes into the cytoplasm may also occur within choroid plexus epithelial cells. Since RFC is expressed at the apical membrane of these cells, it could also serve as a route for folate efflux into the CSF (Y Wang et al., 2001). Inactivating FRa or PCFT mutations can cause severe CSF folate deficiency, affecting progenitor-rich CNS tissues situated in proximity to the cerebral ventricles, and resulting in early childhood neurological disorders such as CFD (Grapp et al., 2012; Steinfeld et al., 2009) and HFM (Qiu et al., 2006; R Zhao et al., 2017), respectively. Although there is an overlap in the neurological defects associated with these two disorders, there are differences in their clinical presentation. HFM is associated with impaired intestinal folate absorption resulting in systemic folate deficiency, as well as impaired folate transport into the CSF causing neural folate deficiency. Signs and symptoms related to HFM such as anemia and immune deficiency occur within a few months of birth, while the neurological consequences usually come later (R Zhao et al., 2017). On the other hand, patients with CFD are normal at birth and throughout infancy and only exhibit neurological deficits at 2-3 years of age (Hyland et al., 2010).

Apart from the choroid plexus, the vascular BBB represents an arterial route that provides folate to a substantial portion of the brain (Araújo et al., 2010; Pardridge, 2012; D. Wu & Pardridge, 1999). There are several reports that 5-methylTHF is transported into the brain via the BBB (Araújo et al., 2010; D. Wu & Pardridge, 1999). There is also evidence confirming the presence of folate transport systems, RFC and PCFT, at the human (Araújo et al., 2010) and rodent BBB (X. Wang et al., 2013; Y Wang et al., 2001). Substantial folate delivery at the vascular BBB is further exemplified by the fact that despite the absence of CSF folate in HFM or CFD, some patients with HFM exhibit minimal neurological defects, and the onset of CFD-related complications only occur a few years after birth. One explanation for why HFM has an earlier onset than CFD is that the extremely low blood folate levels limits the protective effect of folate delivery across the BBB (Rongbao Zhao et al., 2017). In contrast, despite normal blood folate levels in patients with CFD and presumed normal folate uptake at the BBB, the loss of FRa function and consequent low CSF folate levels ultimately lead to neurological deficits, indicating that there may be insufficient folate transport across the BBB to adequately supply neural cells in proximity to the ventricles with folates that are normally supplied via the CSF (Lehtinen et al.,

2011; Lehtinen & Walsh, 2011). Therefore, investigating the functional role of folate transport systems (i.e, RFC) at the level of the BBB and modulating folate uptake though interactions with specific nuclear receptors (i.e., VDR) could potentially increase folate uptake into the CNS, especially when the contribution of folate delivery to the CSF from the choroid plexus is impaired.

In this study, we detected mRNA expression of SLC19A1/Slc19a1 (RFC) and SLC46A1/Slc46a1 (PCFT) in various human and rodent BBB model systems (Figure 6-1A). FOLR1/Folr1 (FRa) mRNA was not detected in human brain microvessel endothelial cells (hCMEC/D3 and hBMEC) but was present in isolated rodent brain capillaries, suggesting species- specific differences in the expression of this receptor. Our gene expression data corroborates previous findings by others in that RFC and PCFT are expressed in both human (Araújo et al., 2010) and rodent BBB (X. Wang et al., 2013; Y Wang et al., 2001). Several studies have also confirmed the absence of FRα in an in vitro model of the human BBB (Araújo et al., 2010) and in different regions of human brain tissues except for the choroid plexus (Steinfeld et al., 2009). Corresponding Western blots confirmed our qPCR results and showed multiple protein bands for RFC (58-75 kDa) and PCFT (50-63 kDa) (Figure 6-1B). Previously, both transporters have been detected at different molecular sizes in various cell lines and species (Bozard et al., 2010; N. A. Wani, Nada, & Kaur, 2011), most of which are within the range of molecular weights reported here. This variability in folate transporter migration could be attributed to the differential glycosylation of these transmembrane proteins at their respective N-linked glycosylation sites (Unal et al., 2008; Wong et al., 1998). Wong et al. have shown that although human RFC only possess a single glycosylation site (Asn 58), it undergoes heavy glycosylation that could result in up to 20 kDa increase in its molecular weight (Wong et al., 1998). Similarly, Unal et al. detected multiple protein bands (~35-55 kDa) for human PCFT in HeLa cells, which they confirmed to be due to variability in the glycosylation of this transporter (Unal et al., 2008).

Although PCFT was detected in all of the BBB model systems, it might be a less relevant transporter at this site due to the neutral pH at the BBB interface and the low pH required for optimal PCFT uptake. Therefore, functional assays focused on the characterization of folate uptake at pH 7.4 that is most likely mediated by RFC. To investigate the contribution of RFC to folate uptake at the BBB, [3H]-methotrexate uptake was investigated using an in vitro model of the human BBB (hCMEC/D3 cells). The hCMEC/D3 cell line was specifically chosen for this work as it retains many of the in vivo morphological and biochemical properties of the human brain 59

microvascular endothelium, including functional expression of tight junction proteins, endothelial cell markers, and drug efflux/influx transporters (B. B. Weksler, 2005). In hCMEC/D3 cells, the specificity of [3H]-methotrexate uptake at pH 7.4 was assessed in the presence 10 μM PT523 or pemetrexed, which are potent inhibitors of RFC with Ki ’s of 0.2 and 1 μM, respectively (Yanhua Wang et al., 2004). These compounds exhibited an inhibitory effect of 50-70% (Figure 6-2), corroborating previous studies demonstrating potent inhibition by PT523 or pemetrexed of methotrexate uptake in HepG2 cells (Yanhua Wang et al., 2004). Taken together, these findings demonstrate functional expression of RFC in an in vitro model of the human BBB, suggesting a potential role for this transporter in mediating folate uptake at this site.

Although transport properties of RFC have been extensively examined in several mammalian cells and tissues, relatively little is known about its regulatory mechanisms. A few reports in the literature suggest that RFC can be modulated by ligand-activated transcriptional factors or nuclear receptors (Gonen & Assaraf, 2010; Halwachs et al., 2010). In HeLa cells, Gonen and Assaraf have shown that NRF-1 could act as an inducible transcriptional regulator of RFC since siRNA silencing of this nuclear receptor resulted in a moderate but significant reduction in SLC19A1 mRNA expression (Gonen & Assaraf, 2010). In contrast, RFC functional expression in rat liver was reported to be downregulated by the aryl hydrocarbon receptor following exposure to its activating ligand, dioxin (Halwachs et al., 2010). Previous work by Eloranta et al. also indicated that VDR could upregulate the expression and functional activity of folate transporters, such as PCFT and potentially RFC, in the intestine (J J Eloranta et al., 2009). At the BBB, VDR expression has been detected in human (hCMEC/D3) and rat (RBE4) brain microvessel endothelial cells as well as in isolated rat brain capillaries (Durk et al., 2012). Activation of VDR by calcitriol (Km in the pM range) has also been shown to regulate the functional expression of a wide range of membrane-associated drug transporters (i.e., P-gp, MRP2, MRP4, OATP1A2) in the brain (Durk et al., 2012, 2014) and other tissues (Jyrki J Eloranta, Hiller, Jüttner, & Kullak-Ublick, 2012). In this study, the potential role of VDR in the context of RFC regulation at the BBB was examined. We confirmed that NRI1I (VDR) mRNA is expressed in our BBB model systems (hCMEC/D3 cells and isolated mouse brain capillaries) (Figure 6-3A). Exposure of hCMEC/D3 cells to 50-500 nM of calcitriol resulted in nearly 50% induction of RFC mRNA and protein expression (Figure 6-3B and C). Transport assays with methotrexate also showed a corresponding increase in RFC functional activity (up to 40%) in hCMEC/D3 cells treated with calcitriol compared to control

(Figure 6-4). To further verify the involvement of VDR, mRNA and protein upregulation of PCFT, a known VDR target, was also documented (Appendix A, Figure A-3) (J J Eloranta et al., 2009). These results are most exciting as they demonstrate for the first time that ligand-dependent activation of VDR through calcitriol treatment can upregulate RFC functional expression in an in vitro human BBB model. Moreover, transfection of hCMEC/D3 cells with VDR-targeting siRNA reduced RFC protein by approximately 40% relative to control, further demonstrating the role of VDR in modulating RFC expression (Figure 6-5). To confirm our in vitro findings, the effect of VDR activation on RFC expression was also investigated in isolated mouse brain capillaries, where treatment with calcitriol resulted in an upregulation of RFC mRNA and protein levels (40- 50%) that are comparable to what was observed in the hCMEC/D3 cell line (Figure 6-6). Taken together, these findings strongly suggest for the first time that VDR is directly involved in the regulation of RFC expression and function in human brain microvessel endothelial cells or mouse brain capillaries that are representative of the BBB.

Modulation of folate uptake at the BBB through RFC may have clinical importance due to the lack of established optimal therapy for childhood neurodegenerative disorders caused by mutations in FRa or PCFT. The standard approach to the treatment of CFD and HFM is to increase folate levels in the CSF through administration of oral folinic acid for the former and parenteral folinic acid for the latter (Hyland et al., 2010; R Zhao et al., 2017). Folinic acid (leucovorin or 5- formyltetrahydrofolate) is a stable form of folate, a good RFC substrate (Km = 2-7 μM), and the preferred treatment for these disorders. Despite significant increases in systemic folate concentrations in response to folinic acid intervention, it is a challenge to achieve even near normal CSF folate levels in infants and children sufficient to correct these neurological complications (Geller et al., 2002; Torres et al., 2015; R Zhao et al., 2017). Folate concentration in the CSF varies with age and is at its highest during infancy and early childhood at ~100-150 nM, then decreasing to ~50-90 nM by the age of six, and >60 nM during puberty (Ormazábal et al., 2011; Perez-Duenas et al., 2011). Thus, our present findings could lead to the development of promising approaches to enhance the delivery of folate to the brain via RFC at the vascular BBB. However, it should also be noted that since VDR is a key regulator of numerous drug transporters (i.e., P-gp, MRP2, MRP4, OATP1A2) and drug-metabolizing enzymes (i.e., cytochrome P450s, SULT2A1) (E. C. Y. Chow et al., 2009; Durk et al., 2012; Jyrki J Eloranta et al., 2012), its activation could also interfere with normal body functions including folate homeostasis. In particular, P-gp and MRPs have been

documented to be low affinity (Km = 0.2-2 mM) efflux transporters of folates and their functional activity could oppose folate uptake mediated by RFC (Assaraf, 2006). Additionally, calcitriol- mediated activation of VDR may further affect calcium homeostasis through transactivation of calcium ion channels in the kidney and intestine (i.e., TRPV5 and TRPV6) that could lead to increased plasma calcium levels (E. C. Y. Chow, Quach, Vieth, & Pang, 2013). Taken together, calcitriol supplementation may require strict monitoring to establish a fine balance between the risks and benefits of VDR activation.

In summary, we detected RFC expression and function in different in vitro systems representative of the BBB, particularly in the hCMEC/D3 cell line. We also provided evidence that activation of VDR through calcitriol exposure upregulates RFC mRNA and protein expression as well as function in hCMEC/D3 cells or isolated mouse brain capillaries. We further demonstrated that RFC expression could be downregulated by VDR-targeting siRNA, further confirming the role of VDR in the direct regulation of this folate transporter. Together, these data suggest that augmenting RFC functional expression through activation of VDR could constitute a novel strategy for enhancing brain folate delivery for the treatment of neurometabolic disorders caused by loss of FRa or PCFT function at the level of the choroid plexus.

6.6 Acknowledgements

We thank Dr. Robert Steinfeld (University Medical Center Göttingen, Göttingen, Germany) and Dr. Bogdan Wlodarczyk (Baylor College of Medicine, Texas, USA) for their insights on this work. We also thank Theodora Bruun, Marc Li, and Adrian Turner for their technical assistance. This research was supported by an operating grant from the Natural Sciences and Engineering Research Council of Canada (NSERC) awarded to Dr. Reina Bendayan. Camille Alam is a recipient of an Ontario Graduate Scholarship, Centre for Pharmaceutical Oncology Scholarship, and Pfizer Canada Graduate Fellowship. Dr. Richard Finnell was supported by NIH grants R01HD081216 and R01HD083809. Dr. I. David Goldman was supported by a National Cancer Institute grant CA082621.

Chapter 3 Upregulation of Reduced Folate Carrier by Vitamin D Enhances Brain Folate Uptake in Mice Lacking Folate Receptor Alpha

This work is published and reproduced in this thesis with permission from: Alam C, Aufreiter S, Georgiou CJ, Hoque MT, Finnell RH, O’Connor DL, Goldman ID, and Bendayan R. (2019). Upregulation of reduced folate carrier by vitamin D enhances brain folate uptake in mice lacking folate receptor alpha. Proceedings of the National Academy of Sciences (PNAS), 116(35):17531-17540.

Author Contributions:

Research design: C Alam (first author), S Aufreiter (collaborator), DL O’Connor (collaborator), R Bendayan (principal investigator)

Experiments and data analysis: C Alam (all figures), S Aufreiter (Figures 7-6, 7-7, 7-8), CJ Georgiou (Figures 7-2 to 7-5), MT Hoque (Figure 7-1), RH Finnell (developed and provided animals for in vivo work), DL O’Connor (contributed reagents and analytical tools), R Bendayan (all figures)

Writing of the manuscript: C Alam (manuscript preparation, revisions, submissions, and responses to reviewers’ comments); S Aufreiter, CJ Georgiou, MT Hoque, RH Finnell, and DL O’Connor (provided comments on initial submission); ID Goldman (editorial review of several manuscript drafts, responses to reviewers’ comments); R Bendayan (overall conceptual and editorial review of several manuscript drafts, responses to reviewers’ comments)

7.1 Abstract

Folates are critical for central nervous system function. Folate transport is mediated by three major pathways, reduced folate carrier (RFC), proton-coupled folate transporter (PCFT), and folate receptor alpha (FRα/Folr1), known to be regulated by ligand-activated nuclear receptors. Cerebral folate delivery primarily occurs at the choroid plexus through FRα and PCFT; inactivation of these transport systems can result in very low folate levels in the cerebrospinal fluid causing childhood neurodegenerative disorders. These disorders have devastating effects in young children, and current therapeutic approaches are not sufficiently effective. Our group has previously reported in vitro that functional expression of RFC at the blood-brain barrier (BBB) and its upregulation by the vitamin D nuclear receptor (VDR) could provide an alternative route for brain folate uptake. In this study, we further demonstrated in vivo, using Folr1 knockout (KO) mice, that loss of FRα led to a substantial decrease of folate delivery to the brain and that pretreatment of Folr1 KO mice with the VDR activating ligand, calcitriol (1,25-dihydroxyvitamin

13 D3), resulted in over 6-fold increase in [ C5]-5-formyltetrahydrofolate (5-formylTHF) concentration in brain tissues, with levels comparable to wild type animals. Brain-to-plasma 13 concentration ratio of [ C5]-5-formylTHF was also significantly higher in calcitriol-treated Folr1 KO mice (15-fold), indicating a remarkable enhancement in brain folate delivery. These findings demonstrate that augmenting RFC functional expression at the BBB could effectively compensate for the loss of Folr1-mediated folate uptake at the choroid plexus, providing a therapeutic approach for neurometabolic disorders caused by defective brain folate transport.

7.2 Significance

Folates are critical for brain development and function. Abnormalities in brain folate transport have been implicated in a number of childhood neurodevelopmental disorders, including cerebral folate deficiency syndrome, hereditary folate malabsorption, and autism spectrum disorders. These disorders have devastating effects in young children and current therapeutic approaches are not sufficiently effective. In this study, we demonstrate that functional expression of the folate transporter, reduced folate carrier (RFC), at the blood-brain barrier (BBB) and its upregulation by the vitamin D nuclear receptor (VDR) can remarkably increase folate transport to the brain. These findings provide a strategy for enhancing brain folate delivery for the treatment of neurometabolic disorders caused by folate transport defects.

7.3 Introduction

Folates are required for key biosynthetic processes in mammalian cells (Tibbetts & Appling, 2010). Mammals must obtain folates from their diet since they lack the enzymatic capacity for folate biosynthesis; maintaining sufficient levels of folates requires effective gastrointestinal absorption and tissue distribution. Folate transport is mediated by three major pathways: folate receptor alpha (FRα; FOLR1), proton-coupled folate transporter (PCFT; SLC46A1), and reduced folate carrier (RFC; SLC19A1). FRα is a cell surface glycoprotein that binds folate with high affinity (Michaelis constant [Km] = 1-10 nM) and facilitates transport through receptor-mediated endocytosis (Elnakat & Ratnam, 2004; Kamen & Smith, 2004). PCFT, which exhibits a lower affinity for folates (Km = 1 μM) compared to FRα, is a proton cotransporter with optimal activity at pH 5.5, and is responsible for the majority of intestinal folate absorption

(Qiu et al., 2006; Visentin et al., 2014). RFC has a comparable affinity as PCFT (Km = 2-7 μM) for reduced folate uptake at physiological pH, and is an antiporter that exchanges folates with intracellular organic phosphates (Rongbao Zhao & Goldman, 2013).

Folates are critical for the development and function of the central nervous system (CNS). Brain folate transport is primarily mediated by concerted actions of FRα and PCFT at the choroid plexus. Folate uptake is initiated by FRα-mediated transcytosis across the choroid plexus epithelium, followed by the export of folates from FRα-containing exosomes directly into the cerebrospinal fluid (CSF) presumably via PCFT (Grapp et al., 2013; R Zhao et al., 2017). PCFT- mediated folate export from acidified endosomes may also occur within the cytoplasm of epithelial cells in order to maintain the function of the choroid plexus (Grapp et al., 2013). Inactivation of FRα causing cerebral folate deficiency (Grapp et al., 2012; Steinfeld et al., 2009), or inactivation of PCFT causing hereditary folate malabsorption (Qiu et al., 2006; R Zhao et al., 2017), through loss-of-function mutations or the presence of FRα autoantibodies (Vincent T Ramaekers et al., 2005) can severely impair brain folate uptake, thereby resulting in very low CSF folate levels and causing early childhood neurodegenerative disorders. While rare, these disorders have devastating effects in young children. Hereditary folate malabsorption presents within a few months after parturition, and is characterized by anemia and failure to thrive, followed by developmental delays, abnormal brain myelination, psychomotor regression, ataxia, and recurrent seizures (Qiu et al., 2006; R Zhao et al., 2017). Loss of FRα shares similar neurological defects as with loss of PCFT, but signs of this disorder only present several years after birth. This delay in onset may be due to 65

normal intestinal absorption and blood folate levels that continuously provide folates to the developing brain, albeit at reduced levels (Grapp et al., 2012; Steinfeld et al., 2009).

The three major folate transport pathways are modulated, in part, by nuclear receptors. These receptors are DNA-binding transcription factors that regulate the functional expression of target genes in the presence of specific ligand activators (Benoit et al., 2006). The vitamin D receptor (VDR), in particular, has been reported to regulate PCFT functional expression in the intestine. Eloranta et al. (J J Eloranta et al., 2009) demonstrated that treatment of intestinal Caco-

2 cells and isolated rat duodenum with the VDR activating ligand, 1,25-dihydroxyvitamin D3

(1,25(OH)2D3 or calcitriol), induced SLC46A1/Slc46a1 (PCFT) messenger RNA (mRNA) expression, which subsequently increased folic acid uptake. Recently, our laboratory has also examined the role of VDR in the regulation of RFC at the blood-brain barrier (BBB) (Alam et al., 2017). We provided evidence that activation of VDR through calcitriol exposure upregulates RFC mRNA and protein expression, as well as function, in immortalized cultures of human brain microvessel endothelial cells (hCMEC/D3) and isolated mouse brain capillaries that are representative of the BBB. Our study also showed that RFC expression was down-regulated by VDR-targeting siRNA, further suggesting a role for VDR in the regulation of this folate transporter.

To date, brain folate transport has been primarily characterized at the choroid plexus. However, in cases where FRα or PCFT function is compromised, functional expression of RFC at the BBB may constitute an alternative pathway for folate delivery into the CNS. Furthermore, induction of RFC via interactions with specific nuclear receptors, such as VDR, may enhance transport of folate into the brain. Several groups, including ours, have confirmed the presence of RFC in human (Alam et al., 2017; Araújo et al., 2010) and rodent (Y Wang et al., 2001) BBB, and demonstrated active transport of folates in various BBB model systems, including isolated human brain capillaries (D. Wu & Pardridge, 1999), immortalized human cell lines (hCMEC/D3), and rat (RBE4) cerebral microvessel endothelium (Alam et al., 2017; Araújo et al., 2010). The objective of the current study was to investigate the role of VDR in the regulation of RFC in vivo, using a FRα/Folr1 knockout (KO) mouse model. Modulating folate uptake at the BBB through RFC could potentially represent a novel strategy for the treatment of neurometabolic disorders resulting from failure of folate transport across the choroid plexus.

7.4 Materials and Methods

7.4.1 Materials

All cell culture reagents were obtained from Invitrogen, unless indicated otherwise. Real- time quantitative polymerase chain reaction (qPCR) reagents, such as reverse transcription complementary DNA (cDNA) kits and qPCR primers, were purchased from Applied Biosystems and Life Technologies, respectively. Primary rabbit polyclonal AE390 anti-RFC antibody was kindly provided by one of the authors (I.D.G.). Primary rabbit polyclonal anti-HCP1 (ab25134) and anti-FBP (ab67422) antibodies were obtained from Abcam. Mouse monoclonal anti-Na+/K+- adenosinetriphosphatase α (ATPase α; sc-58628) antibody was purchased from Santa Cruz Biotechnology. Anti-rabbit Alexa Fluor 594- and anti-mouse Alexa Fluor 488-conjugated secondary antibodies were supplied by Invitrogen. Calcitriol (1,25(OH)2D3) was obtained from 13 Cayman Chemical Company. Isotopically labeled (glutamyl- C5) 5-formyltetrahydrofolate 13 13 13 ([ C5]-5-formylTHF), 5-methyltetrahydrofolate ([ C5]-5-methylTHF), folic acid ([ C5]-folic acid), and corresponding unlabeled folates were synthesized by Merck Eprova AG and generously provided to us by one of the authors (D.L.O). All liquid chromatography-tandem mass spectrometry (LC-MS/MS) reagents and standard laboratory chemicals were purchased from Sigma-Aldrich.

7.4.2 Cell Culture

Primary cultures of mouse brain microvascular endothelial cells (C57BL/6 strain) were kindly provided by I. Aubert (Sunnybrook Health Sciences Centre, Toronto, ON, Canada). Cells (passage 2-4) were cultured in mouse endothelial cell basal medium supplemented with vascular endothelial growth factor, endothelial cell growth supplement, heparin, epidermal growth factor, hydrocortisone, L-glutamine, fetal bovine serum, and antibiotic-antimycotic solution (Cell Biologics). Cells were grown in flasks coated with gelatin-based solution, maintained in a humidified incubator at 37°C, 5% CO2, 95% air atmosphere, and subcultured with 0.25% trypsin- ethylenediaminetetraacetic acid (EDTA) upon reaching 80% confluence. Culture medium was replaced every 2 to 3 d.

7.4.3 Mouse Brain Capillary Isolation

Brain capillaries were isolated from adult male or female LM/Bc mice (8 to 12 wk old) as described previously (Alam et al., 2017; Chan & Cannon, 2017). Mice were anesthetized through isoflurane inhalation and decapitated once a deep anesthetic surgical plane was achieved. Brains were immediately removed, and cortical gray matter was homogenized in ice-cold isolation buffer (i.e., phosphate-buffered saline [PBS] containing calcium and magnesium and supplemented with 5 mM glucose and 1 mM sodium pyruvate). Ficoll solution (30% final concentration) was added to the brain homogenates, and the mixture was centrifuged at 5,800 x g for 20 min at 4°C. The resulting pellet of capillaries was resuspended in ice-cold isolation buffer supplemented with 1% bovine serum albumin (BSA) and filtered through a 300-μm nylon mesh. The filtrate was passed through a 30-μm pluriStrainer and washed with 50 mL isolation buffer containing 1% BSA. Capillaries were harvested with 50 mL isolation buffer and centrifuged at 1,600 x g for 5 min. The resulting pellet containing the capillaries was snap frozen in liquid nitrogen until further analysis.

7.4.4 Immunocytochemical Analysis

Subcellular localization of RFC, PCFT, and FRα proteins was investigated by laser confocal microscopy in primary cultures of mouse brain microvascular endothelial cells representative of the BBB. Cells were grown as a monolayer on gelatin-coated glass coverslips and fixed with 4% paraformaldehyde for 20 min at room temperature (RT). Cells were then washed four times with PBS and permeabilized with 0.1% Triton X-100 for 5 min at RT as described previously (Hoque, Robillard, & Bendayan, 2012). Nonspecific sites were blocked with 0.1% (mass/volume [m/v]) BSA and 0.1% (m/v) skim milk solution for 1 h before incubation with primary antibodies for 1.5 h at 37°C, followed by overnight incubation at 4°C in a humidified condition. Primary rabbit polyclonal AE390 anti-RFC (1:50), anti-HCP1 (1:50), and anti-FBP (1:50) were used to detect RFC, PCFT and FRα proteins, respectively. Mouse monoclonal anti- Na+/K+ ATPase α (1:50) was used to visualize the plasma membrane. Following primary antibody incubation, cells were washed with PBS four times by gentle agitation and incubated with anti- rabbit Alexa Fluor 594- or anti-mouse Alexa Fluor 488-conjugated secondary antibodies (1:500) for 1.5 h at 37°C. Cell staining without primary antibodies was used as a negative control. Following secondary antibody incubation, cells were washed again four times with PBS and mounted on a 76 x 26 microscope slide (VWR) using SlowFade Gold Antifade Mountant with DAPI (S36938; Invitrogen). Cells were visualized using an LSM 700 laser-scanning confocal 68

microscope (Carl Zeiss AG) operated with ZEN software. Three-dimensional colocalization of folate transporters/receptor to the plasma membrane marker, Na+/K+-ATPase α, was quantified using Imaris Bitplane software (Oxford Instruments).

7.4.5 Animal Model

Folr1 KO and wild type (WT) mice of LM/Bc background strain were developed and kindly provided by one of the authors (R.H.F.) (Piedrahita et al., 1999). Functional inactivation of Folr1 produces mouse embryos with severe growth retardation and developmental abnormalities causing in utero death by gestational day 10 (Tang & Finnell, 2003). To prevent embryonic lethality, heterozygous females were supplemented with 40 mg/kg of dietary folic acid starting from 2 wk before mating and continued throughout the gestational period (Research Diets). Following birth, this supplementation was no longer required and mice were maintained on an AIN-93G casein-based semi-purified diet designed to meet all nutritional requirements for the growing mouse, including 2 mg/kg of folic acid and 0.0375 mg/kg of vitamin D. Genotypes of adult mice (2 to 3 wk of age) were determined by PCR analysis of genomic DNA extracted from ear punch samples as described previously (Piedrahita et al., 1999). All mice were housed in clear polycarbonate microisolator cages, allowed free access to food and water, and maintained on a 14:10 light:dark cycle. All experiments, procedures, and animal care were conducted in accordance with the Canadian Council on Animal Care guidelines and approved by the University of Toronto Animal Care Committee.

7.4.6 Calcitriol (1,25(OH)2D3) Treatment in Mice

Calcitriol in powdered form was initially dissolved in anhydrous ethanol, and the concentration was measured spectrophotometrically at 265 nm (UV-1700, Schimadzu Scientific Instruments) before diluting in sterile corn oil for injections. Male or female WT and Folr1 KO mice (8 to 12 wk old) were injected intraperitoneally (i.p.) every other day for 8 d with corn oil (vehicle) or the VDR ligand, calcitriol, at a dose of 2.5 µg/kg. This dosing regimen was demonstrated to effectively elicit the desired VDR activation in vivo while alleviating the hypercalcemic effects of calcitriol (E. C. Y. Chow, Sondervan, Jin, Groothuis, & Pang, 2011). At 24 h following the last injection, mice were anesthetized through isoflurane inhalation, whole blood (1 ml) was collected via cardiac puncture, and animals were decapitated prior to collection of various tissues (brain, liver, kidney, intestine). Blood was mixed with 5 µl EDTA (0.5 M) as an 69

anticoagulant, and plasma was isolated following centrifugation at 1,500 x g for 15 min at 4°C. All samples were stored at -80°C until further analyzed.

7.4.7 Quantification of Calcium, Phosphorus, and Calcitriol in Mouse Plasma

Plasma calcium and phosphorus concentrations were measured by inductively coupled plasma atomic emission spectroscopy (Optima 7300 DV; PerkinElmer). Plasma samples were diluted 800-fold with MilliQ deionized water and filtered through a 0.22-μm filter before measurement. The emission spectra of each element were analyzed at two different wavelengths (calcium: 317.9 and 315.9 nm; phosphorus: 213.6 and 214.9 nm) to ensure accuracy in the measurements. Calcitriol levels in mouse plasma were measured using an enzyme immunoassay kit (catalog no. AC-62F1) manufactured by Immunodiagnostics Systems and purchased from Inter Medico.

7.4.8 Gene Expression Analysis

mRNA expression of specific genes of interest was quantified using qPCR. Total RNA was isolated from mouse tissues (brain capillaries, liver, kidney, intestine) using TRIzol reagent (Invitrogen) and treated with DNase I to remove contaminating genomic DNA. RNA concentration (absorbance at 260 nm) and purity (absorbance ratio 260/280) were assessed using a Beckman Coulter DU Series 700 Scanning UV/VIS Spectrophotometer. The RNA (2 μg) was then reverse transcribed to cDNA using a high-capacity reverse transcription cDNA kit according to the manufacturer’s instructions. Specific mouse primer pairs for Slc19a1 (RFC; Mm00446220_m1), Slc46a1 (PCFT; Mm00546630_m1), Folr1 (FRa; Mm00433355_m1), Abcb1a (P-glycoprotein [P-gp]; Mm00440761_m1), Abcc1 (Mrp1; Mm00456156_m1), Abcc2 (Mrp2; Mm00496899_m1), Abcc3 (Mrp3; Mm00551550_m1), Abcc4 (Mrp4; Mm01226381_m1), Slc22a6 (Oat1; Mm00456258_m1), Slc22a7 (Oat2; Mm00460672_m1), and Slc22a8 (Oat3; Mm00459534_m1) were designed and validated by Life Technologies for use with TaqMan qPCR chemistry. All assays were performed in triplicate with the housekeeping gene for mouse cyclophilin B

(Mm00478295_m1) as internal control. For each gene of interest, the critical threshold cycle (CT) was normalized to cyclophilin B using the comparative CT method. The difference in CT values

(ΔCT) between the target gene and cyclophilin B was then normalized to the corresponding ΔCT -ΔΔCT of the vehicle control (ΔΔCT) and expressed as fold expression (2 ) to assess the relative difference in mRNA expression for each gene. 70

13 7.4.9 Preparation of Intravenous [ C5]-5-formylTHF and Standard LC-MS/MS Solutions

13 Solutions for labeled and unlabeled folate calibration curves and for [ C5]-5-formylTHF injections were prepared by serial dilution of stock solutions according to methods described by Pfeiffer et al. (Christine M. Pfeiffer, Fazili, McCoy, Zhang, & Gunter, 2004). Spectrophotometrically validated 100 µg/mL stock solutions of the crystalline-form folates (i.e.,

13 unlabeled or [ C5]-labeled 5-formylTHF, 5-methylTHF, and folic acid) dissolved in phosphate buffer (pH 7.2) containing 1 g/L cysteine, were aliquoted and stored at -80°C and shielded from light until use. For intravenous (i.v.) injection, stock solutions were further diluted in sterile saline to obtain 60 µg/ml of test solution.

13 Fourteen mixed calibrator solutions containing unlabeled and [ C5]-labeled 5-formylTHF and 5-methylTHF were prepared in 1% ammonium formate buffer containing 0.1% ascorbic acid.

13 [ C5]-folic acid was used as an internal standard. All solutions were stored at -80°C until LC- MS/MS analysis.

13 7.4.10 Distribution of [ C5]-5-formylTHF in Mouse Plasma and Brain Tissue

13 [ C5]-5-formylTHF in sterile saline was administered to male or female WT and Folr1 KO mice (8 to 12 wk old) through i.v. injection (0.25 mg/kg of body weight) via the tail vein.

13 13 Distribution of injected [ C5]-5-formylTHF and its metabolite [ C5]-5-methylTHF, as well as unlabeled 5-formylTHF and 5-methylTHF, was quantified in the collected plasma and brain tissue 13 samples. At specific times following [ C5]-5-formylTHF administration, mice were anesthetized with isoflurane and whole blood (1 ml) was collected by cardiac puncture. Blood was mixed with 5 µl EDTA (0.5 M) and plasma was isolated following centrifugation at 1,500 x g for 15 min at 4°C. Plasma samples were aliquoted into tubes containing sodium ascorbate (0.5% weight/volume) to prevent folate oxidation and stored at -80°C until LC-MS/MS analysis. Following exsanguination, brain tissues were collected, washed, and weighed before tissue processing.

