Decellularization and Recellularization of Porcine Livers and Design and Construction of an Automatic Machine Setup for Liver Perfusion Decellularization

Home , Decellularization

Decellularization and recellularization of porcine livers and design and construction of an automatic machine setup for liver perfusion decellularization

Manuel Lariao˜ de Almeida

Thesis to obtain the Master of Science Degree in Biological Engineering

Supervisors: Dr. Ana Margarida Pires Fernandes-Platzgummer Dr. Pedro Miguel Baptista

Examination Committee

Chairperson: Prof. Gabriel Antonio´ Amaro Monteiro Supervisor: Dr. Ana Margarida Pires Fernandes-Platzgummer Members of the Committee: Prof. Claudia´ Alexandra Martins Lobato da Silva

December 2016

Acknowledgments

I would like to thank my family for their friendship, encouragement and caring over all these years, for always being there for me and for helping me achieve all of that I wish for. I am lucky to have you all . I would also like to thank my second family, which is composed by all of my friends. Without them I wouldn’t be the person I am. I learned and still learn a lot from them everyday. It is said that friends are the one family you get to choose and with that in mind I know mine are the undeniable proof that I really have made a lot of right decisions throughout all of these years. From those that accompanied me all the way from preschool, to those that I met only in the last years, to all of you I give my heartfelt thanks. So again, thanks for helping me with all of the lows, and for being there in all of the highs, for all the tears and all of the laughs that we shared. I would also like to acknowledge my supervisors Pedro and Ana who made this all possible with their insight and support, as well as the fantastic team that received me so well in IACS. They made this journey all the more pleasant and I cannot imagine going through it without them. Even though I was not there for that long you have definitely deserved a special spot in my heart. Thanks for everything and I will see you in another Juepincho for sure! As a final note I would like to thank everyone who contributed for my personal growth throughout the years. I was lucky enough to have some awesome teachers that instilled in me the desire to always know more. Without them I would have never gotten here and so I thank you for tending the flames that drive me in this quest for knowledge. To each and every one of you – Thank you.

Abstract

Liver transplantation remains to date, the only option for patients with end-stage liver disease (ESLD). Several bioengineering approaches have been developed over the years to increase the number of options for these patients, mostly failing due to lack of a vascular network. This work aimed to recellularize the liver vasculature in decellularized porcine livers in order to solve the aforementioned problem. Porcine livers were decellularized through detergent perfusion at constant pressure or ﬂow rate, with results that show the possibility of decellularization without using strong detergents such as sodium dodecyl sulfate (SDS), while seemingly conserving the microarchitecture of the liver extracellular matrix (ECM). The results obtained also suggest that efforts to stabilize the liver in the decellularization process are rewarded with better results. The decellularized liver scaffolds (DLSs) were then recellularized with human mesenchymal stem cells (hMSCs) and human umbilical vein endothelial cells (hUVECs), in a perfusion setup with pressure control. Recellularization results were lacking and, as such, no conclusions were taken at the moment as further testing is needed. In an effort to circumvent some issues related with the decellularization protocol, the construction of a decellularization machine for automating this process has been initiated. As it is still unﬁnished, no decellularization was performed in this setup and no results were collected at the present stage.

Keywords

tissue engineering; regenerative medicine; liver bioengineering; decellularization; recellularization; decellularized liver scaffold

iii

Resumo

Actualmente, o transplante hepatico´ continua a ser a unica´ opçao˜ para pacientes com doenças do foro hepatico´ em estado terminal (ESLD). Varias´ estrategias´ ao n´ıvel da bioengenharia foram desenvolvi- das ao longo dos anos de modo a criar mais opçoes˜ para estes pacientes, sendo que na sua maioria falharam devido a` falta de uma rede vascular. Este trabalho teve como objectivo recelularizar a vas- culatura hepatica´ em f´ıgados descelularizados de porco de modo a resolver o problema mencionado. Os f´ıgados foram descelularizados atraves´ de perfusao˜ de detergentes a pressao˜ ou caudal constante, com resultados indicativos da possibilidade de descelularizaçao˜ evitando a adiçao˜ de detergentes fortes como o dodecil sulfato de sodio´ (SDS), aparentemente mantendo a microarquitetura da matriz ex- tracelular (ECM) do f´ıgado. Os resultados sugerem tambem´ que a estabilizaçao˜ do f´ıgado durante o processo de descelularizaçao˜ promove uma melhoria dos resultados. Posteriormente os scaffolds de f´ıgados descelularizados (DLSs) foram recelularizados com celulas´ estaminais mesenquimais (hMSCs) e celulas´ endoteliais da veia umbilical (hUVECs), por perfusao˜ a pressao˜ controlada. Os resultados recolhidos das recelularizaçoes˜ nao˜ foram, ate´ ja,´ suficientes para tirar conclusoes˜ sendo necessaria´ a realizaçao˜ de mais testes. De modo a evitar alguns dos problemas relacionados com o protocolo de descelularizaçao,˜ começou-se o desenvolvimento de uma maquina´ de modo a automatizar este processo. Uma vez que esta se encontra ainda por terminar ainda nao˜ foram feitas descelularizaçoes˜ neste contexto e por isso nao˜ existem ainda resultados a apresentar neste topico.´

Palavras Chave

engenharia de tecidos; medicina regenerativa; bioengenharia do f´ıgado; descelularizac¸ao;˜ recelularizac¸ao;˜ scaffold de f´ıgado descelularizado

Contents

1 Introduction 1 1.1 The liver...... 3 1.1.1 Structure...... 3 1.1.2 Cells...... 4 1.1.3 Function...... 6 1.2 Liver bioengineering...... 7 1.2.1 Spheroids and Organoids...... 7 1.2.2 Liver tissue engineering...... 9 1.2.3 Liver bioengineering...... 11 1.2.3.1 Scaffold-based strategies...... 11 1.2.3.1.1 Decellularized scaffolds...... 11 1.2.3.1.1.1 Automatic decellularization machines...... 12 1.2.3.1.2 Man-made scaffolds...... 14 1.2.3.1.3 Recellularization of scaffolds...... 14 1.2.3.2 Other strategies...... 15 1.2.4 Future applications...... 15

2 Aim of Studies 17 2.1 Aim of Studies...... 19

3 Materials and Methods 21 3.1 Cell preparation...... 23 3.1.1 Cell lines...... 23 3.1.1.1 Isolation...... 23 3.1.1.1.1 Isolation protocol for hMSCs from lipoaspirate samples..... 23 3.1.1.1.2 Isolation protocol for pig mesenchymal stem cell (pMSC) from pig femur bone marrow...... 24 3.1.1.1.3 Isolation protocol for hUVEC and pig umbilical vein endothelial cell (pUVEC) from umbilical cords...... 24

vii 3.1.1.2 Culture media...... 25 3.1.1.3 Cell culture...... 26 3.1.1.4 Cell cryopreservation...... 27 3.1.1.5 Cell thawing...... 27 3.1.2 Cell tagging...... 27 3.1.2.1 Tag preparation...... 27 3.1.2.2 Transfection...... 28 3.1.2.3 Puriﬁcation...... 28 3.2 Liver preparation...... 29 3.2.1 Liver harvesting...... 29 3.3 Decellularization...... 29 3.3.1 Setup assembly...... 29 3.3.1.1 Material...... 29 3.3.1.2 Assembly...... 30 3.3.2 Decellularization...... 30 3.4 Recellularization...... 31 3.4.1 Preparation...... 31 3.4.1.1 Scaffold sterilization...... 31 3.4.1.2 Cell seed expansion...... 31 3.4.2 Setup assembly...... 32 3.4.2.1 Material...... 32 3.4.2.2 Assembly and priming...... 33 3.4.3 Seeding...... 36 3.4.4 Control Assays...... 37 3.5 Quality testing...... 38 3.5.1 Microscopy Staining...... 38 3.6 Decellularization machine...... 38 3.6.1 Requirements...... 38 3.6.2 Control...... 39 3.6.2.1 Parameters...... 39 3.6.2.2 Planed operation modes...... 40 3.6.2.3 System description and components...... 42

4 Results 45 4.1 Decellularization...... 47 4.1.1 Liver preparation...... 47

viii 4.1.2 Decellularization...... 48 4.1.3 Decellularization machine...... 50 4.2 Recellularization...... 51 4.2.1 Cell seed expansion...... 51 4.2.2 Recellularization...... 51 4.3 Others...... 55

5 Conclusion and Future Work 57 5.1 Conclusion...... 59 5.2 Future Work...... 59

A Medium Formulations 69

B Vessel Design 73

ix x List of Figures

1.1 Diagramatic representation of the liver structure, showing the lobule and the vasculature.3 1.2 Flow pattern inside the liver lobules...... 4 1.3 Cellular population of the hepatic sinusoids and surrounding tissues...... 5 1.4 The metabolic functions of the liver and the ﬂow of products through the portal vein, hepatic artery, inferior vena cava and common bile duct...... 6 1.5 The spatial relationship among the different cell types of the liver...... 10 1.6 Comparison of automated and manual decellularization method in the lung case...... 13

3.1 Scheme of the decellularization setup...... 30 3.2 Recellularization setup...... 35 3.3 Bioreactor vessel in different stress conditions...... 37 3.4 Example interfaces for the automatic decellularization machine...... 42 3.5 Scheme of the current design of the decellularization machine setup...... 43

4.1 Decellularized liver scaffold which was partially hepatectomized prior to decellularization. 47 4.2 Decellularized liver scaffold which was ligated prior to decellularization...... 48 4.3 Livers which were considered well decellularized...... 49 4.4 Liver support structures...... 50 4.5 Recellularization results...... 53 4.6 Hematoxylin and eosin stains of the recellularized scaffolds...... 55 4.7 Adipocyte differentiation assay...... 56

B.1 Representation of the complete vessel without the front panel...... 74 B.2 Representation of the complete vessel without the lid and cut through the center of the front support columns...... 74 B.3 Technical drawing of the upper shelf...... 75 B.4 Technical drawing of the lower shelf...... 75

xi xii List of Tables

3.1 Composition of DMEM/F-12++ ...... 25 3.2 Composition of hUVEC medium...... 25 3.3 Composition of the bioreactor medium...... 26

A.1 Complete formulation for DMEM/F-12...... 70 A.2 Complete formulation for DMEM/F-12 (high glucose)...... 70 A.3 Complete formulation for MCDB 131...... 71

xiii xiv Acronyms

2D two dimensional

3D three dimensional

ABS artiﬁcial biomimetic scaffold

DAPI 4’,6-diamidino-2-phenylindole dH2O distilled water

DLS decellularized liver scaffold

DMEM/F-12 Dulbecco’s modiﬁed Eagle medium/Ham’s nutrient mixture F-12

DMEM/F-12++ Dulbecco’s modiﬁed Eagle medium/Ham’s nutrient mixture F-12++

DMSO dimethylsulfoxide

DNA deoxyribonucleic acid dsDNA double stranded deoxyribonucleic acid

ECM extracellular matrix

EDTA ethylenediamine tetraacetic acid

EGF epidermal growth factor

EGTA ethylene glycol-bis( β-aminoethyl ether)-N,N,N’,N’-tetraacetic acid

ESC embryonic stem cell

ESLD end-stage liver disease

EC endothelial cell

EtOH ethylene oxide

xv FACS ﬂuorescence-activated cell sorting

FBS fetal bovine serum

FGF-2 ﬁbroblast growth factor 2

GAG glycosaminoglycan

GFP green ﬂuorescent protein

GUI graphical user interface

HA hepatic artery

H&E hematoxylin and eosin hMSC human mesenchymal stem cell hUVEC human umbilical vein endothelial cell

HV hepatic vein

IGF-1 insulin-like growth factor type 1 iPSC induced pluripotent stem cell iPSC-HE induced pluripotent stem cell-derived hepatic endoderm cell

IR infrared

IVC infrahepatic vena cava

LDH lactate dehydrogenase

LED light-emission diode

LSEC liver sinusoidal endothelial cell

MCDB 131 Molecular, Cellular and Developmental Biology 131 medium

MSC mesenchymal stem cell

NK natural killer

ORCA Organ Regenerative Control Acquisition

PAA peracetic acid

PBS phosphate buffer saline

xvi PEDOT 3,4-ethylenedioxythiophene

PEG polyethylene glycol

PGA polyglycolic acid

PID proportional–integral–derivative pMSC pig mesenchymal stem cell

P/S penicillin/streptomycin pUVEC pig umbilical vein endothelial cell

PV portal vein

SDS sodium dodecyl sulfate

SMC smooth muscle cell

SVC suprahepatic vena cava

TOM tdTomato

UV ultraviolet

VEGF vascular endothelial growth factor

VC vena cava

xvii xviii 1 Introduction

Contents

1.1 The liver...... 3

1.2 Liver bioengineering...... 7

1 2 1.1 The liver

As the target organ of this thesis, it is important to know what constitutes the liver, how it is structured and how it works.

1.1.1 Structure

The liver is a solid organ that belongs to the digestive system. It is composed by many cell types that are organized in a very deﬁned structure. It is anatomically divided into four lobes, and has a complex vascular system.

Figure 1.1: Diagramatic representation of the liver structure, showing the lobule and the vasculature. In this ﬁgure, the repetitive, modular microarchitecture of the liver is observable, as well as the highly structured vascular network. Retrieved from Martini, Frederic; Nath, Judi; Bartholomew, Edwin. Fundamentals of Anatomy & Physiology, 10th Edition (Pearson, 2015) viewed in http://slideplayer.com/slide/7789218/ at October 28th, 2016.

The internal structure of the liver is made of functional units known as lobules, which resemble hexagons where the hepatic vein (HV) sits in the center, surrounded by six portal triads (also called portal areas) composed of the hepatic artery (HA), the portal vein (PV) and the bile duct, as seen in Figure 1.1. Between them, a mesh of thin-walled hepatic capillary-like microvessels called sinusoids extend from thePVs andHAs to meet theHV in the center, connecting them. In these sinusoids, the oxygen-rich arterial blood (20-40% of total) that ﬂows through theHA and the nutrient-rich venous blood (60-80% of total) that comes from the gastrointestinal tract to the liver, through thePV are mixed together, resulting in a specially rich blood with a distinct ﬂow pattern that allows the liver to carry out all

3 of its functions [1]. After mixing in theHV, the blood eventually drains to the vena cava (VC), serving the HV as the outlet for the blood the enters through both inlets (PV andHA) as presented in Figure 1.2.

Figure 1.2: Flow pattern inside the liver lobules. In this figure a longitudinal cut of a liver lobule and its associated vasculature is represented allowing the visualization of the blood flow and bile flow directions. Retrieved from British Medical Association Complete Home Medical Guide (DK, 2010) viewed in www.aviva.co.uk at November 4th, 2016.

Also important in the hepatic vasculature, the biliary tract that has a similar vascular tree to the circulatory one. In this system, the thin microvessels are called bile canaliculi and they are the recipients of the bile generated in the liver. These canaliculi run parallelly to the sinusoids carrying bile directly to the bile ducts in the opposite direction of what happens with the blood stream, i.e. from the center of the lobule to the bile ducts in the portal triads. Eventually, the bile reaches the common bile duct and then meets the bile from the cystic duct before being released into the duodenum. Apart from the vasculature, the liver is also composed by metabolic units, that make up for the majority of the liver mass. These units are primarily composed of hepatocytes that mainly organize into plates. These are separated from the blood supply by a thin layer of fenestrated liver sinusoidal endothelial cells (LSECs) and stellate cells.

1.1.2 Cells

As mentioned before, the main cellular type found the liver is the hepatocyte. These cells are are cuboidal epithelial cells that line the sinusoids and make up the majority of cells in the liver. Hepatocytes perform most of the metabolic functions of the liver. Separating the hepatocytes from the sinusoidal lumen, the LSECs make up the lining of this fenestrated wall. They play a role on keeping the liver homeostasis [2] and are believed to be key players in the process of liver regeneration [2–5], along with stellate cells.