7.4.11 Tissue Preparation for LC-MS/MS

Isolated mouse brain tissues were homogenized in Wilson-Horne buffer (2% sodium ascorbate in 50 mM Hepes, 50 mM N-cyclohexyl-2-aminoethanesulfonic acid, and 0.2 M 2- 71

mercaptoethanol, pH 7.85) and divided into 0.5 ml aliquots. Samples were then subjected to a trienzyme extraction procedure for hydrolysis of the folate polyglutamate chain (Hyun & Tamura, 2005). Briefly, a mixture of 0.5 ml of tissue extract, 0.25 ml of protease, and 1 ml of phosphate buffer (1% sodium ascorbate, pH 4.1) was incubated at 37°C for 2 h with gentle shaking and shielded from light. The added protease was inactivated by heating for 10 min at 100°C and cooled on ice. Subsequently, 0.5 ml of α-amylase, 0.25 ml of rat serum conjugase, and 1 ml of phosphate buffer (1% sodium ascorbate, pH 6.8) were added to the original mixture, and further incubated at 37°C. After centrifugation at 5,000 x g for 10 min, the supernatant was collected and stored at - 80°C until further analysis.

13 13 7.4.12 Quantification of Plasma and Tissue [ C5]-5-formylTHF or [ C5]-5- methylTHF Concentrations by LC-MS/MS

13 13 Plasma and brain tissue enrichment of injected [ C5]-5-formylTHF or its metabolite [ C5]- 5-methylTHF was determined by LC-MS/MS at the Analytical Facility for Bioactive Molecules (The Hospital for Sick Children), following the protocol of Pfeiffer et al. (Christine M. Pfeiffer et al., 2004), with minor modifications. Before LC-MS/MS analysis, folates were extracted by solid phase extraction (SPE) from plasma samples, enzyme-digested tissues, and calibration standards. Plasma and tissue samples were first mixed with 1% ammonium formate buffer and internal standard solution (770 µl of SPE sample buffer, 275 µl of sample, and 55 µl of internal standard); calibration standards were similarly prepared (495 µl of SPE sample buffer, 275 µl of calibration solution, 55 µl of internal standard, and 275 µl of water) to give a final volume of 1.1 ml, and solutions were allowed to equilibrate at RT. SPE cartridges were conditioned with 2 ml each of acetonitrile, methanol, and 1% ammonium formate buffer (pH 3.2). 1 ml of prepared sample or calibration standards was loaded onto the cartridges and equilibrated for 1 min. The samples were then washed with 3 ml of 0.05% ammonium formate buffer containing 0.1% ascorbic acid at pH 3.4, and folates were eluted with 250 µl of elution solution (1% acetic acid, 49% water, 40% methanol, 10% acetonitrile, 0.5% ascorbic acid). Eluted samples and standards were stored in - 80°C and shielded from light until LC-MS/MS analysis.

Sample extracts (10 µl) were loaded onto a Luna C-8 analytic column for chromatographic separation using an isocratic mobile phase, as described previously (Aufreiter et al., 2009). Mass- to-charge ratios of the transition of interest [(M+5)] were monitored in positive ion mode via turbo 72

ion electrospray on an AB Sciex 5500 triple-quadrupole MS system (Applied Biosystems). The concentration of each analyte in plasma or tissue samples was calculated by interpolation of the absorbed analyte/internal standard peak area ratio into the linear regression line for the calibration curve, which was obtained by plotting peak area ratios versus analyte concentrations. Brain tissue 13 13 concentrations of [ C5]-5-formylTHF, [ C5]-5-methylTHF, or unlabeled 5-formylTHF and 5- methylTHF (ng/ml), as determined by LC-MS/MS, were normalized to the tissue mass and presented as nanograms per gram of tissue.

Two types of certified reference materials were employed to confirm the accuracy of the LC-MS/MS analyses. Analysis of mouse plasma was evaluated using standard reference material SRM 1955 Level II (homocysteine and folate in frozen human plasma) from the National Institute of Standards and Technology (NIST). In our studies, we determined 4.30 ng/ml 5-methylTHF in the NIST 1955 plasma standard, which represents >95% of the expected value of 4.47 ±0.11 ng/ml. We also used certified reference material BCR-487 (pig liver) from the European Commission Institute for Reference Materials and Measurements to test the accuracy of our analysis of digested mouse brain tissues and obtained an average of 104% ±7.7 of the certified value for 5-methylTHF (n = 3 independent experiments).

7.4.13 Pharmacokinetic Analysis

13 Compartmental analyses of [ C5]-5-formylTHF plasma concentration-time profiles from WT mice were performed using PKsolver software for Microsoft Excel (Y. Zhang, Huo, Zhou, & Xie, 2010). Pharmacokinetic parameters were estimated after fitting the concentration data to one- and two-compartment models. The final model was selected based on goodness of fit through visual inspection; residual plot analysis; and statistical parameters, including weighted sum of squares (WSS), Akaike Information Criterion (AIC), and Schwarz Criterion (SC). Various weighting schemes (actual, estimated, 1/concentration, and 1/concentration2) were used, and the 1/concentration2 weighting yielded the highest model selection criterion. The area under the 13 concentration-time curve (AUC) for plasma [ C5]-5-formylTHF was extrapolated to infinity

(AUC0-∞) and estimated using the linear trapezoidal method. The plasma terminal elimination half- life (t1/2b) was determined by linear regression of the natural log-transformed concentration-time plot. Clearance (CL) and volume of distribution (V) were calculated from the equations dose/AUC0-∞ and total amount of drug in body dose (mg/kg)/Cplasma, respectively.

7.4.14 Data Analysis All experiments were repeated at least three times using cells from different passages or different mouse brain capillary preparations. For in vivo experiments, samples were collected from 4-9 animals per treatment group or time point. Results are presented as mean ± S.E.M. All statistical analyses were performed using Prism 6 software (GraphPad Software, Inc.). Statistical significance between two groups was assessed by a two-tailed Student’s t test for unpaired experimental values. Multiple group comparisons were performed using either one-way or two- way analysis of variance (ANOVA) with Bonferroni’s post-hoc test. A p value of <0.05 was considered statistically significant.

7.5 Results

7.5.1 Localization of Folate Receptor/Transporters in Mouse BBB

Our laboratory has previously reported robust expression of RFC, PCFT, and FRα in several in vitro models of human and rodent BBB (Alam et al., 2017). In this study, we further investigated the cellular localization of the three major folate transport pathways in primary cultures of mouse brain microvascular endothelial cells by confocal microscopy (Figure 7-1). RFC, PCFT, and FRα appeared to be primarily localized to the cell plasma membrane, as shown by their similar localization to the plasma membrane marker, Na+/K+-ATPase a. The Pearson’s coefficient values between RFC/PCFT/FRα and Na+/K+-ATPase α were 0.89, 0.85, and 0.68, respectively, indicating robust colocalization between the folate transporters/receptor and the cell membrane marker.

7.5.2 Expression of Folate Receptor/Transporters in WT and Folr1 KO Mice

Relative mRNA expression of major folate transport pathways was determined in WT and Folr1 KO mouse tissues by qPCR (Figure 7-2). We confirmed that Folr1 (FRα) mRNA was not present in various tissues (isolated brain capillaries, liver, kidney) of Folr1 KO mice compared with WT controls. We also did not observe significant differences in the level of Slc19a1 (RFC) and Slc46a1 (PCFT) expression between Folr1 KO and WT animals, suggesting a lack of compensation from these transporters in response to the loss of FRα.

DAPI RFC Na+/K+-ATPase a Merge

DAPI PCFT Na+/K+-ATPase a Merge

DAPI FRα Na+/K+-ATPase a Merge

Figure 7-1. Cellular localization of major folate transport pathways in primary mouse brain microvascular endothelial cells representative of the BBB. Cells were immunostained with the following: (1) DAPI nuclear marker; (2) AE390 anti-RFC (1:50), anti-HCP1/PCFT (1:50), or anti-FBP/FRα (1:50); and (3) anti-Na+/K+-ATPase α plasma membrane marker (1:50). Cells were visualized using confocal microscopy (LSM 700; Carl Zeiss) operated with ZEN software using 40x or 63x objective lens. Scale bar, 50 µm. Three-dimensional colocalization of folate transporters/receptor and the plasma membrane marker, Na+/K+-ATPase α, was quantified using Imaris Bitplane software. The Pearson’s coefficient values between RFC/PCFT/FRα and Na+/K+-ATPase α were 0.89, 0.85, and 0.68, respectively.

A B C FRα (mRNA) RFC (mRNA) PCFT (mRNA) 4 0.8 0.4 3 0.3 2 WT 0.6 WT WT KO KO 1 KO 0.4 0.2 4×10-4

0.2 0.1 2×10-4 (compared to cyclophilin B) (compared to cyclophilin B) (compared to cyclophilin B) 0 0.0 0.0 Relative Folr1 mRNAexpression Folr1 Relative Relative Slc19a1 mRNAexpression Slc19a1 Relative Brain Capillary Liver Kidney Brain Capillary Liver Kidney mRNAexpression Slc46a1 Relative Brain Capillary Liver Kidney Figure 7-2. Relative expression of major folate transport pathways in various tissues of WT and Folr1 KO mice. mRNA expression of mouse Folr1 (A, FRa), Slc19a1 (B, RFC), and Slc46a1 (C, PCFT) genes was determined using TaqMan gene expression assay. Results are presented as mean relative mRNA expression ± S.E.M. normalized to the housekeeping mouse cyclophilin B gene from n = 3 independent experiments (total of 8 animals per group).

7.5.3 Effect of Calcitriol on RFC Expression in WT and Folr1 KO Mice

To examine the effect of VDR activation in the regulation of RFC in vivo, mice were treated i.p. with the VDR ligand, calcitriol (1,25(OH)2D3; 2.5 µg/kg), every other day for 8 d. Significant increases in Slc19a1 (RFC) mRNA were observed in isolated brain capillaries (1.5-fold), liver (2- fold), kidney (1.3-fold), and duodenum (6-fold) of WT or Folr1 KO animals compared with vehicle (Figures 7-3A and 7-4A). We additionally used P-gp as a positive control since the functional expression of this membrane transporter has been extensively demonstrated to be regulated by VDR (E. C. Y. Chow, Durk, Cummins, & Pang, 2011; Durk et al., 2012). As shown in Figures 7- 3B and 7-4B, Abcb1a (P-gp) mRNA levels were significantly elevated in various tissues of WT and Folr1 KO mice following calcitriol treatment. A modest but significant increase in Slc46a1 (PCFT) mRNA expression was also observed in isolated brain capillaries (1.5-fold) and liver (1.4- fold) of calcitriol-treated WT mice; however the contribution of this transporter in mediating folate delivery at the BBB may be less relevant due to the low pH required for optimal PCFT activity (Appendix B, Figure B-1). Furthermore, mRNA expression of several members of the organic anion transporter (OAT) and multidrug resistance-associated protein (MRP) families were determined following calcitriol treatment (Appendix B, Figure B-2). A number of these membrane-associated transporters exhibit relatively modest affinities for folates (OAT1-3, Km =

10-700 μM; MRP1-4, Km = 0.2-2 mM) and could potentially contribute to the observed effect of calcitriol on brain folate levels. Significant induction in Slc22a8 (Oat3) and Abcc3 (Mrp3) mRNA was observed in kidney tissues of calcitriol-treated WT and Folr1 KO mice compared with vehicle (Appendix B, Figure B-2A). Abcc2 (Mrp2) mRNA was also increased in the liver of WT and Folr1 KO mice following calcitriol administration (Appendix B, Figure B-2B). No significant changes in Oat or Mrp mRNA expression was observed in mouse brain capillaries isolated from vehicle and calcitriol-treated WT and Folr1 KO animals (Appendix B, Figure B-2C).

A Isolated Brain Capillaries Liver Kidney Duodenum (RFC mRNA) (RFC mRNA) (RFC mRNA) (RFC mRNA) 2.0 2.0 Vehicle 2.5 **** 3 ** 2.5 g/kg 1,25(OH) D µ 2 3 2.0 1.5 1.5 **** 2 1.5 1.0 1.0 1.0 1 0.5 0.5 (relative to vehicle)

(relative to vehicle) 0.5 (relative to vehicle) (relative to vehicle) Slc19a1 mRNA Expression mRNA Slc19a1 Slc19a1 mRNA Expression mRNA Slc19a1 Slc19a1 mRNA Expression mRNA Slc19a1 0.0 Expression mRNA Slc19a1 0.0 0.0 0

B Isolated Brain Capillaries Liver Kidney Duodenum (P-gp mRNA) (P-gp mRNA) (P-gp mRNA) (P-gp mRNA) 2.0 10.0 ** 10 Vehicle 3 * ** 2.5 µg/kg 1,25(OH)2D3 *** 1.5 8 7.5 6 2 1.0 5.0 4 1 0.5 2.5

(relative to vehicle) 2 (relative to vehicle) (relative to vehicle) (relative to vehicle) Abcb1a mRNAExpression Abcb1a Abcb1a mRNAExpression Abcb1a Abcb1a mRNAExpression Abcb1a 0.0 0.0 mRNAExpression Abcb1a 0 0 Figure 7-3. Effect of calcitriol treatment on RFC and P-gp expression in WT mice. Significant increases in Slc19a1 (A, RFC) and Abcb1a (B, P-gp) mRNA were observed in isolated brain capillaries, liver, kidney, and/or duodenum of mice treated with calcitriol (2.5 µg/kg) compared with vehicle (corn oil). Results are presented as mean ± S.E.M. for n = 3 independent experiments (total of 8-9 animals per group). Asterisks represent data points significantly different from vehicle (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

A Isolated Brain Capillaries Liver Kidney Duodenum (RFC mRNA) (RFC mRNA) (RFC mRNA) (RFC mRNA) 2.0 * Vehicle 2.5 2.0 10 * 2.5 µg/kg 1,25(OH) D *** 1.5 2 3 2.0 1.5 8 1.5 1.0 6 1.0 1.0 4 0.5 0.5 (relative to vehicle) 0.5 (relative to vehicle) (relative to vehicle) 2 (relative to vehicle) Slc19a1 mRNA Expression mRNA Slc19a1 Slc19a1 mRNA Expression mRNA Slc19a1 0.0 Expression mRNA Slc19a1 0.0 0.0 Expression mRNA Slc19a1 0

B Isolated Brain Capillaries Liver Kidney Duodenum (P-gp mRNA) (P-gp mRNA) (P-gp mRNA) (P-gp mRNA) 2.0 15 15 Vehicle **** 4 2.5 µg/kg 1,25(OH)2D3 ** 1.5 * ** 3 10 10 1.0 2 5 5 0.5 1 (relative to vehicle) (relative to vehicle) (relative to vehicle) (relative to vehicle) Abcb1a mRNAExpression Abcb1a Abcb1a mRNAExpression Abcb1a Abcb1a mRNAExpression Abcb1a 0.0 mRNAExpression Abcb1a 0 0 0 Figure 7-4. Effect of calcitriol treatment on RFC and P-gp expression in Folr1 KO mice. Significant increases in Slc19a1 (A, RFC) and Abcb1a (B, P-gp) mRNA were observed in isolated brain capillaries, liver, duodenum, and/or kidney of mice treated with calcitriol (2.5 µg/kg) compared with vehicle (corn oil). Results are presented as mean ± S.E.M. for n = 3 independent experiments (total of 8-9 animals per group). Asterisks represent data points significantly different from vehicle (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

7.5.4 Systemic Effects of Calcitriol Treatment

The body weight of WT and Folr1 KO mice was also monitored prior to and during calcitriol treatments. Changes in body weight relative to Day 0 (time before vehicle/calcitriol injections; initial weight set to 1) was plotted over a course of 6 d (Figure 7-5A). WT and Folr1 KO mice treated with calcitriol showed a statistically significant decrease in body weight (~10%) starting at Day 4 compared to vehicle control. However, this observed weight loss plateaued after prolonged calcitriol administration (i.e., 16 d) (Appendix B, Figure B-3). Since vitamin D is involved in numerous physiological processes particularly calcium uptake and bone metabolism, we also determined plasma levels of calcium and phosphorus in response to calcitriol administration. As shown in Figures 7-5B and 7-5C, the 8-d treatment regimen significantly increased calcitriol (5.5-fold) and calcium (1.5-fold) concentrations in mouse plasma. Phosphorus levels, on the other hand, remained unchanged following calcitriol treatment (Figure 7-5D).

A WT KO

1.1 1.1 ** # **** ** **** # 1.0 1.0

0.9 0.9 # #### #### #### 0.8 0.8 ####

0.7 Vehicle 0.7 Vehicle (relative Day 0) Day (relative (relative Day 0) Day (relative 2.5 µg/kg 1,25(OH) D 2.5 µg/kg 1,25(OH) D 0.6 2 3 0.6 2 3 Ratio of body weight change Ratio of body weight change 0.5 0.5 Day 0 Day 2 Day 4 Day 6 Day 0 Day 2 Day 4 Day 6 B C Plasma Calcium Plasma Calcitriol

500 *** ** 200 * **

400 150

300 100 Untreated Control 200 (mg/L) WT + Vehicle (pmol/L) WT + Vehicle WT + 2.5 µg/kg 1,25(OH)2D3 WT + 2.5 µg/kg 1,25(OH)2D3 50 100 KO + Vehicle KO + Vehicle KO + 2.5 µg/kg 1,25(OH)2D3 KO + 2.5 µg/kg 1,25(OH) D 0 ConcentrationCalcium Plasma 0 2 3 Plasma Calcitriol Concentration Concentration Calcitriol Plasma

D Plasma Phosphorus

200

150

100

(mg/L) WT + Vehicle

WT + 2.5 µg/kg 1,25(OH)2D3 50 KO + Vehicle KO + 2.5 µg/kg 1,25(OH) D 0 2 3 Plasma PhosphorusPlasma Concentration Figure 7-5. Effect of calcitriol treatment on body weight, plasma calcitriol, calcium, and phosphorus in WT and Folr1 KO mice. (A) Significant weight loss was observed at Days 4 and 6 following treatment with calcitriol (2.5 µg/kg) compared with vehicle (corn oil). Calcitriol treatment also increased calcitriol (B) and calcium (C) concentrations in mouse plasma, but not phosphorus (D). Results are presented as mean ± S.E.M. for n = 3 independent experiments, with at least 4 animals per group. Asterisks represent data points significantly different from vehicle control (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001). For weight data, pound symbol represents data points significantly different from Day 0 (time before vehicle or calcitriol injections) (#, p < 0.05; ####, p < 0.0001).

7.5.5 Assessment of Folate Levels (5-formylTHF and 5-methylTHF) in WT and Folr1 KO Mice

Basal levels of reduced folates, 5-formylTHF and 5-methylTHF, were initially determined in plasma and brain tissues of WT and Folr1 KO mice by LC-MS/MS, respectively (Figure 7-6).

We confirmed that the concentrations of 5-formylTHF and 5-methylTHF in the plasma (Cplasma) and brain (Cbrain) of Folr1 KO mice were much lower than in WT animals, suggesting that loss of Folr1 adversely affects folate uptake to the brain as well as folate conservation in the body, which primarily occurs via Folr1-mediated reabsorption in the kidneys.

A B Plasma 5-formylTHF Brain 5-formylTHF ) ) 100 500 brain plasma 80 400

60 300

40 200 (ng/g tissue) (ng/ml plasma) 20 100 ND 0 0 5-formylTHF Concentration (C

WT KO 5-formylTHF Concentration (C WT KO

C D Plasma 5-methylTHF Brain 5-methylTHF ) ) 500 100 ** brain plasma * 80 400

60 300

40 200 (ng/g tissue) (ng/ml plasma) 20 100

0 0 5-methylTHF Concentration (C 5-methylTHF Concentration (C WT KO WT KO

Figure 7-6. Basal folate levels in WT versus Folr1 KO mice. Plasma and brain concentrations of unlabeled 5-formylTHF (A and B) and 5-methylTHF (C and D) were much lower in Folr1 KO mice compared with WT mice. ND, 5-formylTHF levels in brain tissues of Folr1 KO mice were not detected by LC-MS/MS. Results are presented as mean ± S.E.M. (total of 4 animals per group). Asterisks represent data points significantly different from WT (*, p < 0.05; **, p < 0.01).

13 7.5.6 Plasma and Tissue Distribution of [ C5]-5-formylTHF in WT Mice

13 In vivo transport activity of RFC was examined using [ C5]-5-formylTHF, which is a heavy isotope tracer that can be used to evaluate folate distribution in WT and/or Folr1 KO mice.

5-formylTHF (leucovorin or folinic acid) is a known RFC substrate (Km = 2-7 μM) and a stable form of folate primarily used in the treatment of neurodevelopmental disorders, such as hereditary folate malabsorption and cerebral folate deficiency syndrome. Plasma or serum distribution of 13 injected [ C5]-5-formylTHF was previously characterized in human subjects (Aufreiter et al., 2009; Lakoff et al., 2014), but its distribution to biological tissues has yet to be examined. To 13 determine whether [ C5]-5-formylTHF can successfully penetrate across the BBB, a single bolus 13 dose of [ C5]-5-formylTHF (0.25 mg/kg) was initially administered to WT mice through i.v. 13 injection via the tail vein. As shown in Figure 7-7A, the amount of [ C5]-5-formylTHF in WT plasma rose to a maximum concentration of 403 ± 150 ng/ml after i.v. injection, followed by a

13 rapid decline to baseline at 30 to 120 min postinjection. Fitting of plasma [ C5]-5-formylTHF concentration data to a two-compartment i.v. bolus model also yielded the following pharmacokinetic parameter estimates: t1/2b = 25.01 min, V = 396.34 ml/kg, AUC0à¥ = 2,849 ng×min/ml, and CL = 87.76 ml/min/kg (Appendix B, Table B-1). Furthermore, plasma levels of

13 13 [ C5]-5-methylTHF, a metabolite of [ C5]-5-formylTHF and the major circulating form of folates, was measured to assess the conversion between different folate forms. As shown in Figure 13 13 7-7A, a steady increase in plasma [ C5]-5-methylTHF was observed following [ C5]-5- formylTHF administration, which then declined toward baseline after 60 min postinjection. Most importantly, the presence of both folate forms was also detected in brain tissues isolated from 13 treated animals, with [ C5]-5-methylTHF being present at higher concentrations over the 120 min 13 study period compared to [ C5]-5-formylTHF (Figure 7-7B).

A B

) 600 200 ) plasma 500 brain 13 150 [13C ]-5-formylTHF [ C5]-5-formylTHF 5 400 *** 13 [13C ]-5-methylTHF [ C5]-5-methylTHF 5 300 100

200 (ng/g tissue) (ng/ml plasma) 50 100 Brain Concentration (C Concentration Brain

Plasma Concentration (C Concentration Plasma * ** * 0 0 20 40 60 80 100 120 140 2 5 10 30 60 120 Time (min) Time (min)

13 13 Figure 7-7. Plasma and brain distribution of [ C5]-5-formylTHF and its metabolite [ C5]-5- methylTHF in WT mice. Plasma (A, ng/ml) and brain tissue (B, ng/g) concentrations were measured over 13 120 min following an i.v. bolus injection of 0.25 mg/kg [ C5]-5-formylTHF in WT mice. Results are 13 presented as mean ± S.E.M. for each time point (total of 4 animals per time point). Plasma [ C5]-5- formylTHF data were fitted to a two-compartment i.v. bolus pharmacokinetic model. The change in plasma 13 [ C5]-5-formylTHF concentration over 120 min following i.v. injection was statistically significant (p < 13 13 0.001), but the change in plasma [ C5]-5-methylTHF was not. The change in brain [ C5]-5-formylTHF or 13 [ C5]-5-methylTHF concentrations over the 120 min study period was not statistically significant. 13 13 Asterisks represent significant differences between [ C5]-5-formylTHF and [ C5]-5-methylTHF concentrations in plasma or brain tissue for each time point (*, p < 0.05; **, p < 0.01; ***, p < 0.001).

7.5.7 Effect of Calcitriol Treatment on Plasma and Tissue Distribution of 13 [ C5]-5-formylTHF in Folr1 KO Mice

To determine whether in vivo activation of VDR through calcitriol treatment will result in an increase in RFC functional activity, Folr1 KO mice were pretreated with calcitriol (2.5 µg/kg) or vehicle (corn oil) every other day for 8 d. At 24 h following the last calcitriol dose, each mouse 13 received a single bolus dose of [ C5]-5-formylTHF (0.25 mg/kg) by tail vein injection and was subsequently euthanized 5 min postinjection. As shown in Figure 7-8A, plasma levels of injected

13 [ C5]-5-formylTHF (Cplasma) did not differ between the four treatment groups: (1) untreated WT mice, (2) untreated Folr1 KO mice, (3) vehicle-treated Folr1 KO mice, and (4) calcitriol-treated 13 Folr1 KO mice. We also found that brain concentrations of [ C5]-5-formylTHF (Cbrain) was lower in Folr1 KO animals (untreated and vehicle-treated) compared with WT controls, suggesting inefficient folate uptake to the brain with the loss of Folr1 (Figure 7-8B). However, pretreatment 13 with calcitriol resulted in over 6-fold increase in [ C5]-5-formylTHF accumulation in brain tissues 13 of Folr1 KO mice compared to vehicle. In fact, the levels of [ C5]-5-formylTHF found in brain tissues of calcitriol-treated Folr1 KO mice were remarkably comparable to those of WT mice 82

(Figure 7-8B). We also determined brain-to-plasma concentration ratios (Cbrain/Cplasma) of injected 13 [ C5]-5-formylTHF to evaluate brain penetration of this compound (Figure 7-8C). As expected, much lower Cbrain/Cplasma ratios were observed in Folr1 KO mice (untreated and vehicle-treated) compared with WT controls, and pretreatment with calcitriol significantly increased the 13 Cbrain/Cplasma ratio of injected [ C5]-5-formylTHF by approximately 15-fold. Taken together, these results demonstrate a significant and robust enhancement of RFC-mediated brain folate uptake in Folr1 KO mice following VDR activation with calcitriol.

A B C 13 13 13 Brain:Plasma C5-5-formylTHF Plasma C5-5-formylTHF Brain C5-5-formylTHF ) ) plasma

1000 brain 200 0.5 # ## 800 0.4

150 ) * 600 0.3 plasma

100 /C 400 0.2 ]-5-formylTHF ]-5-formylTHF brain 5 (ng/g tissue) C (C (ng/ml plasma)

50 13 200 [ 0.1 ]-5-formylTHF ]-5-formylTHF Concentration (C ]-5-formylTHF ]-5-formylTHF Concentration (C 0 5 0 0.0 5 C

C 3 3 3 D 13 D D WT KO [ WT KO WT KO 13 2 2 2 [

KO + Vehicle KO + Vehicle KO + Vehicle KO + 1,25(OH) KO + 1,25(OH) KO + 1,25(OH)

13 Figure 7-8. Effect of calcitriol treatment on plasma and brain distribution of [ C5]-5-formylTHF in Folr1 KO mice. Plasma (A, ng/ml) and brain tissue (B, ng/g) concentrations were measured over 5 min 13 following an i.v. bolus injection of 0.25 mg/kg [ C5]-5-formylTHF to WT and Folr1 KO mice. (C) Tissue- 13 to-plasma (Cbrain/Cplasma) concentration ratio of injected [ C5]-5-formylTHF was also determined. Results are presented as mean ± S.E.M. (total of 4 animals per group). Asterisk, pound, and dagger symbols represent data points significantly different from untreated WT, untreated Folr1 KO, and vehicle-treated Folr1 KO group, respectively (*, p < 0.05; #, p < 0.05; ##, p < 0.01; ††, p < 0.01).

7.6 Discussion

Folates are essential for proper cognitive function, especially during rapid periods of growth and brain development in infancy and childhood. Folate deficiency has been implicated in a number of neurodevelopmental disorders. Low maternal folate status is an important contributor to the prevalence of congenital malformations like neural tube defects (Blom et al., 2006). Abnormalities in folate transport have also been associated with hereditary folate malabsorption (R Zhao et al., 2017), cerebral folate deficiency syndrome (Steinfeld et al., 2009), and autism spectrum disorders (R. E. Frye et al., 2013). Several lines of evidence indicate that folate

supplementation can somewhat reduce the occurrence of these neurological disorders; however, current therapeutic approaches are not sufficiently effective (Kronn & Goldman, 2017; Torres et al., 2015; R Zhao et al., 2017).

There are two potential routes for folate delivery into the brain: (1) the blood-CSF barrier at the choroid plexus and (2) the cerebral vasculature endothelium or BBB. As previously indicated, brain folate transport primarily occurs at the choroid plexus via FRα and PCFT (Grapp et al., 2013; R Zhao et al., 2017). Loss of FRα or PCFT function can severely impair folate uptake into the CSF, causing neural folate deficiency that manifests into childhood neurodegenerative disorders, such as cerebral folate deficiency (Grapp et al., 2012; Steinfeld et al., 2009) and hereditary folate malabsorption (Qiu et al., 2006; R Zhao et al., 2017), respectively. These two disorders share similar neurological symptoms, but with distinct clinical presentations. Hereditary folate malabsorption occurs within a few months of birth and is characterized by defective intestinal folate absorption causing systemic folate deficiency, as well as impaired brain folate transport resulting in low CSF folate (Qiu et al., 2006; R Zhao et al., 2017). On the other hand, patients with cerebral folate deficiency exhibit normal blood folate levels and only present with neurological deficits after 2 to 3 y of age (Grapp et al., 2012; Steinfeld et al., 2009).

The vascular BBB represents an arterial route for brain folate transport, although the exact mechanisms of folate uptake at this site remain unknown (Araújo et al., 2010; Y Wang et al., 2001; D. Wu & Pardridge, 1999). There is some clinical evidence pertaining to the role of the BBB in folate delivery. In particular, some patients with hereditary folate malabsorption can present with mild neurological deficits despite suboptimal CSF folate levels, suggesting that in the absence of choroid plexus function, brain folate uptake can still occur via the BBB; however, the efficiency of transport varies among patients. Furthermore, the earlier onset of hereditary folate malabsorption compared with cerebral folate deficiency could be due to the presence of severe systemic folate deficiency, which may limit the protective effect of BBB folate transport (R Zhao et al., 2017). In contrast, despite normal systemic folate levels in individuals with cerebral folate deficiency and presumed normal folate uptake at the BBB, the loss of FRa function still results in low CSF folate levels ultimately causing cognitive deficits. These clinical observations suggest that folate transport across the BBB may not adequately supply neural cells with sufficient folates to sustain normal development (Lehtinen et al., 2011; Lehtinen & Walsh, 2011). Thus, modulating folate transport at the level of the BBB could significantly contribute to folate uptake into the CNS. 84

Finding new approaches for enhancing brain folate delivery may have a significant impact on the treatment of childhood neurometabolic disorders caused by folate deficiency. Our group has reported that the vascular BBB could present an alternative route for brain folate transport, especially when inactivation of FRα or PCFT impairs the major route of folate uptake at the choroid plexus (Alam et al., 2017). We further demonstrated in vitro that functional expression of another folate transporter, RFC, and its upregulation by the VDR nuclear receptor could potentially increase folate transport across the BBB (Alam et al., 2017). In the present study, we go on to demonstrate in vivo that loss of Folr1 expression results in a substantial decrease in the delivery of 13 C5-folates to the brain and that delivery is restored by administration of the VDR activating ligand, calcitriol. The data indicate that this is due to the salutary impact of calcitriol on the expression of RFC at the BBB. Since Folr1 is almost exclusively expressed in the choroid plexus, these findings (1) represent a demonstration of the substantial impact of the loss of choroid plexus function on the delivery of folates to the brain, as observed in hereditary folate malabsorption and cerebral folate deficiency syndrome, and (2) suggest a novel adjunct therapy to the treatment of these neurological disorders with calcitriol.

Using immunocytochemical staining and confocal microscopy, we characterized the localization of RFC, PCFT, and FRα in primary mouse microvessel endothelial cells representative of an in vitro BBB system (Figure 7-1). We confirmed that all three folate transport pathways were primarily localized to the plasma membrane, as evident by their similar localization to the plasma membrane marker Na+/K+-ATPase a. Detection of robust expression of folate transporters at the mouse BBB corroborates previous reports (Alam et al., 2017; X. Wang et al., 2013; Y Wang et al., 2001) and provides evidence of a potential role for the BBB in brain folate delivery.

To assess the contribution of folate transporters (i.e., RFC) in facilitating folate uptake across the BBB, we implemented the use of an in vivo mouse model lacking Folr1 (Folr1 KO). Systemic deletion of this receptor was confirmed through qPCR analysis, which showed a lack of Folr1 (FRα) mRNA expression in various tissues (liver, kidney, isolated brain capillaries representative of the BBB) of Folr1 KO mice compared with WT controls (Figure 7-2). We also did not observe significant differences in the level of Slc19a1 (RFC) and Slc46a1 (PCFT) expression between Folr1 KO and WT animals, indicating a lack of compensation from these transporters in response to the loss of Folr1. Given these findings, it is important to note that although PCFT was detected in isolated mouse brain capillaries, this transporter may not be 85

functionally relevant at this site due to the neutral pH of the BBB interface and the low pH required for optimal PCFT uptake. Thus, our subsequent studies focused on characterizing the role of RFC at the level of the BBB.

In vivo regulation of RFC by VDR was examined through i.p. administration of the VDR ligand, calcitriol, to WT and Folr1 KO mice using the dosing regimen specified by Chow et al. (E. C. Y. Chow, Sondervan, et al., 2011). We demonstrated that calcitriol treatment significantly increased Slc19a1 (RFC) mRNA expression in isolated brain capillaries (1.5-fold) of WT and Folr1 KO mice (Figures 7-3A and 7-4A). These results are in agreement with our laboratory’s previous findings that in vitro exposure of hCMEC/D3 cells and mouse brain capillaries to calcitriol resulted in over 50% upregulation of RFC mRNA and protein (Alam et al., 2017). In the present study, Slc19a1 (RFC) mRNA was also induced in other tissues known to express this transporter, such as the liver (2-fold), kidney (1.3-fold), or duodenum (6-fold) (Figures 7-3A and 7-4A). Only a modest increase in Slc46a1 (PCFT) mRNA was observed in isolated brain capillaries (1.5-fold) and liver (1.4-fold) of calcitriol-treated WT mice, but the contribution of this transporter in BBB folate delivery may be less relevant due to the low pH required for optimal PCFT activity (Appendix B, Figure B-1). Furthermore, Abcb1a (P-gp) mRNA was elevated in various tissues of WT and Folr1 KO mice following calcitriol treatment (Figures 7-3B and 7-4B). This was not surprising since VDR is a key regulator of a number of membrane transporters including the drug efflux pump, P-gp (E. Chow, Sun, Khan, Groothuis, & Pang, 2010; Durk et al., 2012). However, induction of this transporter at the BBB may have functional consequences since

P-gp has also been identified to be a low-affinity efflux transporter of folates (Km = 0.2−2 mM) and could potentially oppose folate uptake by RFC (Assaraf, 2006). We further examined the expression of several OAT and MRP transporters that are known to transport folates (OAT1-3, Km

= 10-700 μM; MRP1-4, Km = 0.2-2 mM) and potentially contribute to the observed effect of calcitriol on brain folate levels. As shown in Figure B-2 (Appendix B), substantial increases in Slc22a8 (Oat3), Abcc2 (Mrp2), or Abcc3 (Mrp3) mRNA were observed in kidney or liver tissues of calcitriol-treated mice, but not in isolated mouse brain capillaries. These results suggest that there may be minimal contribution from these transporters in folate uptake at the BBB, further supporting our hypothesis that folate transport at the BBB could be largely mediated by RFC.

Systemic effects of calcitriol administration were also investigated to evaluate the safety of our treatment protocol. As expected, exogenous administration of calcitriol significantly 86

increased calcitriol levels in plasma of treated mice (5.5-fold) as seen in earlier reports by other groups (E. C. Y. Chow et al., 2013) (Figure 7-5B). The alternate-day dosing regimen recommended by Chow et al. (E. C. Y. Chow, Sondervan, et al., 2011) was also shown previously to alleviate the hypercalcemic and weight loss effects of calcitriol, but we observed significant increases in plasma calcium concentrations (1.5-fold) of WT and Folr1 KO mice after the 8-d treatment period in our studies (Figure 7-5C). Phosphorus levels, on the other hand, were unchanged following calcitriol treatment (Figure 7-5D). WT and Folr1 KO mice also exhibited a statistically significant decrease in body weight starting at Day 4 of calcitriol injections (Figure 7- 5A), but this weight loss plateaued after prolonged calcitriol administration (Appendix B, Figure B-3). Hypercalcemia may have been induced by VDR-mediated transactivation of calcium ion channels in the kidney and intestine (i.e., TRPV5 and TRPV6), resulting in volume depletion due to frequent urination (E. C. Y. Chow et al., 2013). Although the observed hypercalcemia and weight loss did not seem to affect the animals’ overall health, exogenous administration of calcitriol still requires strict monitoring to avoid toxicities associated with VDR activation in vivo. Supplementation with other vitamin D forms, such as ergocalciferol and cholecalciferol, may also be helpful in reducing toxicities since these compounds require initial conversion into the active calcitriol form by CYP27B1, an enzyme tightly regulated by serum calcium levels (Anderson, O’ Loughlin, May, & Morris, 2002).

Folate distribution in WT and Folr1 KO mice was subsequently examined to further understand the effect of VDR activation on RFC functional activity, particularly at the BBB. We initially determined that basal concentrations of reduced folates, 5-formylTHF and 5-methylTHF, in the plasma (Cplasma) and brain (Cbrain) of Folr1 KO mice were much lower compared to WT (Figure 7-6). These results suggest that despite supplementing with high doses of folic acid (40 mg/kg) during gestation, loss of Folr1 function continues to affect brain folate uptake as well as folate reabsorption in the kidneys of adult Folr1 KO animals (Birn et al., 2005). We also used 13 [ C5]-labeled 5-formylTHF to evaluate folate disposition in vivo. In our hands, i.v. administration 13 of [ C5]-5-formylTHF in WT mice yielded pharmacokinetic parameters that are comparable to an earlier report by Wu and Pardridge (D. Wu & Pardridge, 1999) (Figure 7-7A and Appendix B, 13 13 Table B-1). Significant levels of injected [ C5]-5-formylTHF and its metabolite, [ C5]-5- 13 methylTHF, were also detected in brain tissues of WT mice, consistent with transport of [ C5]- labeled folates across the BBB (Figure 7-7B). Finally, to verify whether VDR activation will also

induce RFC function in vivo, Folr1 KO mice were subjected to an 8-d pretreatment with calcitriol 13 or vehicle before receiving a single i.v. injection of [ C5]-5-formylTHF. Within the Folr1 KO 13 brain tissues, we observed a significant and remarkable increase of [ C5]-5-formylTHF concentrations (6-fold) and corresponding Cbrain/Cplasma ratio (15-fold) in calcitriol-treated mice compared with the vehicle-treated group (Figure 7-8). These findings are most exciting as they demonstrate that in vivo induction of RFC functional expression through activation of VDR by calcitriol can significantly increase brain folate delivery in Folr1 KO mice. In fact, pretreatment

13 with calcitriol resulted in brain [ C5]-5-formylTHF levels that were comparable to those of WT animals, suggesting that upregulation of RFC could potentially compensate for the loss of Folr1- mediated brain folate uptake.

Beyond abnormalities in folate delivery to the brain due to loss-of-function mutations in FRα and PCFT, low CSF folate level was previously identified in at least two types of autism spectrum disorders, such as Rett syndrome and infantile low-functioning autism (V. T. Ramaekers et al., 2007, 2013). Reduced brain folate transport was linked to the presence of FRα autoantibodies, which inhibit FRα function either by blocking folate binding to the receptor or by binding to an epitope distant from the folate-binding site and disrupting receptor function. A high prevalence of FRα autoantibodies (75%) was also reported in a cohort of children with idiopathic autism spectrum disorders and was found to correlate with low folate concentrations in the CSF (R. E. Frye et al., 2013). The intervention with 5-formylTHF (leucovorin or folinic acid) has shown favorable responses in autism spectrum disorders, especially among patients who are positive for FRα autoantibodies. Ramaekers et al. (V. T. Ramaekers et al., 2007) reported partial recovery of neurological and social impairments in patients with infantile low-functioning autism after oral supplementation with 5-formylTHF. Frye et al. (R. E. Frye et al., 2013) also observed marked improvements in attention, language, and behavior in over 60% of patients treated with high oral doses of 5-formylTHF compared with nontreated controls. In a randomized, double-blind, placebo-controlled trial conducted by the same group, 12-wk high-dose oral 5-formylTHF supplementation significantly improved verbal communication in patients expressing FRα autoantibodies (R. E. Frye et al., 2018). Despite the promising effects of 5-formylTHF intervention in autism spectrum disorders, there is still a need to establish more effective therapies. Frye et al. (R. E. Frye et al., 2013) have reported that prolonged intake of 5-formylTHF at such high doses could result in adverse events, including insomnia, gastroesophageal reflux, and worsening

aggression. Since autism is a lifelong disorder, finding an optimal therapeutic approach with limited adverse effects is important. Our present findings could potentially represent a new strategy for enhancing folate delivery to the brain via RFC at the BBB.

In summary, we have shown in vivo that loss of Folr1 substantially decreases folate delivery to the brain, demonstrating the major contribution of the choroid plexus to brain folate transport. We further provide the first evidence that in vivo activation of VDR by its natural activating ligand, calcitriol, can induce RFC expression in mice lacking Folr1. These findings suggest that augmenting RFC functional expression enhances folate delivery at the BBB and could potentially compensate for the loss of FRα-mediated folate uptake at the choroid plexus. The current therapeutic approach for neurological disorders resulting from folate transport defects is to achieve high 5-formylTHF blood levels in order to deliver sufficient folate to the brain to sustain normal neural development. However, despite achievement of high blood folate levels, CSF folate levels are still below the normal range and neurological signs, particularly seizures, may be difficult to control (Aluri et al., 2018; Geller et al., 2002; Torres et al., 2015; R Zhao et al., 2017). Therefore, modulating folate transport at the BBB through RFC offers a therapeutic approach to improving brain folate delivery for the treatment of neurometabolic disorders caused by loss of FRα or PCFT function.

7.7 Acknowledgements

We thank Dr. Robert Steinfeld (Department of Pediatric Neurology, University Children’s Hospital Zürich, Zürich, Switzerland) for his initial insights on this work. We acknowledge Dr. Bogdan Wlodarczyk (Baylor College of Medicine, Houston, TX) for his excellent advice on establishing the animal breeding colonies. We also thank Adrian Turner for his assistance with the isolation of mouse brain capillaries. This research was supported by an operating grant from the Natural Sciences and Engineering Research Council of Canada (NSERC 498383 to R.B.). C.A. was a recipient of an internal graduate fellowship and Centre for Pharmaceutical Oncology Scholarship from the Leslie Dan Faculty of Pharmacy (University of Toronto). R.H.F. was supported by NIH grants R01HD081216 and R01HD083809. I.D.G. was supported by a National Cancer Institute Grant CA082621. D.L.O. was supported by NSERC Grant 43302.

Chapter 4 Nuclear Respiratory Factor 1 (NRF-1) Upregulates the Expression and Function of Reduced Folate Carrier (RFC) at the Blood-Brain Barrier

This chapter contains unpublished work from: Alam C and Bendayan R. (2019). Nuclear respiratory factor 1 (NRF-1) upregulates the expression and function of reduced folate carrier (RFC) at the blood-brain barrier. Research manuscript in preparation.

Author Contributions:

Research design: C Alam (first author), R Bendayan (principal investigator)

Experiments and data analysis: C Alam (all figures), R Bendayan (all figures)

Writing of the manuscript: C Alam (manuscript preparation and revisions); R Bendayan (overall conceptual and editorial review of several manuscript drafts)

8.1 Abstract

Folates are important for neurodevelopment and cognitive function. Folate transport across biological membranes is mediated by three major pathways: folate receptor alpha (FRα), proton- coupled folate transporter (PCFT), and reduced folate carrier (RFC). Brain folate transport primarily occurs at the choroid plexus through FRα and PCFT; inactivation of these transport systems causes early childhood neurodegenerative disorders. Our group has recently reported that upregulation of RFC at the blood-brain barrier (BBB) through interactions with specific transcription factors could increase brain folate delivery. The objective of this study was to investigate the role of nuclear respiratory factor 1 (NRF-1) in the regulation of RFC at the BBB. We detected robust expression of NRF-1 and its coactivator, peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α), in human (hCMEC/D3 cells) and rodent (mouse brain capillaries) in vitro models of the BBB. Activation of NRF-1/PGC-1α signaling through treatment with pyrroloquinoline quinone (PQQ) significantly induced RFC expression and transport activity in hCMEC/D3 cells. We also demonstrated that RFC functional expression could be downregulated by NRF-1 or PGC-1α targeting siRNA, suggesting that the NRF-1/PGC-1α signaling pathway is involved in the regulation of this transporter. Furthermore, in vivo treatment of wild type mice with PQQ increased RFC expression in isolated mouse brain capillaries. These findings demonstrate that NRF-1/PGC-1α activation by PQQ upregulates RFC functional expression at the BBB and could potentially enhance brain folate uptake.

8.2 Introduction

Folates are essential for key biosynthetic processes in mammalian cells (Tibbetts & Appling, 2010). Folate requirements in mammals are obtained through the diet since they lack the enzymatic capacity for folate biosynthesis. Folate transport across biological membranes is mediated by three major pathways: folate receptor alpha (FRα; FOLR1), proton-coupled folate transporter (PCFT; SLC46A1), and reduced folate carrier (RFC; SLC19A1). FRα is a glycosylphosphatidylinositol-anchored receptor that binds folates with high affinity (Km = 1-10 nM) and facilitates transport though receptor-mediated endocytosis (Elnakat & Ratnam, 2004;

Kamen & Smith, 2004). PCFT is a lower affinity folate transporter (Km = 1 μM) that acts as a proton cotransporter with optimal activity at pH 5.5 (Qiu et al., 2006; Visentin et al., 2014). RFC has a comparable affinity as PCFT (Km = 2-7 μM) for reduced folate uptake at physiological pH,

and is an antiporter that exchanges folates with intracellular organic phosphates (Rongbao Zhao & Goldman, 2013).

Folates play a crucial role in the development and function of the central nervous system (CNS). Brain folate transport primarily occurs at the choroid plexus through concerted actions of FRα and PCFT. FRα constitutes a major transcytosis pathway for folates across the choroid plexus epithelium, while PCFT facilitates folate export from FRα-containing exosomes directly into the cerebrospinal fluid (CSF) (Grapp et al., 2013; R Zhao, Min, et al., 2009). Abnormalities in brain folate transport can lead to suboptimal folate levels in the CSF causing early childhood neurodegenerative disorders. Inactivation of FRα through loss-of-function mutations (Grapp et al., 2012; Steinfeld et al., 2009) or the presence of FRα autoantibodies (Vincent T Ramaekers et al., 2005) causes cerebral folate deficiency, which is characterized by developmental delays, abnormal brain myelination, psychomotor regression, ataxia, and recurrent seizures. Genetic mutations in the SLC46A1 gene which encodes for PCFT can result in hereditary folate malabsorption, a condition that manifests within a few months of birth causing newborns to develop anemia, immune deficiency, and cognitive deficits (Qiu et al., 2006; R Zhao et al., 2017). Low CSF folate has also been documented in several types of autism spectrum disorders (i.e., Rett syndrome and infantile low-functioning autism) due to the high prevalence of FRα autoantibodies in these patients (V. T. Ramaekers et al., 2007, 2013). Several studies have shown that folate supplementation can somewhat reduce the occurrence of these neurodevelopmental disorders; however, current therapeutic approaches are not sufficiently effective (Kronn & Goldman, 2017; Torres et al., 2015; R Zhao et al., 2017).

The cerebral vascular endothelium or blood-brain barrier (BBB) represents an alternative route for folate delivery into the CNS, especially when inactivation of FRα or PCFT impair the major route of folate uptake at the choroid plexus. Several groups, including ours, have confirmed the presence of folate transporters particularly RFC in human (Alam et al., 2017; Araújo et al., 2010) and rodent (Y Wang et al., 2001) BBB, and demonstrated active transport of folates in various BBB model systems, such as isolated human brain capillaries (D. Wu & Pardridge, 1999), and immortalized cell lines of human (hCMEC/D3) and rat (RBE4) cerebral microvessel endothelium (Alam et al., 2017; Araújo et al., 2010). Recent work from our laboratory also examined the role of transcription factors, particularly vitamin D receptor (VDR), in upregulating RFC functional expression at the BBB (Alam et al., 2019, 2017). Transcription factors are proteins 92

that bind to specific recognition sites on DNA and control gene expression by activating or repressing the transcription of targeted DNA sequences (Lambert et al., 2018). We provided evidence that activation of VDR through exposure to its activating ligand, 1,25-dihydroxyvitamin

D3 (1,25(OH)2D3 or calcitriol), induced RFC expression and function in hCMEC/D3 cells and isolated mouse brain capillaries (Alam et al., 2017). We further showed in vivo that treatment with calcitriol enhanced RFC expression at the BBB and effectively restored folate delivery to the brain of Folr1 (FRα) knockout mice (Alam et al., 2019). These findings indicate that modulating folate uptake at the BBB through RFC could represent a novel strategy for enhancing brain folate delivery.

Several other transcription factors have been reported to modulate the expression of major folate transport pathways, although their physiological roles have yet to be determined (Gonen & Assaraf, 2010; Halwachs et al., 2010; Salbaum et al., 2009; Tran et al., 2005). In particular, nuclear respiratory factor 1 (NRF-1) is a transcription factor primarily known for its role in promoting transcription of genes required for mitochondrial biogenesis and respiratory function. NRF-1 can regulate the expression of its target genes by binding to specific DNA recognition sites as a protein homodimer, or through interaction with coactivator proteins such as the peroxisome proliferator- activated receptor-γ coactivator-1α (PGC-1α) (Appendix C, Figure C-1) (Z. Wu et al., 1999). Bioinformatic analyses have identified at least four putative NRF-1 binding sites on the promoter region of the SLC19A1 gene which encodes for RFC, suggesting a potential role for this transcription factor in the regulation of RFC (Gonen & Assaraf, 2010). The objective of the present study was to investigate the involvement of NRF-1 in regulating RFC functional expression at the BBB in vitro using hCMEC/D3 cells and in vivo in mice. Finding new approaches for enhancing brain folate delivery may have significant impact to the treatment of childhood neurometabolic disorders caused by folate deficiency.

8.3 Materials and Methods 8.3.1 Materials All cell culture reagents were obtained from Invitrogen (Carlsbad, CA, USA), unless indicated otherwise. Real-time quantitative polymerase chain reaction (qPCR) reagents, such as reverse transcription cDNA kits and qPCR primers, were purchased from Applied Biosystems (Foster City, CA, USA) and Life Technologies (Carlsbad, CA, USA), respectively. Primary rabbit

polyclonal anti-SLC19A1 (RFC; AV44167) antibody was obtained from Sigma-Aldrich (Oakville, ON, Canada). Mouse monoclonal anti-beta actin (sc-47778), anti-NRF-1 (sc-101102), and anti- PGC-1α (sc-517380) antibodies were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Pyrroloquinoline quinone or PQQ (Item no. 20681) was obtained from Cayman Chemical (Ann Arbor, MI, USA). Pre-designed and validated small interfering RNA (siRNA) against human NRF-1 (sc-38105) or human PGC-1α (sc-38884), and scrambled non-silencing control siRNA (sc- 37007) were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Tritium-labeled [3H]- methotrexate (23.4 Ci/mmol) was purchased from Moravek Biochemicals (Brea, CA, USA). Unlabeled methotrexate and pemetrexed were obtained from Sigma-Aldrich (Oakville, ON, Canada).

8.3.2 Cell Culture Immortalized human cerebral microvessel endothelial cell line (hCMEC/D3) was kindly provided by P.O. Couraud (Institut Cochin, Departement Biologie Cellulaire and INSERM, Paris, France). hCMEC/D3 cells (passage 27-36) were cultured in Endothelial Cell Basal Medium-2 (Lonza, Walkersville, MD, USA), supplemented with vascular endothelial growth factor, insulin- like growth factor 1, epidermal growth factor, fibroblast growth factors, hydrocortisone, ascorbate, GA-1000, heparin and 2.5% fetal bovine serum, and grown on rat tail collagen type I-coated flasks and plates. Cells were maintained in a humidified incubator at 37°C, 5% CO2, and 95% air atmosphere with fresh medium replaced every 2 to 3 days. Cells were subcultured with 0.25% trypsin-EDTA upon reaching 95% confluence. For functional studies, cells were seeded into rat tail collagen-coated 24-well plates at a density of 6.3 x 104 cells/cm2 and subsequently used for experiments upon reaching 100% confluence (4-5 days). Cells used for PQQ treatment were seeded into rat tail collagen-coated 6-well plates at an initial density of 3.3 x 104 cells/cm2 and grown to confluence for 5 days. For siRNA transfection studies, cells were plated in rat tail collagen-coated 6-well plates at an initial density of 7.8 x 104 cells/cm2 and grown to 80% confluence for 1 day before being subjected to transfection. The culture medium of plated cells was replaced every 2 days, and 24 h before experiments.

8.3.3 Mouse Brain Capillary Isolation

Brain capillaries were isolated from adult male LM/Bc mice (8−13 weeks old) as described previously (Alam et al., 2019; Chan & Cannon, 2017). Briefly, animals were anesthetized through

isoflurane inhalation and decapitated once a deep anesthetic surgical plane was achieved. Brains were removed immediately and cortical gray matter was isolated and homogenized in ice-cold isolation buffer (phosphate-buffered saline or PBS containing calcium, magnesium, and supplemented with 5 mM glucose and 1 mM sodium pyruvate). Ficoll solution (30% final concentration) was added to the brain homogenates and the mixture was centrifuged at 5,800 x g for 20 min at 4°C. The resulting pellet of capillaries was resuspended in isolation buffer supplemented with 1% bovine serum albumin (BSA) and filtered through a 300 μm nylon mesh. The filtrate containing the capillaries was passed through a 30 μm pluriStrainer and washed with 50 mL isolation buffer containing 1% BSA. Capillaries were harvested with 50 mL ice-cold isolation buffer and centrifuged at 1,600 x g for 5 min. The resulting pellet containing the capillaries was snap frozen in liquid nitrogen until further analysis.

8.3.4 Gene Expression Analysis

mRNA expression of specific genes of interest was quantified using qPCR. Total RNA was isolated from cell or tissue samples using TRIzol reagent and treated with DNase I to remove contaminating genomic DNA. RNA concentration (absorbance at 260 nm) and purity (absorbance ratio 260/280) was assessed using a NanoDrop. The RNA (2 μg) was then reverse transcribed to cDNA using a high-capacity reverse transcription cDNA kit according to the manufacturer’s instructions. Specific human or mouse primer pairs for: SLC19A1/Slc19a1 (RFC; Hs00953344_m1, Mm00446220_m1), NRF1/Nrf1 (NRF-1; Hs00602161_m1, Mm01135606_m1), PPARGC1A/Ppargc1a (PGC-1α; Hs00173304_m1, Mm01208835_m1), Tfam (Tfam; Hs01082775_m1, Mm00447485_m1), TFB1M/Tfb1m (TFB1M; Hs01084404_m1, Mm00524825_m1), TFB2M/Tfb2m (TFB2M; Hs00915025_m1, Mm01620397_s1), ABCB1 (P- gp; Hs00184500_m1), ABCG2 (BCRP; Hs01053790_m1), ABCC2 (MRP2; Hs00166123_m1), and ABCC3 (MRP3; Hs00978473_m1), were designed and validated by Life Technologies for use with TaqMan qPCR chemistry. All assays were done in triplicates with the housekeeping gene for human or mouse cyclophilin B (Hs00168719_m1, Mm00478295_m1) as internal control. For each gene of interest, the critical threshold cycle (CT) was normalized to cyclophilin B using the comparative CT method. The difference in CT values (ΔCT) between the target gene and cyclophilin

B was then normalized to the corresponding ΔCT of the vehicle control (ΔΔCT) and expressed as fold expression (2-ΔΔCT) to assess the relative difference in mRNA expression for each gene.

8.3.5 Protein Expression Analysis Western blotting was performed according to our published protocol (Alam et al., 2017). Briefly, whole cell lysates were obtained after lysing cells with a modified RIPA buffer containing: 50 mM Tris pH 7.5, 150 mM NaCl, 1 mM EGTA, 1 mM sodium o-vanadate, 0.25% (v/v) sodium deoxycholic acid, 0.1% (v/v) sodium dodecyl sulphate (SDS), 1% (v/v) NP-40, 200 μM PMSF, and 0.1% (v/v) protease inhibitor. Protein concentration of cell lysates was quantified using Bradford’s protein assay (Bio-Rad Laboratories) with BSA as the standard. For each sample, total protein (10 or 50 μg) was mixed in Laemmli buffer and 10% β-mercaptoethanol, separated on 7 or 10% SDS-polyacrylamide gel, and electrotransferred onto a polyvinylidene fluoride membrane overnight. The blots were blocked for 2 h at room temperature in 5% skim milk Tris-buffered saline solution containing 0.1% Tween 20 and incubated overnight at 4°C with primary rabbit polyclonal anti-SLC19A1 (RFC) antibody (1:250), mouse monoclonal anti-NRF-1 antibody (1:200), mouse monoclonal anti-PGC-1α antibody (1:200), or mouse monoclonal anti-beta actin antibody (1:1000). The blots were incubated for 1.5 h with corresponding horseradish peroxidase- conjugated anti-rabbit (1:5000) or anti-mouse (1:5000) secondary antibody. Protein bands were detected using enhanced chemiluminescence SuperSignal West Pico System (Thermo Fisher Scientific) and autoradiographed onto X-ray film.

8.3.6 Transport Assays Functional assays with [3H]-methotrexate were performed following standard procedures from our laboratory (Alam et al., 2017). All experiments were performed using transport buffer consisting of Hanks’ balanced salt solution supplemented with 0.01% BSA and 25 mM HEPES (pH 7.4). Confluent hCMEC/D3 cell monolayers grown on 24-well plates were washed twice with transport buffer (pH 7.4) and then pre-incubated in the same buffer at 37°C for 20 minutes. Transport was initiated by adding 0.5 ml of transport buffer (pH 7.4) containing 50 nM [3H]- methotrexate at 37°C. The inhibitory effect of pemetrexed, a folate analog with high affinity for RFC, was examined by adding the compound to both pre-incubation and radioactive transport buffers. At the desired time interval (i.e., 1 min), the radioactive medium was aspirated and cells were washed twice with ice-cold PBS and solubilized in 1 ml of 1% Triton X-100 at 37°C for 40 min. The content of each well was collected and mixed with 3 ml of PicoFluor 40 scintillation fluid (PerkinElmer Life and Analytical Sciences), and total radioactivity was measured with a Beckman Coulter LS6500 Scintillation counter. For each experiment, correction for nonspecific

binding and variable quench time was conducted by estimating the retention of radiolabeled compound in the cells after a minimum (zero) time of exposure. The “zero time” uptake (background) was determined by removing the radiolabeled solution immediately after its introduction into the well, followed by two washes of ice-cold PBS and collection of cells for liquid scintillation counting. Cellular uptake of radiolabeled methotrexate was normalized to total protein content per well, which was measured using the Bio-Rad DC Protein Assay kit with BSA as the standard.

8.3.7 Pyrroloquinoline Quinone (PQQ) Treatments PQQ is an enzyme cofactor naturally found in soil and plant-derived food. This compound has been shown to indirectly activate NRF-1 by modulating PGC-1α expression via transcriptional and post-translational mechanisms (Appendix C, Figure C-1) (Chowanadisai et al., 2010; Saihara, Kamikubo, Ikemoto, Uchida, & Akagawa, 2017). For in vitro studies, confluent hCMEC/D3 cell monolayers grown on 6- or 24-well plates were treated with either vehicle (DMSO) or PQQ (0.01-5 µM) for a period of 24 or 48 h at 37°C. At the desired time interval, treated cells were collected in TRIzol or RIPA lysis buffer and subsequently processed for gene and protein analyses, respectively. The effect of PQQ on methotrexate uptake was further examined by treating hCMEC/D3 cell monolayers with 1 and 5 µM PQQ for 24 h before conducting transport assays with [3H]-methotrexate at pH 7.4 and 37°C. To ensure that cells remained viable during treatment, all concentrations of PQQ used in this study were tested with tetrazolium salts (MTT) assay. We confirmed that there was no significant reduction in cell viability in the PQQ-treated groups compared to vehicle-treated or untreated controls (Appendix C, Figure C-2).

For in vivo studies, PQQ in powdered form was initially dissolved in DMSO before diluting in a solution containing 5% DMSO, 5% Tween-80, and saline. Male wild type mice of LM/Bc background strain (8-13 weeks old) were subjected to daily intraperitoneal (i.p.) injections with PQQ (10 mg/kg) or vehicle (5% DMSO, 5% Tween-80, 90% saline) for 10 consecutive days. This dosing regimen was previously demonstrated to show neuroprotective effects in mice (C. Yang et al., 2014). At 24 h following the last injection, animals were anesthetized through isoflurane inhalation and decapitated prior to collection of various tissues (brain, liver, kidney).

8.3.8 siRNA Transfection

hCMEC/D3 cells plated in 6- or 24-well plates were subjected to siRNA transfection upon reaching 80% confluence (24 h). The transfection mix was prepared in Opti-MEM (Invitrogen) medium containing control, NRF-1, or PGC-1α siRNA and Lipofectamine RNAi MAX (Invitrogen) according to the manufacturer’s protocol. The final concentration of siRNA and Lipofectamine added to the cells were 100 nM and 2 μl/ml, respectively. Cells were cultured in the presence of transfection mixture for 24 h before replacing with fresh hCMEC/D3 cell medium the following day. Cells were grown for an additional 48 h before being harvested and processed for western blotting analysis to determine NRF-1, PGC-1α, and RFC protein expression, or used for transport assays to determine cellular accumulation of [3H]-methotrexate following siRNA transfection.

8.3.9 Data Analysis

All experiments were repeated at least three times using cells from different passages or different mouse brain capillary preparations. For in vivo experiments, samples were collected from 7-9 animals per treatment group. Results are presented as mean ± S.E.M. All statistical analyses were performed using Prism 6 software (GraphPad Software Inc., San Diego, CA, USA). Statistical significance between two groups was assessed by two-tailed Student’s t test for unpaired experimental values. Multiple group comparisons were performed using either one-way or two- way analysis of variance (ANOVA) with Bonferroni’s post-hoc test. A p value of < 0.05 was considered statistically significant.

8.4 Results 8.4.1 Expression of NRF-1 and PGC-1α at the BBB

Using qPCR analysis, relative expression of NRF-1 and PGC-1α was initially examined in several in vitro systems representative of the human (hCMEC/D3 cell line) and rodent (isolated mouse brain capillaries) BBB. NRF1/Nrf1 (NRF-1) and PPARGC1A/Ppargc1a (PGC-1α) mRNA were expressed in both BBB models, particularly in isolated mouse brain capillaries (Figure 8-1). In mouse brain capillaries, Nrf1 was detected at significantly higher levels compared to Ppargc1a (Figure 8-1).

1.5

****

1.0 ***

0.5 mRNA Expression (relative to (relative cyclophilin B) Relative NRF1/PPARGC1A 0.0 NRF-1 PGC-1α

hCMEC/D3 Mouse Brain Capillary

Figure 8-1. Relative expression of NRF-1 and PGC-1α in human (hCMEC/D3 cell line) and rodent (isolated mouse brain capillaries) in vitro models of the BBB. mRNA expression of human or mouse NRF1/Nrf1 (NRF-1) and PPARGC1A/Ppargc1a (PGC-1α) genes were determined using TaqMan gene expression assays. Results are presented as mean relative mRNA expression ± S.E.M. normalized to the housekeeping human/mouse cyclophilin B gene from n = 4 independent experiments. Asterisks represent significant differences in NRF1/Nrf1 or PPARGC1A/Ppargc1a expression between hCMEC/D3 cells and isolated mouse brain capillaries (***, p < 0.001; ****, p < 0.0001). In isolated mouse brain capillaries, Nrf- 1 mRNA was also expressed at significantly higher levels compared to Ppargc1a (p < 0.0001).

8.4.2 Effect of PQQ Treatment on NRF-1 and PGC-1α Expression

Activation of NRF-1/PGC-1α signaling was performed by exposing hCMEC/D3 cells to various concentrations of PQQ (0.01-5 µM) for 24 and 48 h. PQQ is an enzyme cofactor naturally found in soil and plant-derived food. This compound has received increasing amount of attention in recent years due to its purported health benefits. In rodent models, PQQ administration showed improvements in reproduction, early embryonic development, growth and immune function, and protective effects towards neuronal cell death and cardiac damage (Rucker, Chowanadisai, & Nakano, 2009). Several studies have attributed these health benefits to PQQ’s antioxidant and anti- inflammatory properties (C. Yang et al., 2014), and potentially its ability to enhance mitochondrial biogenesis and function (Chowanadisai et al., 2010; Saihara et al., 2017). PQQ has been shown to indirectly activate NRF-1 by modulating PGC-1α expression via transcriptional and post- translational mechanisms (Appendix C, Figure C-1) (Chowanadisai et al., 2010; Saihara et al., 2017). In the present study, treatment with PQQ did not alter NRF1 expression in hCMEC/D3 cells 99

(Figure 8-2A), but a significant increase in PPARGC1A mRNA (up to 10-fold) was observed in cells exposed to 5 µM PQQ at 48 h (Figure 8-2B). The expression of known NRF-1 target genes, including mitochondrial transcription factors A (Tfam) and B (TFB1M, TFB2M), in response to PQQ treatment was also observed. These genes contain response elements for NRF-1 and they encode for transcription factors that are crucial to the stimulation (Tfam) and initiation (TFB1M and TFB2M) of mitochondrial DNA transcription (Scarpulla, 2008). As shown in Figure 8-3A, hCMEC/D3 cells treated with increasing concentrations of PQQ (0.01-5 µM) exhibited significantly higher levels of Tfam mRNA (up to 50%) compared to vehicle (DMSO) at 48 h. A similar 50% induction in TFB1M and TFB2M mRNA expression was observed after exposure to PQQ for 24 h (Figures 8-3B and 8-3C).