4 Figure 1.3: Cellular population of the hepatic sinusoids and surrounding tissues. In this ﬁgure an hepatic sinusoid is schematized and the main cell types of this liver microarchitecture represented. Adapted from Thomson et al. [1]

Stellate cells are, in turn, the predominant extracellular matrix (ECM) molecule producers, synthesiz- ing a plethora of components including collagen, proteoglycan, and adhesive glycoproteins. They are also involved in the production of ECM degrading metalloproteases in hepatic parenchyma, all of which shows that they are a central player when it comes to ECM remodeling in both normal and pathological conditions. [6,7]

Moving into the sinusoidal lumen, a number of different cell types can be found. From these, some relevance can be given to Kupffer cells which comprise the largest population of resident tissue macrophages in the body at about 80-90% of the total number. [8,9] They are responsible for for breaking down and recycling old, worn out red blood cells passing through the sinusoids, as well as for phagocytizing pathogens, particulates and other immunoreactive material entering from the portal or arterial circulation. Due to that, they can be considered the ﬁnal component in gut barrier function, and are the main line of defense that prevents the passage of all gut-derived substances past the hepatic sinusoids. [8]

Additionally to the Kupffer cells, dendritic cells, natural killer (NK) cells, NKT cells and lymphocytes populate the sinusoidal lumen (Figure 1.3).

In regards to the vasculature, the circulatory tree is lined with several layers called tunicae where ECM molecules like collagen and elastin are meshed with ﬁbroblasts, smooth muscle cells (SMCs) and endothelial cells (ECs) in a very deﬁned manner. As for the biliary tract the resident lining cells are called cholangiocytes.

Starting by the circulatory vasculature, the inner layer (tunica intima) is made of a thin sheet of ECs and connective tissue, the medial layer (tunica media) is composed by concentric layers of SMCs, collagen and elastin and the outer layer (tunica adventitia) is made mainly out of collagen and elastic ﬁbers, and is populated mainly by ﬁbroblasts. Depending on the exact type of vessel, the proportions

5 between the different layers and sometimes the constitution may vary.

The cholangiocytes are a heterogeneous, highly dynamic population of epithelial cells, derived from the hepatoblasts, responsible by the modiﬁcation of hepatic canalicular bile as it is transported along the biliary tree. [10]

1.1.3 Function

Overall the liver has a detoxifying and metabolic function in the human body, being a key player in maintaining homeostasis of the human body.

It is responsible for the metabolism of carbohydrates, lipids and proteins and the control of the levels of monomers of said compounds.

It is also the main hub for the degradation of toxins and waste products. Due to that it is also one of the most important organs involved in the metabolism of drugs, and as such of extreme interest in the pharmaceutical industry.

It is responsible for maintaining the quality of the red blood cells, as mentioned before and it is also where most hormones are broken down.

Additionally, it is in the liver that the bile, cholesterol and most plasma proteins are produced, while in its fetal form it is the main site for erythropoiesis.

Figure 1.4: The metabolic functions of the liver and the ﬂow of products through the portal vein, hepatic artery, inferior vena cava and common bile duct. Adapted from Thomson et al. [1]

6 1.2 Liver bioengineering

It is also worth mentioning how has this speciﬁc area of bioengineering evolved. For that, an overview of the progression of liver engineering is presented.

1.2.1 Spheroids and Organoids

In 1961, Moscona et al. [11] described how from individual embryonic stem cells (ESCs) it is possible to generate in vitro tissue-like constructions under standard controlled conditions. The term ”aggregation pattern” began being used to describe the capacity of some cell types to give rise to aggregates within 24 hour if in the right conditions. Despite this, it was not until the 1980s, that the word spheroids1 started being used by Landry et al. [12] to describe these 3D cellular aggregates. In this work, when liver cells isolated from rats were prevented from attaching to a solid surface, they re-aggregated to form structures very similar to those found in vivo, where cells produce their own ECM and hepatocytes can preserve their metabolic functions [12]. When the case is that of primary hepatocytes that are capable of forming these structures, they may also be called hepatospheres. In these the majority of cell-cell and cell-ECM attachments are preserved, being this essential to maintain both hepatic differentiation and functionality [13]. Studies in primary rat hepatocyte spheroids have demonstrated that they are able to recreate the microanatomy of the liver [14]. When considering spheroids, it has been proved that one of the most important aspects in their formation is size [15]. Glickis et al. [16] found that there is a decline of cell viability with increasing spheroid size. Based on their observations, they created a mathematical model and predicted that the optimal hepatocyte spheroids size was 100µm in diameter, which would maintain cell viability above 90%. As for hepatocyte spheroids larger than that, the mortality rate would increase rapidly, being that around 250µm the mortality rate would have already escalated to 50%. This relationship between mortality and spheroid size can be explained by the hypoxic conditions the cells on the core of the spheroid are subject to, due to poor oxygen diffusion and lack of any vascular system that could supply oxygen to those cells. This occurrence is normally referred to as necrotic core or necrotic center. The formation of these spheroids can be explained when the fact that these cells are anchorage- dependent is kept in mind. As these cells need to adhere, if they have no more space to expand into, then they adhere to one another, which triggers an incessant chain reaction of adhesions that culminates by giving rise to the spheroid. As a ﬁrst step, the spheroids are stimulated by integrin-ECM binding originating small cell aggregates. These multiple multicellular aggregates give rise to a spheroid

1Spheroids are spontaneous, spherical non-adherent aggregates of cells that form a 3D tissue construct in culture retaining tissue-speciﬁc functions.

7 via cadherin-cadherin interactions. Hence, spheroid assembly represents the most energy efﬁcient structure by minimizing their surface.

The key to spheroid formation is to discover a reproducible protocol capable of rebuilding, in the case of the liver, the hepatic tissue. Presently, several techniques are capable of achieving this. These include non-adherent dishes under static conditions, agitation cultures and hanging drops.

The simplest way is to seed the hepatocytes in a low adherent well. After an initial attachment to the surface, the hepatocytes give rise to a monolayer that little by little separates from the dish forming spheroids. In addition, different conditions like uncoated plates with a positive surface charge, coated dishes with albumin, or the simple elimination of serum factors have been demonstrated to be use- ful in spheroid formation [17]. By contrast, coatings with collagens, ﬁbronectin, laminin or other ECM molecules inhibit spheroid formation since they support hepatocyte adhesion.

Besides stationary conditions, agitation cultures such as rocked and rotary cultures in tissue culture plates or bioreactors have been developed in order to improve spheroid formation. One example of this was the development of an innovative bioreactor in 2005 that rapidly gives rise to spheroids when loaded with porcine hepatocytes [17]. Compared with monolayer cultures, hepatocyte spheroids from this bioreactor showed less cell death and increased metabolic functions [17]. Though, recently it has been demonstrated that rocked cultures increase spheroid formation due to increment of the number of times hepatocytes clash compared with rotary cultures [18].

However, all techniques described above have several drawbacks, among them it is important to mention the necessity of manually selecting homogeneous populations, since aggregates with irregular geometry are generally obtained.

Kelm et al. [19] described an universal method to form hanging drops applicable to a lot of cell lines. This culture method consists of a small amount of cells in suspension seeded upside down in the lid of a culture dish. The hepatospheres formed have high size reproducibility with variations lower than 10% of the average value.

As mentioned before, due to oxygen demands, it is extremely important to construct a functional vascular network, being that one of the biggest challenges in tissue engineering. For this reason, the introduction of endothelial cells in the hepatocyte spheroids production has emerged as a possible solution [20], not only due to the intended need of angiogenesis, but also to increase cell functionality by adding a non-parenchymal cell population.

Additionally,as mentioned in subsection 1.1.1, stellate cells also have an important role in regeneration after liver injury as they secrete ECM components like laminin which in is needed for the formation of structures like the hepatic sinusoids. Hence, spheroids formed by hepatocytes and stellate cells are also an interesting in vitro system that harbors great potential in drug discovery and many other applications [21].

8 The search for more biologically meaningful systems is making scientists more aware of the importance of these hepatic non-parenchymal cell populations when assembling these cellular structures, as usually these populations are the biological solution to most problems faced by systems that do not include these cells.

1.2.2 Liver tissue engineering

Up to date, the treatment of a huge variety of end-stage liver disease (ESLD) can only be achieved through liver transplantation. Due to the shortage of liver donors, hepatic tissue engineering has become a promising strategy for the treatment of these patients. In this quest, a large effort has been dedicated to the development of suitable supporting biomaterials that mimic the liver ECM and that allow for reliable cell growth, maintenance of their differentiation and metabolic functions, hepatic tissue organization and microarchitecture, and the mechanical and biological properties observed in vivo. As mentioned above, it is fundamental to identify the most appropriate cells, in order to recapitulate in vitro the natural liver microarchitecture. Under speciﬁc stimuli, these cells should interact with neighbor- ing cells and the ECM/ECM-mimic and form liver parenchymal tissue, which could then be transplanted into patients to repair damaged tissue and increase liver function. From this point of view, these tissue engineered liver constructs are also excellent biological surrogates for a myriad of biomedical and pharmaceutical applications. Up to date, different types of biomaterials have been used, being that alginate-based hydrogels con- stitute one of the most widely used in tissue engineering due to its hydrophilic properties, porosity, weak adhesive properties and exceptional tissue compatibility. Some studies suggest that alginate hydrogels loaded with hepatocytes or mesenchymal stem cells (MSCs) increase the survival of animal models with 70-80% partial hepatectomy [22, 23]. Lin et al. [24] also reported that bone marrow-derived MSCs can be differentiated into hepatocyte-like cells. In this scenario, alginate hydrogels could also be used not for direct implantation, but for encapsulation of these cells, which represents a new source of hepatic cells required for liver tissue engineering, alongside options like ESCs and induced pluripotent stem cells (iPSCs). Additionally, they can also be used to induce hepatocyte differentiation in vitro [25]. Another biomaterial commonly used to produce scaffolds is chitosan, which is a linear amino hetero- polysaccharide derived from chitin with interesting characteristics for tissue engineering applications. These include its low cytotoxicity, high biocompatibility and high biodegradability, with the added bonus of having a structure that is very similar to the glycosaminoglycans (GAGs) present in the liver ECM. Examples of chitosan applications include the study of Shang et al. [26] where they were able to build a hybrid sponge made of galactosylated chitosan and hyaluronic acid mimicking the hepatic microenvironment and seeded hepatocytes andECs. Also a relevant biomaterial, type I collagen has also been used extensively for hepatocyte in vitro

9 models. Because these cells lose their differentiated functions in 2D cultures, mainly due to the loss of orientation needed for these polar cells, 3D supporting culture conditions are needed to maintain hepatic function. Such conditions can be found in collagen sandwich constructs, which consists of culturing the cells between two layers of gelled collagen, serving this protein as matrix for cell attachment allowing hepatic polarity and maintenance of their differentiated functions [27]. This is due to the capacity of sandwiches to mimic liver microenvironment, promoting cell-cell and cell-ECM interactions [28], and for allowing the hepatocytes to keep correct orientation patterns (Figure 1.5), respecting their natural polarities, which in turn maintains their functions [29] and makes the formation of microstructures such as the bile canaliculi possible. Because of this, collagen sandwiches have proven to be a good tool for the pharmaceutical industry as a liver tissue model [30, 31].

Figure 1.5: The spatial relationship among the different cell types of the liver. In this ﬁgure hepatocytes and hepatic sinusoid are schematized giving some detail on the involvement of cell-cell interactions between hepatocytes in the formation of bile canaliculi. Additionally, the polarity of said cells is represented in regards to their relative orientation towards the sinusoids or other hepatocytes. It is important to note that, despite not being represented in this ﬁgure, since most hepatocytes are organized into plates/sheets, the pattern is repeated in a way that the faces of said sheet have only basolateral membranes while sideways, to form the sheet, all cells display the apical membrane, giving rise to the bile canaliculi. Retrieved from http://clinicalgate.com/liver-physiology-and-energy-metabolism/ at November 5th, 2016.

In other studies, like Melgar-Lesmes et al. [32], people have used collagen constructs to seed endothelial cells. These matrices were then transplanted into living animals, showing liver damage repara- tion, suggesting that endothelial cells play a critical role in hepatic repair. Ranucci et al. [33] bet on the utilization of void size collagen foams to induce rat hepatocyte differentiation, suggesting that pore sizes of the substrate (collagen I in this case) is quite relevant for selective cell morphogenesis. Hyaluronic acid is yet another example. It consists of one of the main components of the ECM, and as most, plays an important role in cell proliferation and migration. It is also commonly used for liver

10 tissue engineering as scaffold for cell growth [34, 35]. Not only naturally-derived materials have been used in liver tissue engineering efforts. Due to histor- ical reasons, polyglycolic acid (PGA) scaffolds have been used extensively at the beginning of this ﬁeld of knowledge to generate hepatic tissue when seeded with primary hepatocytes, showing some albumin and urea secretion capability [36, 37]. Hydroxyapatite [38] and polyethylene glycol (PEG)[39] have also been used to create hepatic tissue constructs. More complex composite biomaterials have been also designed. For instance, combinations of some of the compounds described above (chitosan, gelatin, type I collagen and hyaluronic acid) together with conducting polymers such as 3,4-ethylenedioxythiophene (PEDOT) have also been used [40]. The rationale behind the use of this polymer lies with the creation of local electrical ﬁelds inside of the scaffold so as to improve cell attachment, proliferation and protein expression of the seeded cells. Recently, another innovative technique is to use acellular matrix derived from cells in culture. Kanni- nen et al. [41] demonstrated that after seeding iPSCs in a HepaRG-derived acellular matrix, this matrix induced hepatic commitment of the iPSCs, suggesting the importance of HepaRG acellular matrix in hepatic differentiation and maturation. Tiwari et al. [42] also used this type of acellular scaffold, in this case to expand hematopoietic stem/progenitor cells. Finally, through the multitude of strategies chosen by the multiple authors described above, most of the generated hepatic tissues reported have shown some degree of functionality, either in transplantation or in vitro assays. Despite all of this, the end goal of generating reliable hepatic tissue in vitro with the complexity observed in vivo is still distant in most of the presented cases.

1.2.3 Liver bioengineering

In the past years, organ bioengineering has ﬂourished and several techniques have proved to be suitable candidates for the job at hand. Most strategies have relied on scaffolds with increasing complexity in order to better replicate the liver microarchitecture and niche, but there is also some work done in scaffold-free organogenesis focused approaches. More recently, there has been an alternative to the approaches mentioned above, in which instead of trying to produce liver tissue or liver niche cell cultures and co-cultures, there has been a shift towards the more complex alternative of bioengineering whole livers or physiologically relevant liver structures like liver lobes, liver buds or liver vasculature and ducts.

1.2.3.1 Scaffold-based strategies

1.2.3.1.1 Decellularized scaffolds

11 So far, the most widely described technique for liver bioengineering is the use of decellularized livers as scaffold for liver regeneration. In this technique the rationale is that if the ECM remains in a good condition after decellularization, then the different components of the ECM will guide the different cell types into their proper positions and serve as anchors for them to attach, aiding not only in the actual attachment but also in the formation of the various structures that characterize the liver. This kind of approach is based on a two-step process, the decellularization step and the recellularization step. For the first step, the decellularization, as the objective is to remove all cells and cellular material, protocols are based on the combination of various cell-damaging factors such as freezing/thawing cycles, isotonic stress, enzymes, shear stress and chemical action being that the chemical action of detergent solutions is the most widely described for dense non-hollow organs such as the liver. These solutions are always perfused throughout the organ vasculature in order to detach the cellular material from the ECM so that only the structured ECM remains. As far as these solutions go, there is a tendency for the use of detergent solutions such as Triton X-100 and sodium dodecyl sulfate (SDS)[43–47], but there have been also various other papers using solutions ranging from enzymes such as trypsin to chelating agents such as EDTA or EGTA[48, 49]. Additionally there have been successful attempts while using as inlet theVC[50], thePV[51] and theHA[47], as well as using fixed flow [45], fixed pressure [46] or even oscillating pressure conditions [47, 52], even existing video articles on the subject [52]. As for validation of the process, even though there is not any standard quantitative definition of decellularization, Crapo et al. [53] suggests that minimal criteria that would suffice in order to classify a tissue as decellularized would be the ones which yield results where there is constructive remodeling response and adverse cell and host responses are avoided. For that, in this review three criteria are presented, these being:

1. <50 ng dsDNA per mg ECM dry weight;

2. <200 bp DNA fragment length;

3. Lack of visible nuclear material in tissue sections stained with DAPI or hematoxylin and eosin (H&E).

1.2.3.1.1.1 Automatic decellularization machines

Currently, the best option on the market for automatic decellularization of solid organs is the Organ Regenerative Control Acquisition (ORCA) bioreactor and system from Harvard Apparatus, which mon- itors and records a number of parameters important for the control of the decellularization process.