A NRF1 (24h mRNA) NRF1 (48h mRNA) 2.0 2.0

1.5 1.5

1.0 1.0

0.5 0.5 (relative to (relative vehicle) (relative to (relative vehicle) NRF1 mRNA Expression NRF1 mRNA Expression

0.0 0.0 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5

PQQ Concentration (µM) PQQ Concentration (µM)

B PGC-1α (24h mRNA) PGC-1α (48h mRNA) 12 15 * 10

8 10

4 5 (relative to (relative vehicle) (relative to (relative vehicle) 2 PPARGC1A mRNA Expression mRNA PPARGC1A 0 Expression mRNA PPARGC1A 0 Vehicle 0.01 0.05 1 5 Vehicle 0.01 0.1 1 5

PQQ Concentration (µM) PQQ Concentration (µM)

Figure 8-2. Effect of PQQ treatment on NRF-1 and PGC-1α expression in hCMEC/D3 cells. (A) NRF1 (NRF-1) mRNA expression was unchanged following PQQ treatment (0.01-5 µM) for 24 and 48 h compared to vehicle (DMSO) control. (B) Significant increase in PPARGC1A (PGC-1α) mRNA was observed in hCMEC/D3 cells treated with 5 µM PQQ for 48 h. Results are presented as mean ± S.E.M. for n = 3-4 independent experiments. Asterisks represent data points significantly different from vehicle (DMSO) control (*, p < 0.05).

100

A Tfam (24h mRNA) Tfam (48h mRNA) 2.0 2.0 ** * ** * 1.5 1.5

1.0 1.0

0.5 0.5 (relative to (relative vehicle) (relative to (relative vehicle) Tfam mRNA Expression mRNA Tfam Tfam mRNA Expression mRNA Tfam

0.0 0.0 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5

PQQ Concentration (µM) PQQ Concentration (µM)

B TFB1M (24h mRNA) TFB1M (48h mRNA) 2.5 2.0 * 2.0 1.5

1.5 1.0 1.0

0.5 (relative to (relative vehicle) 0.5 to (relative vehicle) TFB1M mRNA Expression mRNA TFB1M TFB1M mRNA Expression mRNA TFB1M 0.0 0.0 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5

PQQ Concentration (µM) PQQ Concentration (µM)

C TFB2M (24h mRNA) TFB2M (48h mRNA) 2.0 2.0 * * *

1.5 1.5

1.0 1.0

0.5 0.5 (relative to (relative vehicle) (relative to (relative vehicle) TFB2M mRNA Expression mRNA TFB2M TFB2M mRNA Expression mRNA TFB2M 0.0 0.0 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5

PQQ Concentration (µM) PQQ Concentration (µM)

Figure 8-3. Effect of PQQ treatment on the expression of NRF-1 target genes in hCMEC/D3 cells. Significant increases in Tfam (Tfam) (A), TFB1M (TFB1M) (B), and TFB2M (TFB2M) (C) mRNA were observed following PQQ treatment (0.01-5 µM) for 24 or 48 h compared to vehicle (DMSO) control. Results are presented as mean ± S.E.M. for n = 3-4 independent experiments. Asterisks represent data points significantly different from vehicle (DMSO) control (*, p < 0.05; **, p < 0.01).

101

8.4.3 Effect of PQQ Treatment on RFC Functional Expression We further investigated the effect of NRF-1/PGC-1α activation on RFC expression at the BBB by treating hCMEC/D3 cells with PQQ. SLC19A1 (RFC) mRNA expression was induced by over 50% following exposure to increasing PQQ concentrations (0.01-5 µM) for 24 and 48 h (Figure 8-4A). A corresponding increase in RFC protein expression (over 2-fold) was also observed in PQQ-treated cells compared to vehicle (DMSO) control (Figure 8-4B). To determine whether the upregulation of RFC expression would lead to an enhancement in function, transport

3 assays were performed using [ H]-methotrexate. Methotrexate is an established RFC substrate (Km ~ 1-7 μM) and has been used in numerous studies to assess RFC-mediated transport (Alam et al., 2017; Matherly & Hou, 2008; Rongbao Zhao & Goldman, 2013). As shown in Figure 8-4C, cellular accumulation of [3H]-methotrexate was significantly higher in hCMEC/D3 cells exposed to PQQ (up to 60%) compared to vehicle (DMSO). [3H]-methotrexate uptake in vehicle and PQQ- treated cells was also significantly reduced (up to 70%) in the presence of the RFC inhibitor, pemetrexed, suggesting that transport was specifically mediated by RFC (Figure 8-4C). Together, these results indicate a potential role for NRF-1 and PGC-1α in the regulation of RFC.

102

A RFC (24h mRNA) RFC (48h mRNA)

2.0 2.0 **** **** **** **** **** **** *** 1.5 ** 1.5

1.0 1.0

0.5 0.5 (relative to (relative vehicle) (relative to (relative vehicle) SLC19A1 mRNA Expression mRNA SLC19A1 SLC19A1 mRNA Expression mRNA SLC19A1 0.0 0.0 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5

PQQ Concentration (µM) PQQ Concentration (µM)

B RFC (24h Protein)

300 *** PQQ (!M) ) * Vehicle 1 5 HepG2 200 -actin 75 β RFC 63 (65 kDa) 100 normalized to to normalized

48 RFC Protein Expression

"-actin to (relative DMSO control (42 kDa) 0 0 1 5

PQQ Concentration (µM) C 200 **

150

100 ** ** **** (% (% Control) 50 H]-Methotrexate Uptake 3 [ 0

Vehicle Vehicle 1 µM PQQ5 µM PQQ 1 µM PQQ5 µM PQQ

10 µM PMX

Figure 8-4. Effect of PQQ treatment on RFC functional expression in hCMEC/D3 cells. (A) SLC19A1 (RFC) mRNA was significantly increased following PQQ treatment (0.01-5 µM) for 24 and 48 h compared to vehicle (DMSO) control. (B) Corresponding increase in RFC protein (65 kDa) was observed in hCMEC/D3 cells treated with PQQ (1 and 5 µM) for 24 h. The HepG2 cell line served as positive control while actin was used as a loading control. 50 µg of protein was loaded for treated samples, and positive control was loaded at 10 µg. (C) Cellular uptake of [3H]-methotrexate (50 nM), measured after 1 min at pH 7.4 and 37°C, was significantly higher in cells treated with PQQ (5 µM) for 24 h compared to vehicle (DMSO). [3H]-methotrexate uptake was reduced in the presence of the RFC inhibitor, pemetrexed (PMX; 10 µM). In the absence of PMX, results are expressed as the mean percent change in [3H]-methotrexate accumulation relative to vehicle (DMSO). In the presence of PMX, results are expressed as the mean percent change in [3H]-methotrexate accumulation normalized to the corresponding treatment without PMX. Results are presented as mean ± S.E.M. for n = 3-4 independent experiments. Asterisks represent data points significantly different from vehicle (DMSO) control (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

103

8.4.4 Effect of PQQ Treatment on Efflux Transporters

Members of the ATP-binding cassette (ABC) superfamily of efflux transporters, particularly P-glycoprotein (P-gp; ABCB1), breast cancer resistance protein (BCRP; ABCG2) and multidrug resistance-associated protein (MRP; ABCC), exhibit relatively modest affinities for folates (Km = 0.2-2 mM) and could potentially oppose RFC-mediated folate uptake at the BBB (Assaraf, 2006). To determine whether NRF-1 or PGC-1α can modulate ABC transporter expression, mRNA levels of several efflux transporters were determined in hCMEC/D3 cells following exposure to PQQ or vehicle (DMSO) control. As shown in Figure 8-5, expression of ABCB1, ABCG2, ABCC2, and ABCC3 mRNA were unchanged following in vitro treatments with PQQ (0.01-5 µM) for 24 and 48 h. These findings suggest that NRF-1/PGC-1α activation is not involved in the regulation of these membrane transporters.

A B P-gp (24h mRNA) P-gp (48h mRNA) BCRP (24h mRNA) BCRP (48h mRNA)

2.0 2.0 2.0 2.0

1.5 1.5 1.5 1.5

1.0 1.0 1.0 1.0

0.5 0.5 0.5 0.5 (relative to (relative vehicle) (relative to (relative vehicle) (relative to (relative vehicle) (relative to (relative vehicle) ABCB1 mRNA Expression ABCG2 mRNA Expression ABCG2 mRNA Expression ABCB1 mRNA Expression 0.0 0.0 0.0 0.0 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5

PQQ Concentration (µM) PQQ Concentration (µM) PQQ Concentration (µM) PQQ Concentration (µM)

C MRP2 (24h mRNA) MRP2 (48h mRNA) D MRP3 (24h mRNA) MRP3 (48h mRNA) 2.0 2.0 2.0 2.0

1.5 1.5 1.5 1.5

1.0 1.0 1.0 1.0

0.5 0.5 0.5 0.5 (relative to (relative vehicle) (relative to (relative vehicle) (relative to (relative vehicle) (relative to (relative vehicle) ABCC3 mRNA Expression ABCC3 mRNA Expression ABCC2 mRNA Expression ABCC2 mRNA Expression 0.0 0.0 0.0 0.0 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5 Vehicle 0.01 0.1 1 5 PQQ Concentration (µM) PQQ Concentration (µM) PQQ Concentration (µM) PQQ Concentration (µM)

Figure 8-5. Effect of PQQ treatment on the expression of ABC membrane transporters in hCMEC/D3 cells. There were no changes in ABCB1 (P-gp) (A), ABCG2 (BCRP) (B), ABCC2 (MRP2) (C), and ABCC3 (MRP3) (D) mRNA following PQQ treatment (0.01-5 µM) for 24 and 48 h compared to vehicle (DMSO) control. Results are presented as mean ± SEM for n = 3-4 independent experiments.

104

8.4.5 Downregulation of RFC Expression by NRF-1 and PGC-1α siRNA

To investigate the direct involvement of NRF-1/PGC-1α signaling in the regulation of RFC at the BBB, hCMEC/D3 cells were transfected with NRF-1 or PGC-1α targeting siRNA. Transfection of cells with NRF-1 siRNA resulted in approximately 50% downregulation of NRF- 1 protein expression and a corresponding 50% decrease in RFC protein (Figure 8-6A). Significant decreases in PGC-1α (up to 50%) and RFC (up to 40%) protein were also observed in PGC-1α siRNA-transfected cells compared to scrambled siRNA-treated or untreated control groups (Figure 8-6B). Cellular accumulation of the RFC substrate, [3H]-methotrexate, was significantly decreased (30-40%) following transfection with NRF-1 or PGC-1α siRNA, indicating a decrease in RFC transport activity (Figure 8-6C). We also showed that PQQ treatment (5 µM) for 24 h did not induce RFC protein expression in cells transfected with NRF-1 or PGC-1α siRNA, further demonstrating the role of NRF-1/PGC-1α signaling in modulating RFC functional expression (Figures 8-6D and 8-6E).

105

A siRNA ** ** Untreated 150 Control Control NRF-1 HepG2 # # 75 NRF-1 (68 kDa) 63 100 75 RFC (65 kDa) 63 50 48 Untreated !-actin

Protein Expression Expression Protein Control siRNA

(42 kDa) (% Untreated Control) NRF-1 siRNA 0 NRF-1 RFC

B siRNA C **** **** 125 Untreated Mouse 150 Control Control PGC-1" Kidney #### ## 100 100 PGC-1" ## ### (91 kDa) 100 75 ** *** 75 75 RFC 50 63 (65 kDa) 50 Untreated 25 (% Untreated Control) H]-Methotrexate Uptake Uptake H]-Methotrexate

48 3 Protein Expression Expression Protein Control siRNA !-actin [ (% Untreated Control) (42 kDa) PGC-1α siRNA 0 0 PGC-1α RFC Untreated Control NRF-1 PGC-1α

siRNA D E Control Control NRF-1 Control Control PGC-1! siRNA siRNA siRNA siRNA siRNA siRNA + + + + + + Vehicle PQQ PQQ Vehicle PQQ PQQ 75 **** 100 **** ### 150 #### *** PGC-1! 150 *** NRF-1 #### (91 kDa) #### (68 kDa) 75 100 RFC 100 75 (65 kDa) RFC (65 kDa) 50 50 50 "-actin 50 Protein Expression Expression Protein Protein Expression Expression Protein (% Vehicle Control) !-actin (% Vehicle Control) (42 kDa) 37 (42 kDa) 37 0 0 NRF-1 RFC PGC-1α RFC

Control siRNA + Vehicle Control siRNA + Vehicle Control siRNA + PQQ Control siRNA + PQQ NRF-1 siRNA + PQQ PGC-1α siRNA + PQQ

Figure 8-6. Effect of NRF-1 and PGC-1α knockdown on RFC functional expression in hCMEC/D3 cells. (A, B) Significant reduction in RFC protein expression was observed in cells transfected with NRF- 1 and PGC-1α siRNA compared to untreated or scrambled siRNA-treated controls. (C) Cellular uptake of [3H]-methotrexate (50 nM), measured after 1 min at pH 7.4 and 37°C, was significantly decreased following transfection with NRF-1 or PGC-1α targeting siRNA. (D, E) Treatment with PQQ (5 µM) for 24 h did not result in changes to RFC protein expression in cells transfected with NRF-1 or PGC-1α siRNA compared to vehicle (DMSO) control. Relative levels of NRF-1, PGC-1α, and RFC protein expression were determined by densitometric analysis. The HepG2 cell line and mouse kidney lysates served as positive controls for NRF-1 and PGC-1α expression, respectively. Actin was used as a loading control. 50 µg of protein was loaded for untreated/siRNA-treated samples, while positive controls were loaded at 10 µg. Results are expressed as percent change normalized to untreated/vehicle-treated control and reported as mean ± SEM for n = 3 independent experiments. Asterisks and pound symbols represent data points significantly different from scrambled siRNA-treated and untreated controls, respectively (**, p < 0.01; ***, p < 0.001; ****, p < 0.0001; #, p < 0.05; ##, p < 0.01; ###, p < 0.001; ####, p < 0.0001).

106

8.4.6 Effect of PQQ Treatment on RFC Expression in Mice

The effect of NRF-1/PGC-1α activation on RFC expression was further investigated in vivo using wild type mice. Mice (8-13 weeks old) were treated i.p. with PQQ (10 mg/kg) or vehicle for 10 consecutive days. This dosing regimen was previously shown to exhibit neuroprotective effects in a mouse model of lipopolysaccharide-induced brain inflammation, as well as upregulate the expression of NRF-1 target genes (C. Yang et al., 2014). As shown in Figure 8-7A, body weight of PQQ- and vehicle-treated animals were closely monitored prior to and during treatments to assess potential toxicities. Mice treated with PQQ did not show significant changes in body weight compared to vehicle control. For each treatment group, we also did not see significant differences in body weight over the course of the 10-day treatment period (Figure 8-7A). Furthermore, Slc19a1 (RFC) mRNA levels were approximately 50% higher in isolated brain capillaries and liver of mice treated with PQQ compared to vehicle (Figure 8-7B). A similar induction in Tfam and Tfb2m mRNA was observed in brain capillaries or liver tissues of PQQ- treated animals (Figures 8-8A and 8-8C). Tfb1m expression, on the other hand, was unchanged following PQQ treatment (Figures 8-8B). Together, these findings demonstrate that NRF-1/PGC- 1α activation by PQQ can also induce RFC expression in vivo.

107

A 1.1

1.0

0.9

0.8

(relative Day 0) Day (relative 0.7 Vehicle 10 mg/kg PQQ Ratio of body weight change 0.6

0.5

Day 0 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9 Day 10

B Isolated Brain Capillaries Liver Kidney (RFC mRNA) (RFC mRNA) (RFC mRNA)

2.0 1.5 2.0 Vehicle * 10 mg/kg PQQ *** 1.5 1.5 1.0

1.0 1.0

0.5 0.5 0.5 (relative to vehicle) (relative to vehicle) (relative to vehicle) Slc19a1 mRNA Expression mRNA Slc19a1 Slc19a1 mRNA Expression mRNA Slc19a1 0.0 0.0 Expression mRNA Slc19a1 0.0

Figure 8-7. Effect of in vivo PQQ treatment in wild type mice. (A) Changes in body weight relative to Day 0 (time before vehicle/PQQ injections; initial weight set to 1) was plotted over the course of 10 days. There were no changes in body weight following PQQ treatment (10 mg/kg; daily i.p. for 10 days) compared to vehicle (5% DMSO, 5% Tween-80, 90% saline) control. (B) Slc19a1 (RFC) mRNA expression was increased in isolated brain capillaries and liver of mice treated with PQQ (10 mg/kg; daily i.p. for 10 days) compared to vehicle. Results are presented as mean ± S.E.M. Liver and kidney tissues were isolated from a total of 7-9 animals per treatment group. Data for mouse brain capillaries were obtained from n = 3 independent experiments, where each experiment contained pooled brain tissues from 3 animals per group. Asterisks represent data points significantly different from vehicle (*, p < 0.05; ***, p < 0.001).

108

A Isolated Brain Capillaries Liver Kidney (Tfam mRNA) (Tfam mRNA) (Tfam mRNA)

2.0 Vehicle 1.5 1.5 10 mg/kg PQQ * 1.5 1.0 1.0

1.0

0.5 0.5 0.5 (relative to vehicle) (relative to vehicle) (relative to vehicle) Tfam mRNA Expression Tfam mRNA Expression Tfam mRNA Expression 0.0 0.0 0.0

B Isolated Brain Capillaries Liver Kidney (TFB1M mRNA) (TFB1M mRNA) (TFB1M mRNA)

2.0 Vehicle 1.5 1.5 10 mg/kg PQQ 1.5 1.0 1.0

1.0

0.5 0.5 0.5 (relative to vehicle) (relative to vehicle) (relative to vehicle) Tfb1m mRNA Expression Tfb1m mRNA Expression 0.0 Tfb1m mRNA Expression 0.0 0.0

C Isolated Brain Capillaries Liver Kidney (TFB2M mRNA) (TFB2M mRNA) (TFB2M mRNA)

2.0 1.5 Vehicle 1.5 *** * 10 mg/kg PQQ 1.5 1.0 1.0

1.0

0.5 0.5 0.5 (relative to vehicle) (relative to vehicle) (relative to vehicle) Tfb2m mRNA Expression Tfb2m mRNA Expression 0.0 Tfb2m mRNA Expression 0.0 0.0

Figure 8-8. Effect of in vivo PQQ treatment on expression of NRF-1 target genes in wild type mice. (A) Tfam (Tfam) mRNA expression was increased in isolated brain capillaries following PQQ treatment (10 mg/kg; daily i.p. for 10 days) compared to vehicle (5% DMSO, 5% Tween-80, 90% saline) control. (B) Tfb1m (TFB1M) mRNA levels were unchanged. (C) Tfb2m (TFB2M) mRNA was significantly induced in isolated brain capillaries and liver of mice treated with PQQ (10 mg/kg; daily i.p. for 10 days) compared to vehicle. Results are presented as mean ± S.E.M. Liver and kidney tissues were isolated from a total of 7-9 animals per treatment group. Data for mouse brain capillaries were obtained from n = 3 independent experiments, where each experiment contained pooled brain tissues from 3 animals per group. Asterisks represent data points significantly different from vehicle (*, p < 0.05; ***, p < 0.001).

109

8.5 Discussion

Folates are critical for neurodevelopment and cognitive function. Abnormalities in brain folate transport have been implicated in a number of childhood neurological disorders, including cerebral folate deficiency (Grapp et al., 2012; Vincent T Ramaekers et al., 2005; Steinfeld et al., 2009), hereditary folate malabsorption (Qiu et al., 2006; R Zhao et al., 2017), and autism (V. T. Ramaekers et al., 2007, 2013). These disorders have devastating effects in young children and current therapeutic approaches are not sufficiently effective. In recent years, our group has reported that modulating folate uptake at the BBB could enhance folate permeability into the CNS. Using different BBB model systems, we showed that the functional expression of the folate transporter, RFC, is inducible by the VDR transcription factor (Alam et al., 2017). In vivo activation of VDR through administration of its activating ligand, calcitriol (1,25(OH)2D3), resulted in the upregulation of RFC expression at the BBB and significantly increased folate accumulation in brain tissues of mice (Alam et al., 2019). However, despite these promising findings, we did observe several systemic toxicities associated with calcitriol treatment, such as weight loss and hypercalcemia (Alam et al., 2019). Thus, exploring the role of other transcription factors in the regulation of RFC can present a significant advantage. Furthermore, knowledge of the role of transcription factors in modulating RFC expression and/or function remains limited; hence, additional research is needed.

The aim of the present study was to examine the involvement of another transcription factor, NRF-1, in regulating RFC functional expression at the BBB. A previous report by Gonen and Assaraf suggested that NRF-1 could act as an inducible transcriptional regulator of RFC (Gonen & Assaraf, 2010). Bioinformatic analyses revealed that the SLC19A1 gene promoter contains several binding sites for NRF-1 (Gonen & Assaraf, 2010). Additionally, siRNA-mediated knockdown of NRF-1 resulted in a significant reduction of SLC19A1 mRNA in HeLa cells (Gonen & Assaraf, 2010). In this report, we confirmed robust mRNA expression of NRF1 (NRF-1) and its coactivator, PPARGC1A (PGC-1α), in human (hCMEC/D3 cell line) and rodent (isolated mouse brain capillaries) in vitro systems representative of the BBB (Figure 8-1). The activation of NRF- 1/PGC-1α signaling pathway was then performed through treatment of hCMEC/D3 cells with a quinone compound known as PQQ. PQQ is an enzyme cofactor that has been shown to activate NRF-1 by modulating PGC-1α expression via transcriptional or post-translational mechanisms (Chowanadisai et al., 2010; Saihara et al., 2017). We demonstrated that exposure of hCMEC/D3 110

cells to PQQ did not alter NRF1 expression, but a significant increase in PPARGC1A mRNA (up to 10-fold) was observed in cells treated with 5 µM PQQ at 48 h (Figure 8-2). The mRNA expression of mitochondrial transcription factors and NRF-1 target genes, Tfam, TFB1M and TFB2M, were also induced by approximately 50% following PQQ treatment for 24 or 48 h (Figure 8-3). These results corroborate previous reports that demonstrated upregulation of the same mitochondrial transcription factors following PQQ-mediated activation of the NRF-1/PGC-1α signaling pathway, using mouse liver and fibroblast cell lines (Chowanadisai et al., 2010; Saihara et al., 2017).

Next, we sought to assess the effect of NRF-1/PGC-1α activation on RFC functional expression. Significant increases in RFC mRNA (1.5-fold) and protein (2-fold) levels were observed in hCMEC/D3 cells treated with increasing concentrations of PQQ (0.01-5 µM) (Figures 8-4A and 8-4B). The cellular accumulation of [3H]-methotrexate, an established RFC substrate, was also significantly higher in cells exposed to PQQ compared to vehicle (DMSO) control (Figure 8-4C). These data suggest that NRF-1/PGC-1α signaling may be involved in the regulation of RFC expression and transport activity in an in vitro model of the human BBB.

In the same set of experiments, we also documented the effect of PQQ treatment on the expression of several ABC efflux transporters that are known to transport folates and potentially oppose RFC activity (Assaraf, 2006). We determined that the levels of ABCB1 (P-gp), ABCG2 (BCRP), ABCC2 (MRP2), and ABCC3 (MRP3) mRNA were unchanged following in vitro treatments with PQQ (0.01-5 µM) for 24 and 48 h (Figure 8-5). These findings indicate that NRF- 1/PGC-1α activation is not likely involved in the regulation of these membrane transporters.

Additional studies using NRF-1 or PGC-1α targeting siRNA were conducted to confirm the involvement of these transcription regulators on RFC expression at the BBB. As expected, NRF-1 and PGC-1α protein levels were downregulated following transfection of hCMEC/D3 cells with NRF-1 or PGC-1α targeting siRNA, respectively (Figures 8-6A and 8-6B). Knockdown of NRF-1 or PGC-1α also reduced RFC protein levels by 40-50%, compared to scrambled siRNA- treated or untreated control groups (Figures 8-6A and 8-6B). A corresponding decrease in [3H]- methotrexate accumulation was observed in cells transfected with NRF-1 or PGC-1α siRNA, indicating a decrease in RFC functional activity (Figure 8-6C). Furthermore, we demonstrated that siRNA silencing of NRF-1 or PGC-1α did not induce RFC protein expression following

111

exposure to PQQ (Figures 8-6D and 8-6E), although it should be noted that these studies did not include a corresponding vehicle- and siRNA-treated control group, thereby limiting our ability to interpret the data.

To verify our in vitro findings, the effect of NRF-1/PGC-1α activation on RFC expression was also determined in vivo using wild type mice. Our 10-day in vivo treatment regimen (10 mg/kg PQQ, i.p.) was similar to that of Yang et al., where they showed that this PQQ dose exhibits neuroprotective effects against lipopolysaccharide-induced brain inflammation in mice, and upregulates the expression of NRF-1 target genes (C. Yang et al., 2014). In our studies, we found that PQQ administration did not result in systemic toxicity as evident by the lack of weight loss among PQQ-treated animals (Figure 8-7A). In contrast to our previous study with VDR (Alam et al., 2019), exogenous administration of calcitriol resulted in significant weight loss and hypercalcemia due to VDR-mediated transactivation of renal and intestinal calcium ion channels (E. C. Y. Chow et al., 2013). Most importantly, we determined that Slc19a1 (RFC) mRNA expression was significantly induced in various tissues of mice treated with PQQ, particularly in isolated brain capillaries that are representative of the BBB (Figure 8-7B). Furthermore, a modest but significant increase in the expression of NRF-1 target genes, Tfam and Tfb2m, was observed in mouse brain capillaries after the 10-day PQQ treatment (Figure 8-8). Together, these findings demonstrate that NRF-1/PGC-1α activation by PQQ can also induce RFC expression in vivo, which could potentially facilitate increased folate uptake across the BBB.

To our knowledge, the regulation of RFC by NRF-1 at the level of the BBB has not been addressed. This report demonstrates for the first time that in vitro and in vivo activation of NRF- 1/PGC-1α signaling through PQQ treatment can induce RFC functional expression in various BBB model systems (i.e., hCMEC/D3 cell line and isolated mouse brain capillaries). It should also be noted that PGC-1α is an important transcriptional coactivator that interacts with many transcription factors to regulate cellular processes, including mitochondrial respiration (via NRF-1, NRF-2, and estrogen-related receptor alpha or ERRa), reactive oxygen species defense system (via NRF-2), and fatty acid metabolism (via PPARa, PPARγ, and HNF4a) (Scarpulla, 2008, 2011). Hence, future studies may need to examine whether the inductive effect of PQQ treatment on RFC can occur through other PGC-1α signaling pathways independent of NRF-1.

112

Our current data suggest that upregulating RFC through NRF-1/PGC-1α activation could constitute an alternative approach for enhancing brain folate delivery for the treatment of neurometabolic disorders caused by loss of FRa or PCFT function. The findings of this study may have broader clinical applications due to the growing body of evidence associating mitochondrial dysfunction to neurological disorders. The CNS accounts for approximately 20% of the body’s metabolic demand and neurons depend on oxidative phosphorylation for critical developmental processes (Valenti, de Bari, De Filippis, Henrion-Caude, & Vacca, 2014); therefore, the developing brain is vulnerable to alterations in mitochondrial bioenergetics. Several lines of evidence suggest that mitochondrial dysfunction can contribute to impaired brain folate transport. In Kearns-Sayre Syndrome, vast deletions in mitochondrial DNA have been linked to distinct alterations in choroid plexus morphology (i.e, appearance of enlarged and granular epithelial cells) (Tanji et al., 2000), potentially causing defective uptake of folates into the CNS (Spector & Johanson, 2010). Individuals with Kearns-Sayre Syndrome exhibit neurologic dysfunction resulting from low CSF folate concentrations (Garcia-Cazorla et al., 2008; Serrano et al., 2010; Shemesh & Margolin, 2018), in addition to a number of other symptoms affecting the eye (external ophthalmoplegia and pigmentary retinopathy), abnormalities in cardiac conduction, muscle weakness, and endocrine disorders (Shemesh & Margolin, 2018). Suboptimal CSF folate was additionally documented in autism spectrum disorders, such as Rett syndrome and infantile low- functioning autism (V. T. Ramaekers et al., 2007, 2013). Apart from having high levels of FRα autoantibodies that inhibit FRα transport function (V. T. Ramaekers et al., 2007, 2013), some autistic patients also exhibit impairments in mitochondrial energy production, which could consequently alter brain bioenergetics and neuronal function (Richard E. Frye & Rossignol, 2011). This was evident in frontal cortex biopsies of Rett syndrome patients, which revealed abnormal mitochondria morphology and dysfunction in mitochondrial enzymes, including NADH cytochrome c reductase, succinate cytochrome c reductase, and cytochrome c oxidase (Gibson et al., 2010). Interestingly, some autistic patients also present with a metabolic profile commonly observed in mitochondrial diseases, such as elevated plasma levels of lactate, pyruvate and alanine, as well as abnormal levels of organic acids in the urine (Weissman et al., 2008). Taken together, the activation of NRF-1/PGC-1α signaling by PQQ could potentially represent a novel therapeutic approach for childhood neurological disorders, such as autism or Kearns-Sayre Syndrome, by improving mitochondrial function and enhancing brain folate delivery through RFC.

113

8.6 Acknowledgements

We thank Dr. Richard H. Finnell (Baylor College of Medicine, Houston, TX) for providing animals for the in vivo portion of this work. We also thank Constantine J. Georgiou and Misaki Kondo for their assistance with literature review and transport assays, respectively. This research was supported by an operating grant from the Natural Sciences and Engineering Research Council of Canada awarded to Dr. Reina Bendayan (NSERC 498383). Camille Alam was a recipient of an internal Graduate Fellowship and Centre for Pharmaceutical Oncology Scholarship from the Leslie Dan Faculty of Pharmacy.

114

Chapter 5 Discussion, Limitations, Future Directions, Conclusion

Overall Discussion

Maintaining sufficient levels of folate is important for the development and function of vital tissues, such as the brain. Nutritional folate deficiency was initially implicated in the development of NTDs (Blom et al., 2006), which prompted the mandatory folic acid fortification of grain products in North America and other developed countries. Abnormalities in folate transport have also been linked to childhood neurological disorders, particularly cerebral folate deficiency (Steinfeld et al., 2009), hereditary folate malabsorption (R Zhao et al., 2017), and autism spectrum disorders (R. E. Frye et al., 2013). Folate delivery to the brain primarily occurs at the choroid plexus, through concerted functions of FRα and PCFT (Grapp et al., 2013; R Zhao, Min, et al., 2009). Inactivation of these transport systems can result in very low folate levels in the CSF causing early childhood neurodegeneration. Several lines of evidence indicate that folate supplementation can somewhat reduce the occurrence of neurological disorders associated with folate deficiency; however, current therapeutic approaches are not sufficiently effective (Kronn & Goldman, 2017; Torres et al., 2015; R Zhao et al., 2017).

To date, the mechanisms of brain folate transport have been mainly investigated at the choroid plexus through FRα and PCFT; however, little is known on the functional contribution and regulation of folate transporters (i.e., RFC) in other brain regions, particularly the cerebral vascular endothelium or BBB. Earlier publications have documented the expression or function of folate transporters in human and rodent BBB. Wu and Pardridge demonstrated active transport of 5-methylTHF into isolated human brain capillaries (D. Wu & Pardridge, 1999). Araújo et al. observed a saturable, time-dependent uptake of folic acid and 5-methylTHF in immortalized cultures of rat brain microvessel endothelial (RBE4) cells (Araújo et al., 2010). Several groups also confirmed the presence of folate transporters, PCFT and RFC, in human (Araújo et al., 2010) and rodent (X. Wang et al., 2013; Y Wang et al., 2001) BBB systems. The objective of this PhD thesis was to characterize RFC-mediated folate transport at the BBB, using in vitro primary or immortalized cultures of brain microvessel endothelial cells, or in vivo in mice. We further examined the role of specific transcription factors in the regulation of RFC in the same BBB model

115

systems. Modulating folate uptake at the BBB by augmenting RFC functional expression could potentially increase folate delivery into the CNS, especially when the major route of brain folate transport at the choroid plexus is impaired.

In Chapter 2 of this thesis, we examined the contribution of RFC in folate uptake at the BBB and its potential regulation by the VDR nuclear receptor (Alam et al., 2017). We initially confirmed through qPCR and western blotting that RFC is robustly expressed in various in vitro and ex vivo BBB models (i.e., immortalized human (hCMEC/D3) and rat (RBE4) brain microvessel endothelial cell lines, primary cultures of human brain-derived microvascular endothelial cells (hBMEC), and isolated rat and mouse brain capillaries) (Figure 6-1). These data corroborate previous findings by others in that RFC gene or protein expression has been detected in both human (Araújo et al., 2010) and rodent (X. Wang et al., 2013; Y Wang et al., 2001) BBB. RFC functional activity was subsequently assessed by measuring [3H]-methotrexate accumulation in hCMEC/D3 cells. We observed that in the presence of RFC inhibitors, PT523 and pemetrexed (Yanhua Wang et al., 2004), cellular uptake of [3H]-methotrexate was significantly reduced, suggesting that RFC is functionally active and could potentially mediate folate uptake in this human BBB cell system (Figure 6-2). Furthermore, to investigate the effect of VDR activation on RFC functional expression, hCMEC/D3 cells were exposed to increasing concentrations of the

VDR activating ligand, 1,25-dihydroxyvitamin D3 or calcitriol. Treatment with calcitriol significantly induced RFC mRNA and protein expression as well as function in hCMEC/D3 cells (Figures 6-3 and 6-4). Using the same cell system, we also demonstrated that RFC expression could be downregulated by VDR-targeting siRNA (Figure 6-5), further confirming the role of VDR in the regulation of this folate transporter. These results were most exciting as they demonstrate for the first time that ligand-dependent activation of VDR can upregulate RFC functional expression at the BBB. Finally, to confirm our in vitro findings, calcitriol treatments were performed in isolated mouse brain capillaries, which are considered to be a more robust ex vivo model of the BBB. As expected, RFC mRNA and protein levels were upregulated in isolated mouse brain capillaries following exposure to calcitriol (Figure 6-6). Together, these data strongly suggest that VDR is involved in modulating RFC expression and transport function in human brain microvessel endothelial cells or mouse brain capillaries representative of the BBB.

For Chapter 3 of this thesis, we further investigated the role of VDR in the regulation of RFC in vivo, using Folr1 (FRα) KO mice and corresponding WT controls (Alam et al., 2019). 116

Viable Folr1 KO animals were obtained through dietary supplementation of heterozygous females with folic acid (40 mg/kg) starting two weeks before mating and continued throughout the gestational period (Piedrahita et al., 1999). Systemic deletion of FRα was confirmed through qPCR analysis, which showed a lack of Folr1 mRNA expression in various tissues of Folr1 KO mice (i.e., liver, kidney, and isolated brain capillaries representative of the BBB) (Figure 7-2). We also did not observe significant differences in the level of Slc19a1 (RFC) and Slc46a1 (PCFT) expression between Folr1 KO and WT animals, indicating a lack of compensation from these transporters in response to the loss of Folr1 (Figure 7-2). In vivo activation of VDR was performed through administration of calcitriol to WT and Folr1 KO mice using the dosing regimen specified by Chow et al. (2.5 µg/kg calcitriol, i.p., every other day for 8 days) (E. C. Y. Chow, Sondervan, et al., 2011). We demonstrated that calcitriol treatment significantly increased Slc19a1 (RFC) mRNA expression in brain capillaries of WT and Folr1 KO mice (Figures 7-3A and 7-4A), supporting our in vitro findings from Chapter 2 (Alam et al., 2017). Since VDR is a key regulator of a number of membrane transporters (E. Chow et al., 2010; Durk et al., 2012), we observed that Abcb1a (P-gp) mRNA was also elevated in brain capillaries of WT and Folr1 KO mice following calcitriol treatment (Figures 7-3B and 7-4B). To evaluate the safety of our treatment protocol, we further examined the systemic effects of calcitriol administration. A 10% decrease in body weight was observed in WT and Folr1 KO mice starting at Day 4 of calcitriol injections (Figure 7-5A). Additionally, plasma calcium concentrations were significantly induced in calcitriol-treated animals causing hypercalcemia (Figure 7-5C). This was surprising because the alternate-day dosing regimen was previously shown to alleviate the hypercalcemic and weight loss effects of calcitriol (E. C. Y. Chow, Sondervan, et al., 2011). Hypercalcemia may have been induced by VDR-mediated transactivation of calcium ion channels in the kidney and intestine (i.e., TRPV5 and TRPV6), resulting in volume depletion due to frequent urination (E. C. Y. Chow et al., 2013). There is also evidence demonstrating that high levels of vitamin D or calcium intake can induce apoptosis of mouse adipose tissues via activation of calcium-dependent proteases, such as calpain and caspase-12 (Sergeev & Song, 2014). These results suggest that exogenous administration of calcitriol requires strict monitoring to avoid toxicities associated with VDR activation in vivo.

Folate distribution in WT and Folr1 KO mice was subsequently examined to further understand the effect of VDR activation on RFC functional activity, particularly at the BBB (Alam et al., 2019). We showed that basal concentrations of reduced folates, 5-formylTHF and 5-

117

methylTHF, were much lower in brain and plasma of Folr1 KO mice compared to WT controls (Figure 7-6). These results suggest that loss of FRα adversely affects folate uptake to the brain as well as folate conservation in the body, which primarily occurs via Folr1-mediated reabsorption in the kidneys (Birn et al., 2005). To verify whether VDR activation will induce RFC function in vivo, Folr1 KO mice were subjected to an 8-day pretreatment with calcitriol or vehicle before 13 receiving a single i.v. injection of [ C5]-5-formylTHF. Pretreatment with calcitriol resulted in over 13 6-fold increase of [ C5]-5-formylTHF accumulation in brain tissues of Folr1 KO mice, with levels

13 comparable to WT animals (Figure 7-8B). Brain-to-plasma concentration ratio of [ C5]-5- formylTHF was also 15-fold higher in calcitriol-treated Folr1 KO mice compared to vehicle (Figure 7-8C), indicating a remarkable enhancement in brain folate delivery. These findings provide novel compelling evidence that loss of Folr1 expression results in a substantial decrease in the delivery of folates (i.e., 5-formylTHF or 5-methylTHF) to the brain; however, brain folate transport is restored through activation of VDR with calcitriol. Our data indicate that this is due to the salutary impact of calcitriol on the expression of RFC at the BBB.

The results from Chapter 3 strongly demonstrate that modulating folate transport at the BBB through RFC could provide an alternative route for effective folate delivery to the brain. However, activation of VDR with calcitriol can cause systemic toxicities, such as weight loss and hypercalcemia (Alam et al., 2019). Although these factors did not seem to affect the overall health of calcitriol-treated mice, strict monitoring of exogenous calcitriol administration is required to establish a fine balance between the risks and benefits of VDR activation. Furthermore, since VDR is a key regulator of a wide range of transporters (i.e., P-gp, MRP2-4, OATP1A3) and drug- metabolizing enzymes (i.e., cytochrome P450 enzymes 3A4, 2B6, 2C9; sulfotransferase-2A1), its activation could interfere with normal body functions particularly in the metabolism and elimination of specific drug substrates, therefore causing potential drug-drug interactions and toxicities (E. C. Y. Chow et al., 2009; Durk et al., 2012; Jyrki J Eloranta et al., 2012). Folate homeostasis could also be affected since P-gp and MRPs have been identified as low affinity efflux transporters of folates (Km = 0.2-2 mM) (Assaraf, 2006), which could oppose folate uptake by RFC especially after prolonged treatments with calcitriol (E. C. Y. Chow et al., 2009; E. C. Y. Chow, Durk, et al., 2011; Durk et al., 2012). For these reasons, exploring the role of other transcription factors in the regulation of RFC can present a significant advantage. Furthermore, the regulatory

118

mechanisms governing RFC functional expression at the BBB is not well-documented; hence, additional research is needed.

In Chapter 4 of this thesis, we examined the role of NRF-1 transcription factor in the regulation of RFC at the BBB (Alam & Bendayan, 2019). NRF-1 is primarily known for its role in promoting transcription of genes required for mitochondrial biogenesis and respiratory function. Once activated by PQQ, an enzyme cofactor naturally found in soil and plant-derived food (Chowanadisai et al., 2010; Saihara et al., 2017), NRF-1 can regulate the expression of its target genes by binding to specific DNA recognition sites as a protein homodimer, or through interaction with coactivator proteins such as PGC-1α (Z. Wu et al., 1999). At least four putative NRF-1 binding sites have been identified on the promoter region of the SLC19A1 (RFC) gene (Gonen & Assaraf, 2010), suggesting a potential role for NRF-1 in modulating RFC transcription. In our studies, robust expression of NRF-1 and PGC-1α was detected in human (hCMEC/D3 cells) and rodent (mouse brain capillaries) in vitro models of the BBB (Figure 8-1). Activation of NRF- 1/PGC-1α signaling through treatment with PQQ significantly increased RFC mRNA and protein expression as well as function in hCMEC/D3 cells (Figure 8-4). Interestingly, we also found that the levels of ABCB1 (P-gp), ABCG2 (BCRP), ABCC2 (MRP2), and ABCC3 (MRP3) mRNA were unchanged following in vitro treatments with PQQ (Figure 8-5), suggesting that NRF-1/PGC-1α activation is not involved in the regulation of these efflux transporters. Furthermore, we showed that RFC protein expression and transport activity could be downregulated by NRF-1 or PGC-1α targeting siRNA (Figure 8-6), demonstrating the involvement of NRF-1/PGC-1α signaling in the modulation of RFC functional expression. Most importantly, administration of PQQ (10 mg/kg, i.p., daily for 10 days) to WT mice induced SLC19A1 (RFC) mRNA expression in various tissues of treated animals, particularly in isolated brain capillaries that are representative of the BBB (Figure 8-7B). In contrast to our previous study with VDR (Alam et al., 2019), PQQ administration did not result in systemic toxicity as evident by the lack of weight loss among PQQ- treated animals (Figure 8-7A). Together, these results demonstrate that NRF-1/PGC-1α activation upregulates RFC and could potentially facilitate increased folate uptake across the BBB.

In this research thesis, we identified a potential novel approach for increasing folate delivery into the brain by augmenting RFC functional expression at the BBB through interactions with specific transcription factors (i.e., VDR, NRF-1). To the best of our knowledge, regulation of RFC by VDR or NRF-1 has not been addressed at the level of the BBB. We demonstrated for the 119

first time that in vitro and in vivo activation of these transcription factors can upregulate RFC in several BBB systems (i.e., hCMEC/D3 cell line and isolated mouse brain capillaries). Recent publications on mouse brain transcriptome reported minimal expression of the three major folate transport pathways in brain capillary endothelial cells which constitute the BBB (Vanlandewijck et al., 2018; Zeisel et al., 2018). Data from single-cell RNA sequencing determined that FRα is exclusively expressed in choroid plexus epithelial cells (Zeisel et al., 2018). PCFT was abundantly expressed in brain parenchymal cells, such as astrocytes, microglia, and specific neuronal subsets (Zeisel et al., 2018). Additionally, RFC was highly expressed in pericytes and smooth muscle cells of the mouse cerebral vasculature, with much lower expression in brain capillary endothelial cells (Vanlandewijck et al., 2018). These findings further demonstrate that the presence of RFC at the BBB, although very limited, could constitute an alternative pathway for brain folate delivery. Our results have shown that upregulation of RFC can effectively enhance folate permeability into the brain, especially when loss of FRα function impair the major route of folate uptake at the choroid plexus (Alam et al., 2019). We believe that our research has significantly contributed to the limited body of knowledge on brain folate delivery. Further studies are required to investigate the role of other brain regions (i.e., arachnoid barrier, perivascular spaces) to folate distribution within the CNS. Interestingly, new transcriptome data in mice have confirmed the presence of RFC in leptomeningeal cells of the arachnoid barrier (Zeisel et al., 2018), indicating another potential transport route for folates into the CSF.

Modulating brain folate transport, particularly at the BBB, may have clinical importance due to the lack of established optimal therapy for childhood neurodegenerative disorders caused by inactivation of FRa (i.e., cerebral folate deficiency) or PCFT (i.e., hereditary folate malabsorption) (Grapp et al., 2012; R Zhao et al., 2017). These disorders have devastating effects in young children as they can develop cognitive deficits as early as a few months of birth. To date, the standard treatment approach is to increase folate levels in the CSF through administration of oral or parenteral 5-formylTHF (leucovorin or folinic acid) (Hyland et al., 2010; R Zhao et al., 2017). Despite significant increases in systemic folate concentrations following 5-formylTHF intervention, it is a challenge to achieve sufficient CSF folate levels to relieve the neurological complications (i.e., abnormal brain myelination, psychomotor regression, ataxia, seizures) (Geller et al., 2002; Torres et al., 2015; R Zhao et al., 2017). Folate concentration in the CSF tends to vary with age and is at its highest during infancy and early childhood at ~100-150 nM, then decreasing

120

to ~50-90 nM by the age of six, and >60 nM during puberty (Ormazábal et al., 2011; Perez-Duenas et al., 2011). Thus, our present findings could lead to the development of promising new approaches to enhance folate delivery to the brain via RFC at the BBB.

Beyond abnormalities in brain folate uptake due to loss-of-function mutations in folate transport pathways, low CSF folate has also been identified in other neurological disorders, such as autism spectrum disorders (i.e., Rett syndrome and infantile low-functioning autism (V. T. Ramaekers et al., 2007, 2013) and Kearns-Sayre syndrome (Garcia-Cazorla et al., 2008; Serrano et al., 2010). In autism, reduced brain folate transport was linked to the presence of FRα autoantibodies, which inhibit FRα function either by blocking folate binding to the receptor or by binding to an epitope distant from the folate-binding site and disrupting receptor function (R. E. Frye et al., 2013). Multiple studies have now confirmed that supplementation with 5-formylTHF results in partial recovery of neurological and social impairments of autistic patients (i.e., verbal communication, stereotypical behaviour, attention) (R. E. Frye et al., 2018, 2013; V. T. Ramaekers et al., 2007); however, there is still a need to establish more effective therapies. Prolonged intake of high 5-formylTHF doses could result in adverse events, including insomnia, gastroesophageal reflux, and worsening aggression (R. E. Frye et al., 2013). Since autism is a lifelong disorder, finding an optimal therapeutic approach with limited adverse effects is important. To date, our group has demonstrated that the BBB could present an alternative route for effective brain folate transport by upregulating RFC functional expression through interactions with specific transcription factors (Alam et al., 2019, 2017). These findings could provide a novel treatment strategy for neurometabolic disorders associated with folate deficiency in the CNS.

Limitations

Immortalized cell culture systems (i.e., hCMEC/D3 cells) are widely used as an in vitro model for studying BBB physiology and function. The hCMEC/D3 cell line represents a stable, easily grown, and reproducible population of human brain microvessel endothelial cells (B. Weksler, Romero, & Couraud, 2013). Although this cell line retains many of the unique morphological and biochemical properties of the human brain microvascular endothelium, including functional expression of tight junction proteins (JAM-A, ZO-1, claudin-5), drug metabolizing enzymes (cytochrome P450 enzymes 1A1 and 1B1), transporters (P-gp, BCRP, MRPs, OCTs, Glut-1) and receptors (transferrin and insulin receptors) (Dauchy et al., 2009;

121

Ohtsuki et al., 2012; B. B. Weksler, 2005), it remains an artificial system that may not fully recapitulate in vivo BBB conditions (Helms et al., 2016). Since its development, optimizing hCMEC/D3 barrier integrity has been a challenge. Biemans et al. reported high permeability of small dextran molecules (~4 kDa) across hCMEC/D3 monolayers, which is indicative of paracellular leakage and suboptimal barrier function (Biemans, Jäkel, de Waal, Kuiperij, & Verbeek, 2017). Studies have also shown that under static conditions, hCMEC/D3 cells develop very low transendothelial electrical resistance (TEER) values of 30-50 Ω∙cm2, reflecting high ionic permeability (Eigenmann et al., 2013; B. Weksler et al., 2013). Our laboratory has also measured medium-level TEER values ranging from 150-250 Ω∙cm2 (unpublished work). In contrast, mammalian BBB systems (i.e., rat BBB) are known to exhibit TEER values well above 1000 Ω∙cm2 (Butt, Jones, & Abbott, 1990). For these reasons, the hCMEC/D3 cell line may not fully recapitulate the restrictive barrier function of the human brain endothelium. This cell system may also not represent a robust BBB model when conducting in vitro transport assays, as it could enable paracellular entry of various compounds, including folate. Culturing hCMEC/D3 cells for several days after reaching confluence may increase TEER, but it also results in overgrowth (Urich, Lazic, Molnos, Wells, & Freskgård, 2012). Higher TEER values (~300 Ω∙cm2) have been observed with the addition of hydrocortisone to the culture medium, potentially due to the ability of corticosteroids to modulate the expression of tight junction proteins, such as occludin and claudin- 5 (Förster et al., 2008). Co-culturing with astrocytes and pericytes also showed slight improvements in hCMEC/D3 barrier function (TEER of 60 Ω∙cm2), but it remains nowhere close to in vivo BBB conditions (Hatherell, Couraud, Romero, Weksler, & Pilkington, 2011). To date, the most promising method for increasing barrier tightness is the application of flow-based shear stress to hCMEC/D3 monolayers. This method has reportedly increased TEER values to 1000- 1200 Ω∙cm2, which is more reflective of mammalian BBB physiology and may be a necessary step for improving our in vitro studies (Poller et al., 2008).

Primary cultures of mouse brain microvessel endothelial cells and isolated mouse brain capillaries were also used as in vitro models of the rodent BBB. Coisne et al. developed a well- differentiated primary model of mouse endothelial cells with classic BBB characteristics, such as the expression of P-gp and tight junction proteins (occludin, claudin-3, claudin-5); however, reports from multiple laboratories have indicated suboptimal TEER values of 100-300 Ω∙cm2 for this cell system (Coisne et al., 2005; Helms et al., 2016). The use of isolated mouse brain capillaries

122

may also be limited by contamination with non-endothelial cells originating from the basement membrane of cerebral blood vessels, particularly pericytes, smooth muscle cells, and to a lesser degree, astrocyte foot processes (Helms et al., 2016). Furthermore, primary cultures of mouse BBB cells and brain capillaries both require prior isolation from fresh brain tissues, which often result in low sample yield and could show variability between preparations (Helms et al., 2016).

To mitigate the limitations associated with in vitro BBB cell culture systems, we implemented the use of a conventional Folr1 (FRα) KO mouse model to delineate the role of RFC in brain folate uptake as well as its regulation by transcriptions factors (i.e., VDR). Functional inactivation of the Folr1 gene produces mouse embryos with severe growth retardation and multiple developmental abnormalities, including neural (Piedrahita et al., 1999; Tang & Finnell, 2003; Tang et al., 2005), craniofacial (Spiegelstein et al., 2004; Tang & Finnell, 2003; Tang et al., 2005) and cardiac (H. Zhu, Cabrera, et al., 2007; H. Zhu, Wlodarczyk, et al., 2007) defects, leading to in utero death by gestational day 10. To prevent embryonic lethality of Folr1 KO animals, female heterozygous mice were supplemented with a high dose of dietary folic acid (40mg/kg) starting from two weeks before mating and throughout gestation (Piedrahita et al., 1999). Systemic deletion of Folr1 in vivo enabled us to determine the role of RFC in brain folate delivery while eliminating the contribution of FRα-mediated transport. Interestingly, the Folr1 KO mice did not exhibit similar phenotype as patients with cerebral folate deficiency or hereditary folate malabsorption (i.e., psychomotor regression, ataxia, seizures) (Piedrahita et al., 1999; Spiegelstein et al., 2004; H. Zhu, Wlodarczyk, et al., 2007). Therefore, it was not possible to determine whether the upregulation of RFC functional expression through VDR activation by calcitriol could also alleviate the cognitive and/or behavioural deficits associated with folate deficiency.

Lastly, determination of CSF folate levels in Folr1 KO mice following calcitriol treatment would have further completed the findings of this thesis. As mentioned previously, impairments in folate transport across the choroid plexus following inactivation of FRα or PCFT can lead to very low CSF folate concentrations, ultimately causing cognitive dysfunction due to the inability to sustain normal neural development (Lehtinen et al., 2011; Lehtinen & Walsh, 2011). In our in vivo studies, it was not possible to collect CSF samples from vehicle- or calcitriol-treated mice without blood contamination. Additionally, folate quantification in fluid samples (i.e., whole blood, plasma, or CSF) via LC-MS/MS required at least 275 µl of sample (Christine M. Pfeiffer et al., 2004), which is a much larger CSF volume than what can be collected in a single mouse. 123

For these reasons, we were unable to determine whether calcitriol administration could substantially increase folate concentrations in the CSF of Folr1 KO animals and effectively compensate for the loss of FRα function.

Future Directions 11.1 Transcriptional Regulation of RFC by VDR or NRF-1

Our work has demonstrated that RFC is regulated by transcription factors, VDR and NRF- 1, in various models of the BBB (i.e., immortalized human brain microvessel endothelial cell lines and isolated mouse brain capillaries) (Alam et al., 2019; Alam & Bendayan, 2019; Alam et al., 2017). We showed that ligand-mediated activation of VDR or NRF-1 in vitro and in vivo, resulted in the upregulation of RFC expression and/or transport function (Alam et al., 2019, 2017). However, identifying the exact molecular mechanisms involving RFC regulation by these transcription factors is needed. Our laboratory has conducted preliminary work investigating the direct involvement of VDR in the transcriptional regulation of human SLC19A1 gene (unpublished work). It has been reported that VDR:RXR heterodimers preferentially bind to DR-3 motifs within the promoter region of their target genes (Colnot, Lambert, Blin, Thomasset, & Perret, 1995). Colnot et al. determined the following consensus sequence as the most preferred DNA-binding site for VDR:RXR heterodimers: 5’-AGGTCA(n)3AGGTCA-3’ (Colnot et al., 1995). Using NUBIscan (Podvinec, Kaufmann, Handschin, & Meyer, 2002), a nuclear receptor binding site predictor software, we identified three DR-3-like motifs located between -10.5 to -3.5 kb from the transcriptional start site (TSS; position 0) of the human SLC19A1 promoter (Figure 11-1). As for NRF-1, previous bioinformatic analyses have identified at least four putative NRF-1 binding sites, i.e., (T/C)GCGCA(C/T)GCGC(A/G), on the promoter region of the SLC19A1 gene (Gonen & Assaraf, 2010). Future work will need to document the recruitment and interaction of VDR or NRF-1 to their respective binding sites on the SLC19A1 promoter using chromatin immunoprecipitation (ChIP) and luciferase reporter assays (Hoque et al., 2012; Patel et al., 2015).

124

Consensus 1 Consensus 2 (-10,236/-10,221) (-3565/-3550) SLC19A1

-10,584 Consensus 3 0 3763 (-7381/-7366) (TSS)

Site Sequence (5’à3’) Threshold Score

VDR consensus AGGTCAnnnAGGTCA

Human SLC19A1 AGATCAcatAGTTCA 0.92 (-10,236/-10,221)

Human SLC19A1 AGGCCAggcAGGCCA 0.82 (-3565/-3550) Human SLC19A1 TGTTCAgttAGCTCA 0.77 (-7381/-7366)

Figure 11-1. Putative VDR response elements within the human SLC19A1 (RFC) promoter. Alignment of identified DR-3-like motifs with known VDR DR-3 response element was performed with NUBIscan using a threshold score above 0.75. The numbering of the nucleotides is relative to the transcriptional start site (TSS) of the human SLC19A1 gene.

11.2 Effect of PQQ-mediated NRF-1 Activation on RFC Functional Expression In Vivo

The findings from Chapter 4 present promising new insights on the regulatory role of NRF- 1 on RFC functional expression. In particular, we showed that in vivo administration of PQQ, an enzyme cofactor that activates NRF-1 signaling, significantly induced Slc19a1 (RFC) mRNA expression in isolated mouse brain capillaries that are representative of the BBB (Alam & Bendayan, 2019). These data suggest that augmenting RFC functional expression through NRF-1 could enhance folate delivery to the brain. Future in vivo studies will need to verify whether the upregulation of RFC expression in mouse brain capillaries would also result in increased brain folate delivery. Similar to our previous study with VDR (Alam et al., 2019), wild type mice could undergo a 10-day pretreatment with PQQ (10 mg/kg, i.p., daily for 10 days) or vehicle, before

13 receiving a single i.v. injection of [ C5]-5-formylTHF. Plasma and brain tissue concentrations of 13 [ C5]-5-formylTHF will be measured by LC-MS/MS and compared in animals pretreated with PQQ or vehicle. We anticipate that if an upregulation of RFC functional expression occurs, a

125

13 higher concentration of [ C5]-5-formylTHF will be achieved in the brain of mice treated with PQQ as a result of increased brain folate uptake. These experiments will allow us to delineate whether the induction of RFC through NRF-1 activation will have a significant effect on folate accumulation in the brain. This work will also add to the limited body of knowledge regarding folate transporter (i.e., RFC) regulation at the BBB.

Furthermore, our in vivo data only demonstrated modest increases (40-50%) in the mRNA expression of RFC and NRF-1 target genes (i.e., Tfam, TFB2M) following PQQ treatment. This could be attributed to the fact that animal sacrifices and tissue collection were performed 24 hours after the last PQQ dose. Since the effect of NRF-1 activation on RFC expression presumably occurs at the transcriptional level, collecting tissues at much earlier time points (i.e., 3-6 hours post-injection) would potentially generate a more robust response. Alternatively, the PQQ dose could be increased, since several groups have reported a wide dosing range for PQQ in mice (up to 20 mg/kg) without observing toxic effects (Kumar & Kar, 2015; C. Yang et al., 2014; Zhou, Chen, Hu, Mao, & Kong, 2014). Lastly, there is very little data demonstrating PQQ pharmacokinetics in rodent models (i.e., volume of distribution, clearance, half-life) (Smidt, Unkefer, Houck, & Rucker, 1991); hence, a full pharmacokinetic study of PQQ plasma and tissue distribution should be performed.

11.3 Clinical Significance of NRF-1 Activation for the Treatment of Childhood Neurological Disorders

Modulating RFC functional expression at the BBB through NRF-1 may have broader clinical applications due to the growing body of evidence associating mitochondrial diseases to neurological disorders. NRF-1 is primarily known for its role in promoting the transcription of genes required for mitochondrial biogenesis and respiratory function; therefore, activation of NRF- 1 signaling could present significant therapeutic benefits (Chowanadisai et al., 2010; Saihara et al., 2017). In the context of autism spectrum disorders, apart from having high levels of systemic FRα autoantibodies that inhibit FRα transport and cause suboptimal CSF folate concentration (V. T. Ramaekers et al., 2007, 2013), autistic patients also exhibit impairments in mitochondrial energy production, which alter brain bioenergetics and neuronal function (Richard E. Frye & Rossignol, 2011). This was evident in frontal cortex biopsies of Rett syndrome patients, which revealed abnormal mitochondria morphology as well as dysfunction in several mitochondrial enzymes (i.e., 126

NADH cytochrome c reductase, succinate cytochrome c reductase, cytochrome c oxidase) (Gibson et al., 2010). Weissman et al. further reported that some autistic patients present with metabolic profiles commonly observed in mitochondrial diseases, including elevated plasma levels of lactate, pyruvate and alanine, as well as abnormal levels of organic acids in the urine (Weissman et al., 2008).

Implementing an in vivo rodent model of autism could be helpful in investigating the clinical benefits of NRF-1 activation, in terms of improving mitochondrial function and enhancing brain folate uptake in animals that exhibit an autistic phenotype. In rats, Desai et al. demonstrated that exposure to FRα autoantibodies during gestation and preweaning (i.e., postnatal days 10-12) resulted in cognitive and behavioural deficits that closely mimic autism symptoms, including decreased social interaction and communication as well as impairments in memory and learning (Desai et al., 2017; Sequeira et al., 2016). Interestingly, they also provided evidence that supplementation with 5-formylTHF or folinic acid was able to attenuate the observed impairments (Desai et al., 2017), suggesting that increasing brain folate permeability during neurodevelopment could be helpful in addressing autism-related behaviour. Other rodent models of autism are also available, particularly Mecp2-deficient rats or mice, which exhibit similar phenotypic and behavioural characteristics as patients with Rett syndrome. Mecp2−/Y is a rat model of Rett syndrome that shows growth retardation, breathing impairments or apnea, and severe deficits in motor function, locomotion, and social interaction (Y. Wu et al., 2016). Various types of Mecp2 null mice (i.e., Mecp2tm1.1Bird, Mecp2neoTam, Mecp2-/y) also present comparable phenotypic manifestations as human Rett patients, such as autistic behaviour, motor problems, breathing disorders, and cardiac arrythmias (Pelka et al., 2006; Viemari et al., 2005). Implementation of any of these autism rodent models in our laboratory will permit us to delineate whether in vivo activation of NRF-1 by PQQ could also provide valuable insight to the treatment of another debilitating and highly prevalent childhood neurological disorder.

127

Conclusion

Folates are critical for proper brain development and function. Abnormalities in brain folate transport (i.e., loss-of-function mutations in FRα and PCFT) have been implicated in a number of childhood neurodevelopmental disorders, including NTDs, cerebral folate deficiency, and hereditary folate malabsorption. These disorders have devastating effects in young children and current therapeutic approaches are not sufficiently effective. Our work has demonstrated that the vascular BBB could present an alternative route for brain folate delivery, especially when inactivation of FRα or PCFT impairs the major route of folate uptake at the choroid plexus. We showed in vitro that functional expression of another folate transporter, RFC, and its upregulation by the VDR transcription factor could increase folate transport across the BBB. In vivo treatment with the VDR activating ligand, calcitriol, also enhanced RFC expression at the BBB and effectively restored folate delivery to the brain of Folr1 (FRα) knockout mice. Additionally, we demonstrated for the first time the role of another transcription factor, NRF-1, in the upregulation of RFC in several in vitro and in vivo BBB model systems. These findings suggest that augmenting RFC functional expression through interactions with specific transcription factors could effectively compensate for impaired folate uptake at the choroid plexus. Future studies are needed to determine whether modulating folate uptake at the BBB through RFC could provide new treatment strategies for other debilitating childhood neurological disorders, such as autism spectrum disorders.

128

References

Abe, T., Unno, M., Onogawa, T., Tokui, T., Noriko Kondo, T., Nakagomi, R., … Matsuno, S. (2001). LST-2, a human liver-specific organic anion transporter, determines methotrexate sensitivity in gastrointestinal cancers. Gastroenterology, 120(7), 1689–1699.

Al-Baradie, R. S., & Chaudhary, M. W. (2014). Diagnosis and management of cerebral folate deficiency. A form of folinic acid-responsive seizures. Neurosciences (Riyadh), 19(4), 312– 316.

Alam, C., Aufreiter, S. A., Georgiou, C. J., Hoque, M. T., Finnell, R. H., O’Connor, D. L., … Bendayan, R. (2019). Upregulation of reduced folate carrier by vitamin D enhances brain folate uptake in mice lacking folate receptor alpha. Proc Natl Acad Sci USA, 116(35), 17531– 17540.

Alam, C., & Bendayan, R. (2019). Nuclear respiratory factor 1 (NRF-1) upregulates the expression and function of reduced folate carrier (RFC) at the blood-brain barrier [unpublished manuscript].

Alam, C., Hoque, M. T., Finnell, R. H., Goldman, I. D., & Bendayan, R. (2017). Regulation of reduced folate carrier (RFC) by vitamin D receptor at the blood-brain barrier. Mol Pharm, 14(11), 3848–3858.

Allen, L. H. (2008). Causes of vitamin B12 and folate deficiency. Food Nutr Bull, 29(2), S20-34.

Allen, R. J., DiMauro, S., Coulter, D. L., Papadimitriou, A., & Rothenberg, S. P. (1983). Kearns- Sayre syndrome with reduced plasma and cerebrospinal fluid folate. Ann Neurol, 13(6), 679– 682.

Aluri, S., Zhao, R., Lubout, C., Goorden, S. M. I., Fiser, A., & Goldman, I. D. (2018). Hereditary folate malabsorption due to a mutation in the external gate of the proton-coupled folate transporter SLC46A1. Blood Adv, 2(1), 61–68.

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th Ed). Washington, DC: American Psychiatric Association Publishing.

129

Anderson, P. H., O’ Loughlin, P. D., May, B. K., & Morris, H. A. (2002). Quantification of mRNA for the vitamin D metabolizing enzymes CYP27B1 and CYP24 and vitamin D receptor in kidney using real-time reverse transcriptase- polymerase chain reaction. J Mol Endocrinol, 31(1), 123–132.

Antony, A. C. (1996). Folate receptors. Annu Rev Nutr, 16, 501–521.

Araújo, J. R., Gonçalves, P., & Martel, F. (2010). Characterization of uptake of folates by rat and human blood-brain barrier endothelial cells. BioFactors, 36(3), 201–209.

Arth, A., Tinker, S., Moore, C., Canfield, M., Agopian, A., Reefhuis, J., & Centers for Disease Control and Prevention. (2015). Supplement use and other characteristics among pregnant women with a previous pregnancy affected by a neural tube defect - United States, 1997- 2009. MMWR Morb Mortal Wkly Rep, 64(1), 6–9.

Ashraf, T., Kao, A., & Bendayan, R. (2014). Functional expression of drug transporters in glial cells: potential role on drug delivery to the CNS. In Adv Pharmacol (Vol. 71, pp. 45–111).