Some examples of these monitored parameters include pressure, temperature, pH, CO2 and CO2 concentrations. The ORCA system offers high reliability and versatility for both large and small organs, the

12 trade-off being the fact that the system is not portable and very expensive, as the complete setup can cost upwards of 30,000C. Additionally, this setup does not have any notiﬁcation system, nor any way to stop according to decellularization status even though it is equipped with microscope cameras which constantly monitor the liver in theIR, visible andUV spectra. Even though there is not much literature on the subject, some examples can be found in other orgnas and structures such as the lungs [54] and arteries [55]. The lung example, published by Price et al. [54], clearly shows some of the advantages of the transition to automatic systems. Not only did this help in terms of astonishingly reducing the complexity of the process since no longer must the different steps be done manually (which is also a big source of variability) but it also reduced the time needed for the decellularization process (Figure 1.6).

Figure 1.6: Comparison of automated and manual decellularization method in the lung case. From Price et al. [54]

Another advantage of automatic machines is that they can potentially elicit the scaling of the process giving it medical relevance. Additional work of relevance includes the one done at Worcester Polytechnic Institute, where devices for the decellularization of hearts and vascular structures were also developed as low-cost options.

13 [56, 57]

1.2.3.1.2 Man-made scaffolds

A very similar approach to the one mentioned before relies on the substitution of the decellularized liver scaffold (DLS) with an artificial biomimetic scaffold (ABS). By using an ABS, some issues pertaining the use of an animal-derived scaffold can be eliminated, such as the possibility of transmission of zoonotic diseases and the vast ethical constraints associated with products of animal origin. As the objective is to create a structured microenvironment that resembles the natural liver ECM in which the cells are able to generate functional, structured and vascularized liver tissue, the artificial scaffold has to provide both support and a plethora of different cues in order to direct the cells towards the desired goal, liver organogenesis. So as to do that, a biomimicry approach reliant on biofabrication techniques such as 3D printing, electrospinning and other methodologies, can be used either to generate the artificial scaffold which can then be seeded with the desired cell types or to generate an already seeded scaffold/tissue if the cells are present in the printing solution (bioprinting) [58–62].

1.2.3.1.3 Recellularization of scaffolds

Excluding the artiﬁcial bioprinted scaffolds (which are seeded in the fabrication process), both the natural and the artiﬁcial scaffolds must then be seeded. This process is called recellularization in the case of the DLS seeding, as said scaffolds are effectively being returned cells to substitute those that were removed. In order to recellularize a tissue, one must repopulate it either with all the cell types that naturally populate said tissue or repopulate it with progenitors of said cells and then induce some degree of expansion and differentiation in order to replicate as close as possible the natural tissue. As discussed in section 1.1, the liver is composed of a large amount of cell types, organized in a very structured environment. In that environment, one can clearly distinguish various different structures, with different cell types. As such, through perfusion, hepatocytes, stellate cells, endothelial cells and various other cells can be used to repopulate the obtained scaffold. Some examples include the repopulation with mesenchymal and endothelial cell lines for vascular regeneration [63] or the repopulation with hepatic cell lines for metabolic, viability or functional assessments [44, 64]. Another option deserving further thought is the recellularization with more primitive versions of said cells, such as hepatoblasts, from which both the hepatocytes and the cholangyocytes are derived. Strategies like this could potentially simplify the seeding strategies, by reducing the complexity involved with expanding multiple cell lines and mixing them in the appropriate proportions for seeding.

14 1.2.3.2 Other strategies

A different tactic when compared to the previous ones is to rely on the multipotency of progenitor cells and their ability of self-organizing into complex structures. In this way, through the study of developmental biology, protocols could be designed to mimic the natural conditions that lead to liver organogenesis. So far, the most relevant example is the liver bud experiment [65, 66], in which by controlling the set of conditions to which a 2D co-culture of human mesenchymal stem cells (hMSCs) and human umbilical vein endothelial cells (hUVECs) and induced pluripotent stem cell-derived hepatic endoderm cell (iPSC-HE) is exposed to, this culture contracts into a 3D budding structure reminiscent of a liver. It is worth mentioning that in this trial, the liver bud was able to form a vascular network and rapidly emulated an adult liver when connected to a working vasculature (when transplanted). Additionally it has been shown to be functional as it helped rescue drug-induced lethal liver failure models. Other related experiments, pertain to the creation of hepatic/liver organoids from hepatic cell lines that even without having all of the deﬁned liver structure, still show function and regenerative capabilities, such as the ability to generate new bile ducts or the actual organoid development from a single cell [67]. Liver organoids have been obtained also by using hydrogels in microﬂuidic settings [68] and similar results like the development of bile duct have also been obtained through encapsulation in alginate [69].

1.2.4 Future applications

Regarding applications of liver bioengineering strategies, the most obvious (and most distant) one is transplantation. When considering this target, one must consider all of the safety concerns common for all medical applications and for that each different technique shows its strengths. When thinking of an off-the-shelf product for transplantation the fastest option would be to already have a full liver ready for transplant, which is incompatible with an autologous and theoretically safer transplant. This fact leads to two routes, one being the production of sterile scaffolds ready to be seeded with the patient’s cells in order to generate a functional, transplantable liver or the production of allogenic bioengineered livers. In terms of immunogenicity the main concern are the cells since natural scaffolds seem to trigger low or moderate immune responses [70] existing options to further reduce said issue [71]. Furthermore, in the long term, they are completely substituted by the patient’s own ECM (according to Faulk et al. [72] scaffolds are bioabsorbed completely in 90 days) and artiﬁcial scaffolds meant for this aim will have to be designed and prepared so that they pass all quality controls for medical use. With that said, two distinct options arise when considering the recellularization of the scaffolds. On one hand, the scaffolds could be seeded with user speciﬁc iPSC-derived cells, which would have the advantage of theoretically not triggering any immune response in the patient, the trade-off being the risks associated with the possible contamination of the liver with highly primitive cells that could cause teratomas. On the other hand, adult

15 stem cells from donors or a cell bank could be used, which would avoid both the ethical constraints of using ESCs/iPSCs and the possible medical related issues. This option however would not produce a perfect match for the patient in terms of immune response, which would mean that the patient would have to use immune suppressants throughout their entire lives, such as what happens with patients that receive healthy livers as transplant. Other more easily achieved, and just as important applications, include disease modeling, and drug testing. For these applications the liver buds appear to be very good candidates as if the study is about a disease that causes liver malformation it can be followed during the development of the liver bud. Other diseases can also be easily followed and studied over time with this strategy. Drugs could also be tested in this system to study not only the normal parameters such as efﬁcacy/toxicity and drug metabolism, but also the effect of drugs in liver organogenesis. Other liver bioengineering strategies like biofabrication [58, 73] and liver organoids [68] can also be used for this kind of applications.

16 2 Aim of Studies

Contents

2.1 Aim of Studies...... 19

17 18 2.1 Aim of Studies

As the primary ”waste treatment facility” in the human body, the liver is especially susceptible to damage caused by excessive intake of most kinds of toxins in our bodies. Adding to that, the liver is a target of diseases such as hepatitis B and C. These reasons among others place patients of liver-related conditions are the second largest group of people waiting for an organ in waiting lists, all over the world. This thesis is focused on the potential of tissue engineering as a tool for biomedical applications such as regenerative medicine. More precisely, studies are made in a strategy for the regeneration of livers fit for transplantation or to help patients on the waiting list survive until they are called to receive a donated one. This project is inserted in a larger one where it is expected that an alternative to donated livers is achieved so that the pool of livers available for transplantation is expanded, hopefully reducing the waiting lists for liver transplantation and saving/enhancing people’s lives. For this, DLSs are to be obtained from isolated porcine livers, which can then be seeded with healthy human cells in order to create humanized livers which could be transplanted. It is important saying that even though the final objective of creating a real alternative to donated livers fails, in the case that these humanized livers cannot last for a lifetime, they can still be of medical use if they last enough to provide a better life for severe patients while they wait for an healthy donated liver. Besides the aforementioned aim, as the DLS could potentially be seeded with disease-carrying cells (for genetic diseases) or even infected after being seeded. These diseased livers present good candidates for disease modeling or drug development and testing. With the project detailed in this thesis, the first steps to this greater goal are the aim. These first steps are the optimization of the decellularization of the livers and recellularization of the vascular tree. In addition to that, in this project, the development of an automatic decellularization machine was also started, with the aim of developing a low-cost, portable and customizable option for organ decellularization. This aim was defined later since the there were some evidences that the decellularization process benefits from a more controlled and static setup. Combining this with the advantages of an automatic machine was therefore convenient in terms of reliability and reproducibility. At the same time, there was also the aim of avoiding big expenditure which contributed for the creation of this project.

19 20 3 Materials and Methods

Contents

3.1 Cell preparation...... 23

3.2 Liver preparation...... 29

3.3 Decellularization...... 29

3.4 Recellularization...... 31

3.5 Quality testing...... 38

3.6 Decellularization machine...... 38

21 22 In this section, the materials and methodologies used will be detailed, starting with all those related to cell culture and tagging, decellularization and recellularization. Afterwards, details on the construction and planning that went behind the decellularization machine will be given.

3.1 Cell preparation

3.1.1 Cell lines

The cell lines used for the recellularization project were hUVEC, hMSC, pig umbilical vein endothelial cell (pUVEC) and pig mesenchymal stem cell (pMSC). The ﬁrst two cell lines (human) were used from isolation up to recellularization, while the last two (porcine) were just isolated, cultured, tagged and stored for a latter stage of the project.

3.1.1.1 Isolation

In terms of origin, the hUVECs were extracted from umbilical cords of newborn babies provided by Hospital Universitario Miguel Servet under informed consent of the parents. Similarly, the hMSCs were isolated from human lipoaspirate samples ceded by the Renobell aesthetics clinic in Zaragoza with informed consent of the patients. Both of these isolation protocols are approved by the Comite´ Etico´ de Investigacion´ Cl´ınica de Aragon´ (CEICA), and are described below. Regarding the porcine cell lines, the pUVECs were isolated from the bone marrow of pigs euthanized in production farms due to trauma or malformation, and the pMSCs from the umbilical cords of spontaneous abortions also in animal production farms. Later, pUVECs were also extracted from the bone marrow from piglets sacriﬁced for the harvesting of livers for this exact project. The isolation protocols are also described below.

3.1.1.1.1 Isolation protocol for hMSCs from lipoaspirate samples

1. Prepare a 50 mL falcon tube (Corning Life Sciences) of 1×collagenase I (ThermoFisher Scien- tific), dissolving it in a fetal bovine serum (FBS)-free, medium like Dulbecco’s modified Eagle medium/Ham’s nutrient mixture F-12 (DMEM/F-12) (Sigma-Aldrich) and supplement it with 1% penicillin/streptomycin (P/S)(ThermoFisher Scientific);

2. Centrifuge the lipoaspirate samples (which come in 50mL falcon tubes) at 1200×g for 10 minutes;

3. From the three layers formed, collect the upper (fat-rich) and the lower (stromal) into the collagenase I falcon tube and discard the central (aqueous) layer.

4. Mix gently and then centrifuge the sample ﬁrst at 37°C and 300 rpm for 30 minutes and then at 1200×g for 10 minutes;

23 5. Aspirate the fatty layer and the supernatant

6. Resuspend the pellet in Dulbecco’s modiﬁed Eagle medium/Ham’s nutrient mixture F-12++ (DMEM/F-12++) and then centrifuge at 400×g for 5 minutes, repeating this step at least twice and until the supernatant is clear;

7. Resuspend the pellet in 10ml of expansion media (DMEM/F-12++), ﬁlter with a cell strainer and seed in a 10cm dish coated with 0.2% gelatin (ThermoFisher Scientiﬁc);

8. Gently remove non-adherent and dead cells 2 days after plating.

3.1.1.1.2 Isolation protocol for pMSC from pig femur bone marrow

1. Carefully dissect and clean the pig femurs from all muscle/ligament tissue;

2. Cut out both epiphysis with a surgical saw;

3. Flush out the bone marrow in several washes of the diaphysis with DMEM/F-12 with 1% P/S and collect in a 15cm tissue culture dish (Corning Life Sciences);

4. Collect and centrifuge the cell suspension at 300×g for 5 min and then reconstitute in DMEM with 1% P/S;

5. Filter the cell suspension sequentially through 100µm and 40 µm cell strainers to eliminate cell aggregates and bone pieces;

6. Isolate the mononuclear cells from this ﬁltered cell suspension using Ficoll density gradient cen- trifugation (density = 1.077g/mL; Sigma-Aldrich)

7. Culture on the necessary number of 10cm COSTAR tissue culture dishes (Corning Life Sciences) coated with a solution of 0.2% bovine gelatin (ThermoFisher Scientiﬁc) at the density of 1.6×105 cells/cm2 in DMEM/F-12 supplemented with 10% FBS (ThermoFisher Scientiﬁc), 1% GlutaMax

(ThermoFisher Scientiﬁc) and 1% P/S at 5% CO2 and 37°C.

8. Gently remove non-adherent cells 2 days after plating.

3.1.1.1.3 Isolation protocol for hUVEC and pUVEC from umbilical cords

1. Locate the umbilical vein, cannulate with 18-gauge (18G) cannula (Terumo) and wash it with phosphate buffer saline (PBS) (Sigma-Aldrich);

2. Inject trypsin (ThermoFisher Scientiﬁc) through the umbilical vein. Recover the trypsin and re-inject it, repeating this for about 7 minutes;

24 3. Neutralize the enzyme adding any media containing FBS and P/S;

4. Centrifuge at 400×g for 5 minutes;

5. Resuspend the pellet in expansion medium (hUVEC medium) for endothelial cells (hUVEC 2x) and seed in dishes with coated with a solution of 0.2% gelatin

6. Gently remove non-adherent cells 3-4 days after plating.

3.1.1.2 Culture media

For culturing the cells, three media were used. DMEM/F-12++ was used to grow both mesenchymal cell lines (hMSC and pMSC), while hUVEC medium was for both endothelial cell lines (hUVEC and pUVEC). The third medium, the bioreactor culture medium, was used for the recellularization of DLSs, and is almost a direct mixture of the other two. Every media supplementation is described below (3.1, 3.2 and 3.3) and the basal media compositions are in the AppendixA.

Table 3.1: Composition of DMEM/F-12++. DMEM/F-12 - Dulbecco’s modiﬁed Eagle medium/Nutrient mixture F-12 Ham; FBS - Fetal bovine serum; P/S - Penicillin/Streptomycin

Work Solution Concentration DMEM/F-12 — FBS 5 % PS 1 %

Table 3.2: Composition of hUVEC medium. MCDB 131 - Molecular, Cellular and Developmental Biology 131 medium; FBS - Fetal bovine serum; P/S - Penicillin/Streptomycin; VEGF - Vascular endothelial growth factor; EGF - Epidermal growth factor; IGF-1 - Insulin-like growth factor type 1; FGF-2 - Fibroblast growth factor 2

Work Solution Concentration MCDB 131 — FBS 5 % PS 1 % L-Glutamine 2 mM Insulin 5 µg/L Transferrin 10 µg/L VEGF 50 ng/L EGF 40 ng/L IGF-1 40 ng/L FGF-2 40 ng/L

Regarding source, MCDB 131 was purchased from ThermoFisher Scientiﬁc, L-Glutamine, insulin and transferrin from Sigma-Aldrich, and lastly VEGF, EGF, IGF-1 and FGF-2 were purchased from Peprotech.

25 Table 3.3: Composition of the bioreactor medium. DMEM/F-12 - Dulbecco’s modiﬁed Eagle medium/Nutrient mixture F-12 Ham; MCDB 131 - Molecular, Cellular and Developmental Biology 131 medium; FBS- Fetal bovine serum; P/S - Penicillin/Streptomycin; VEGF - Vascular endothelial growth factor; EGF- Epidermal growth factor; IGF-1 - Insulin-like growth factor type 1; FGF-2 - Fibroblast growth factor 2

Work Solution Concentration DMEM/F-12 —1 MCDB 131 —1 FBS 5% PS 1% L-Glutamine 2 mM Insulin 5 µg/L Transferrin 10 µg/L VEGF 50 ng/L EGF 40 ng/L IGF-1 40 ng/L FGF-2 40 ng/L

3.1.1.3 Cell culture

To start culturing the cells, firstly tissue culture plates were selected on a case-by-case scenario, depending on the desired number and type of cells, and on the protocol (isolation, expansion, tagging, thawing, etc...). For both mesenchymal cell lines, tissue culture plates were selected so that cells were seeded at 5.000-6.000 cells/cm2, while for the endothelial cell lines, cells were always seeded in a 5-fold increase ratio of surface area. Gelatin was used to coat the tissue culture plates in order to serve as adhesion substrate for cellular growth. The coatings were always prepared before seedings by loading 0.2% gelatin into the desired tissue culture plate in a quantity that totally covered the bottom. After at least 20 min, the gelatin solution was removed and the cells were seeded with the adequate amount of medium, depending on tissue culture plate size. So as to passage the cells, these were washed with PBS without Ca2+ and Mg2+, so as to remove all loosely attached cells, i.e. dead cells, and to weaken the attachment of the live cells to the gelatin coating. After removing the PBS, trypsin is added and the tissue culture plates are incubated at 37°C for 4 minutes. The cells are observed in the microscope and if too attached, can be incubated again. If floating or loosely attached, then they are flushed in a up-and-down motion, so as to collect and suspend them in trypsin before transferring them into a falcon tube. Afterwards, the tissue culture plate is further flushed with DMEM/F-12++ in equal quantity to the trypsin used, to collect the remaining cells, which are added to the falcon tube, neutralizing the trypsin due to the presence of FBS in the medium. Afterwards, the cell suspension is centrifuged at 1500 rpm for 5 minutes in a centrifuge 5804 (eppendorf). The supernatant then gets discarded and, if more than one falcon was used, the cell pellets all added to the same falcon tube and suspended in the appropriate expansion medium. Afterwards the cells are counted in a Neubauer chamber. After cell counting, the cell culture dishes are coated with 0.2% gelatin and labeled. After at least 20 minutes, the cell suspension is then divided equally between the tissue culture dishes and if needed more expansion medium is added. The tissue culture dishes are then

26 gently mixed so as to homogenize the cell suspension and then left to incubate at 37°C with 5% CO2 and normoxia in the hUVEC and pUVEC case or 3% hypoxia in the hMSC and pMSC case. For cell counting, the same amount of cell suspension and trypan blue (Sigma-Aldrich) are mixed together and then diluted with PBS so as to achieve 20µL of the cell sample in the desired dilution. After achieving an homogeneous mixture, it was loaded into a Neubauer chamber, and observed through an inverted optical microscope, counted and then the cell density calculated.

3.1.1.4 Cell cryopreservation

So as to store cellular samples, after cell counting, 1mL of cryopreservation medium is prepared with 90% growth medium and 10% dimethylsulfoxide (DMSO) (Sigma-Aldrich) per each two million cells counted. Afterwards, cryovials (Sigma-Aldrich) were labeled and opened while the cells underwent cen- trifugation for 5 minutes at 1500×g. Afterwards, the supernatant was removed, the cryopreservation solution transferred into the falcon tube, and the cells resuspended in said medium. The cell suspension was then transferred into the cryovials (1mL per cryovial) which were then stored inside a Frosty (ThermoFisher Scientiﬁc) at -80°C over night before being moved into liquid nitrogen (-196°C).

3.1.1.5 Cell thawing

To culture previously cryopreseved cells, cryovials containing 1mL of 2 million cell/mL samples were retrieved from the liquid nitrogen tanks, and heated in a 37°C water bath until thawed. Immediately after, each sample was transferred with a micropipette into 15 mL Falcon tubes (Corning Life Sciences) pre- ﬁlled with 10 mL of the cell’s own growth medium, which were subsequently centrifuged for 5 minutes at 1500×g. Afterwards, the supernatant was discarded and cells were resuspended in 10 mL of the appropriate growth medium and seeded onto a gelatin-coated 15 cm tissue culture dish. They were then further cultured as described above.

3.1.2 Cell tagging

The target cells were also tagged with green ﬂuorescent protein (GFP) and tdTomato (TOM), so as to later track them on the DLS.

3.1.2.1 Tag preparation

To start, three different media were prepared. The ﬁrst was the usual expansion media, which consists of a supplementation of 10%FBS, 1%P/S and 1% L-Glutamine in DMEM/F-12 (high glucose). The second was OptiMEM which is DMEM/F-12, without FBS and P/S supplementation. The third and ﬁnal

27 one was the packaging medium in which DMEM/F-12 (high glucose) was supplemented with 5%FBS, 1%P/S, 1% L-Glutamine and 1mM sodium pyruvate. Afterwards, the protocol for lentiviral production using Lipofectamine 3000 reagent was followed. At day 0, 293T cells were seeded in a 10cm tissue culture plate with usual expansion media. After reaching 95% confluence, the usual expansion media was removed and replaced by 12ml of packaging medium. At day 1, two mixtures were prepared. Mix A (1.5mL OptiMEM media with 41µL of Lipo- fectamine 3000 reagent) and Mix B (1.5mL Optimem media with 13µg of PAK, 13µg of ENV, 4.3µg of expression vector and 35µL of P3000 reagent), which were afterwards mixed in a 1:1 ratio, and left to incubate for 20 minutes forming the DNA-lipid complex. After removing 6mL of medium from the 293T tissue culture plate, the DNA-lipid complex was added and then incubated at 37°C for 6 hours. Afterwards the medium was removed and replaced with 12mL of packaging medium. At day 2, the first batch of viral supernatant was collected (entirety of the medium from the 293T tissue culture plate) and 12mL more of packaging medium were added. The viral supernatant was stored at 4°C. At day 3, the second batch of viral supernatant was collected and combined with the first. Afterwards it was centrifuged at 2000 rpm for 10 minutes, and then filtered through a 45µm pore size filter. The virus was then titered, aliquoted and stored at -80°C.

3.1.2.2 Transfection

To start the transfection, the target cells were ﬁrst expanded in a 10cm tissue culture dish until it was 95% conﬂuent. When that stage was reached, the cells were tagged with GFP and TOM, using the lentiviral supernatants obtained with the protocol described above, to get two differently tagged populations of both hMSC and hUVEC.

3.1.2.3 Puriﬁcation

Following the transfection, the tagged cells were prepared to be separated by fluorescence-activated cell sorting (FACS) in order to get purified samples of tagged cells. For that, the cells were trypsinized and collected into 15mL falcon tubes. Then they were centrifuged and the supernatant removed. Sub- sequently, the cell pellet was resuspended in PBS. Furthermore, 15 mL falcon tubes were filled with the adequate growth medium in order to serve as recipient vessels for the target cell sample. With the cellular material prepared, the samples were sequentially separated using a FACS ARIA II cell sorter (BD Biosciences). A positive selection method was used, since the cells that have shown to have been tagged with the fluorescent GFP and TOM tags were the ones collected as part of the sample and those that did not show any fluorescence were discarded.

28 3.2 Liver preparation

3.2.1 Liver harvesting

The livers were collected initially from 2Kg, and afterwards from 5Kg (18-23 days old), Landrace swine piglets which were heparinized (150UI/Kg), euthanized and bled by a veterinarian. Afterwards, in some piglets, the liver was readily harvested while in others, the piglets were frozen at -20°C before harvesting. Firstly, an incision was made longitudinally through the abdominal wall so as to expose the liver. Afterwards, the suprahepatic vena cava (SVC) was transected. Carefully, theHA and thePV were identified, dissected and transected and all their ramifications ligated. Next up, the infrahepatic vena cava (IVC) was also transected, and the diaphragm and any remaining attachments to the liver were dissected freeing up the liver. Urethras were also retrieved as part of another project. After removing the liver, theHA was cannulated with a 24G cannula (Terumo) and thePV with either a 14G cannula (Terumo) or a Masterflex 1/4” straight barbed fitting (Cole-Palmer Instrument Co.) depending on whether the piglet had 2Kg or 5Kg. The livers were then partially hepatectomized, decellularized or frozen at - 20°C for later decellularization.

3.3 Decellularization

3.3.1 Setup assembly

3.3.1.1 Material

With all the different components ready to be used for the recellularization protocol, all that is missing is the setup that allows for the procedure. First of all as vessel for the liver, various glass lab ware and lastly a custom designed acrylic vessel was used (AppendixB). The liquid tubing lines used were Masterflex platinum-cured silicon tubing L/S 14, 16 or 17 (Cole-Palmer Instrument Co.), depending on thePV andHA diameters. For connections, male-lock, luer 4-way stopcocks were used as well as Mas- terflex luer fittings (Cole-Palmer Instrument Co.), 1/16”, 1/8” or 1/4”, male and female variants (according to the tubing size). For the pumping action, two Masterflex L/S model 7519-06 peristaltic pumps with Masterflex L/S easy load pump head (Cole-Palmer Instrument Co.) were used. These pumps were connected with two Hugo Sachs Elektronik servo controller for perfusion and transducer amplifier modules (Harvard Apparatus) which were used to control the pressure measured at the pressure sensor in order to maintain it at a constant user-selected level. To eliminate the pulsatile action of the peristaltic pumps, two Masterflex pulse dampeners (Cole-Palmer Instrument Co.) were also used. Lastly, depending on the diameter of the target vasculature, either 24G cannulae (Terumo), 14G cannulae (Terumo) or a 1/4” straight barbed fitting (Cole-Palmer Instrument Co.) were used.

29 So as to prepare all of the equipment for the recellularization, all tubing lengths and ﬁttings are

cleaned with distilled water (dH2O). Sterilization is not necessary at this stage since the liver is only sterilized after decellularization as a preparatory step for recellularization.

3.3.1.2 Assembly

So as to decellularize, ﬁrstly the setup is assembled either in the laboratory or in the cold room, according to the scheme below (Figure 3.1), excluding the liver. The 4-way stopcocks should all be completely open, except for the one that connects to the tanks, as that one should be open to the

passage of dH2O and not of decellularization solution. Following this, the tanks are ﬁlled up with the respective liquids, being that the decellularization solution used was either 1% triton X-100 (Panreac) +

0.1% ammonium hydroxide (Panreac) in dH2O or 1% SDS (Sigma-Aldrich) + 0.1% ammonium hydroxide

in dH2O.

Figure 3.1: Scheme of the decellularization setup. In this scheme, the wider line represents thePV line and the narrower line the representsHA line. The blue boxes represent the decellularization solution and the dH2O tanks, respectively.

3.3.2 Decellularization

To start the decellularization the tubing was primed with dH2O and then connected into the liver, through cannulae or ﬁttings using thePV and theHA as inlets. The liver was then lowered into the respective dH2O ﬁlled vessel. Afterwards, the perfusion of dH2O started in order to completely clean the vasculature of remaining blood and to start the decellularization process by submitting the cells to

30 osmotic stress. As for this step, it is performed until there is no visible exit of blood from the outlets (IVC and SVC). The next step is to turn the valves so that the decellularization solution is perfused instead of the dH2O. In this step, both triton X-100 and SDS were tested. The tests consisted of perfusion of just one of the detergent solutions or perfusion of both, starting with SDS and ﬁnishing with triton X-100. This perfusion step was performed either until the liver was fully decellularized, or until no improvement was achieved from further detergent perfusion. After detergent perfusion, the liver was once again perfused with dH2O for double the volume of that used of detergents, in order to clean the obtained DLS. If the decellularization was successful then the liver was stored in little plastic boxes submerged in water with two-three droplets of bleach while awaiting sterilization. If not, then samples were taken for DNA assays and histology.

3.4 Recellularization

After the decellularization step, the next logical step is the recellularization of the DLS. However, so as to recellularize them, a series of preparation steps must be followed in order to make possible the recellularization process. These steps are comprised of guaranteeing that the DLS, the bioreactor and the cells are all ready for this procedure.

3.4.1 Preparation

3.4.1.1 Scaffold sterilization

At last, the DLS must be sterilized if not yet sterile. Several protocols for the sterilization of DLS exist. This step can be done with the perfusion of peracetic acid (PAA), ethylene oxide (EtOH), or gamma irradiation among others, being that combination methods like PAA or EtOH + gamma irradiation appears to give the best outcome. [74] In the present case, none of these methods were followed due to time and material constraints, instead being sterilization achieved with X-ray irradiation. The DLS was placed in a 15cm tissue culture plate which was then half-filled with dH2O. Then the lid was closed, sealed with parafilm and finally sealed in an autoclave bag. Then it was sent for X-ray sterilization using the max parameters possible with the available equipment (160kV, 6.3mA, 99.9 min). Following the sterilization, the DLS was stored at 4ºC while waiting for the recellularization.

3.4.1.2 Cell seed expansion

In this project, instead of recellularizing the whole liver, the ﬁrst objective was set to be the recellularization of the vascular network in a way that would prevent blood clotting, since this phenomenon would

31 indicate that the ECM is not thoroughly covered as it is triggered by the detection of certain components of the ECM, namely collagen, by the platelets. As such, to meet that objective, the cells chosen for the recellularization process were the hUVECs with the addition of hMSCs with more of a supporting role. Despite being considered for the objective, up to this point, SMC were still not included. An estimation based on previous work and the average porcine liver weight was made and a quantity of 100 million hUVECs and 20 million hMSCs was set as a starting point for the seedings on the recellularization attempts. Starting from cryopreserved samples of approximately 2 million tagged cells, these were thawed and cultured as described in subsections 3.1.1.3 and 3.1.1.5. An average of 3 passages was needed to reach the required cell numbers, for both the hUVECs and the hMSCs.

3.4.2 Setup assembly

3.4.2.1 Material

With all the different components ready to be used for the recellularization protocol, all that is missing is the setup that allows for the procedure. First of all a spinner flask is used as vessel, over a Isotemp Heated Magnetic Stirrer (Fisher Scientific) with a magnet providing agitation. The liquid tubing lines used were Masterflex platinum-cured silicon tubing L/S 14, 16 or 17 (Cole-Palmer Instrument Co.), depending on thePV andHA diameters. For connections, male-lock, luer 4-way stopcocks were used as well as Masterflex luer fittings (Cole-Palmer Instrument Co.), 1/16”, 1/8” or 1/4”, male and female variants (according to the tubing size). For the pumping action, two Masterflex L/S model 7519-06 peristaltic pumps with Masterflex L/S easy load pump head (Cole-Palmer Instrument Co.) were used. These pumps were connected with two Hugo Sachs Elektronik servo controller for perfusion and transducer amplifier modules (Harvard Apparatus) which were used to control the pressure measured at the pressure sensor in order to maintain it at a constant user-selected level. To eliminate the pulsatile action of the peristaltic pumps, two Masterflex pulse dampeners (Cole-Palmer Instrument Co.) were also used. Three Masterflex female x male luer- lock smart site connections (Cole-Palmer Instrument Co.), were used in key location in order to aid on the priming process (described below), to seed the bioreactor and to collect samples for further testing and control. Additionally empty spare bottles of PBS, DMEM/F-12++ and hUVEC medium were used as well. Also, the spare bottles’ and the spinner flask’s caps were drilled so as to make tight entrances for the tubing. Lastly, depending on the diameter of the target vasculature, either 24G cannulae (Terumo), 14G cannulae (Terumo) or a 1/4” straight barbed fitting (Cole-Palmer Instrument Co.) were used. So as to prepare all of the equipment for the recellularization, everything that comes in contact

32 with the cells must be sterilized. Excluding the pressure sensor which is sterilized with a biocide, the stopcocks where EtOH is used and the filters which are already sterile (individual sterile package), everything is autoclaved. Additionally, the sterile DLS and the bioreactor medium are also needed. With everything ready to start the recellularization, the first step is the assembly of all the recellularization setup. For that, all of the components that do not come in contact with the cells are placed in their respective positions, either inside (pumps and magnetic stirrer) or outside (pressure controller) of the incubator. Afterwards, in the laminar flow hood, the inside of the pressure sensor is sterilized with biocide and set aside for later.