Assaraf, Y. G. (2006). The role of multidrug resistance efflux transporters in antifolate resistance and folate homeostasis. Drug Resist Updat, 9(4–5), 227–246.

Aufreiter, S., Gregory, J. F., Pfeiffer, C. M., Fazili, Z., Kim, Y. I., Marcon, N., … O’Connor, D. L. (2009). Folate is absorbed across the colon of adults: evidence from cecal infusion of 13C- labeled [6S]-5-formyltetrahydrofolic acid. Am J Clin Nutr, 90(1), 116–123.

Badagnani, I., Castro, R., Taylor, T., Brett, C., Huang, C., Stryke, D., … Giacomini, K. (2006). Interaction of methotrexate with organic-anion transporting polypeptide 1A2 and its genetic variants. J Pharmacol Exp Ther, 318(2), 521–529.

Badovinac, R. L., Werler, M. M., Williams, P. L., Kelsey, K. T., & Hayes, C. (2007). Folic acid- containing supplement consumption during pregnancy and risk for oral clefts: a meta- analysis. Birth Defects Res A Clin Mol Teratol, 79(1), 8–15.

Bailey, L. B., Rampersaud, G. C., & Kauwell, G. P. A. (2003). Folic acid supplements and fortification affect the risk for neural tube defects, vascular disease and cancer: evolving science. J Nutr, 133(6), 1961S–1968S. 130

Bailey, R. L., Mills, J. L., Yetley, E. A., Gahche, J. J., Pfeiffer, C. M., Dwyer, J. T., … Picciano, M. F. (2012). Serum unmetabolized folic acid in a nationally representative sample of adults ≥60 years in the United States, 2001–2002. Food Nutr Res, 56, 10–12.

Bailey, R. L., West, K. P., & Black, R. E. (2015). The epidemiology of global micronutrient deficiencies. Ann Nutr Metab, 66(Suppl 2), 22–33.

Benoit, R., Cooney, A., Giguere, V., Ingraham, H., Lazar, M., & Muscat, G. (2006). International Union of Pharmacology. LXVI. Orphan Nuclear Receptors. Pharmacol Rev, 58(4), 798–836.

Bianchi, E., Doe, B., Goulding, D., & Wright, G. J. (2014). Juno is the egg Izumo receptor and is essential for mammalian fertilization. Nature, 508(7497), 483–487.

Biemans, E. A. L. M., Jäkel, L., de Waal, R. M. W., Kuiperij, H. B., & Verbeek, M. M. (2017). Limitations of the hCMEC/D3 cell line as a model for Aβ clearance by the human blood- brain barrier. J Neurosci Res, 95(7), 1513–1522.

Birn, H., Spiegelstein, O., Christensen, E. I., & Finnell, R. H. (2005). Renal tubular reabsorption of folate mediated by folate binding protein 1. J Am Soc Nephrol, 16(3), 608–615.

Blom, H. J., Shaw, G. M., Den Heijer, M., & Finnell, R. H. (2006). Neural tube defects and folate: case far from closed. Nat Rev Neurosci, 7(9), 724–731.

Borst, P., Evers, R., Kool, M., & Wijnholds, J. (2000). A family of drug transporters: the multidrug resistance-associated proteins. J Natl Cancer Inst, 92(16), 1295–1302.

Botto, L. D., & Yang, Q. (2000). 5,10-Methylenetetrahydrofolate reductase gene variants and congenital anomalies: a HuGE review. Am J Epidemiol, 151(9), 862–877.

Bozard, B. R., Ganapathy, P. S., Duplantier, J., Mysona, B., Ha, Y., Roon, P., … Smith, S. B. (2010). Molecular and biochemical characterization of folate transport proteins in retinal Müller cells. Invest Ophthalmol Vis Sci, 51(6), 3226–3235.

Bruni, F., Polosa, P. L., Gadaleta, M. N., Cantatore, P., & Roberti, M. (2010). Nuclear respiratory factor 2 induces the expression of many but not all human proteins acting in mitochondrial DNA transcription and replication. J Biol Chem, 285(6), 3939–3948.

131

Butt, B. Y. A. M., Jones, H. C., & Abbott, N. J. (1990). Electrical resistance across the blood-brain barrier in anaesthetized rats: a developmental study. J Physiol, 429, 47–62.

Cao, W., & Matherly, L. H. (2004). Analysis of the membrane topology for transmembrane domains 7-12 of the human reduced folate carrier by scanning cysteine accessibility methods. Biochem J, 378(Pt 1), 201–206.

Cario, H., Bode, H., Debatin, K.-M., Opladen, T., & Schwarz, K. (2009). Congenital null mutations of the FOLR1 gene: a progressive neurologic disease and its treatment. Neurology, 73(24), 2127–2129.

Cha, S. H., Sekine, T., Fukushima, J. I., Kanai, Y., Kobayashi, Y., Goya, T., & Endou, H. (2001). Identification and characterization of human organic anion transporter 3 expressing predominantly in the kidney. Mol Pharmacol, 59(5), 1277–1286.

Chan, G. N. Y., & Cannon, R. E. (2017). Assessment of ex vivo transport function in isolated rodent brain capillaries. Curr Protoc Pharmacol., 76, 7.16.1-7.16.16.

Chancy, C. D., Kekuda, R., Huang, W., Prasad, P. D., Kuhnel, J. M., Sirotnak, F. M., … Smith, S. B. (2000). Expression and differential polarization of the reduced-folate transporter-1 and the folate receptor α in mammalian retinal pigment epithelium. J Biol Chem, 275(27), 20676– 20684.

Chen, C., Ke, J., Zhou, X. E., Yi, W., Brunzelle, J. S., Li, J., … Melcher, K. (2013). Structural basis for molecular recognition of folic acid by folate receptors. Nature, 500(7463), 486–489.

Chen, C., Scott, D., Hanson, E., Franco, J., Berryman, E., Volberg, M., & Liu, X. (2003). Impact of Mrp2 on the biliary excretion and intestinal absorption of furosemide, probenecid, and methotrexate using Eisai hyperbilirubinemic rats. Pharm Res, 20(1), 31–37.

Chen, Z. S., Lee, K., Walther, S., Raftogianis, R. B., Kuwano, M., Zeng, H., & Kruh, G. D. (2002). Analysis of methotrexate and folate transport by multidrug resistance protein 4 (ABCC4): MRP4 is a component of the methotrexate efflux system. Cancer Res, 62(11), 3144–3150.

Chow, E. C. Y., Durk, M. R., Cummins, C. L., & Pang, K. S. (2011). 1Alpha,25-dihydroxyvitamin D3 up-regulates P-glycoprotein via the vitamin D receptor and not farnesoid X receptor in 132

both fxr(-/-) and fxr(+/+) mice and increased renal and brain efflux of digoxin in mice in vivo. J Pharmacol Exp Ther, 337(3), 846–859.

Chow, E. C. Y., Maeng, H.-J., Liu, S., Khan, A. A., Groothuis, G. M. M., & Pang, K. S. (2009). 1α,25-dihydroxyvitamin D3 triggered vitamin D receptor and Farnesoid X Receptor-like effects in rat intestine, liver, and kidney in vivo. Biopharm Drug Dispos, 30, 457–475.

Chow, E. C. Y., Quach, H. P., Vieth, R., & Pang, K. S. (2013). Temporal changes in tissue 1α,25- dihydroxyvitamin D3, vitamin D receptor target genes, and calcium and PTH levels after 1,25(OH)2D3 treatment in mice. Am J Physiol Endocrinol Metab, 304(9), E977–E989.

Chow, E. C. Y., Sondervan, M., Jin, C., Groothuis, G. M. M., & Pang, K. S. (2011). Comparative effects of doxercalciferol (1α-hydroxyvitamin D2) versus calcitriol (1α,25-dihydroxyvitamin D3) on the expression of transporters and enzymes in the rat in vivo. J Pharm Sci, 100(4), 1594–1604.

Chow, E., Sun, H., Khan, A., Groothuis, G., & Pang, K. (2010). Effects of 1α,25-dihydroxyvitamin D3 on transporters and enzymes of the rat intestine and kidney in vivo. Biopharm Drug Dispos, 31(1), 91–108.

Chowanadisai, W., Bauerly, K. A., Tchaparian, E., Wong, A., Cortopassi, G. A., & Rucker, R. B. (2010). Pyrroloquinoline quinone stimulates mitochondrial biogenesis through cAMP response element-binding protein phosphorylation and increased PGC-1a expression. J Biol Chem, 285(1), 142–152.

Coisne, C., Dehouck, L., Faveeuw, C., Delplace, Y., Miller, F., Landry, C., … Dehouck, B. (2005). Mouse syngenic in vitro blood-brain barrier model: a new tool to examine inflammatory events in cerebral endothelium. Lab Invest, 85(6), 734–746.

Colapinto, C. K., O’Connor, D. L., & Tremblay, M. S. (2011). Folate status of the population in the Canadian Health Measures Survey. CMAJ, 183(2), E100-6.

Colnot, S., Lambert, M., Blin, C., Thomasset, M., & Perret, C. (1995). Identification of DNA sequences that bind retinoid X receptor-1,25(OH)2D3-receptor heterodimers with high affinity. Mol Cell Endocrinol, 113(1), 89–98.

133

Cooray, H. C., Blackmore, C. G., Maskell, L., & Barrand, M. A. (2002). Localisation of breast cancer resistance protein in microvessel endothelium of human brain. Neuroreport, 13(16), 2059–2063.

Cravo, M. L., Gloria, L. M., Selhub, J., Nadeau, M. R., Camilo, M. E., Resende, M. P., … Mira, F. C. (1996). Hyperhomocysteinemia in chronic alcoholism: correlation with folate, vitamin B-12, and vitamin B-6 status. Am J Clin Nutr, 63(2), 220–224.

Crider, K. S., Zhu, J., Hao, L., Yang, Q., Yang, T. P., Gindler, J., … Berry, R. J. (2011). MTHFR 677C->T genotype is associated with folate and homocysteine concentrations in a large, population-based, double-blind trial of folic acid supplementation. Am J Clin Nutr, 93(6), 1365–1372.

Czeizel, A. E. (1993). Prevention of congenital abnormalities by periconceptional multivitamin supplementation. BMJ, 306(6893), 1645–1648.

Czeizel, A. E., Tímár, L., & Sárközi, A. (1999). Dose-dependent effect of folic acid on the prevention of orofacial clefts. Pediatrics, 104(6), e66.

Dauchy, S., Miller, F., Couraud, P.-O., Weaver, R. J., Weksler, B., Romero, I.-A., … Declèves, X. (2009). Expression and transcriptional regulation of ABC transporters and cytochromes P450 in hCMEC/D3 human cerebral microvascular endothelial cells. Biochem Pharmacol, 77(5), 897–909.

De La Vega, M. J., Santolaria, F., González-Reimers, E., Alemán, M. R., Milena, A., Martínez- Riera, A., & González-García, C. (2001). High prevalence of hyperhomocysteinemia in chronic alcoholism: the importance of the thermolabile form of the enzyme methylenetetrahydrofolate reductase (MTHFR). Alcohol, 25(2), 59–67.

Dean, M. (2002). The Human ATP-Binding Cassette (ABC) Transporter Superfamily. Bethesda: National Center for Biotechnology Information (US).

Dean, M., Rzhetsky, A., & Allikmets, R. (2001). The human ATP-binding cassette (ABC) transporter superfamily. Genome Res, 11(7), 1156–1166.

Desai, A., Sequeira, J. M., & Quadros, E. V. (2017). Prevention of behavioral deficits in rats 134

exposed to folate receptor antibodies: implication in autism. Mol Psychiatry, 22(9), 1291– 1297.

Dev, S., Ahmad Wani, N., & Kaur, J. (2011). Regulatory mechanisms of intestinal folate uptake in a rat model of folate oversupplementation. Br J Nutr, 105(6), 827–835.

Devilbiss, E. A., Gardner, R. M., Newschaffer, C. J., & Lee, B. K. (2015). Maternal folate status as a risk factor for autism spectrum disorders : a review of existing evidence. Br J Nutr, 114(5), 663–672.

Dougados, M., Zittoun, J., Laplane, D., & Castaigne, P. (1983). Folate metabolism disorder in Kearns-Sayre syndrome. Ann Neurol, 13(6), 687.

Durk, M. R., Chan, G. N. Y., Campos, C. R., Peart, J. C., Chow, E. C. Y., Lee, E., … Pang, K. S. (2012). 1α,25-Dihydroxyvitamin D3-liganded vitamin D receptor increases expression and transport activity of P-glycoprotein in isolated rat brain capillaries and human and rat brain microvessel endothelial cells. J Neurochem, 123(6), 944–953.

Durk, M. R., Han, K., Chow, E. C. Y., Ahrens, R., Henderson, J. T., Fraser, P. E., & Pang, K. S. (2014). 1α,25-Dihydroxyvitamin D3 reduces cerebral amyloid-β accumulation and improves cognition in mouse models of Alzheimer’s disease. J Neurosci, 34(21), 7091–7101.

Eichner, E. R., & Hillman, R. S. (1973). Effect of alcohol on serum folate level. J Clin Invest, 52(3), 584–591.

Eigenmann, D. E., Xue, G., Kim, K. S., Moses, A. V., Hamburger, M., & Oufir, M. (2013). Comparative study of four immortalized human brain capillary endothelial cell lines, hCMEC/D3, hBMEC, TY10, and BB19, and optimization of culture conditions, for an in vitro blood-brain barrier model for drug permeability studies. Fluids Barriers CNS, 10(1), 33.

Elnakat, H., & Ratnam, M. (2004). Distribution, functionality and gene regulation of folate receptor isoforms: Implications in targeted therapy. Adv Drug Deliv Rev, 56(8), 1067–1084.

Eloranta, J. J., Hiller, C., Jüttner, M., & Kullak-Ublick, G. a. (2012). The SLCO1A2 gene, encoding human organic anion-transporting polypeptide 1A2, is transactivated by the vitamin D receptor. Mol Pharmacol, 82(1), 37–46. 135

Eloranta, J. J., Zaır, Z. M., Hiller, C., Hausler, S., Stieger, B., & Kullak-Ublick, G. A. (2009). Vitamin D3 and its nuclear receptor increase the expression and activity of the human proton- coupled folate transporter. Mol Pharmacol, 76(5), 1062–1071.

Ferguson, P. L., & Flintoff, W. F. (1999). Topological and functional analysis of the human reduced folate carrier by hemagglutinin epitope insertion. J Biol Chem, 274(23), 16269– 16278.

Fernandez-Zapico, M., Lomberk, G., Tsuji, S., DeMars, C., Bardsley, M., Lin, Y., … Urrutia, R. (2011). A functional family-wide screening of SP/KLF proteins identifies a subset of suppressors of KRAS-mediated cell growth. Biochem J, 435(2), 529–537.

Förster, C., Burek, M., Romero, I. A., Weksler, B., Couraud, P. O., & Drenckhahn, D. (2008). Differential effects of hydrocortisone and TNFα on tight junction proteins in an in vitro model of the human blood-brain barrier. J Physiol, 586(7), 1937–1949.

Frosst, P., Blom, H. J., Milos, R., Goyette, P., Sheppard, C. A., Matthews, R. G., … Al., E. (1995). A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet, 10, 111–113.

Frye, R. E., Delhey, L., Slattery, J., Tippett, M., Wynne, R., Rose, S., … Quadros, E. (2016). Blocking and binding folate receptor alpha autoantibodies identify novel autism spectrum disorder subgroups. Front Neurosci, 10, 1–10.

Frye, R. E., & Rossignol, D. A. (2011). Mitochondrial dysfunction can connect the diverse medical symptoms associated with autism spectrum disorders. Pediatr Res, 69(5 Pt 2), 41R–47R.

Frye, R. E., Sequeira, J. M., Quadros, E. V., James, S. J., & Rossignol, D. A. (2013). Cerebral folate receptor autoantibodies in autism spectrum disorder. Mol Psychiatry, 18(3), 369–381.

Frye, R. E., Slattery, J., Delhey, L., Furgerson, B., Strickland, T., Tippett, M., … Quadros, E. V. (2018). Folinic acid improves verbal communication in children with autism and language impairment: a randomized double-blind placebo-controlled trial. Mol Psychiatry, 23(2), 247– 256.

Furumiya, M., Inoue, K., Ohta, K., Hayashi, Y., & Yuasa, H. (2013). Transcriptional regulation of 136

PCFT by KLF4, HNF4α, CDX2 and C/EBPα: implication in its site-specific expression in the small intestine. Biochem Biophys Res Commun, 431(2), 158–163.

Galvão, T. C., & Thomas, J. O. (2005). Structure-specific binding of MeCP2 to four-way junction DNA through its methyl CpG-binding domain. Nucleic Acids Res, 33(20), 6603–6609.

Garcia-Cazorla, A., Quadros, E. V., Nascimento, A., Garcia-Silva, M. T., Briones, P., Montoya, J., … Ramaekers, V. T. (2008). Mitochondrial diseases associated with cerebral folate deficiency. Neurology, 70, 1360–1362.

Geller, J., Kronn, D., Jayabose, S., & Sandoval, C. (2002). Hereditary folate malabsorption: family report and review of the literature. Medicine (Baltimore), 81(1), 51–68.

Gibson, J., Slobedman, B., Williamson, S., Minchenko, D., El-Osta, A., Stern, J., & Christodoulou, J. (2010). Downstream targets of methyl CpG binding protein 2 and their abnormal expression in the frontal cortex of the human Rett syndrome brain. BMC Neurosci, 11, 53.

Gnana-Prakasam, J. P., Reddy, S. K., Veeranan-Karmegam, R., Smith, S. B., Martin, P. M., & Ganapathy, V. (2011). Polarized distribution of heme transporters in retinal pigment epithelium and their regulation in the iron-overload disease hemochromatosis. Invest Ophthalmol Vis Sci, 52(12), 9279–9286.

Goldman, I., Lichtenstein, N., & Oliverio, V. (1968). Carrier-mediated transport of the folic acid analogue, methotrexate, in the L1210 leukemia cell. J Biol Chem, 243(19), 5007–5017.

Gonen, N., & Assaraf, Y. G. (2010). The obligatory intestinal folate transporter PCFT (SLC46A1) is regulated by nuclear respiratory factor 1. J Biol Chem, 285(44), 33602–33613.

Gonen, N., & Assaraf, Y. G. (2012). Antifolates in cancer therapy: structure, activity and mechanisms of drug resistance. Drug Resist Updat, 15(4), 183–210.

Goyette, P., Sumner, J. S., Milos, R., Duncan, A. M., Rosenblatt, D. S., Matthews, R. G., & Rozen, R. (1994). Human methylenetetrahydrofolate reductase: isolation of cDNA, mapping and mutation identification. Nat Genet, 7, 195–200.

Grapp, M., Just, I. A., Linnankivi, T., Wolf, P., Lücke, T., Häusler, M., … Steinfeld, R. (2012).

137

Molecular characterization of folate receptor 1 mutations delineates cerebral folate transport deficiency. Brain, 135(7), 2022–2031.

Grapp, M., Wrede, A., Schweizer, M., Hüwel, S., Galla, H.-J., Snaidero, N., … Steinfeld, R. (2013). Choroid plexus transcytosis and exosome shuttling deliver folate into brain parenchyma. Nat Commun, 4, 2123.

Håberg, S. E., London, S. J., Stigum, H., Nafstad, P., & Nystad, W. (2009). Folic acid supplements in pregnancy and early childhood respiratory health. Arch Dis Child, 94(3), 180–184.

Hagberg, B., Goutières, F., Hanefeld, F., Rett, A., & Wilson, J. (1985). Rett syndrome: criteria for inclusion and exclusion. Brain Dev, 7(3), 372–373.

Halwachs, S., Lakoma, C., Gebhardt, R., Schafer, I., Seibel, P., & Honscha, W. (2010). Dioxin mediates downregulation of the reduced folate carrier transport activity via the arylhydrocarbon receptor signalling pathway. Toxicol Appl Pharmacol, 246(1–2), 100–106.

Hamid, A., Kiran, M., Rana, S., & Kaur, J. (2009). Low folate transport across intestinal basolateral surface is associated with down-regulation of reduced folate carrier in in vivo model of folate malabsorption. IUBMB Life, 61(3), 236–243.

Hao, H., Qi, H., & Ratnam, M. (2003). Modulation of the folate receptor type beta gene by coordinate actions of retinoic acid receptors at activator Sp1/ets and repressor AP-1 sites. Blood, 101(11), 4551–4560.

Harrington, J. W., & Allen, K. (2014). The clinician’s guide to autism. Pediat Rev, 35(2), 62–78.

Hatherell, K., Couraud, P. O., Romero, I. A., Weksler, B., & Pilkington, G. J. (2011). Development of a three-dimensional, all-human in vitro model of the blood-brain barrier using mono-, co- , and tri-cultivation Transwell models. J Neurosci Meth, 199(2), 223–229.

He, L., Vasiliou, K., & Nebert, D. W. (2009). Analysis and update of the human solute carrier (SLC) gene superfamily. Hum Genomics, 3(2), 195–206.

Health Canada. (2007). Eating well with Canada’s food guide. Ottawa, Ontario.

138

Helms, H. C., Abbott, N. J., Burek, M., Cecchelli, R., Couraud, P. O., Deli, M. A., … Brodin, B. (2016). In vitro models of the blood-brain barrier: an overview of commonly used brain endothelial cell culture models and guidelines for their use. J Cereb Blood Flow Metab, 36(5), 862–890.

Ho, R. H., & Kim, R. B. (2005). Transporters and drug therapy: implications for drug disposition and disease. Clin Pharmacol Ther, 78(3), 260–277.

Hollenstein, K., Frei, D. C., & Locher, K. P. (2007). Structure of an ABC transporter in complex with its binding protein. Nature, 446(7132), 213–216.

Hooijberg, J. H., Peters, G. J., Assaraf, Y. G., Kathmann, I., Priest, D. G., Bunni, M. A., … Jansen, G. (2003). The role of multidrug resistance proteins MRP1, MRP2 and MRP3 in cellular folate homeostasis. Biochem Pharmacol, 65(5), 765–771.

Hoque, M. T., Robillard, K. R., & Bendayan, R. (2012). Regulation of breast cancer resistant protein by peroxisome proliferator-activated receptor α in human brain microvessel endothelial cells. Mol Pharmacol, 81(4), 598–609.

Hoque, M. T., Shah, A., More, V., Miller, D. S., & Bendayan, R. (2015). In vivo and ex vivo regulation of breast cancer resistant protein (Bcrp) by peroxisome proliferator-activated receptor alpha (Pparα) at the blood-brain barrier. J Neurochem, 135(6), 1113–1122.

Hou, Z., & Matherly, L. H. (2014). Biology of the major facilitative folate transporters SLC19A1 and SLC46A1. Current Topics in Membranes (1st ed., Vol. 73). Elsevier Inc.

Hsueh, C.-T., & Dolnick, B. J. (1993). Altered folate-binding protein mRNA stability in KB cells grown in folate-deficient medium. Biochem Pharmacol, 45(12), 2537–2545.

Huang, T., Chen, Y., Yang, B., Yang, J., Wahlqvist, M. L., & Li, D. (2012). Meta-analysis of B vitamin supplementation on plasma homocysteine, cardiovascular and all-cause mortality. Clin Nutr, 31(4), 448–454.

Huo, Y., Li, J., Qin, X., Huang, Y., Wang, X., Gottesman, R. F., … Hou, F. F. (2015). Efficacy of folic acid therapy in primary prevention of stroke among adults with hypertension in China: the CSPPT randomized clinical trial. JAMA, 313(13), 1325–1335. 139

Hyland, K., Shoffner, J., & Heales, S. J. (2010). Cerebral folate deficiency. J Inherit Metab Dis, 33(5), 563–570.

Hyun, T. H., & Tamura, T. (2005). Trienzyme extraction in combination with microbiologic assay in food folate analysis: an updated review. Exp Biol Med (Maywood), 230(7), 444–454.

Inoue, K., Nakai, Y., Ueda, S., Kamigaso, S., Ohta, K., Hatakeyama, M., … Yuasa, H. (2008). Functional characterization of PCFT/HCP1 as the molecular entity of the carrier-mediated intestinal folate transport system in the rat model. Am J Physiol Gastrointest Liver Physiol, 294(3), G660-668.

Institute of Medicine; Food and Nutrition Board. (1998). Dietary reference intakes: thiamin, riboflavin, vitamin B6, folate, vitamin B12, pantothenic acid, biotin and choline. National Academies Press. Washington, DC: National Academy Press.

Kalenik, J. L., Chen, D., Bradley, M. E., Chen, S., & Lee, T. (1997). Yeast two-hybrid cloning of a novel zinc finger protein that interacts with the multifunctional transcription factor YY1. Nucleic Acids Res, 25(4), 843–849.

Kamen, B. A., & Smith, A. K. (2004). A review of folate receptor alpha cycling and 5- methyltetrahydrofolate accumulation with an emphasis on cell models in vitro. Adv Drug Deliv Rev, 56(8), 1085–1097.

Kane, M. A., Elwood, P. C., Portillo, R. M., Antony, A. C., Najfeld, V., Finley, A., … Kolhouse, J. F. (1988). Influence on immunoreactive folate-binding proteins of extracellular folate concentration in cultured human cells. J Clin Invest, 81(5), 1398–1406.

Kelley, K. M. M., Rowan, B. G., & Ratnam, M. (2003). Modulation of the folate receptor a gene by the estrogen receptor: mechanism and implications in tumor targeting. Cancer Res, 63(11), 2820–2828.

Kiefte-de Jong, J. C., Timmermans, S., Jaddoe, V. W. V., Hofman, A., Tiemeier, H., Steegers, E. a., … Moll, H. a. (2012). High circulating folate and vitamin B-12 concentrations in women during pregnancy are associated with increased prevalence of atopic dermatitis in their offspring. J Nutr, 142(4), 731–738.

140

Kim, T. H., Yang, J., Darling, P. B., & O’Connor, D. L. (2004). A large pool of available folate exists in the large intestine of human infants and piglets. J Nutr, 134(6), 1389–1394.

Kim, Y. (1999). Folate and carcinogenesis: evidence, mechanisms, and implications. J Nutr Biochem, 10(2), 66–88.

Kim, Y. (2003). Role of folate in colon cancer development and progression. J Nutr, 133(11 Suppl 1), 3731–3739.

Kis, O., Zastre, J. a, Ramaswamy, M., & Bendayan, R. (2010). pH dependence of organic anion- transporting polypeptide 2B1 in Caco-2 cells: potential role in antiretroviral drug oral bioavailability and drug-drug interactions. J Pharmacol Exp Ther, 334(3), 1009–1022.

Kitamura, Y., Hirouchi, M., Kusuhara, H., Schuetz, J. D., & Sugiyama, Y. (2008). Increasing systemic exposure of methotrexate by active efflux mediated by multidrug resistance- associated protein 3 (mrp3/abcc3). J Pharmacol Exp Ther, 327(2), 465–473.

Kitamura, Y., Kusuhara, H., & Sugiyama, Y. (2010). Basolateral efflux mediated by multidrug resistance-associated protein 3 (Mrp3/Abcc3) facilitates intestinal absorption of folates in mouse. Pharm Res, 27(4), 665–672.

Kneuer, C., Honscha, K. U., & Honscha, W. (2005). Rat reduced-folate carrier-1 is localized basolaterally in MDCK kidney epithelial cells and contributes to the secretory transport of methotrexate and fluoresceinated methotrexate. Cell Tissue Res, 320(3), 517–524.

Kronn, D., & Goldman, I. D. (2017). Hereditary Folate Malabsorption. Gene Reviews. Seattle, WA: University of Washington, Seattle.

Kumar, N., & Kar, A. (2015). Pyrroloquinoline quinone ameliorates oxidative stress and lipid peroxidation in the brain of streptozotocin-induced diabetic mice. Can J Physiol Pharmacol, 93(1), 71–79.

Kusuhara, H., Han, Y. H., Shimoda, M., Kokue, E., Suzuki, H., & Sugiyama, Y. (1998). Reduced folate derivatives are endogenous substrates for cMOAT in rats. Am J Physiol, 275(4 Pt 1), G789-96.

141

Lakoff, A., Fazili, Z., Aufreiter, S., Pfeiffer, C. M., Connolly, B., Gregory, J. F., … O’Connor, D. L. (2014). Folate is absorbed across the human colon: evidence by using enteric-coated caplets containing 13C-labeled [6S]-5-formyltetrahydrofolate. Am J Clin Nutr, 100(5), 1278– 1286.

Lambert, S. A., Jolma, A., Campitelli, L. F., Das, P. K., Yin, Y., Albu, M., … Weirauch, M. T. (2018). The human transcription factors. Cell, 172(4), 650–665.

Latchman, D. (1997). Transcription Factors: an overview. Int J Biochem Cell Biol, 29(12), 1305– 1312.

Lehtinen, M. K., & Walsh, C. A. (2011). Neurogenesis at the brain–cerebrospinal fluid interface. Annu Rev Cell Dev Biol, 27, 653–679.

Lehtinen, M. K., Zappaterra, M. W., Chen, X., Yang, Y. J., Hill, A., Lun, M., … Walsh, C. A. (2011). The cerebrospinal fluid provides a proliferative niche for neural progenitor cells. Neuron, 69(5), 893–905.

Levine, S. Z., Kodesh, A., Viktorin, A., Smith, L., Uher, R., Reichenberg, A., & Sandin, S. (2018). Association of maternal use of folic acid and multivitamin supplements in the periods before and during pregnancy with the risk of autism spectrum disorder in offspring. JAMA Psychiatry, 75(2), 176–184.

Lin, L., Yee, S. W., Kim, R. B., & Giacomini, K. M. (2015). SLC transporters as therapeutic targets: emerging opportunities. Nat Rev Drug Discov, 14(8), 543–560.

Lindzon, G., & O’Connor, D. L. (2007). Folate during reproduction: the Canadian experience with folic acid fortification. Nutr Res Pract, 1(3), 163–174.

Liu, M., Ge, Y., Cabelof, D. C., Aboukameel, A., Heydari, A. R., Mohammad, R., & Matherly, L. H. (2005). Structure and regulation of the murine reduced folate carrier gene: identification of four noncoding exons and promoters and regulation by dietary folates. J Biol Chem, 280(7), 5588–5597.

Liu, M., Whetstine, J. R., Payton, S. G., Ge, Y., Flatley, R. M., & Matherly, L. H. (2004). Roles of USF, Ikaros and Sp proteins in the transcriptional regulation of the human reduced folate 142

carrier B promoter. Biochem J, 383(Pt 2), 249–257.

Liu, X. Y., & Matherly, L. H. (2002). Analysis of membrane topology of the human reduced folate carrier protein by hemagglutinin epitope insertion and scanning glycosylation insertion mutagenesis. Biochim Biophys Acta, 1564(2), 333–342.

Lucock, M. (2000). Folic acid: nutritional biochemistry, molecular biology, and role in disease processes. Mol Genet Metab, 71(1–2), 121–138.

Masuda, M., I’izuka, Y., Yamazaki, M., Nishigaki, R., Kato, Y., Ni’inuma, K., … Sugiyama, Y. (1997). Methotrexate is excreted into the bile by canalicular multispecific organic anion transporter in rats. Cancer Res, 57(16), 3506–3510.

Masuda, S. (2003). Functional characteristics and pharmacokinetic significance of kidney-specific organic anion transporters, OAT-K1 and OAT-K2, in the urinary excretion of anionic drugs. Drug Metab Pharmacokinet, 18(2), 91–103.

Matherly, L. H., & Hou, Z. (2008). Structure and function of the reduced folate carrier: a paradigm of a major facilitator superfamily mammalian nutrient transporter. Vitam Horm, 79, 145–184.

Matherly, L. H., Wilson, M. R., & Hou, Z. (2014). The major facilitative folate transporters solute carrier 19A1 and solute carrier 46A1: biology and role in antifolate chemotherapy of cancer. Drug Metab Dispos, 42(4), 632–649.

McHugh, M., & Cheng, Y. C. (1979). Demonstration of high affinity folate binder in human cell membranes and its characterization in cultured human KB cells. J Biol Chem, 254(22), 11312–11318.

McLean, E., de Benoist, B., & Allen, L. H. (2008). Review of the magnitude of folate and vitamin B12 deficiencies worldwide. Food Nutr Bull, 29(2 Suppl), S38–S51.

McMartin, K., Collins, T., Eisenga, B., Fortney, T., Bates, W., & Bairnsfather, L. (1989). Effects of chronic ethanol and diet treatment on urinary folate excretion and development of folate deficiency in the rat. J Nutr, 119(10), 1490–1497.

Mellén, M., Ayata, P., Dewell, S., Kriaucionis, S., & Heintz, N. (2012). MeCP2 Binds to 5hmC

143

enriched within active genes and accessible chromatin in the nervous system. Cell, 151(7), 1417–1430.

MRC Vitamin Study Research Group. (1991). Prevention of neural tube defects: results of the Medical Research Council Vitamin Study. Lancet, 338(8760), 131–137.

Neul, J. L., Kaufmann, W. E., Glaze, D. G., Christodoulou, J., Clarke, A. J., Bahi-Buisson, N., … RettSearch Consortium. (2010). Rett syndrome: revised diagnostic criteria and nomenclature. Ann Neurol, 68(6), 944–950.

Ohtsuki, S., Ikeda, C., Uchida, Y., Sakamoto, Y., Miller, F., Glacial, F., … Terasaki, T. (2012). Quantitative targeted absolute proteomic analysis of transporters, receptors and junction proteins for validation of human cerebral microvascular endothelial cell line hCMEC/D3 as a human blood-brain barrier model. Mol Pharm, 10(1), 289–296.

Orjuela, M., Cabrera-Muñoz, L., Paul, L., Ramirez-Ortiz, M. A., Liu, X., Chen, J., … Ponce- Castañeda, M. V. (2012). Risk of retinoblastoma is associated with a maternal polymorphism in dihydrofolatereductase (DHFR) and prenatal folic acid intake. Cancer, 118(23), 5912– 5919.

Ormazábal, A., Perez-Dueñas, B., Sierra, C., Urreitzi, R., Montoya, J., Serrano, M., … Artuch, R. (2011). Folate analysis for the differential diagnosis of profound cerebrospinal fluid folate deficiency. Clin Biochem, 44(8–9), 719–721.

Pardridge, W. M. (2012). Drug transport across the blood–brain barrier. J Cereb Blood Flow Metab, 32(11), 1959–1972.

Parker, N., Jo, M., Westrick, E., Je, V., Lewis, D., Low, P. S., & Leamon, C. P. (2005). Folate receptor expression in carcinomas and normal tissues determined by a quantitative radioligand binding assay. Anal Biochem, 338(2), 284–293.

Patanwala, I., King, M. J., Barrett, D. A., Rose, J., Jackson, R., Hudson, M., … Jones, D. E. (2014). Folic acid handling by the human gut: implications for food fortification. Am J Clin Nutr, 100(2), 593–599.

Patel, R., Bookout, A. L., Magomedova, L., Owen, B. M., Consiglio, G. P., Shimizu, M., … 144

Cummins, C. L. (2015). Glucocorticoids regulate the metabolic hormone FGF21 in a feed- forward loop. Mol Endocrinol, 29(2), 213–223.

Patrick, T. A., Kranz, D. M., van Dyke, T. A., & Roy, E. J. (1997). Folate receptors as potential therapeutic targets in choroid plexus tumors of SV40 transgenic mice. J Neurooncol, 32(2), 111–123.

Payton, S. G., Liu, M., Ge, Y., & Matherly, L. H. (2005). Transcriptional regulation of the human reduced folate carrier A1/A2 promoter: identification of critical roles for the USF and GATA families of transcription factors. Biochim Biophys Acta, 1731(2), 115–124.

Payton, S. G., Whetstine, J. R., Ge, Y., & Matherly, L. H. (2005). Transcriptional regulation of the human reduced folate carrier promoter C: synergistic transactivation by Sp1 and C/EBP beta and identification of a downstream repressor. Biochim Biophys Acta, 1727(1), 45–57.

Pelka, G. J., Watson, C. M., Radziewic, T., Hayward, M., Lahooti, H., Christodoulou, J., & Tam, P. P. L. (2006). Mecp2 deficiency is associated with learning and cognitive deficits and altered gene activity in the hippocampal region of mice. Brain, 129(4), 887–898.

Perez-Duenas, B., Ormazabal, A., Toma, C., Torrico, B., Cormand, B., Serrano, M., … Artuch, R. (2011). Cerebral folate deficiency syndromes in childhood: clinical, analytical, and etiologic aspects. Arch Neurol, 68(5), 615–621.

Pfeiffer, C. M., Fazili, Z., McCoy, L., Zhang, M., & Gunter, E. W. (2004). Determination of folate vitamers in human serum by stable-isotope-dilution tandem mass spectrometry and comparison with radioassay and microbiologic assay. Clin Chem, 50(2), 423–432.

Pfeiffer, C. M., Hughes, J. P., Lacher, D. A., Bailey, R. L., Berry, R. J., Zhang, M., … Johnson, C. L. (2012). Estimation of trends in serum and RBC folate in the U.S. population from pre- to postfortification using assay-adjusted data from the NHANES 1988-2010. J Nutr, 142(5), 886–893.

Pfeiffer, C. M., Sternberg, M. R., Fazili, Z., Yetley, E. A., Lacher, D. A., Bailey, R. L., & Johnson, C. L. (2015). Unmetabolized folic acid is detected in nearly all serum samples from US children, adolescents, and adults. J Nutr, 145(3), 520–531.

145

Piedrahita, J. A., Oetama, B., Bennett, G. D., van Waes, J., Kamen, B. A., Richardson, J., … Finnell, R. H. (1999). Mice lacking the folic acid-binding protein Folbp1 are defective in early embryonic development. Nat Genet, 23(2), 228–232.

Pitkin, R. M. (2007). Folate and neural tube defects. Am J Clin Nutr, 85(1), 285S–288S.

Podvinec, M., Kaufmann, M. R., Handschin, C., & Meyer, U. A. (2002). NUBIScan, an in silico approach for prediction of nuclear receptor response elements. Mol Endocrinol, 16(6), 1269– 1279.

Poller, B., Gutmann, H., Krähenbühl, S., Weksler, B., Romero, I., Couraud, P. O., … Huwyler, J. (2008). The human brain endothelial cell line hCMEC/D3 as a human blood-brain barrier model for drug transport studies. J Neurochem, 107(5), 1358–1368.

Qiu, A., Jansen, M., Sakaris, A., Min, S. H., Chattopadhyay, S., Tsai, E., … Goldman, I. D. (2006). Identification of an intestinal folate transporter and the molecular basis for hereditary folate malabsorption. Cell, 127(5), 917–928.

Qiu, A., Min, S. H., Jansen, M., Malhotra, U., Tsai, E., Cabelof, D. C., … Goldman, I. D. (2007). Rodent intestinal folate transporters (SLC46A1): secondary structure, functional properties, and response to dietary folate restriction. Am J Physiol Cell Physiol, 293(5), C1669–C1678.

Ramaekers, V. T., & Blau, N. (2004). Cerebral folate deficiency. Dev Med Child Neurol, 46(12), 843–851.

Ramaekers, V. T., Blau, N., Sequeira, J. M., Nassogne, M. C., & Quadros, E. V. (2007). Folate receptor autoimmunity and cerebral folate deficiency in low-functioning autism with neurological deficits. Neuropediatrics, 38(6), 276–281.

Ramaekers, V. T., Hansen, S. I., Holm, J., Opladen, T., Senderek, J., Häusler, M., … Blau, N. (2003). Reduced folate transport to the CNS in female Rett patients. Neurology, 61(4), 506– 515.

Ramaekers, V. T., Quadros, E. V., & Sequeira, J. M. (2013). Role of folate receptor autoantibodies in infantile autism. Mol Psychiatry, 18(3), 270–271.

146

Ramaekers, V. T., Rothenberg, S. P., Sequeira, J. M., Opladen, T., Blau, N., Quadros, E. V, & Selhub, J. (2005). Autoantibodies to folate receptors in the cerebral folate deficiency syndrome. N Engl J Med, 352(19), 1985–1991.

Ramaekers, V. T., Sequeira, J. M., Artuch, R., Blau, N., Temudo, T., Ormazabal, A., … Quadros, E. V. (2007). Folate receptor autoantibodies and spinal fluid 5-methyltetrahydrofolate deficiency in Rett syndrome. Neuropediatrics, 38(4), 179–183.

Ramsey, L., Panetta, J., Smith, C., Yang, W., Fan, Y., Winick, N., … Relling, M. (2013). Genome- wide study of methotrexate clearance replicates SLCO1B1. Blood, 121(6), 898–905.

Ratnam, M., Marquardt, H., Duhring, J. L., & Freisheim, J. H. (1989). Homologous membrane folate binding proteins in human placenta: cloning and sequence of a cDNA. Biochemistry, 28(20), 8249–8254.

Ravaglia, G., Forti, P., Maioli, F., Martelli, M., Servadei, L., Brunetti, N., … Licastro, F. (2005). Homocysteine and folate as risk factors for dementia and Alzheimer disease. Am J Clin Nutr, 82(3), 636–643.

Ray, J. G., Singh, G., & Burrows, R. F. (2004). Evidence for suboptimal use of periconceptional folic acid supplements globally. BJOG, 111(5), 399–408.

Reddy, J. A., Haneline, L. S., Srour, E. F., Antony, A. C., Clapp, D. W., & Low, P. S. (1999). Expression and functional characterization of the β-isoform of the folate receptor on CD34+ cells. Blood, 93(11), 3940–3948.

Rogers, L. M., Cordero, A. M., Pfeiffer, C. M., Hausman, D. B., Tsang, B. L., De-Regil, L. M., … Bailey, L. B. (2018). Global folate status in women of reproductive age: a systematic review with emphasis on methodological issues. Ann N Y Acad Sci, 1431, 35–57.

Ross, J. F., Wang, H., Behm, F. G., Mathew, P., Wu, M., Booth, R., & Ratnam, M. (1999). Folate receptor type beta is a neutrophilic lineage marker and is differentially expressed in myeloid leukemia. Cancer, 85(2), 348–357.

Roth, M., Obaidat, A., & Hagenbuch, B. (2012). OATPs, OATs and OCTs: the organic anion and cation transporters of the SLCO and SLC22A gene superfamilies. Br J Pharmacol, 165(5), 147

1260–1287.

Rucker, R., Chowanadisai, W., & Nakano, M. (2009). Potential physiological importance of pyrroloquinoline quinone. Alternative Medicine Review, 14(3), 268–277.

Russell, R., Rosenberg, I., Wilson, P., Iber, F., Oaks, E., Giovetti, A., … Press, A. (1983). Increased urinary excretion and prolonged turnover time of folic acid during ethanol ingestion. Am J Clin Nutr, 38(1), 64–70.

Sadasivan, E., Cedeno, M., & Rothenberg, S. (1994). Characterization of the gene encoding a folate-binding protein expressed in human placenta. Identification of promoter activity in a G-rich SP1 site linked with the tandemly repeated GGAAG motif for the ets encoded GA- binding protein. J Biol Chem, 269(7), 4725–4735.

Said, H. M., Chatterjee, N., Haq, R. U., Subramanian, V. S., Ortiz, A., Matherly, L. H., … Rubin, S. a. (2000). Adaptive regulation of intestinal folate uptake: effect of dietary folate deficiency. Am J Physiol Cell Physiol, 279(6), C1889–C1895.

Saihara, K., Kamikubo, R., Ikemoto, K., Uchida, K., & Akagawa, M. (2017). Pyrroloquinoline quinone, a redox-active o-quinone, stimulates mitochondrial biogenesis by activating the SIRT1/PGC-1α signaling pathway. Biochemistry, 56(50), 6615–6625.

Saikawa, Y., Price, K., Hance, K., Chen, T., & Elwood, P. (1995). Structural and functional analysis of the human KB cell folate receptor gene P4 promoter: cooperation of three clustered Sp1-binding sites with initiator region for basal promoter activity. Biochemistry, 34(31), 9951–9961.

Salbaum, J. M., Finnell, R. H., & Kappen, C. (2009). Regulation of folate receptor 1 gene expression in the visceral endoderm. Birth Defects Res A Clin Mol Teratol, 85(4), 303–313.

Salojin, K. V, Cabrera, R. M., Sun, W., Chang, W. C., Lin, C., Duncan, L., … Oravecz, T. (2011). A mouse model of hereditary folate malabsorption: deletion of the PCFT gene leads to systemic folate deficiency. Blood, 117(18), 4895–4904.

Scarpulla, R. C. (2008). Transcriptional paradigms in mammalian mitochondrial biogenesis and function. Physiol Rev, 88(2), 611–638. 148

Scarpulla, R. C. (2011). Metabolic control of mitochondrial biogenesis through the PGC-1 family regulatory network. Biochim Biophys Acta, 1813(7), 1269–1278.

Schmidt, R. J., Tancredi, D. J., Ozonoff, S., Hansen, R. L., Hartiala, J., Allayee, H., … Hertz- Picciotto, I. (2012). Maternal periconceptional folic acid intake and risk of autism spectrum disorders and developmental delay in the CHARGE (CHildhood Autism Risks from Genetics and Environment) case-control study. Am J Clin Nutr, 96(1), 80–89.

Selhub, J., & Franklin, W. A. (1984). The folate-binding protein of rat kidney. Purification, properties, and cellular distribution. J Biol Chem, 259(10), 6601–6606.

Selhub, J., Nakamura, S., & Carone, F. A. (1987). Renal folate absorption and the kidney folate binding protein. II. Microinfusion studies. Am J Physiol, 252(4 Pt 2), F757-60.

Sequeira, J. M., Desai, A., Berrocal-Zaragoza, M. I., Murphy, M. M., Fernandez-Ballart, J. D., & Quadros, E. V. (2016). Exposure to folate receptor alpha antibodies during gestation and weaning leads to severe behavioral deficits in rats: a pilot study. PLoS ONE, 11(3), e0152249.

Sergeev, I. N., & Song, Q. (2014). High vitamin D and calcium intakes reduce diet-induced obesity in mice by increasing adipose tissue apoptosis. Mol Nutr Food Res, 58(6), 1342–1348.

Serrano, M., García-Silva, M. T., Martin-Hernandez, E., O’Callaghan, M. del M., Quijada, P., Martinez-Aragón, A., … Pineda, M. (2010). Kearns-Sayre syndrome: Cerebral folate deficiency, MRI findings and new cerebrospinal fluid biochemical features. Mitochondrion, 10(5), 429–432.

Seshadri, S., Beiser, A., Selhub, J., Jacques, P. F., Rosenberg, I. H., D’Agostino, R. B., … Wolf, P. A. (2002). Plasma homocysteine as a risk factor for dementia and Alzheimer’s disease. N Engl J Med, 346(7), 476–483.

Shafizadeh, T. B., & Halsted, C. H. (2007). Gamma-glutamyl hydrolase, not glutamate carboxypeptidase II, hydrolyzes dietary folate in rat small intestine. J Nutr, 137, 1149–1153.

Shafizadeh, T. B., & Halsted, C. H. (2009). Postnatal ontogeny of intestinal GCPII and the RFC in pig. Am J Physiol Gastrointest Liver Physiol, 296(3), G476-481.

149

Shahbazian, M. D., Antalffy, B., Armstrong, D. L., & Zoghbi, H. Y. (2002). Insight into Rett syndrome: MeCP2 levels display tissue- and cell-specific differences and correlate with neuronal maturation. Hum Mol Gen, 11(2), 115–124.

Shayeghi, M., Latunde-Dada, G. O., Oakhill, J. S., Laftah, A. H., Takeuchi, K., Halliday, N., … McKie, A. T. (2005). Identification of an intestinal heme transporter. Cell, 122(5), 789–801.

Shemesh, A., & Margolin, E. (2018). Kearns Sayre Syndrome. Retrieved from https://www.ncbi.nlm.nih.gov/books/NBK482341/

Shen, F., Wu, M., Ross, J. F., Miller, D., & Ratnam, M. (1995). Folate receptor type gamma is primarily a secretory protein due to lack of an efficient signal for glycosylphosphatidylinositol modification: protein characterization and cell type specificity. Biochemistry, 34(16), 5660–5665.

Shin, D. S., Mahadeo, K., Min, S. H., Diop-Bove, N., Clayton, P., Zhao, R., & Goldman, I. D. (2011). Identification of novel mutations in the proton-coupled folate transporter (PCFT- SLC46A1) associated with hereditary folate malabsorption. Mol Genet Metab, 103(1), 33– 37.

Shin, D. S., Zhao, R., Yap, E. H., Fiser, A., & Goldman, I. D. (2012). A P425R mutation of the proton-coupled folate transporter causing hereditary folate malabsorption produces a highly selective alteration in folate binding. Am J Physiol Cell Physiol, 302(9), C1405-1412.

Sid, V., Siow, Y. L., Shang, Y., Woo, C. W., & Karmin, O. (2018). High-fat diet consumption reduces hepatic folate transporter expression via nuclear respiratory factor-1. J Mol Med (Berl), 96(11), 1203–1213.

Sissung, T., Troutman, S., Campbell, T., Pressler, H., Sung, H., Bates, S. E., & Figg, W. D. (2012). Transporter pharmacogenetics: transporter polymorphisms affect normal physiology, diseases, and pharmacotherapy. Discov Med, 13(68), 19–34.

Smidt, C., Unkefer, C., Houck, D., & Rucker, R. (1991). Intestinal absorption and tissue distribution of [14C]pyrroloquinoline quinone in mice. Proc Soc Exp Biol Med, 197(1), 27– 31.

150

Smith, A. D., Kim, Y.-I., & Refsum, H. (2008). Is folic acid good for everyone? Am J Clin Nutr, 87(3), 517–533.

Solanky, N., Requena Jimenez, A., D’Souza, S. W., Sibley, C. P., & Glazier, J. D. (2010). Expression of folate transporters in human placenta and implications for homocysteine metabolism. Placenta, 31(2), 134–143.

Song, J., Medline, A., Mason, J. B., Gallinger, S., & Kim, Y. I. (2000). Effects of dietary folate on intestinal tumorigenesis in the apcMin mouse. Cancer Res, 60(19), 5434–5440.

Song, J., Sohn, K. J., Medline, A., Ash, C., Gallinger, S., & Kim, Y. I. (2000). Chemopreventive effects of dietary folate on intestinal polyps in Apc+/-Msh2-/- mice. Cancer Res, 60(12), 3191–3199.

Sowers, R., Toguchida, J., Qin, J., Meyers, P. A., Healey, J. H., Huvos, A., … Gorlick, R. (2003). mRNA expression levels of E2F transcription factors correlate with dihydrofolate reductase, reduced folate carrier, and thymidylate synthase mRNA expression in osteosarcoma. Mol Cancer Ther, 2(6), 535–541.

Spector, R., & Johanson, C. E. (2010). Choroid plexus failure in the Kearns-Sayre syndrome. Cerebrospinal Fluid Res, 7, 14.

Spiegelstein, O., Mitchell, L. E., Merriweather, M. Y., Wicker, N. J., Zhang, Q., Lammer, E. J., & Finnell, R. H. (2004). Embryonic development of folate binding protein-1 (Folbp1) knockout mice: effects of the chemical form, dose and timing of maternal folate supplementation. Dev Dyn, 231(1), 221–231.

Stark, M., Gonen, N., & Assaraf, Y. G. (2009). Functional elements in the minimal promoter of the human proton-coupled folate transporter. Biochem Biophys Res Commun, 388(1), 79–85.

Steinfeld, R., Grapp, M., Kraetzner, R., Dreha-Kulaczewski, S., Helms, G., Dechent, P., … Gärtner, J. (2009). Folate receptor alpha defect causes cerebral folate transport deficiency: a treatable neurodegenerative disorder associated with disturbed myelin metabolism. Am J Hum Genet, 85(3), 354–363.

Stevens, V. L., McCullough, M. L., Sun, J., Jacobs, E. J., Campbell, P. T., & Gapstur, S. M. (2011). 151

High levels of folate from supplements and fortification are not associated with increased risk of colorectal cancer. Gastroenterology, 141(1), 98–105.

Surén, P., Roth, C., Bresnahan, M., Haugen, M., Hornig, M., Hirtz, D., … Stoltenberg, C. (2013). Association between maternal use of folic acid supplements and risk of autism spectrum disorders in children. JAMA, 309(6), 570–577.

Surtees, R. (2001). Cobalamin and folate responsive disorders. Vitamin responsive conditions on pediatric neurology, International review of child neurology series. (P. Baxter, Ed.). London: Mac Keith Press.

Sweeney, M. R., McPartlin, J., Weir, D. G., Daly, L., & Scott, J. M. (2006). Postprandial serum folic acid response to multiple doses of folic acid in fortified bread. Br J Nutr, 95(01), 145.

Sweeney, M. R., Mcpartlin, J., Weir, D. G., & Scott, J. M. (2003). Measurements of sub-nanomolar concentrations of unmetabolised folic acid in serum. J Chromatogr B Analyt Technol Biomed Life Sci, 788(1), 187–191.

Tachikawa, M., Watanabe, M., Hori, S., Fukaya, M., Ohtsuki, S., Asashima, T., & Terasaki, T. (2005). Distinct spatio-temporal expression of ABCA and ABCG transporters in the developing and adult mouse brain. J Neurochem, 95(1), 294–304.

Takeda, M., Khamdang, S., Narikawa, S., Kimura, H., Hosoyamada, M., Cha, S. H. O., … Endou, H. (2002). Characterization of methotrexate transport and its drug interactions with human organic anion transporters. J Pharmacol Exp Ther, 302(2), 666–671.

Tamura, T., & Halsted, C. (1983). Folate turnover in chronically alcoholic monkeys. J Lab Clin Med, 101(4), 623–628.

Tamura, T., & Picciano, M. F. (2006). Folate and human reproduction. Am J Clin Nutr, 83(5), 993–1016.

Tamura, T., Romero, J., Watson, J., Gong, E., & Halsted, C. (1981). Hepatic folate metabolism in the chronic alcoholic monkey. J Lab Clin Med, 97(5), 654–661.

Tang, L. S., & Finnell, R. H. (2003). Neural and orofacial defects in Folbp1 knockout mice. Birth

152

Defects Res A Clin Mol Teratol, 67(4), 209–218.

Tang, L. S., Santillano, D. R., Wlodarczyk, B. J., Miranda, R. C., & Finnell, R. H. (2005). Role of Folbp1 in the regional regulation of apoptosis and cell proliferation in the developing neural tube and craniofacies. Am J Med Genet C Semin Med Genet, 135C(1), 48–58.

Tanji, K., Schon, E. A., Dimauro, S., & Bonilla, E. (2000). Kearns-Sayre syndrome: oncocytic transformation of choroid plexus epithelium. J Neurol Sci, 178(1), 29–36.

Tibbetts, A. S., & Appling, D. R. (2010). Compartmentalization of mammalian folate-mediated one-carbon metabolism. Annu Rev Nutr, 30, 57–81.

Toffoli, G., Cernigoi, C., Russo, A., Gallo, A., Bagnoli, M., & Biocchi, M. (1997). Overexpression of folate binding protein in ovarian cancers. Int J Cancer, 74(2), 193–198.

Torres, A., Newton, S. A., Crompton, B., Borzutzky, A., Neufeld, E. J., Notarangelo, L., & Berry, G. T. (2015). CSF 5-methyltetrahydrofolate serial monitoring to guide treatment of congenital folate malabsorption due to proton-coupled folate transporter (PCFT) deficiency. JIMD Reports, 24, 91–96.

Tran, T., Shatnawi, A., Zheng, X., Kelley, K. M. M., & Ratnam, M. (2005). Enhancement of folate receptor alpha expression in tumor cells through the glucocorticoid receptor: a promising means to improved tumor detection and targeting. Cancer Res, 65(10), 4431–4442.

Treviño, L. R., Shimasaki, N., Yang, W., Panetta, J. C., Cheng, C., Pei, D., … Relling, M. V. (2009). Germline genetic variation in an organic anion transporter polypeptide associated with methotrexate pharmacokinetics and clinical effects. J Clin Oncol, 27(35), 5972–5978.

Ueland, P. M., Hustad, S., Schneede, J., Refsum, H., & Vollset, S. E. (2001). Biological and clinical implications of the MTHFR C677T polymorphism. Trends Pharmacol Sci, 22(4), 195–201.

Unal, E., Zhao, R., Qiu, A., & Goldman, I. (2008). N-linked glycosylation and its impact on the electrophoretic mobility and function of the human proton-coupled folate transporter (HsPCFT). Biochim Biophys Acta, 1778(6), 1407–1414.

153

Urich, E., Lazic, S. E., Molnos, J., Wells, I., & Freskgård, P. O. (2012). Transcriptional profiling of human brain endothelial cells reveals key properties crucial for predictive in vitro blood- brain barrier models. PLoS ONE, 7(5), e38149.

Urquhart, B. L., Gregor, J. C., Chande, N., Knauer, M. J., Tirona, R. G., & Kim, R. B. (2010). The human proton-coupled folate transporter (hPCFT): modulation of intestinal expression and function by drugs. Am J Physiol Gastrointest Liver Physiol, 298(2), G248-254.

Uwai, Y., Taniguchi, R., Motohashi, H., Saito, H., Okuda, M., & Inui, K. (2004). Methotrexate- loxoprofen interaction: involvement of human organic anion transporters hOAT1 and hOAT3. Drug Metab Pharmacokinet, 19(5), 369–374.

Valenti, D., de Bari, L., De Filippis, B., Henrion-Caude, A., & Vacca, R. A. (2014). Mitochondrial dysfunction as a central actor in intellectual disability-related diseases: an overview of Down syndrome, autism, Fragile X and Rett syndrome. Neurosci Biobehav Rev, 46(P2), 202–217. van de Steeg, E., van der Kruijssen, C., Wagenaar, E., Burggraaff, J., Mesman, E., Kenworthy, K., & Schinkel, A. (2009). Methotrexate pharmacokinetics in transgenic mice with liver-specific expression of human organic anion-transporting polypeptide 1B1 (SLCO1B1). Drug Metab Dispos, 37(2), 277–281. van der Put, N. M. J., Gabreëls, F., Stevens, E. M. B., Smeitink, J. A. M., Trijbels, F. J. M., Eskes, T. K. A. B., … Blom, H. J. (1998). A second common mutation in the methylenetetrahydrofolate reductase gene: an additional risk factor for neural-tube defects? Am J Hum Genet, 62(5), 1044–1051. van der Put, N. M., Steegers-Theunissen, R. P., Frosst, P., Trijbels, F. J., Eskes, T. K., van den Heuvel, L. P., … Blom, H. J. (1995). Mutated methylenetetrahydrofolate reductase as a risk factor for spina bifida. Lancet, 346(8982), 1070–1071.

Vanlandewijck, M., He, L., Mäe, M. A., Andrae, J., Ando, K., Del Gaudio, F., … Betsholtz, C. (2018). A molecular atlas of cell types and zonation in the brain vasculature. Nature, 554(7693), 475–480.

Viemari, J. C., Roux, J. C., Tryba, A. K., Saywell, V., Burnet, H., Peña, F., … Hilaire, G. (2005).

154

Mecp2 deficiency disrupts norepinephrine and respiratory systems in mice. J Neurosci, 25(50), 11521–11530.

Villanueva, J. A., Devlin, A. M., & Halsted, C. H. (2001). Reduced folate carrier: tissue distribution and effects of chronic ethanol intake in the micropig. Alcohol Clin Exp Res, 25(3), 415–420.

Virk, J., Liew, Z., Olsen, J., Nohr, E. A., Catov, J. M., & Ritz, B. (2016). Preconceptional and prenatal supplementary folic acid and multivitamin intake and autism spectrum disorders. Autism, 20(6), 710–718.

Visentin, M., Chang, M.-H., Romero, M. F., Zhao, R., & Goldman, I. D. (2012). Substrate- and pH-specific antifolate transport mediated by organic anion-transporting polypeptide 2B1 (OATP2B1-SLCO2B1). Mol Pharmacol, 81(2), 134–142.

Visentin, M., Diope-Bove, N., Zhao, R., & Goldman, I. D. (2014). The intestinal absorption of folates. Annu Rev Physiol, 76, 251–274.

Vlaming, M. L. H., Van Esch, A., Van De Steeg, E., Pala, Z., Wagenaar, E., Van Tellingen, O., & Schinkel, A. H. (2011). Impact of Abcc2 [multidrug resistance-associated protein (Mrp) 2], Abcc3 (Mrp3), and Abcg2 (breast cancer resistance protein) on the oral pharmacokinetics of methotrexate and its main metabolite 7-hydroxymethotrexate. Drug Metab Dispos, 39(8), 1338–1344.

Wang, H., Ross, J. F., & Ratnam, M. (1998). Structure and regulation of a polymorphic gene encoding folate receptor type γ/γ′. Nucleic Acids Res, 26(9), 2132–2142.

Wang, H., Zheng, X., Behm, F. G., & Ratnam, M. (2000). Differentiation-independent retinoid induction of folate receptor type β, a potential tumor target in myeloid leukemia. Blood, 96(10), 3529–3537.

Wang, X., Cabrera, R. M., Li, Y., Miller, D. S., & Finnell, R. H. (2013). Functional regulation of P-glycoprotein at the blood-brain barrier in proton-coupled folate transporter (PCFT) mutant mice. FASEB Journal, 27(3), 1167–1175.

Wang, Y., Zhao, R., & Goldman, I. D. (2004). Characterization of a folate transporter in HeLa 155

cells with a low pH optimum and high affinity for pemetrexed distinct from the reduced folate carrier. Clin Cancer Res, 10(18 I), 6256–6264.

Wang, Y., Zhao, R., Russell, R. G., & Goldman, I. D. (2001). Localization of the murine reduced folate carrier as assessed by immunohistochemical analysis. Biochim Biophys Acta, 1513(1), 49–54.

Wani, N. A., & Kaur, J. (2011). Reduced levels of folate transporters (PCFT and RFC) in membrane lipid rafts result in colonic folate malabsorption in chronic alcoholism. J Cell Physiol, 226(3), 579–587.

Wani, N. A., Nada, R., & Kaur, J. (2011). Biochemical and molecular mechanisms of folate transport in rat pancreas; interference with ethanol ingestion. PLoS ONE, 6(12), e28599.

Wani, N., Hamid, A., & Kaur, J. (2012). Alcohol-associated folate disturbances result in altered methylation of folate-regulating genes. Mol Cell Biochem, 363(1–2), 157–166.

Weisberg I, Tran P, Christensen B, Sibani S, R. R. (1998). A second genetic polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with decreased enzyme activity. Mol Genet Metab, 64(3), 169–172.

Weissman, J. R., Kelley, R. I., Bauman, M. L., Cohen, B. H., Murray, K. F., Mitchell, R. L., … Natowicz, M. R. (2008). Mitochondrial disease in autism spectrum disorder patients: a cohort analysis. PLoS ONE, 3(11), e3815.

Weitman, S. D., Frazier, K. M., & Kamen, B. A. (1994). The folate receptor in central nervous system malignancies of childhood. J Neurooncol, 21(2), 107–112.

Weksler, B. B. (2005). Blood-brain barrier-specific properties of a human adult brain endothelial cell line. The FASEB Journal, 19(13), 1872–1874.

Weksler, B., Romero, I. A., & Couraud, P.-O. (2013). The hCMEC/D3 cell line as a model of the human blood brain barrier. Fluids Barriers CNS, 10(1), 16.

Whetstine, J. R., Flatley, R. M., & Matherly, L. H. (2002). The human reduced folate carrier gene is ubiquitously and differentially expressed in normal human tissues : identification of seven

156

non-coding exons and characterization of a novel promoter. Biochem J, 367(Pt 3), 629–640.

Whetstine, J. R., & Matherly, L. H. (2001). The basal promoters for the human reduced folate carrier gene are regulated by a GC-box and a cAMP-response element/AP-1-like element. Basis for tissue-specific gene expression. J Biol Chem, 276(9), 6350–6358.

Whetstine, J. R., Witt, T. L., & Matherly, L. H. (2002). The human reduced folate carrier gene is regulated by the AP2 and sp1 transcription factor families and a functional 61-base pair polymorphism. J Biol Chem, 277(46), 43873–43880.

Wielinga, P., Hooijberg, J. H., Gunnarsdottir, S., Kathmann, I., Reid, G., Zelcer, N., … Borst, P. (2005). The human multidrug resistance protein MRP5 transports folates and can mediate cellular resistance against antifolates. Cancer Res, 65(10), 4425–4430.

Wien, T. N., Pike, E., Wisløff, T., Staff, A., Smeland, S., & Klemp, M. (2012). Cancer risk with folic acid supplements: a systematic review and meta-analysis. BMJ Open, 2(1), e000653.

Wilcken, B., Bamforth, F., Li, Z., Zhu, H., Ritvanen, A., Renlund, M., … Botto, L. D. (2003). Geographical and ethnic variation of the 677C>T allele of 5,10 methylenetetrahydrofolate reductase (MTHFR): findings from over 7000 newborns from 16 areas world wide. J Med Genet, 40(8), 619–625.

Williams, J., Mai, C. T., Mulinare, J., Isenburg, J., Flood, T., Ethen, M., … Centers for Disease Control and Prevention. (2015). Updated estimates of neural tube defects prevented by mandatory folic Acid fortification - United States, 1995-2011. MMWR Morb Mortal Wkly Rep, 64(1), 1–5.

Wilson, R., Genetics Committee, & Motherisk. (2007). Pre-conceptional vitamin/folic acid supplementation 2007: the use of folic acid in combination with a multivitamin supplement for the prevention of neural tube defects and other congenital anomalies. J Obstet Gynaecol Can, 29(12), 1003–1013.

Wollack, J. B., Makori, B., Ahlawat, S., Koneru, R., Picinich, S. C., Smith, A., … Kamen, B. (2008). Characterization of folate uptake by choroid plexus epithelial cells in a rat primary culture model. J Neurochem, 104(6), 1494–1503.