3.4.2.2 Assembly and priming

After that, the setup is assembled and primed in sterile conditions, according to the scheme below (Figure 3.2), by following the following protocol. Working in sterile conditions entails not only working in the laminar ﬂow hood but also using a sterile laboratory coat, sterile gloves, and a face mask.

1. Connect all the parts, excluding the pressure sensors;

2. Turn the stopcocks 1 and 2 so that the stopcock 2 is diagonal (fully closed) and the stopcock 1 is open towards the tubing A;

3. Prime the tubing A through the smart connector, with biorreactor medium (table 3.3);

4. Connect either the cannula or the ﬁtting on the DLS to the tubing A carefully to avoid the entrance of bubbles (Note 1);

5. Turn the stopcock 1, so as to open the smart connector towards stopcock 2. By doing so, the DLS and tubing A are already primed closed off, so that the DLS cannot be disturbed.

6. Insert the liver in the vessel;

7. Disconnect the stopcocks 1 and 2 and close the top cap of the spinner ﬂask, through where the liver entered (Note 2);

8. Turn the stopcock 2 so that the pressure sensor spot is opened towards the tubing B;

9. Guarantee that the stopcock 3 is open from the tubing B towards either the tubing D or H;

10. From the sensor position, prime the line until the pulse dampener, then ﬁll the pulse dampener until the required level (or slightly upwards). Finally prime the tubing until the stopcock 3;

11. Close of the tubing length C by turning stopcock 3 so that the tubing H is connected to the tubing D;

33 12. Reconnect the stopcocks 1 and 2;

13. Turn the stopcock 2 so that there is an open line between both stopcocks connecting the smart site connector to the place where the pressure sensor connects;

14. Guarantee that the system is not tangled. If it is tangled, it is possible to disconnect the stopcocks 5 and 6 to disentangle the lines. The ﬁttings in the stopcock 4 can also be disconnected temporarily to remove tangles (Note 3);

15. After guaranteeing everything is connected and closed besides the pressure sensor, ﬁnally connect it in its place, prime it through the smart site connector and close it so as to avoid leaks and contamination (Note 4);

16. Double-check everything is closed and either open the connected medium bottle to ﬁll it up or exchange it for a ﬁlled one if the bottle is empty;

17. Move the setup into its place inside of the CO2 incubator (which should be at 37ºC and 5% CO2) and attach the tubing to its respective pumps;

18. Turn the stopcocks 4, 5 and 6 so that there is an open path all the way from the medium bottle, through the pump, and up until the spinner ﬂask (tubing lengths D, H, F and G);

19. Program the pump so that it pumps from the medium bottle to the spinner ﬂask and ﬁll the vessel up to the desired level;

20. Reverse the pumping direction and turn the stopcocks 1, 2, and 3 so as to open all of the main line, the smart site connector branch in the stopcock 1 and the pressure sensor branches in the stopcock 2 (only the branch that connects to the medium bottle in the stopcock 3 should be closed). At this stage, there should be an uninterrupted pathway that connects the tubing lengths G, F, H, C, B and A, completing the bioreactor recirculation loop;

21. Lastly, input the desired parameters and start the pump. Wait for a minimum of 4 hours (overnight would be better) before starting the seeding.

Note 1: The entrance of bubbles can possibly result in clogging part of the vasculature, greatly diminishing the quality of the recellularized liver if not ﬂat out rendering it worthless, hence the need of care on this topic. Note 2: The stopcocks 1 and 2 are disconnected for three reasons, the ﬁrst being so that when closing the cap the tubing doesn’t get tangled, the second being that when performing the rest of the priming, the vessel and the DLS are safe from pulls and further possibly catastrophic manipulations and thirdly so that the priming of the remaining lines and branches is done more easily.

34 Figure 3.2: Recellularization setup. In the left ﬁgure, a scheme of the recellularization setup can be seen. In this scheme, in thePV line (wider) the tubing lengths are labeled and on theHA line (narrower) the 4-way stopcocks are labeled. In the right ﬁgure the actual setup is displayed, assembled inside of the CO2 incubator.

Note 3: As with the stopcocks 1 and 2, the stopcocks 5 and 6 are meant to be disconnected when closing the correspondent spinner flask cap so that they don’t get tangled. If any of the tubing lengths gets tangled, they can strangle enough that it affects the flow and pressure. Possibly worse is that, if tangled the tubing may bend enough so that it stops being tightly fitted in the cap entry holes which effectively exposes the inside of the bioreactor and can result in a contamination. Note 4: As the only non-sterile part (the outside is not sterile), the pressure sensor should be the last piece assembled in order to minimize contamination risks. The sensor is only attached in its position after all of the setup is closed, not counting the site where the pressure sensor connects, which includes the liver being cannulated and inside the vessel. Note 5: In the first part, when using a syringe, the L/s 17 lines (1/4” internal diameter tubing) should be primed from the lowest to the highest part by lifting the tubing and parts so that the medium never flows down during the priming. This is so that bubbles don’t form inside. The tubing is wide enough for the medium to flow inside without pushing all of the air out and if this method is followed no bubbles are

35 formed and that way a lot of work and trouble is avoided. Note 6: It is suggested that the tubing for each line (PV andHA) are autoclaved in separate bags so that each line can be assembled separately to avoid confusion and easen the assembly process. Additionally, for the same reason, all the tubing lengths should be labeled.

3.4.3 Seeding

After waiting at the very least four hours, the DLS can ﬁnally be seeded. The ﬁrst step of the seeding is to calculate how much and how often can the seedings take place. For that, the amount of time it takes for the entire volume inside the system to go 5 times through the liver was calculated with equation 3.1, being that value used as time between seedings.

5 × Vmedium tseeding = (3.1) Fpump tseeding is the time between seedings, in min;

Vmedium is the volume of medium in the bioreactor (vessel + tubing), in mL;

Fpump is the sum of the ﬂow rate in both lines (HA+PV), in mL/min;

As a reference value from previous work developed in the laboratory, a maximum of 5 million cell/min was admitted for seeding, since the entry of too many cells at once in the DLS can cause clogging. Taking this into consideration, the number of cells per seeding was calculated with equation 3.2

Max 5 × Fcells Nseeding = (3.2) Fpump

Nseeding is the maximum number of seedable cells per seeding;

Max Fcells is the maximum number of cells seeded, in cell/min;

Fpump is the sum of the ﬂow rate in both lines (HA+PV) in mL/min;

With these data, the DLS can finally be seeded through the smart connector that leads directly into the vessel. With two different cell samples (equal types, but distinct tags) the seedings for the first condition are performed until the sample is all used up, and then the pumps are turned off to let the cells adhere better without a shear stress source. Four hours later the pumps are re-started and the next seeding cycle can start. As the conditions may have varied greatly due to the colonization of the DLS by the first cells seeded, it is advisable that the seeding parameters are recalculated.

36 The different cell tags were used since the two seeding cycles were performed with different set- points in the pressure control, to assess the pressure that would obtain an optimal penetration to cell death ratio of the seeded cells inside the DLS.

3.4.4 Control Assays

In terms of bioreactor monitoring, apart from the control over the pressure and flow which is part of the control loop, and the control over temperature, pH, and dissolved gases that exists due to the placement of the bioreactor in a CO2 incubator, additional mechanisms are used to monitor the recellularization process. Firstly, due to the presence of phenol red indicator in the medium, one can program the medium changes depending on the color of the bioreactor medium. Due to the acidifying effect of the cellular metabolism, as the consumption of nutrients in the medium advances, the medium gets increasingly acid and that is visible as the medium starts shifting from a pink-reddish to an orange and finally to a yellow hue. Secondly, since the objective entails the culture of anchorage-dependent animal cells and medium is by itself clear, increase of turbidity of the medium is an indicator of possible contamination, that if detected early, can be treated in the case of microbial origin. Fungal contaminations in this stage are extremely difficult to remove.

Figure 3.3: Bioreactor vessel in different stress conditions. From the leftmost to the rightmost picture, three different stages of the bioreactor are shown, in increasing order of degree of nutrient consumption as evidenced by the color of the medium. Additionally, it is noticeable that in the bioreactor pictured in the center and rightmost pictures a contamination has occurred due to the turbidity of the medium.

Additionally glucose and lactate dehydrogenase concentrations are used as indicators of nutrient consumption and cell metabolism. For this, collected samples were centrifuged at 3000 rpm for 5 minutes and then the supernatants transferred to falcon tubes. These were stored at -20°C until the end of the respective experiment so that all samples could be processed simultaneously. The glucose concentration was assessed with blood glucose test strips and the lactate concentration

37 was determined with the Pierce lactate dehydrogenase (LDH) Cytotoxicity Assay Kit (ThermoFisher Scientiﬁc). The protocol for determining the glucose consists of placing a droplet of the relevant sample on a clean surface and then measuring the glucose concentration with the glucometer. The protocol used for the LDH measurements was the one described in its user manual without any alteration.

3.5 Quality testing

3.5.1 Microscopy Staining

So as to assess the extent of both the decellularization and the recellularization, DLSs and recellularized livers were ﬁxated by perfusing for 15 to 20 minutes and then immersing them in formaldehyde overnight. Afterwards they were sent for tissue processing and H&E staining.

3.6 Decellularization machine

3.6.1 Requirements

First of all, a list of requirements was made for the device and project to answer to:

1. The device must free up all of the parts and equipment which are currently being used in both decellularization and recellularization protocols, so as to make possible running both protocols in tandem;

2. The device must allow the user to set conditions for ﬂow rate or pressure. This possibility would allow for a better control of the process and would enable reproducible results;

3. The device must combine in one assembly all of the system, including the solution tanks, the pumping station, the control system, a validation system and a way to record all data;

4. The system must be mobile, enabling studies both at room temperature or ate lower temperatures in a cold room;

5. The data recorded by the system must be easily accessible/outputted in a common format;

6. The device must automatically protect the liver from common risks, stopping in case of issues such as empty tanks or bubbles inside the tubing;

38 7. The device must alert the user remotely in case of any problem, or if nothing goes wrong, when the program ends;

8. The system must be able to receive user inputs and pause/stop if the user gives that command.

Additionally, to make it possible branding it as a low-cost alternative, it was decided that the entire project should cost less then a single Masterﬂex L/S model 7519-06 peristaltic pump (Cole-Palmer In- strument Co.), which are the currently used ones. These pumps, are currently sold for about 2500C, so that price was set as a limit.

3.6.2 Control

3.6.2.1 Parameters

The first two important control parameters are the low rate and the pressure, as they are the targets of the desired control loop. On one hand, if the pumping action is determined by the flow rate then it is important to be able to measure directly or indirectly that parameter, on the other hand if a pressure- based pump control then the pressure at the relevant points (PV andHA inlets) must be known. Regarding the flow rate, if a flow meter (preferred option since the reading gives a direct measurement) is used, then it must be assembled in the main line. This doesn’t happen with the pressure sensor which can be installed either in the main line or in a side branch, which leads to pros and cons depending on the chosen option. If the sensor is installed in the main line, then the reading is as precise as it can be, since it is the closest possible to the inlets2. Additionally, sudden pressure drops can be used as an indication of bubbles in the system, which can be used as a programming cue for pump shut down, to prevent the entry of said bubbles in the liver, preventing possible embolization of the blood vessels. Considering the other option, the installation of the sensor in a side branch, the main advantage would be the easier accessibility to the sensor in case of malfunction, and the fact that it is more protected from damage extending its life. For both parameters, not only are they needed for the respective control loops, but also they are measured so that they can be plotted, as from these plots some information about the decellularization is obtained. If a constant flow rate is used, then it is expected that as the decellularization proceeds, the inlet pressure decreases, since there is a decreasing amount of cellular material generating resistance to the perfusion. If a pressure-based control is used then it is expected that the flow rate increases over time for the reason mentioned above, as there is less cellular material over time, then the resistance generated is decreasing and as such, to generate a constant pressure then the flow rate must increase accordingly. With these plots, there is a very direct indication of how the decellularization is progressing.

2The pressure sensor should be the part that sits closest to the inlets, so that the reading is as signiﬁcant as possible. If the sensor is far, then transfer phenomena like friction will dictate that the value measured is lower than the real inlet values.

39 Other use for the flow rate measurement is the calculation of the volumes spent in the decellularization, as this parameter can be continually integrated over the decellularization time. This leads us to the following two important control parameters, which are the time elapsed and the volumes spent. These parameters complement the last two as they provide a scale for the evaluation of the evolution of the decellularization. As such, they are mainly used to support the calculation of some parameters and to plot the various graphs. They are also important for the definition of cycles as these are defined by the perfusion of a determined quantity of perfusate or by the perfusion of a perfusate for a defined amount of time. Additionally they are of utmost importance to evaluate the success and scalability of the process as the total time and perfusate volumes invested are the main costs associated with running a decellularization experiment. Also, if translated to clinical practice, it is important that the machine is as quick as possible while simultaneously not wasting the decellularization solution(s). Finally, for the evaluation of the decellularization, the liver opacity and the protein and DNA content on the effluent will also be controlled throughout the decellularization. When these parameters stabilize, then it can be inferred that the perfusates no longer decellularize the liver, meaning that the decellularization reached its end, either because the liver is completely decellularized or because the remaining cellular material is too attached or trapped in a way that makes it impossible for the perfusate to remove it without further handling (gentle massage in poorly decellularized regions may sometimes promote further decellularization).

3.6.2.2 Planed operation modes

Knowing the parameters, the next step is to define the operation protocols for the pumps. As the pumping action will be determined by a control loop either driven by pressure or flow rate at the liver inlets, the possible protocols will be pressure or flow rate functions3. With this said, the three protocols that are being implemented are the constant, the oscillating, and the ramp protocol. The first protocol, the constant pressure or flow rate (Eq.3.3), is the simplest one and the type of pumping pattern most often seen in the literature as most authors focus their studies in the solutions used for decellularization and not in the flow patterns. Due to that, most often the pumps are left running in a constant flow, or with a pressure sensor maintaining a constant pressure. Since it is a user favorite, it is one of the most important protocols to be implemented. As the simplest protocol it is also represented by the simplest function,

SPA = X (3.3)

3Possibly, with further knowledge in the subject, a function of both the ﬂow rate and pressure would be an even better candidate for driving the pumping action

40 4 SPA is the set-point of the system (in mmHg or mL/min) ;

X is the user inputted set-point(in mmHg or mL/min)4;

The second planned protocol, is the oscillating protocol (Eq.3.4). This protocol has been planned so as to replicate the physiological state of the arteries, in which the pressure has a typical pattern of oscillation derived from the pumping action of the heart. For this reason, the oscillating pattern is interesting, in order to study how maintaining the conditions naturally present in these tissues may help keep the integrity of the scaffold and/or produce better decellularization results. Since this protocol aims to emulate the arterial pressure, the function must represent what is the natural physiological oscillations and for that, a Fourier Series was considered in order to represent this as an approximation,

!! A sin 1×2πωt + 2.750 sin 2×2πωt + 4.745 sin 3×2πωt + 0.598 SP = × −0.228+0.721× − 60 + 60 − 60 (3.4) B 2 1 2 3

5 SPB is the set-point of the system(in mmHg or mL/min) ;

A is the oscillation amplitude (in mmHg or mL/min)5;

ω is the pulse frequency i.e. heart rate (in bpm);

t is the simulation time (in s);

The third and as of now last is the ramp protocol (Eq.3.5), in which the ﬂow rate/pressure increases or decreases from the previous to a new set-point throughout a set amount of time. This serves mainly for maintenance and security reasons, and is to be implemented just so users can select this between cycles where the conditions are too distinct or to start out the decellularization without jumping straight from the stopped condition to high ﬂow rate or pressures. The function which describes this pattern is the following,

t SPC = ∆ × (3.5) tf

5 SPC is the set-point of the system (in mmHg or mL/min) ;

∆ is the desired set-point variation (in mmHg or mL/min)5; tf is the duration of the cycle (in s); t is the simulation time (in s);

4Depending on the control setup. Either mmHg for pressure control or mL/min for ﬂow rate control. 5Depending on the control setup. Either mmHg for pressure control or mL/min for ﬂow rate control.