157

Wong, S. C., Zhang, L., Proefke, S. A., & Matherly, L. H. (1998). Effects of the loss of capacity for N-glycosylation on the transport activity and cellular localization of the human reduced folate carrier. Biochim Biophys Acta, 1375(1–2), 6–12.

World Health Organization; Department of Nutrition for Health and Development. (2012). Serum and red blood cell folate concentrations for assessing folate status in populations. Geneva, Switzerland.

Wu, D., & Pardridge, W. M. (1999). Blood-brain barrier transport of reduced folic acid. Pharm Res, 16(3), 415–419.

Wu, M., Gunning, W., & Ratnam, M. (1999). Expression of folate receptor type alpha in relation to cell type, malignancy, and differentiation in ovary, uterus, and cervix. Cancer Epidemiol Biomarkers Prev, 8(9), 775–782.

Wu, Y., Zhong, W., Cui, N., Johnson, C. M., Xing, H., Zhang, S., & Jiang, C. (2016). Characterization of Rett Syndrome-like phenotypes in Mecp2-knockout rats. J Neurodev Disord, 8, 23.

Wu, Z., Puigserver, P., Andersson, U., Zhan, C., Adelmant, G., Mootha, V., … Spiegelman, B. M. (1999). Mechanisms controlling mitochondrial biogenesis and respiration through the thermogenic coactivator PGC-1. Cell, 98(1), 115–124.

Yajnik, C. S., Deshpande, S. S., Jackson, A. A., Refsum, H., Rao, S., Fisher, D. J., … Fall, C. H. D. (2008). Vitamin B12 and folate concentrations during pregnancy and insulin resistance in the offspring: the Pune Maternal Nutrition Study. Diabetologia, 51(1), 29–38.

Yamaguchi, T., Hirota, K., Nagahama, K., Ohkawa, K., Takahashi, T., Nomura, T., & Sakaguchi, S. (2007). Control of immune responses by antigen-specific regulatory T cells expressing the folate receptor. Immunity, 27(1), 145–159.

Yang, C., Yu, L., Kong, L., Ma, R., Zhang, J., Zhu, Q., … Hao, D. (2014). Pyrroloquinoline quinone (PQQ) inhibits lipopolysaccharide induced inflammation in part via downregulated NF-kB and p38/JNK activation in microglial and attenuates microglia activation in lipopolysaccharide treatment mice. PLoS ONE, 9(10), e109502.

158

Yang, W., Carmichael, S. L., & Shaw, G. M. (2016). Folic acid fortification and prevalences of neural tube defects, orofacial clefts, and gastroschisis in California, 1989 to 2010. Birth Defects Res A Clin Mol Teratol, 106(12), 1032–1041.

Yasuda, S., Hasui, S., Yamamoto, C., Yoshioka, C., Kobayashi, M., Itagaki, S., … Iseki, K. (2008). Placental folate transport during pregnancy. Biosci Biotechnol Biochem, 72(9), 2277–2284.

Zeisel, A., Hochgerner, H., Lönnerberg, P., Johnsson, A., Memic, F., van der Zwan, J., … Linnarsson, S. (2018). Molecular architecture of the mouse nervous system. Cell, 174(4), 999–1014.e22.

Zhang, Y., Huo, M., Zhou, J., & Xie, S. (2010). PKSolver: An add-in program for pharmacokinetic and pharmacodynamic data analysis in Microsoft Excel. Computer Methods and Programs in Biomedicine, 99(3), 306–314.

Zhang, Z., Burch, P. E., Cooney, A. J., Lanz, R. B., Pereira, F. A., Wu, J., … Wheeler, D. A. (2004). Genomic analysis of the nuclear receptor family: new insights into structure, regulation, and evolution from the rat genome. Genome Res, 14(4), 580–590.

Zhao, R., Aluri, S., & Goldman, I. (2017). The proton-coupled folate transporter (PCFT- SLC46A1) and the syndrome of systemic and cerebral folate deficiency of infancy: hereditary folate malabsorption. Mol Aspects Med, 53, 57–72.

Zhao, R., Diop-Bove, N., Visentin, M., & Goldman, I. D. (2011). Mechanisms of membrane transport of folates into cells and across epithelia. Annu Rev Nutr, 31, 177–201.

Zhao, R., Gao, F., Wang, Y., Diaz, G. A., Gelb, B. D., & Goldman, I. D. (2001). Impact of the reduced folate carrier on the accumulation of active thiamin metabolites in murine leukemia cells. J Biol Chem, 276(2), 1114–1118.

Zhao, R., & Goldman, I. D. (2013). Folate and thiamine transporters mediated by facilitative carriers (SLC19A1-3 and SLC46A1) and folate receptors. Mol Aspects Med, 34(2–3), 373– 385.

Zhao, R., Matherly, L. H., & Goldman, I. D. (2009). Membrane transporters and folate homeostasis: intestinal absorption and transport into systemic compartments and tissues. 159

Expert Rev Mol Med, 11, e4.

Zhao, R., Min, S. H., Qiu, A., Sakaris, A., Goldberg, G. L., Sandoval, C., … Goldman, I. D. (2007). The spectrum of mutations in the PCFT gene, coding for an intestinal folate transporter, that are the basis for hereditary folate malabsorption. Blood, 110(4), 1147–1152.

Zhao, R., Min, S. H., Wang, Y., Campanella, E., Low, P. S., & Goldman, I. D. (2009). A role for the proton-coupled folate transporter (PCFT-SLC46A1) in folate receptor-mediated endocytosis. J Biol Chem, 284(7), 4267–4274.

Zhao, R., Qiu, A., Tsai, E., Jansen, M., Akabas, M. H., & Goldman, I. D. (2008). The proton- coupled folate transporter: impact on pemetrexed transport and on antifolates activities compared with the reduced folate carrier. Mol Pharmacol, 74(3), 854–862.

Zhao, R., Russell, R. G., Wang, Y., Liu, L., Gao, F., Kneitz, B., … Goldman, I. D. (2001). Rescue of embryonic lethality in reduced folate carrier-deficient mice by maternal folic acid supplementation reveals early neonatal failure of hematopoietic organs. J Biol Chem, 276(13), 10224–10228.

Zhao, R., Unal, E. S., Shin, D. S., & Goldman, I. D. G. (2010). Membrane topological analysis of the proton-coupled folate transporter (PCFT-SLC46A1) by the substituted cysteine accessibility method. Biochemistry, 49(13), 2925–2931.

Zhou, X., Chen, Q., Hu, X., Mao, S., & Kong, Y. (2014). Pyrroloquinoline quinone prevents MK- 801-induced stereotypical behavior and cognitive deficits in mice. Behav Brain Res, 258, 153–159.

Zhu, H., Cabrera, R. M., Wlodarczyk, B. J., Bozinov, D., Wang, D., Schwartz, R. J., & Finnell, R. H. (2007). Differentially expressed genes in embryonic cardiac tissues of mice lacking Folr1 gene activity. BMC Dev Biol, 7, 128.

Zhu, H., Wlodarczyk, B. J., Scott, M., Yu, W., Merriweather, M., Waes, J. G. Van, … Finnell, R. H. (2007). Cardiovascular abnormalities in Folr1 knockout mice and folate rescue. Birth Defects Research Part A - Clinical and Molecular Teratology, 79(4), 257–268.

Zhu, W., Alliegro, M., & Melera, P. (2001). The rate of folate receptor alpha (FR alpha) synthesis 160

in folate depleted CHL cells is regulated by a translational mechanism sensitive to media folate levels, while stable overexpression of its mRNA is mediated by gene amplification and an increase in tr. J Cell Biochem, 81(2), 205–219.

161

Appendices Appendix A: Supplemental Data for Chapter 2

150

100

50 Cell Viability Viability Cell (% Untreated Control) 0

50 10 100 500 750 1000 Vehicle Untreated

1,25(OH)2D3 (nM)

Figure A-1. Effect of calcitriol treatment on hCMEC/D3 cell viability. A tetrazolium salts (MTT) assay was used to confirm that exposure to calcitriol (10-1000 nM for 24 h) did not affect the viability of hCMEC/D3 cells compared to untreated or vehicle (ethanol)-treated controls. Results are presented as mean ± S.E.M. for n = 3 independent experiments. Data were generated and analyzed by C Alam in the laboratory of Dr. R Bendayan.

162

2.0 **** ** ** * * * ** ** * 1.5

1.0

0.5 (relative to ethanol control)

SLC19A1 mRNA Expression Expression mRNA SLC19A1 0.0 1 5 10 25 50 0.1 100 250 500 Vehicle

1,25(OH)2D3 Concentration (nM)

Figure A-2. Dose-dependent effect of calcitriol treatment on RFC expression in hCMEC/D3 cells. Significant increases in SLC19A1 mRNA were observed in hCMEC/D3 cells treated with a wide range of calcitriol concentrations (0.1-500 nM) for 24 h. In all of the tested concentrations, SLC19A1 mRNA was induced by 40-50% compared to vehicle (ethanol) control. Results are presented as mean ± S.E.M. for n = 3-4 independent experiments. Asterisks represent data points significantly different from vehicle (*, p < 0.05; **, p < 0.01; ****, p < 0.0001). Data were generated and analyzed by C Alam in the laboratory of Dr. R Bendayan.

163

A PCFT (mRNA)

1,25(OH)2D3 3 Vehicle ** * 50 nM 100 nM 2 500 nM

1 (relative to ethanol control) SLC46A1 mRNA Expression Expression mRNA SLC46A1 0

6 hr 24 hr Treatment Duration

B 1,25(OH)2D3 (nM) 0 50 100 500 Rat Kidney PCFT 63 kDa 50 kDa

β-Actin 42 kDa

PCFT Protein (50kDa) PCFT Protein (63kDa)

200 300 1,25(OH) D 1,25(OH)2D3 **** 2 3 Vehicle Vehicle 150 *** 500 nM 500 nM 200

100

100 50 PCFT Protein Expression Expression Protein PCFT PCFT Protein Expression Expression Protein PCFT (relative to ethanol control) 0 (relative to ethanol control) 0

24 hr 24 hr Treatment Duration Treatment Duration

Figure A-3. Effect of calcitriol treatment on PCFT expression in hCMEC/D3 cells. Significant increases in SLC46A1 mRNA (A) and PCFT protein (B) expression were observed in hCMEC/D3 cells treated with calcitriol (50-500 nM) for 6 or 24 h compared to vehicle (ethanol) control. Multiple protein bands for PCFT (50 and 63 kDa) are indicative of differential glycosylation of the transmembrane protein. Rat kidney lysates served as positive control, while actin was used as a loading control. Results are presented as mean ± S.E.M. for n = 3-4 independent experiments. Asterisks represent data points significantly different from vehicle (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001). Data were generated and analyzed by C Alam in the laboratory of Dr. R Bendayan.

164

A siRNA

Control VDR

PCFT 63 kDa

50 kDa

VDR 48 kDa

β-Actin 42 kDa

B 150 Control siRNA VDR siRNA

100

* ** *

(% Control) 50 PCFT Protein Expression Expression Protein PCFT

VDR

PCFT (50 kDa) PCFT (63 kDa)

Figure A-4. Effect of VDR downregulation on PCFT expression in hCMEC/D3 cells. (A) Significant decreases in VDR and PCFT protein expression were observed in cells transfected with VDR siRNA compared to scrambled siRNA-treated control. Multiple protein bands for PCFT (50 and 63 kDa) are indicative of differential glycosylation of the transmembrane protein. Actin was used as a loading control. (B) Relative levels of VDR and PCFT expression were determined by densitometric analyses. Results are expressed as percentage change normalized to control siRNA and reported as mean ± S.E.M. for n = 3 independent experiments. Asterisks represent data points significantly different from control siRNA (*, p < 0.05; **, p < 0.01). Experiments were performed by Dr. MT Hoque and C Alam in the laboratory of Dr. R Bendayan. Data were analyzed by C Alam.

165

Appendix B: Supplemental Data for Chapter 3

13 Table B-1. Pharmacokinetic parameters for [ C5]-5-formylTHF in WT mice estimated from plasma concentration data fitted to a two-compartment single i.v. bolus pharmacokinetic model. Data were generated and analyzed by C Alam in the laboratory of Dr. R Bendayan.

13 Pharmacokinetic [ C5]-5-formylTHF Parameters

-1 k12 (min ) 0.055 ± 0.017 -1 k21 (min ) 0.036 ± 0.019 -1 k10 (min ) 0.221 ± 0.118 V (ml/kg) 396.34 CL (ml/min/kg) 87.76

t1/2b (min) 25.01

AUC0-∞ (ng/ml • min) 2,849 Analysis of model fit WSS 0.10 R2 0.98 AIC -6.13 SC -6.96

166

A Isolated Brain Capillaries Liver Kidney Duodenum (PCFT mRNA) (PCFT mRNA) (PCFT mRNA) (PCFT mRNA)

2.0 2.0 2.0 8 Vehicle * 2.5 µg/kg 1,25(OH)2D3 1.5 1.5 *** 1.5 6

1.0 1.0 1.0 4

0.5 0.5 0.5 2 (relative to (relative vehicle) (relative to (relative vehicle) (relative to (relative vehicle) (relative to (relative vehicle) Slc46a1 mRNA Expression mRNA Slc46a1 Slc46a1 mRNA Expression mRNA Slc46a1 Slc46a1 mRNA Expression mRNA Slc46a1 Slc46a1 mRNA Expression mRNA Slc46a1 0.0 0.0 0.0 0

B Isolated Brain Capillaries Liver Kidney Duodenum (PCFT mRNA) (PCFT mRNA) (PCFT mRNA) (PCFT mRNA) 2.0 2.0 2.0 Vehicle 6

2.5 µg/kg 1,25(OH)2D3 1.5 1.5 1.5 4

1.0 1.0 1.0

2 0.5 0.5 0.5 (relative to (relative vehicle) (relative to (relative vehicle) (relative to (relative vehicle) (relative to (relative vehicle) Slc46a1 mRNA Expression mRNA Slc46a1 Slc46a1 mRNA Expression mRNA Slc46a1 Slc46a1 mRNA Expression mRNA Slc46a1 Slc46a1 mRNA Expression mRNA Slc46a1 0.0 0.0 0.0 0

Figure B-1. Effect of calcitriol treatment on PCFT expression. (A) Significant increases in Slc46a1 (PCFT) mRNA was observed in isolated brain capillaries and liver of WT mice treated with calcitriol (2.5 µg/kg) compared to vehicle (corn oil). (B) There were no changes in Slc46a1 (PCFT) mRNA following treatment of Folr1 KO mice with calcitriol. Results are presented as mean ± S.E.M. for n = 3 independent experiments (total of 8-9 animals per group). Asterisks represent data points significantly different from vehicle (*, p < 0.05; ***, p < 0.001). Experiments were performed by CJ Georgiou and C Alam in the laboratory of Dr. R Bendayan. Data were analyzed by C Alam.

167

A WT Kidney KO Kidney

3 **** **** 8 ****

6 *** 2 * 4

1 2 mRNA Expression mRNA Expression (relative to (relative vehicle) (relative to (relative vehicle)

0 0 Oat1 Oat2 Oat3 Mrp1 Mrp2 Mrp3 Mrp4 Oat1 Oat2 Oat3 Mrp1 Mrp2 Mrp3 Mrp4

Vehicle Vehicle 2.5 µg/kg 1,25(OH) D 2 3 2.5 µg/kg 1,25(OH)2D3

B WT Liver KO Liver ** 3 2.5

2.0 2 1.5

1.0 1 mRNA Expression (relative to (relative vehicle) mRNA Expression (relative to (relative vehicle) 0.5

0 0.0 Oat2 Mrp2 Mrp3 Mrp4 Oat2 Mrp2 Mrp3 Mrp4

Vehicle Vehicle 2.5 µg/kg 1,25(OH) D 2.5 µg/kg 1,25(OH)2D3 2 3

C WT Isolated Brain Capillaries KO Isolated Brain Capillaries

2.5 2.5

2.0 2.0

1.5 1.5

1.0 1.0 mRNA Expression mRNA Expression (relative to (relative vehicle) (relative to (relative vehicle) 0.5 0.5

0.0 0.0 Oat3 Mrp1 Mrp2 Oat3 Mrp1 Mrp2

Vehicle Vehicle 2.5 µg/kg 1,25(OH) D 2 3 2.5 µg/kg 1,25(OH) D 2 3

Figure B-2. Effect of calcitriol treatment on Oat and Mrp expression in WT and Folr1 KO mice. (A) Significant increases in Slc22a8 (Oat3) and Abcc3 (Mrp3) mRNA were observed in kidney of WT and Folr1 KO mice treated with calcitriol (2.5 µg/kg) compared to vehicle (corn oil). (B) Significant induction of Abcc2 (Mrp2) mRNA was also observed in liver of WT mice following calcitriol administration. (C) There were no significant changes in Oat or Mrp mRNA following treatment of WT and Folr1 KO mice with calcitriol. Results are presented as mean ± S.E.M. for n = 3 independent experiments (total of 5-6 animals per group). Asterisks represent data points significantly different from vehicle (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001). Experiments were performed by CJ Georgiou and C Alam in the laboratory of Dr. R Bendayan. Data were analyzed by C Alam.

168

1.1

* ** **** **** **** **** 1.0 ## #

0.9 ## ## ## ### #### #### 0.8 ####

####

(relative Day 0) Day (relative #### #### #### 0.7 Vehicle Ratio of body weight change

0.6 2.5 µg/kg 1,25(OH)2D3

0.5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Time (Day) Figure B-3. Long term effect of calcitriol treatment on body weight of WT mice. Significant weight loss was observed in mice treated with calcitriol (2.5 µg/kg) compared to vehicle (corn oil), but this weight loss seemed to plateau at the beginning of Day 12. Results are presented as mean ± S.E.M. (total of 3 animals per group). Asterisk and pound symbols represent data points significantly different from vehicle control and Day 0 (time before vehicle or calcitriol injections), respectively (*, p < 0.05; **, p < 0.01; ****, p < 0.0001; #, p < 0.05; ##, p < 0.01; ###, p < 0.001; ####, p < 0.0001). Experiments were performed by CJ Georgiou in the laboratory of Dr. R Bendayan.

Pgp (mRNA) 10 ns 8

6 WT 4 KO 2 ns 0.04

0.02 ns (compared to(compared cyclophilin B)

Relative Abcb1a mRNA expression 0.00 Brain Capillary Liver Kidney Figure B-4. Relative expression of P-gp in various tissues of WT and Folr1 KO mice. There were no differences in Abcb1a (P-gp) mRNA expression between WT and Folr1 KO mice, particularly in isolated mouse brain capillaries, liver, and kidney tissues. mRNA levels were determined using TaqMan gene expression assay. Results are presented as mean relative mRNA expression ± S.E.M. normalized to the housekeeping mouse cyclophilin B gene (total of 3-4 animals per group). ns, not significant. Data were generated and analyzed by C Alam in the laboratory of Dr. R Bendayan.

169

A DAPI RFC Na+/K+-ATPase a Merge

B DAPI RFC Na+/K+-ATPase a Merge (with peptide)

C DAPI RFC Na+/K+-ATPase a Merge (no primary antibody) (no primary antibody)

Figure B-5. Expression of RFC in mouse lymphocytic leukemia (L1210) cell line. The specificity of primary rabbit polyclonal AE390 anti-RFC antibody to murine RFC was validated in L1210 cells, which are known to express high levels of RFC. (A) Cellular localization of RFC in L1210 cells was determined through immunostaining with: i) DAPI nuclear marker, ii) AE390 anti-RFC antibody (1:50), and iii) anti- Na+/K+-ATPase α plasma membrane marker (1:50). Three-dimensional colocalization of RFC and the Na+/K+-ATPase α plasma membrane marker was quantified using Imaris Bitplane software. A Pearson’s coefficient colocalization value of 0.83 was determined for RFC and Na+/K+-ATPase α. (B) Co-incubation with a highly purified custom made synthetic peptide (>98% purity; ThermoFisher Scientific) spanning the distal C-terminus of murine RFC from Met499 through Ala512 (epitope for AE390) significantly decreased RFC fluorescence intensity. (C) Cell staining without primary antibodies was used as a negative control. Cells were visualized using confocal microscopy (Carl Zeiss LSM 700) operated with ZEN software using 63x objective lens. Scale bar, 50 µm. Experiments were performed by Dr. MT Hoque and C Alam in the laboratory of Dr. R Bendayan.

170

Primary mouse brain brain mouse Primary cells endothelial cells L1210

10 50 10 50 μg/lane 75

Mouse RFC (58 kDa) 63

!-Actin (42 kDa)

37 kDa

Figure B-6. RFC protein expression in mouse brain microvascular endothelial cells and mouse lymphocytic leukemia (L1210) cell line. The specificity of primary rabbit polyclonal AE390 anti-RFC antibody to murine RFC was validated in L1210 cells, which are known to express high levels of RFC. Whole cell lysates (WCL) (10 or 50 µg/lane) were separated on 10% SDS-PAGE gel and electrotransferred onto PVDF membrane. The membranes were incubated overnight with primary rabbit polyclonal AE390 anti-RFC antibody (1:1000) or murine monoclonal anti-b-actin (1:1000) antibodies, followed by incubation with respective horseradish peroxidase-conjugated anti-rabbit (1:5000) or anti-mouse (1:5000) secondary antibodies. When probing for RFC, a predominant band at ~58 kDa was detected in both cell systems. b- actin was used as a loading control. Experiments were performed by Dr. MT Hoque in the laboratory of Dr. R Bendayan.

171

A Hoechst PCFT Aqp1 Merge

20 μM

B Hoechst FR! Aqp1 Merge

20 μM

Figure B-7. Expression of PCFT (A) and FRα (B) in mouse choroid plexus epithelium. The specificity of primary rabbit polyclonal anti-HCP1 and anti-FBP antibodies (Abcam) to murine PCFT and FRα, respectively, were validated in mouse choroid plexus, which has been documented to show robust expression of these folate transport systems. Fourth ventricle choroid plexus was immunostained with: i) Hoechst 33342 nuclear marker, ii) anti-HCP1/PCFT (1:100) or anti-FBP/FRα (1:100), and iii) anti- Aquaporin 1 (Aqp1) choroid plexus marker (1:100). The choroid plexus epithelium was visualized using confocal microscopy (Carl Zeiss LSM 700) operated with ZEN software using 63x objective lens. Scale bar, 20 µm. Experiments were performed by ML Shannon, Dr. RM Fame, and C Alam in the laboratory of Dr. MK Lehtinen (Harvard Medical School). C Alam visited Dr. Lehtinen’s laboratory in September 2018 to learn choroid plexus isolation from mouse brain.

172

Appendix C: Supplemental Data for Chapter 4

TRANSCRIPTIONAL PARADIGMS IN MAMMALIAN MITOCHONDRIA 625

PQQ

phosphorylation

PQQ ↑ NAD+ deacetylation SIRT1

phosphorylation

Low Glucose ↑ AMP/ATP ratio

FIG. 9. Illustration summarizing PGC-1␣-mediated pathways governing mitochondrial biogenesis and function. Depicted in the nucleus (shaded sphere) are the key transcription factors (NRF-1, NRF-2, ERR␣, PPAR␣, and MEF-2) that are PGC-1␣ targets and act on nuclear genes governing the indicated mitochondrial functions. Some of the physiological effector pathways mediating changes in the transcriptional expression or function Figureof PGC-1 C-␣1.are Illustration also shown. The CREBsummarizing activation of PGC-1 the␣ gene role transcription of perox in responseisome to cold proliferator (thermogenesis),-activated fasting (gluconeogenesis), receptor and-γ coactivatorexercise has-1α been (PGC well documented.-1α) in Thegoverning physiological mitochondrial mechanisms of PGC-1 biogenesis␣ induction by nitricand oxide function are not established. PGC-1α but can may involveinteract the production of endogenous nitric oxide by eNOS. A potential pathway of retrograde signaling through calcium is also included. with a range of transcription factors (i.e., NRF-1, NRF-2, ERRα, PPARα, and MEF-2) to regulate the expression of specific nuclear genes which encode for important components of mitochondrial respiratory function.transcriptional There are expression at least oftwo PGC-1 proposed␣,whichinturnserves mechanisms forof the an NO-dependentactivation of pathwayPGC-1α of and mitochondrial nuclear respiratory biogenesis as a coactivator of gluconeogenic gene expression (93). awaits confirmation of these findings. factor 1 The (NRF cAMP-dependent-1) signaling pathway by pyrroloquinoline is one of several quinone that (PQQ).A number At of th reportse transcriptional link the expression level, of PQQPGC-1 ␣ canto upregulatedirect thePGC induction-1α expression of PGC-1 by␣ in inducing a number the of phosphorylation tissues exercise-induced of cAMP mitochondrial response element biogenesis binding in skeletalprotein (CREB).(126). Phosphorylated PGC-1␣ along with CREB Tfam and subsequently NRF-1 are induced binds via to a musclepotent (3,CRE 15, binding 76, 154, 178,site 207–209,in the promoter 233). In eachregion case, of the PPARGC1AcGMP-dependent (PGC signaling-1α) resulting gene, resulting from elevated in transcriptional levels PGC-1␣ activation.mRNA and/or Increased protein increase PGC-1α as an expression adaptive enableofs nitric binding oxide to (NO) NRF (150)-1 and (Fig. subsequent 9). The NO activation induction of specificresponse target to endurance genes. At exercise the post of- varyingtranslational intensity level, and PGC-1␣ is correlated with increased mitochondrial bio- duration.+ These findings are satisfying+ in the context of PQQgenesis exposure in several can catalyze cell lines. the The oxidation induced of mitochondrial NADH to NADthe long-standing. High levels relationship of NAD enable between deacetylation endurance exer- of + PGCmass-1α by is accompaniedthe NAD -dependent by increased enzyme, oxidative Sirtuin phosphory- 1 (SIRT1),cise and allowing enhanced for mitochondrial the translocation respiratory of PGC chain-1α ex- to the nucleuslation-coupled where respiration it can then consistent bind to withNRF an-1 increase and activate in pression mitochondrial and function target (95). genes. Both NRF-1 Other and physiologic NRF-2 DNAal factorsfunctional that media mitochondriate changes (151). in PGC These-1α results expression were ob- or functionbinding activities are also were indicated elevated (i.e., along exposure with the expression to cold, tained by pharmacological increases in either NO or of several NRF target genes in rat skeletal muscle follow- fasting,cGMP, and raising energy the deprivation). question of whether Figure physiological modified with fluc- permissioning an exercise from regimen (Scarpulla, (15). Similar 2008) changes. in the PGC- tuations in endogenous NO can regulate mitochondrial 1␣-NRF pathway were mimicked in cultured myotubules content (114). Interestingly, tissue mitochondria in mice in response to increased calcium levels, suggesting that with a homozygous disruption of the gene encoding en- PGC-1␣ signaling contributes to adaptive changes in mi- / dothelial NO synthase (eNOSϪ Ϫ) were somewhat smaller tochondrial biogenesis in skeletal muscle cells (154). Ex- and less densely packed than in wild-type mice (150, 151). ercise and other neuromuscular activity may also lead to These changes were accompanied by reductions in en- the activation of the p38 mitogen-activated protein kinase ergy expenditure and mRNA levels for PGC-1␣, Tfam, and (MAPK) pathway resulting in PGC-1␣ induction through NRF-1, arguing that basal mitochondrial content is af- ATF2 and MEF2 (3). There is evidence that depletion of fected by the loss of eNOS-generated NO. Establishment ATP during exercise leads to elevated AMP/ATP ratios,

Physiol Rev • VOL 88 • APRIL 2008 • www.prv.org

Downloaded from www.physiology.org/journal/physrev by ${individualUser.givenNames} ${individualUser.surname} (142.150.147.096) on January 8, 2019.173

150

100 **

50 Cell Viability Viability Cell (% Untreated Control) 0 1 5 10 30 50 0.01 0.05 0.1 Vehicle Untreated PQQ (µM)

Figure C-2. Effect of PQQ treatment on hCMEC/D3 cell viability. A tetrazolium salts (MTT) assay was used to confirm that exposure to a wide range of PQQ concentrations (0.01-30 μM for 48 h) did not affect the viability of hCMEC/D3 cells compared to untreated or vehicle (DMSO)-treated controls. Results are presented as mean ± S.E.M. for n = 3 independent experiments. Asterisks represent data points significantly different from vehicle (**, p < 0.01). Data were generated and analyzed by C Alam in the laboratory of Dr. R Bendayan.

174

List of Relevant Publications

Peer-Reviewed Research Publications • Alam, C., & Bendayan, R. (2019). Nuclear respiratory factor 1 (NRF-1) upregulates the expression and function of reduced folate carrier (RFC) at the blood-brain barrier. Manuscript in preparation.

• Alam, C., Aufreiter, S., Georgiou C.J., Hoque, M.T., Finnell, R.H., O’Connor, D.L., Goldman, I.D., & Bendayan, R. (2019). Upregulation of reduced folate carrier by vitamin D enhances brain folate uptake in mice lacking folate receptor alpha. Proceedings of the National Academy of Sciences, 116(35): 17531-17540.

• Alam, C., Hoque, M.T., Finnell, R.H., Goldman, I.D., & Bendayan, R. (2017). Regulation of reduced folate carrier (RFC) by vitamin D receptor at the blood-brain barrier. Molecular Pharmaceutics, 14(11): 3848-3858.

Review Articles and Book Chapters • Alam, C., Kondo, M., Goldman, I.D., O’Connor, D.L., & Bendayan, R. Clinical implications of folate transport in the central nervous system. Manuscript in preparation; to be submitted to Trends in Pharmacological Sciences for publication consideration.

• Turner, A.P., Alam, C., & Bendayan, R. (2020). Chapter 1 - Efflux transporters in cancer resistance: molecular and functional characterization of P-glycoprotein. Drug efflux pumps in cancer resistance pathways: from molecular recognition and characterization to possible inhibition strategies in chemotherapy, 7: 1-30.

• Alam, C., Whyte-Allman, S.K., Omeragic, A., & Bendayan, R. (2016). Role and modulation of drug transporters in HIV-1 therapy. Advanced Drug Delivery Reviews, 103: 121-143.

Conference Abstracts • Alam, C., Aufreiter, S., Georgiou, C.J., Hoque, M.T., Finnell, R.H., O’Connor, D.L., Goldman, I.D., & Bendayan, R. (2019). Vitamin D enhances brain folate uptake in mice lacking folate receptor alpha (Folr1): a novel strategy for the treatment of cerebral folate deficiency. Poster presentation at the AAPS/IBBS workshop on Novel Approaches Targeting Brain Barriers for Effective Delivery of Therapeutics, Herndon, VA, USA.

• Alam, C., Georgiou, C., Finnell, R.H., Goldman, I.D., & Bendayan, R. (2018). Upregulation of reduced folate carrier (RFC) by vitamin D enhances folate uptake at the blood-brain barrier. Poster award winner at the 2018 CSPS Conference, Toronto, ON, Canada.

• Georgiou, C., Alam, C., Finnell, R.H., Goldman, I.D., & Bendayan, R. (2018). Characterizing the upregulation of reduced folate carrier (RFC) by vitamin D receptor at the blood-brain

175

barrier. Best undergraduate poster award winner at the 2018 Leslie Dan Faculty of Pharmacy Research Symposium for Undergraduate Students, Toronto, ON, Canada (Undergraduate Student Trainee).

• Alam, C., Hoque, M.T., Goldman, I.D., & Bendayan, R. (2017). Folate transport at the blood- brain barrier: a potential novel strategy for enhanced folate delivery into the brain. Oral presentation at the International Conference on Cerebral Vascular Biology, Melbourne, Australia.

• Alam, C., Li, M., Hoque, T., & Bendayan, R. (2016). Characterization of low affinity folate transporters at the blood-brain barrier. Poster presentation at the 2016 AAPS Annual Meeting and Exposition, Denver, CO, USA.

• Alam, C., Li, M., Hoque, T., Finnell, R.H., Steinfeld, R., & Bendayan, R. (2016). A potential role for proton-coupled folate transporter in folate uptake at the blood-brain barrier. Travelship award winner; oral presentation at the 6th International Symposium on Folate Receptors and Transporters, Breckenridge, CO, USA.

• Alam, C., Li, M., Hoque, T., & Bendayan, R. (2016). Characterization of folate (Vitamin B9) transport at the blood-brain barrier. Poster presentation at the 2016 GRIP Symposium, Toronto, ON, Canada.

• Alam, C., Bruun, T., Hoque, T., & Bendayan, R. (2015). Folate transport at the blood-brain barrier. Poster presentation at the 2015 AAPS Annual Meeting and Exposition, Orlando, FL, USA.

• Li, M., Alam, C., Hoque, T., & Bendayan, R. (2015). Characterization of low affinity folate transporters at the blood-brain barrier. Best undergraduate poster award winner at the 2015 Leslie Dan Faculty of Pharmacy Research Symposium for Undergraduate Students, Toronto, ON, Canada (Undergraduate Student Trainee).

• Alam, C., Bruun, T., Hoque, T., & Bendayan, R. (2015). Folate transport at the blood-brain barrier. Poster presentation at the 11th International Conference on Cerebral Vascular Biology, Paris, France.

176

Copyright acknowledgement for publications or figures has been indicated at the beginning of each chapter or within figure legends.

177