41 Finally, the program combines all of the actuation modes selected, creating a set-point value which is the direct sum of the different modes (Eq. 3.6). This means that actually having, for instance, the oscillation mode active with A = 0, or having the mode turned off yields exactly the same result. This mode of action provides a very malleable way of representing the desired proﬁles while maintaining some simplicity since the implemented proﬁles represent simple concepts with different utilities.

SP = SPA + SPB + SPC (3.6)

With these protocols in the backgroud, graphical user interfaces (GUIs) such as the ones represented in ﬁgure 3.4 could then be used to provide the users means of easily creating an entire set of perfusion steps that would then be recreated by the pumps through a PID control loop feedback mechanism.

Figure 3.4: Example interfaces for the automatic decellularization machine.

3.6.2.3 System description and components

In the present iteration, this control system is being designed to run simultaneously in Arduino and MATLAB. The objective of having both systems working together is to have the computer as the controller of the operation through MATLAB, with the Arduino responsible for the communication between it and the rest of the system. Following With this design, the computer would be where all the different parameters are stored and plotted as well as processed in real-time, functioning as the controller in the control loop. The Arduino would then function as the connection between the computer and the sensors and actuators, effectively closing the loops by receiving, transforming and transmitting the data from the sensors to the computer and then correcting the actuation of the pumps accordingly as well as triggering all sort of activities such as alarms and opening/closing of valves. The option of using a computer and not just the Arduino provides a direct way of controlling how the data will be outputted since it is processed directly in the computer. It also allows us to avoid storing

42 large amounts of data in the Arduino, which would be unfeasible since these systems do not have large memories.

Apart from the control setup, there are also two 20L tanks (LaboAragon)´ which house dH2O and decellularization solution respectively and two electrically actuated valves (Sirai) to control which fluid flows through the system. This setups leaves the possibility of adding more tanks (for different decellularization solutions or PAA for sterilization) by adding along with the extra tank one more valve, 4-way stopcock and L/S 17 tubing length. The fluids flow through two tubing lines, one that leads to thePV through L/S 17 tubing and on that leads to theHA through L/S 14 tubing. In these lines, the fluids pass sequentially through a peristaltic pump (Williamson Manufacturing Company Ltd), a pulse dampener and a 4-way stopcock with a pressure sensor (Pendotech) mounted in a side branch before finally entering the liver through the canullated inlets (PV andHA). After exiting the liver chamber the fluid is then directed through a quartz chamber where its DNA and protein contents are continuously monitored through A260/A280 ratio, making it possible to program an auto-stop feature when this value stabilizes, which means the decellularization is no longer being effective.6

Figure 3.5: Scheme of the current design of the decellularization machine setup. In this scheme, the wider line represents thePV line and the narrower line the representsHA line. The dotted lines represent the control setup.

As a further decellularization follow-up system, the opacity of the liver will be continuously monitored by the use of white light-emission diodes (LEDs) and light sensors in opposite sides of the liver. As in the

6Since the decellularization solution may have components that interfere with the absorbance ratio, such as Triton X-100, it is advised that the cell is calibrated with the solution beforehand

43 case of theUV light system, this system can be used as an auto-stop system, when the the opacity no longer changes, as this would mean, that further perfusion no longer contributes for the decellularization of the liver. So as to support the liver during the process the acrylic vessel described before was selected. A schematic representation of the device is portrayed in Figure 3.5 Additional features include the assembly of the entire decellularization system in a mobile structure, such as a service cart. Another intended feature is programming the machine so that it can produce a range of different notiﬁcations which could then be sent via internet.

44 4 Results

Contents

4.1 Decellularization...... 47

4.2 Recellularization...... 51

4.3 Others...... 55

45 46 4.1 Decellularization

4.1.1 Liver preparation

To prepare the DLS for recellularization, two steps may have to be followed depending on the protocol followed. Firstly, considering the objective, a partial hepatectomy may have to be done to the DLS. The existence of the partial hepatectomy in the protocol is dictated by the final objective of the project. As in this project, the final objective is to transplant only the right lobe of the liver, the procedure must be integrated into the protocol. On another note, the partial hepatectomy does not need to be performed post-decellularization. It can also be performed pre-decellularization, which leads to some advantages and disadvantages, when we compare both approaches. On one hand, if the partial hepatectomy is performed first then the liver submitted for decellularization is smaller which helps save time and reagents in the process. On the other hand if the decellularization is performed first then all of the vasculature is visible and as such the partial hepatectomy is easier. Additionally, as the decellularization is done with an intact liver there is no risk diffusional issues that would result in poorly decellularized patches. This issues are related with the possible vascular damage the liver may suffer during decellularization (which is minimized with the post- decellularization approach) and with possible strangulation resulting from the sutures. This resistance to decellularization can easily be seen in figure 4.1 where it is clear how the tightness of the suture can greatly limit the reach of the detergents perfusion in the sutured area.

Figure 4.1: Decellularized liver scaffold which was partially hepatectomized prior to decellularization. For reference, the DLS is placed in a 15cm tissue culture plate.

Additionally, a mixed approach was tried, in which the main vascular branch connecting the right lobe and the right central lobe was ligated without performing partial hepatectomy. The objective of this approach was reaping the advantages of performing a partial hepatectomy while leaving the vast majority of the liver untouched to prevent the disadvantages associated with suturing. The result reached with this approach is evidenced in Figure 4.2. In this picture, apart from the effect of the ligation in the

47 decellularization pattern, the effect of immersion is also observable in the edges of the non-decellularized liver lobes. This image makes it clear why a perfusion-based approach is needed in thick solid organs such as the liver and why one reliant only on diffusion would not cut it. Additionally, as this particular liver was poorly extracted, having theHA being cut too short for canullation, the decellularization was performed using thePV as the only inlet in the procedure. As the result was satisfactory it suggests that for decellularization there is no need forHA canullation and perfusion, which is advantageous as it allows simpliﬁcation of both the procedure and design/construction of the automatic machine. With that said, more experiments would need to be conducted to validate these results, and the structural integrity of these scaffolds would need to be tested to assess the viability of the methods.

Figure 4.2: Decellularized liver scaffold which was ligated prior to decellularization. For reference, the DLS is placed in a 15cm tissue culture plate.

4.1.2 Decellularization

As a non-ionic detergent triton X-100 is preferred since it has been shown to be less damaging than SDS. When both were used sequentially, triton X-100 was used after so that it could help clean up the remaining SDS, which dissolves very poorly in water and as such is difﬁcult to remove. To counteract the denaturing effect of the detergents in the ECM the ammonium hydroxide was added since it has been shown to help stabilize said structure. In total, 21 decellularizations were performed, 10 with livers extracted from 2Kg piglets and 11 from 5Kg piglets. The exact protocols varied in time, volumes perfused, pressures used and detergent compositions and concentrations. From those 21, 5 yielded good decellularization results (Figure 4.3). These were characterized as good as they were transparent after the procedure. As these DLSs were later used for recellularization experiments1, no H&E stains or other immunohistochemistry assays of well decellularized scaffolds were 1The ﬁfth one, was not recellularized as thePV was damaged during isolation. It was instead used for partial hepatectomy

48 performed to verify the decellularization.

Figure 4.3: Livers which were considered well decellularized. (a) First well decellularized liver and the only one extracted from a 2Kg piglet. It was used for the ﬁrst recellularization experiment having been the only full liver recellularized. (b) Liver which was partly hepatectomized prior to decellularization having the left and central left lobes removed. After decellularization the central right lobe was also removed, removing the badly decellularized part where it was sutured (Same liver as in Figure 4.1). (c) and (d) represent livers which were also partly hepatectomized prior to and post decellularization, the difference being that, these livers were not sutured for decellularization. (e) represents a liver that was only decellularized through thePV, with the circulation between the right lobe and the rest of the liver stopped by ligation of the main blood vessel (See Figure 4.2). For reference, all DLSs are placed in a 15cm tissue culture plate.

Even though the details on the protocol followed varied and not enough replicas were performed in any set of conditions to allow for conclusive results some general ideas can still be captured from the experiments performed:

1. As the ﬁrst fully decellularized liver was decellularized with just seven liters of 1% triton X-100 and 0.1% ammonium hydroxide solution, it is deﬁnitely possible to decellularize without using more damaging detergents like SDS;

2. The results were better on a overall analysis when several conditions started being better controlled and stabilized. For instance, the usage of support structures (Figure 4.4) that prevent the movement of the liver during decellularization (such as that caused by ﬂotation) aided in getting more reliable results; practice

49 3. Good bleeding is a fundamental step. Blood clots/cells can severely hinder the decellularization process;

4. If the liver has badly decellularized patches, gentle massage can improve the results by promoting redistribution and eliminating the preferential pathways that may have been formed;

5. The last experiment suggests that the best option in terms of liver preparation is the ligation of the blood vessels between the right and right central lobes as this process appeared to garner good results with the most simple protocol.

Figure 4.4: Liver support structures. (a) (b) First structure made with the aim of supporting and immobilizing the liver during the decelularization procedure. It was made with four 5mL pipettes and three tops from 1mL micropipette tip boxes. (c) (d) Second structure built for liver immobilization. It was built out of acrylic, stainless steel (SAE grade 316) and glued with a silicone-based glue.

4.1.3 Decellularization machine

Regarding the decellularization machine, as it has not yet been built no conclusions can be made as of this point regarding its efﬁcacy. As of now, its budget is calculated in approximately 1750C. This value includes all of the material pre-existant in the lab that did not have to be bought, which includes the pulse dampeners and the tubing. On the other hand, the UV system is yet to be assembled and is estimated in about 500C, which were still not included in the total cost. Some parts like the pulse dampeners, can possibly be constructed by a much lower price, so pottentially a lower cost is achievable without compromising the setup, and eliminating capabilities.

50 4.2 Recellularization

4.2.1 Cell seed expansion

It is worthy of being noted that in the biorreactor stage, the 100M mark for the hUVEC was never reached, since cells started to expand slower overall. This can probably be due to some degree of cell senescence as the samples were already in high passages (passage 10-11) when seeded for expansion. In the last two recellularization experiments it was found out that there was also some hMSC contamination in some of the hUVEC samples, which can also contribute for the disparity between the objective and the results as these cells are bigger and as such don’t reach densities as high as the hUVECs do.

4.2.2 Recellularization

Regarding just the vascular recellularization, it is important to notice that, as mentioned in subsection 1.1.1, SMCs also form an important layer in blood vessels and as such, the addition of said cells could be advantageous for vascular regeneration. As of this stage, SMCs were not used in the seeding. As the project is focused in vascular regeneration, the usage of hepatocytes and other important liver cells such as the ones mentioned in subsection 1.1.2 was also still unconsidered, being hUVECs and hMSCs the only seeded cell types. In total, 4 recellularizations were performed. The first one, was done with a DLS obtained from a liver extracted from a 2Kg piglet while the others were done with DLSs obtained from livers extracted from 5Kg piglets. These DLSs correspond to those portrayed in Figure 4.3 (a), (b), (c) and (d). For the first recellularization, the DLS portrayed in Figure 4.3 (a) was used. This was the only recellularization performed on a whole liver, as well as the only performed on a DLS from a 2Kg piglet, being this the reason for the utilization of the whole liver. Additionally, it was also the only DLS that went through a 7-days protocol as it was used to practice and identify possible problems with said protocol. Seeding was done in two days. On the first, GFP-labeled cells were seeded at high pressure, and on the second TOM-labeled ones were seeded at low pressure. Due to mismanagement of the cell lines in the last days before seeding, high cell mortality was observed and as such, only 37.0 million hUVECs and 24.2 million hMSCs were seeded on the first day, while only 73.8 million hUVECs and 28.0 million hMSCs were seeded on the second day. The pressures programmed for the seedings were of 50mmHg and 25mmHg in theHA andPV, respectively, for the first seeding and 25mmHg and 10mmHg in theHA andPV, respectively, for the second seeding. Between seedings, the bioreactor was set at a maintenance pressure of 4mmHg in both inlets, in order to promote the adhesion of the cells to the DLS by decreasing the shear stress to which they were submitted to. What effectively happened was that, at the end of the first day of seeding theHA canulla

51 fell off and so the protocol was followed only for portal recellularization and the pump associated to the HA line was stopped. At the end of the seedings, the system was left in the maintenance pressure of 4mmHg and manually increased over time at an approximate rate of 2mmHg per day until a pressure of 10mmHg was reached. This gradual increase is performed to avoid sudden drastic increases in shear stress, which can be harmful to the cells, DLS or may flush the cells out from the scaffold. At day 5, thePV canulla also fell of leaving the liver suspended. Due to that, the experiment was temporarily stopped, and the liver recanullated. Afterwards the portal pressure was set at 10mmHg, as it was the value before the recanullation and the arterial pressure was initially set at 50mmHg, which proved unsuccessful as the pumping action varied wildly leading to the entry of some air into the scaffold. At this point the flow rate seemed to be unrelated to the inlet pressure2, and due to that, a direct flow rate control was used to obtain a flow rate between 3 and 4 times smaller than that observed in the portal setup. Medium samples were collected daily for LDH measurements, which were used as means of tracking cellular stress. Microscopy observation of the H&E stains (Figure 4.6 (a), (b) and (c)) revealed that they were mostly empty with some populated areas with high cell density. The cells were in their majority hMSCs, with only residual amounts of hUVECs. High cell mortality was observed, which suggests that the pressures used were too high. From this first experiment, it was observable that pressures like the ones used are damaging to the cells, especially the hUVECs, which indicates that lower pressures should be used. For the second, third and fourth recellularizations, the DLS portrayed in Figure 4.3 (c) (b) and (d) were used, respectively. As mentioned before, the protocol followed for these recellularizations was a 24h protocol, instead of the 7-day one used for the first DLS. For the second recellularization, the planned conditions were 16mmHg and 30mmHg, for thePV and HA respectively in the first seeding and 8mmHg and 15mmHg, for thePV andHA respectively in the second seeding. For maintenance, a constant pressure of 4mmHg was selected. When starting the first seeding, it was noticed that so as to maintain the pressures planned the correspondent flow rates were very high and possibly damaging to the cells in the case of thePV(256ml/min at 8mL/min) and very low in the case of theHA (0-0.04 ml/min at 50mmHg). Due to this, the actual pressures used were adjusted to 4mmHg and 50mmHg, for thePV andHA respectively in the first seeding and 2mmHg and 25mmHg, for thePV andHA respectively in the second seeding and for maintenance pressure. Even so, a flow rate of 180mL/min was obtained with the 4mmHg in thePV.

2Over a 42 hour interval, the pressure varied over 30mmHg (from 14mmHg to 52,7mmHg) in a constant ﬂow setup (2mL/min), being that when the pumps were stopped the pressure did not decrease to 0. Additionally, after recanullation, the initial pressure measured was of -15.7mmHg for a 3.1 mL/min ﬂow rate.

52 Figure 4.5: Recellularization results. (a) First recellularized DLS. (b) Floating structures found in bioreactors 2 and 3. (c) Second recellularized DLS. (d) Third recellularized DLS. (e) Fourth recellularized DLS.

Seeding was done in two days. On the first, GFP-labeled cells were seeded at high pressure, and on the second TOM-labeled ones were seeded at low pressure. 50.0 million hUVECs and 36.4 million hMSCs were seeded on the first day, while only 39.8 million hUVECs and 23.3 million hMSCs were seeded on the second day. Contamination of the hUVEC tissue culture plates by hMSCs was observed, which can partially justify the low cell numbers seeded, as these cells are considerably larger than hUVECs and as such, occupy more space leading to lower cell numbers when confluence is reached. After the second seeding, and stopping the pumps for 4 hours to promote attachment, the bioreactor was left to run at maintenance conditions for 24h before retrieving and fixating the liver for staining. Just before retrieving the liver, at the 24 hour mark, a medium sample was collected for LDH analysis. Additionally to the scaffold, some floating structures which appeared to be either cell aggregates or DLS fragments (Figure 4.5 b). When observed under the microscope after H&E staining (Figure 4.6 (d), (e) and (f)), the DLS was mostly empty yet dotted with some high density populated areas, as was observed in the first bioreactor. Lower cell mortality was observed when compared with the first bioreactor, but still the cell mortality was

53 high enough to suggest that in the third experiment the pressures should be decreased even further. It is worth noting that all the differences in results seen between the ﬁrst and second recellularizations can be explained by the fact that the protocols differed in length of time being the second one a shorter experiment which would inevitably lead to less expansion after adhesion in the scaffold.

For the third recellularization, the planned conditions were 8mmHg and 50mmHg, for thePV and HA respectively in the ﬁrst seeding and 4mmHg and 25mmHg, for thePV andHA respectively in the second seeding. For maintenance, a constant pressure of 2mmHg in thePV and 10mmHg in theHA was selected.

In the ﬁrst seeding, 76.3 million hUVECs and 26.3 million hMSCs were seeded. The second seeding was not done since when it was to be performed the bioreactor was noted to be contaminated with bacteria. Instead, the liver was ﬁxated with formaldehyde skipping the remainder of the experiment entirely.

When observed under the microscope, the DLS was ﬁlled with bacteria which appeared to be from the bacillus genus. Apart from that, minor amounts of hUVECs and hMSC could also be seen, even though most of them appeared to be dead (Figure 4.6 (g), (h) and (i)).

For the fourth recellularization, the planned conditions were 8mmHg and 50mmHg, for thePV and HA respectively in the ﬁrst seeding and 4mmHg and 25mmHg, for thePV andHA respectively in the second seeding. For maintenance, a constant pressure of 2mmHg in thePV and 8mmHg in theHA was selected.

In the first seeding, 70.5 million hUVECs and 20.4 million hMSCs were seeded. After the last seeding the pumps were stopped as per protocol for 4 hours. When they were going to be turned on, it was noticed that even though the pumps were stopped, the pressures registered in the sensors, were of 5.2mmHg and 14.8mmHg in thePV andHA respectively, instead of the expected 0mmHg. since these are higher than the maintenance conditions, to ensure that the cells would not be without perfusion overnight, fixed flows of 4mL/min and 1mL/min were set for thePV andHA respectively. In the next morning, due to the fixed flow conditions, the pressures observed were of 15mmHg and 100mmHg. So as to try to lower these pressures, the medium flow was changed so that it passed for a few minutes by the lower bottle, which is connected to the atmosphere by a filter. Even though this did lower the pressures, these rapidly rose when the original pumping path was restored, stabilizing at 10mmHg and 95mmHg. Using the same line of thought as before, the second seeding was also performed using fixed flow rates and not fixed pressures.

The ﬂow rates used were 8mL/min and 2mL/min for thePV andHA respectively, and 43.6 million hUVECs and 17.2 million hMSCs were seeded.

At seeding the pressures registered were 13.8mmHg and 78.5mmHg, at the last seeding before stopping the pumps the pressures were 14.5mmHg and 82.6mmHg.

54 After 24 hours, medium was recovered for LDH testing and the scaffold was ﬁxated in formaldehyde and sent for H&E staining. Like what happened in the second bioreactor experiment, some ﬂoating structures were found and these too were sent for staining.

Figure 4.6: H&E stains of the four recellularized scaffolds. (a) Staining of the hilum of the first recellularized DLS . (b) Staining of the left lobe of the first recellularized DLS. (c) Staining of the central right lobe of the first recellularized DLS. (d) Staining of the hilum of the second recellularized DLS. (e) Staining of the left lobe of the first recellularized DLS. (f) Staining of the fragments observed on the second experiment. (g) Staining of the right lobe of the third recellularized DLS. (h) Staining of the right lobe of the third recellularized DLS. (i) Staining of the right lobe of the third recellularized DLS.

One further detail worth mentioning that could improve the results is the method of calculation used for the seedings. As of these experiments, the weight of the livers was not taken into account, which is an easy alteration that could lead to more reliable results than those obtained.

4.3 Others

At a certain time point, the pMSC population which was being expanded seemed to be growing at a gradually slower rate than what was expected, which led to suspicions about possible cell senescence

55 and functionality of the mesenchymal lines. Due to that, those cell lines, as well as the ones presently used in the recellularization experiments were subjected to an adipocyte differentiation protocol in order to prove their ability to give rise to that cell type.3 The results obtained (Figure 4.7) suggest that the pMSC lines were senescent while the hMSC ones were healthy and functional, but as mentioned before other differentiation protocols are needed as well as functional assays before conclusions can be taken.

Figure 4.7: Adipocyte differentiation assay. (a) GFP-tagged hMSC p12 stained with red oil. (b)GFP-tagged hMSC p12 stained with red oil. (c)TOM-tagged pMSC p13 stained with red oil. (d)TOM-tagged pMSC p13 stained with red oil.

Additionally to the GFP and TOM, azurite was also used as fluorescent tag to mark cells with blue fluorescence. This was done so as to try an extra tag, since there is a possibility of seeding extra cell types which would require more tags to be used. The protocol followed for this was the same one used for the other two tags, but despite this the results of cell tagging were much less satisfactory, with only faint blue fluorescence being observed. Presently, no further azurite tagging attempts were made.

3Chondrocyte and/or osteocyte differentiation was also programmed but not performed as of this stage.

56 5 Conclusion and Future Work

Contents

5.1 Conclusion...... 59

5.2 Future Work...... 59

57 58 5.1 Conclusion

In conclusion, even though the data collected is not nearly enough to make any statements, there are several leads that can be taken from the experiments performed. In terms of the decellularization of porcine livers, a 1% triton X-100 and 0.1% ammonium hydroxide solution seems enough to decellularize a liver extracted from a 2Kg piglet, which suggests that the use of strong detergent solutions is not needed. Also, the stabilization of the liver seems to be important for the process as well. Even though 21 livers were decellularized, only 5 had good results. Nevertheless, the degree of decellularization improved over time, as the process control and stabilization improved. So as to further improve the decellularization protocol as well as amplify the reproducibility of the results, another project stemmed from this one, in which the construction of an automatic decellularization machine was the aim. Regarding that project, no conclusions are made as it is not advanced enough to do so. Regarding the recellularization of the DLSs, no major conclusions can be made but the results obtained suggest that the pressures used were to high and could cause cell damage. Since there were some issues pertaining to how the ﬂow rates developed and how different they were from liver to liver even in similar conditions, further One of the major shortcomings in this project was the lack of replicates, which is something that will necessarily be dealt with as the project continues and more results are collected. Summing up, even though the results obtained were lacking in number and no conclusions can be extracted from them, they give us an ample room to adjust the protocols used, and to better understand the processes of decellularization and recellularization. It is also important to mention that as part of a larger project, the main objective of this thesis was building enough knowledge that could be used as a stepping stone for the continuation of the project.

5.2 Future Work

As far as future work goes, the most important thing is to get more data in order to get statistical relevance for the results obtained in the decellularization and recellularization processes. Apart from that, a number of things can beneﬁt from further improvement. The decellularization process, as mentioned before, would beneﬁt from the automatization of the process, being the conclusion of the machine needed for the relevant tests. For that a number of tasks are yet to be completed. These tasks include:

1. Finishing the electrical wiring design;

59 2. Designing the light system (opacity reader);

3. Constructing and testing theUV system;

4. Assembling the whole setup;

5. Adjusting the PID control parameters to avoid control related issues, such as derivative kicks;

6. Adjust and calibrate all of the sensors and control loops;

7. Perform a statically relevant quantity of decellularizations;

Additionally to these tasks, some further research is needed and some design options will have to be taken regarding the positioning, assembly of parts such as the light system. Regarding the actual decellularization process, apart from the need for more data, and statistical relevance, some further work of optimization in the conditions used is also needed Regarding the recellularization process, more studies must be done in terms of selecting the best seeding population for meeting the end goal. Even if the recellularization of the vascular network is successful, maybe by recellularizing these vessels, the subsequent recellularization of the functional units, the sinusoids and the bile ducts and canaliculi is then hindered. So, in the future, the protocol for recellularization of the whole liver has to be studied for it may be possibly done in a step-wise approach or by seeding all the relevant cell types simultaneously. Other concern that has to be looked at is exactly which cells and in what relative quantities will they be seeded. For this evaluation the pros and cons of adding said cell type will have to be weighted in terms of how much does the recellularization beneﬁt from the existence of said cell type versus how much does cultivating and seeding said cell type add in terms of cost and complexity of the protocol. This process would also beneﬁt from the usage of a bioreactor for cell expansion, so as to get a more uniform population and to help reach, in a more reproducible way, the number of cells needed for recellularization. Additionally, in the future the cells should be characterized, and some further control implemented in order to guarantee or verify the maintenance of their functions and that they are not senescent by the time they reach the recellularization step. Some further possible optimization of the protocols, include the transition of FBS-enriched media to xeno-free options without animal-derived components which would improve the reproducibility of the results as well as diminish the risk (even though marginally) of contaminating the liver with diseases that may be carried by animal-derived products.

60 Bibliography

[1] Thomson, A. W. & Knolle, P.A. Antigen-presenting cell function in the tolerogenic liver environment. Nature reviews. Immunology 10, 753–766 (2010). URL http://dx.doi.org/10.1038/nri2858.

[2] Maslak, E., Gregorius, A. & Chlopicki, S. Liver sinusoidal endothelial cells (LSECs) function and NAFLD; NO-based therapy targeted to the liver (2015).

[3] DeLeve, L. D. Liver sinusoidal endothelial cells and liver regeneration. The Journal of clinical investigation 123, 1861–6 (2013). URL http://www.pubmedcentral.nih.gov/articlerender. fcgi?artid=3635729{&}tool=pmcentrez{&}rendertype=abstract.

[4] Wang, L. L. et al. Liver sinusoidal endothelial cell progenitor cells promote liver regeneration in rats. Journal of Clinical Investigation 122, 1567–1573 (2012). URL ://WOS: 000302281800045.

[5] Moniaux, N. & Faivre, J. Key role of sinusoidal endothelial cells in the triggering of liver regeneration (2011).

[6] Senoo, H. Structure and function of hepatic stellate cells. Medical electron microscopy : ofﬁcial journal of the Clinical Electron Microscopy Society of Japan 37, 3–15 (2004). URL http://www. ncbi.nlm.nih.gov/pubmed/15057600.

[7] Sato, M., Suzuki, S. & Senoo, H. Hepatic stellate cells: unique characteristics in cell biology and phenotype. Cell structure and function 28, 105–12 (2003). URL http://www.ncbi.nlm.nih.gov/ pubmed/12808230.

[8] Dixon, L. J., Barnes, M., Tang, H., Pritchard, M. T. & Nagy, L. E. Kupffer cells in the liver. Comprehensive Physiology 3, 785–97 (2013). URL http://www.ncbi.nlm.nih.gov/pubmed/ 23720329http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC4748178.

[9] Bilzer, M., Roggel, F. & Gerbes, A. L. Role of Kupffer cells in host defense and liver disease. Liver international : ofﬁcial journal of the International Association for the Study of the Liver 26, 1175–86 (2006). URL http://www.ncbi.nlm.nih.gov/pubmed/17105582.

61 [10] Tabibian, J. H., Masyuk, A. I., Masyuk, T. V., O’Hara, S. P. & LaRusso, N. F. Physiology of cholangiocytes. Comprehensive Physiology 3, 541–65 (2013). URL http://www.ncbi.nlm.nih.gov/pubmed/23720296http://www.pubmedcentral.nih.gov/ articlerender.fcgi?artid=PMC3831353.

[11] Moscona, A. Rotation-mediated histogenetic aggregation of dissociated cells: A quantiﬁable approach to cell interactions in vitro. Experimental Cell Research 22, 455–475 (1961).

[12] Landry, J., Bernier, D., Ouellet, C., Goyette, R. & Marceau, N. Spheroidal aggregate culture of rat liver cells: histotypic reorganization, biomatrix deposition, and maintenance of functional activities. J Cell Biol 101, 914–923 (1985). URL http://www.ncbi.nlm.nih.gov/pubmed/2411740.

[13] Godoy, P. et al. Recent advances in 2D and 3D in vitro systems using primary hepatocytes, alternative hepatocyte sources and non-parenchymal liver cells and their use in investigating mechanisms of hepatotoxicity, cell signaling and ADME. Arch Toxicol 87, 1315–1530 (2013). URL http://www.ncbi.nlm.nih.gov/pubmed/23974980.

[14] Abu-Absi, S. F., Friend, J. R., Hansen, L. K. & Hu, W.-S. Structural polarity and functional bile canaliculi in rat hepatocyte spheroids. Exp Cell Res 274, 56–67 (2002). URL http://www.ncbi. nlm.nih.gov/pubmed/11855857.

[15] Lin, R.-Z. & Chang, H.-Y. Recent advances in three-dimensional multicellular spheroid culture for biomedical research. Biotechnol J 3, 1172–1184 (2008). URL http://www.ncbi.nlm.nih.gov/ pubmed/18566957.

[16] Glicklis, R., Merchuk, J. C. & Cohen, S. Modeling mass transfer in hepatocyte spheroids via cell viability, spheroid size, and hepatocellular functions. Biotechnol Bioeng 86, 672–680 (2004). URL http://www.ncbi.nlm.nih.gov/pubmed/15137079.

[17] Koide, N. et al. Formation of multicellular spheroids composed of adult rat hepatocytes in dishes with positively charged surfaces and under other nonadherent environments. Experimental Cell Research 186, 227–235 (1990). URL http://www.ncbi.nlm.nih.gov/pubmed/2298241.

[18] Brophy, C. M. et al. Rat hepatocyte spheroids formed by rocked technique maintain differentiated hepatocyte gene expression and function. Hepatology 49, 578–586 (2009). URL http://www. ncbi.nlm.nih.gov/pubmed/19085959.

[19] Kelm, J. M., Timmins, N. E., Brown, C. J., Fussenegger, M. & Nielsen, L. K. Method for generation of homogeneous multicellular tumor spheroids applicable to a wide variety of cell types. Biotechnology and Bioengineering 83, 173–180 (2003). URL http://www.ncbi.nlm.nih.gov/pubmed/12768623.

62 [20] Inamori, M., Mizumoto, H. & Kajiwara, T. An approach for formation of vascularized liver tissue by endothelial cell-covered hepatocyte spheroid integration. Tissue Eng Part A 15, 2029–2037 (2009). URL http://www.ncbi.nlm.nih.gov/pubmed/19320555.

[21] Abu-Absi, S. F., Hansen, L. K. & Hu, W.-S. Three-dimensional co-culture of hepatocytes and stellate cells. Cytotechnology 45, 125–140 (2004). URL http://www.ncbi.nlm.nih.gov/pubmed/ 19003250.

[22] Shteyer, E. et al. Reduced liver cell death using an alginate scaffold bandage: A novel approach for liver reconstruction after extended partial hepatectomy. Acta Biomaterialia 10, 3209–3216 (2014). URL http://dx.doi.org/10.1016/j.actbio.2014.02.047.

[23] Lin, J. et al. Use an alginate scaffold-bone marrow stromal cell (BMSC) complex for the treatment of acute liver failure in rats. International journal of clinical and experimental medicine 8, 12593– 600 (2015). URL http://www.ncbi.nlm.nih.gov/pubmed/26550170http://www.pubmedcentral. nih.gov/articlerender.fcgi?artid=PMC4612855.

[24] Lin, N. et al. Differentiation of bone marrow-derived mesenchymal stem cells into hepatocyte-like cells in an alginate scaffold. Cell Prolif 43, 427–434 (2010). URL http://www.ncbi.nlm.nih.gov/ pubmed/20887549.

[25] Dvir-Ginzberg, M., Elkayam, T. & Cohen, S. Induced differentiation and maturation of newborn liver cells into functional hepatic tissue in macroporous alginate scaffolds. FASEB J 22, 1440–1449 (2008). URL http://www.ncbi.nlm.nih.gov/pubmed/18070820.

[26] Shang, Y. et al. Hybrid sponge comprised of galactosylated chitosan and hyaluronic acid mediates the co-culture of hepatocytes and endothelial cells. J Biosci Bioeng 117, 99–106 (2014). URL http://www.ncbi.nlm.nih.gov/pubmed/23896016.

[27] Chien, H.-W., Lai, J.-Y. & Tsai, W.-B. Galactosylated electrospun membranes for hepatocyte sandwich culture. Colloids and surfaces. B, Biointerfaces 116, 576–81 (2014). URL http: //www.ncbi.nlm.nih.gov/pubmed/24583260.

[28] Berthiaume, F., Moghe, P., Toner, M. & Yarmush, M. Effect of extracellular matrix topology on cell structure, function, and physiological responsiveness: hepatocytes cultured in a sandwich conﬁgu- ration. FASEB journal : ofﬁcial publication of the Federation of American Societies for Experimental Biology 10, 1471–84 (1996). URL http://www.ncbi.nlm.nih.gov/pubmed/8940293.

[29] Dunn, J., Tompkins, R. & Yarmush, M. Hepatocytes in collagen sandwich: evidence for transcriptional and translational regulation. The Journal of cell biology 116, 1043–53

63 (1992). URL http://www.ncbi.nlm.nih.gov/pubmed/1734019http://www.pubmedcentral.nih. gov/articlerender.fcgi?artid=PMC2289333.

[30] Yang, K. et al. Sandwich-Cultured Hepatocytes as a Tool to Study Drug Disposi- tion and Drug-Induced Liver Injury. Journal of pharmaceutical sciences 105, 443–59 (2016). URL http://www.ncbi.nlm.nih.gov/pubmed/26869411http://www.pubmedcentral. nih.gov/articlerender.fcgi?artid=PMC4894499.

[31] De Bruyn, T. et al. Sandwich-cultured hepatocytes: utility for in vitro exploration of hepatobiliary drug disposition and drug-induced hepatotoxicity. Expert Opin Drug Metab Toxicol 9, 589–616 (2013). URL http://www.ncbi.nlm.nih.gov/pubmed/23452081.

[32] Melgar-Lesmes, P., Balcells, M. & Edelman, E. R. Implantation of healthy matrix-embedded endothelial cells rescues dysfunctional endothelium and ischaemic tissue in liver engraftment. Gut (2016). URL http://www.ncbi.nlm.nih.gov/pubmed/26851165.

[33] Ranucci, C. S., Kumar, A., Batra, S. P. & Moghe, P. V. Control of hepatocyte function on collagen foams: sizing matrix pores toward selective induction of 2-D and 3-D cellular morphogenesis. Biomaterials 21, 783–793 (2000). URL http://www.ncbi.nlm.nih.gov/pubmed/10721747.

[34] Turner, W. S. et al. Human hepatoblast phenotype maintained by hyaluronan hydrogels. Journal of biomedical materials research. Part B, Applied biomaterials 82, 156–68 (2007). URL http: //www.ncbi.nlm.nih.gov/pubmed/17183559.

[35] Katsuda, T., Takumi, T., Ochiya, T. & Sakai, Y. Transplantation of a fetal liver cell-loaded hyaluronic acid sponge onto the mesentery recovers a Wilson’s disease model rat. J Biochem 148, 281–288 (2010). URL http://www.ncbi.nlm.nih.gov/pubmed/20562412.

[36] Kim, S. S. et al. Dynamic Seeding and in Vitro Culture of Hepatocytes in a Flow Perfusion System. TISSUE ENGINEERING 6 (2000).

[37] Kaihara, S. et al. Survival and function of rat hepatocytes cocultured with nonparenchymal cells or sinusoidal endothelial cells on biodegradable polymers under ﬂow conditions. Journal of pediatric surgery 35, 1287–90 (2000). URL http://www.ncbi.nlm.nih.gov/pubmed/10999680.

[38] Finoli, A., Schmelzer, E., Over, P., Nettleship, I. & Gerlach, J. C. Open-Porous Hydroxyap- atite Scaffolds for Three-Dimensional Culture of Human Adult Liver Cells. BioMed research international 2016, 6040146 (2016). URL http://www.ncbi.nlm.nih.gov/pubmed/27403430http: //www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC4925947.

64 [39] Liu Tsang, V. et al. Fabrication of 3D hepatic tissues by additive photopatterning of cellular hydrogels. FASEB journal : ofﬁcial publication of the Federation of American Societies for Experimental Biology 21, 790–801 (2007). URL http://www.ncbi.nlm.nih.gov/pubmed/17197384.

[40] Rad, A. T. et al. Conducting scaffolds for liver tissue engineering. J Biomed Mater Res A 102, 4169–4181 (2014). URL http://www.ncbi.nlm.nih.gov/pubmed/24443356.

[41] Kanninen, L. K. et al. Hepatic differentiation of human pluripotent stem cells on human liver progenitor HepaRG-derived acellular matrix. Exp Cell Res 341, 207–217 (2016). URL http: //www.ncbi.nlm.nih.gov/pubmed/26854693.

[42] Tiwari, A., Tursky, M. L., Kirkland, M. A. & Pande, G. Expansion of human hematopoietic stem/progenitor cells on decellularized matrix scaffolds. Current Protocols in Stem Cell Biology 1, 1–15 (2014).

[43] Baptista, P. M. et al. The use of whole organ decellularization for the generation of a vascularized liver organoid. Hepatology 53, 604–617 (2011). URL http://www.ncbi.nlm.nih.gov/pubmed/ 21274881.

[44] Sabetkish, S. et al. Whole-organ tissue engineering: decellularization and recellularization of three- dimensional matrix liver scaffolds. J Biomed Mater Res A 103, 1498–1508 (2015). URL http: //www.ncbi.nlm.nih.gov/pubmed/25045886.

[45] Wang, Y. et al. Method for perfusion decellularization of porcine whole liver and kidney for use as a scaffold for clinical-scale bioengineering engrafts. Xenotransplantation 22, 48–61 (2015). URL http://www.ncbi.nlm.nih.gov/pubmed/25291435.

[46]B uhler,¨ N. E. M., Schulze-Osthoff, K., Konigsrainer,¨ A. & Schenk, M. Controlled processing of a full-sized porcine liver to a decellularized matrix in24h. Journal of Bioscience and Bioengineering 119, 609–613 (2015). URL http://www.ncbi.nlm.nih.gov/pubmed/25468420.

[47] Struecker, B. et al. Porcine liver decellularization under oscillating pressure conditions: a technical reﬁnement to improve the homogeneity of the decellularization process. Tissue Eng Part C Methods 21, 303–313 (2015). URL http://www.ncbi.nlm.nih.gov/pubmed/25164028.

[48] Soto-Gutierrez, A. et al. A whole-organ regenerative medicine approach for liver replacement. Tissue Eng Part C Methods 17, 677–686 (2011). URL http://www.ncbi.nlm.nih.gov/pubmed/ 21375407.

[49] Nari, G. A. et al. Preparation of a three-dimensional extracellular matrix by decellularization of rabbit livers. Rev Esp Enferm Dig 105, 138–143 (2013). URL http://www.ncbi.nlm.nih.gov/pubmed/ 23735020.

65 [50] Shupe, T., Williams, M., Brown, A., Willenberg, B. & Petersen, B. E. Method for the decellularization of intact rat liver. Organogenesis 6, 134–6 (2010). URL http://www.ncbi.nlm.nih.gov/pubmed/ 20885860http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC2901817.

[51] Uygun, B. E. et al. Decellularization and recellularization of whole livers. Journal of visualized experiments : JoVE (2011). URL http://www.ncbi.nlm.nih.gov/pubmed/21339718http://www. pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC3197392.

[52] Hillebrandt, K. et al. Procedure for Decellularization of Rat Livers in an Oscillating-pressure Perfusion Device. Journal of Visualized Experi- ments e53029–e53029 (2015). URL http://www.jove.com/video/53029/ procedure-for-decellularization-rat-livers-an-oscillating-pressure.

[53] Crapo, P. M., Gilbert, T. W. & Badylak, S. F. An overview of tissue and whole organ decellularization processes. Biomaterials 32, 3233–3243 (2011). URL http://dx.doi.org/10.1016/j. biomaterials.2011.01.057.

[54] Price, A. P. et al. Automated Decellularization of Intact, Human-Sized Lungs for Tissue Engineering. Tissue Engineering Part C: Methods 21, 94–103 (2015). URL http://online.liebertpub.com/ doi/abs/10.1089/ten.tec.2013.0756.

[55] Pellegata, A. F. et al. Arterial Decellularized Scaffolds Produced Using an Innovative Automatic System. Cells Tissues Organs 200, 363–373 (2015).

[56] Amend, M. L., Marengo, K. & Perreault, L. Development of a Perfusion Based Decellularization System for Cardiovascular Research (2015).

[57] Gotts, L. & Thyden, R. 2016 Diagnostics : Organs on a Vine (2016).

[58] Chang, R., Emami, K., Wu, H. & Sun, W. Biofabrication of a three-dimensional liver micro-organ as an in vitro drug metabolism model. Biofabrication 2, 45004 (2010). URL http://www.ncbi.nlm. nih.gov/pubmed/21079286.

[59] Faulkner-Jones, A. et al. Bioprinting of human pluripotent stem cells and their directed differentiation into hepatocyte-like cells for the generation of mini-livers in 3D. Biofabrication 7, 44102 (2015). URL http://www.ncbi.nlm.nih.gov/pubmed/26486521.

[60] Gong, H., Agustin, J., Wootton, D. & Zhou, J. G. Biomimetic design and fabrication of porous chitosan-gelatin liver scaffolds with hierarchical channel network. J Mater Sci Mater Med 25, 113– 120 (2014). URL http://www.ncbi.nlm.nih.gov/pubmed/24121873.

66 [61] Ma, X. et al. Deterministically patterned biomimetic human iPSC-derived hepatic model via rapid 3D bioprinting. Proceedings of the National Academy of Sciences 201524510 (2016). URL http: //www.pnas.org/lookup/doi/10.1073/pnas.1524510113.

[62] Skardal, A. et al. A hydrogel bioink toolkit for mimicking native tissue biochemical and mechanical properties in bioprinted tissue constructs. Acta Biomaterialia 25, 24–34 (2015). URL http://dx. doi.org/10.1016/j.actbio.2015.07.030.

[63] Baptista, P. M. et al. Fluid Flow Regulation of Revascularization and Cellular Organiza- tion in a Bioengineered Liver Platform. Tissue engineering. Part C, Methods 22, 199– 207 (2016). URL http://www.ncbi.nlm.nih.gov/pubmed/26772270http://www.pubmedcentral. nih.gov/articlerender.fcgi?artid=PMC4863717.

[64] Mazza, G. et al. Decellularized human liver as a natural 3D-scaffold for liver bioengineering and transplantation. Scientiﬁc reports 5, 13079 (2015). URL http://www.ncbi.nlm.nih.gov/pubmed/ 26248878.

[65] Takebe, T. et al. Vascularized and functional human liver from an iPSC-derived organ bud transplant. Nature 499, 481–484 (2013). URL http://www.ncbi.nlm.nih.gov/pubmed/23823721.

[66] Takebe, T. et al. Generation of a vascularized and functional human liver from an iPSC-derived organ bud transplant. Nat Protoc 9, 396–409 (2014). URL http://www.ncbi.nlm.nih.gov/pubmed/ 24457331.

[67] Huch, M. et al. In vitro expansion of single Lgr5+ liver stem cells induced by Wnt-driven regeneration. Nature 494, 247–50 (2013). URL http://www.pubmedcentral.nih.gov/articlerender. fcgi?artid=3634804{&}tool=pmcentrez{&}rendertype=abstract. 9809069v1.

[68] Nantasanti, S. et al. Disease modeling and gene therapy of copper storage disease in canine hepatic organoids. Stem Cell Reports 5, 895–907 (2015). URL http://dx.doi.org/10.1016/j. stemcr.2015.09.002.

[69] Cheng, N., Wauthier, E. & Reid, L. Mature human hepatocytes from ex vivo differentiation of alginate-encapsulated hepatoblasts. Tissue engineering. Part A 14, 1–7 (2008). URL http://www. ncbi.nlm.nih.gov/pubmed/18333800.

[70] Mirmalek-Sani, S. H., Sullivan, D. C., Zimmerman, C., Shupe, T. D. & Petersen, B. E. Immunogenic- ity of decellularized porcine liver for bioengineered hepatic tissue. American Journal of Pathology 183, 558–565 (2013). URL http://dx.doi.org/10.1016/j.ajpath.2013.05.002.

67 [71] Wang, Y. et al. Genipin crosslinking reduced the immunogenicity of xenogeneic decellularized porcine whole-liver matrices through regulation of immune cell proliferation and polarization. Sci- entiﬁc Reports 6, 24779 (2016). URL http://www.nature.com/articles/srep24779.

[72] Faulk, D. M., Wildemann, J. D. & Badylak, S. F. Decellularization and cell seeding of whole liver biologic scaffolds composed of extracellular matrix. Journal of Clinical and Experimen- tal Hepatology 5, 69–80 (2015). URL http://dx.doi.org/10.1016/j.jceh.2014.03.043http: //www.ncbi.nlm.nih.gov/pubmed/25941434.

[73] Skardal, A., Devarasetty, M., Soker, S. & Hall, A. R. In situ patterned micro 3D liver constructs for parallel toxicology testing in a ﬂuidic device. Biofabrication 7, 31001 (2015). URL http://www. ncbi.nlm.nih.gov/pubmed/26355538.

[74] Kajbafzadeh, A.-M., Javan-Farazmand, N., Monajemzadeh, M. & Baghayee, A. Determining the optimal decellularization and sterilization protocol for preparing a tissue scaffold of a human-sized liver tissue. Tissue engineering. Part C, Methods 19, 642–651 (2013). URL http://www.ncbi.nlm.nih. gov/pubmed/23270591http://online.liebertpub.com/doi/abs/10.1089/ten.tec.2012.0334.

68 A Medium Formulations

69 Table A.1: Complete formulation for DMEM/F-12. From www.sigmaaldrich.com/catalog/product/sigma/d8062

Component g/L Component g/L Inorganic Salts Calcium chloride (dihydrate) 0.1545 Copper(II) sulfate (pentahydrate) 0.0000013 Iron (III) nitrate (nonahydrate) 0.00005 Iron(II) sulfate (heptahydrate) 0.000417 Magnesium chloride (hexahydrate) 0.0612 Magnesium sulfate 0.04884 Potassium chloride 0.3118 Sodium bicarbonate 1.2 Sodium chloride 0.3118 Monosodium phosphate 0.0543 Disodium phosphate 0.07102 Zinc(II) sulphate (heptahydrate) 0.000432 Amino Acids L-alanine 0.00445 L-arginine monohydrochloride 0.1475 L-asparagine (monohydrate) 0.0075 L-aspartic acid 0.00665 L-cysteine dihydrochloride 0.03129 L-cysteine monohydrochloride (monohydrate) 0.01756 L-glutamic acid 0.00735 L-glutamine 0.365 Glycine 0.01875 L-histidine monohydrochloride (monohydrate) 0.03148 L-isoleucine 0.05447 L-leucine 0.05905 L-lysine monohydrochloride 0.09125 L-methionine 0.01724 L-phenylalanine 0.03548 L-proline 0.01725 L-serine 0.02625 L-threonine 0.05345 L-tryptophan 0.00902 L-tyrosine disodium salt (dihydrate) 0.05579 L-valine 0.05285 Vitamins D-biotin 0.0000035 Choline chloride 0.00898 Folic Acid 0.00266 myo-inositol 0.0126 Niacinamide 0.00202 D-pantothenic acid hemicalcium salt 0.00224 Pyridoxal monohydrochloride 0.002 Pyridoxine monohydrochloride 0.002031 Riboﬂavin 0.000219 Thiamine monohydrochloride 0.00217 Vitamin B12 0.00068 Other D-glucose 3.15 Hypoxanthine 0.00244 Linoleic acid 0.000042 Phenol red monosodium salt 0.00863 Putrescine dihydrochloride 0.000081 Sodium pyruvate 0.055 DL-thioctic acid 0.000105 Thymidine 0.000365

Table A.2: Complete formulation for DMEM/F-12 (high glucose). From www.sigmaaldrich.com/catalog/product /sigma/d6546

Component g/L Component g/L Inorganic Salts Calcium chloride 0.2 Iron (III) nitrate (nonahydrate) 0.0001 Magnesium sulfate 0.09767 Potassium chloride 0.4 Sodium bicarbonate 3.7 Sodium chloride 6.4 Monosodium phosphate 0.109 Amino Acids L-arginine monohydrochloride 0.1475 L-cysteine dihydrochloride 0.0626 Glycine 0.03 L-histidine monohydrochloride (monohydrate) 0.042 L-isoleucine 0.105 L-leucine 0.105 L-lysine monohydrochloride 0.146 L-methionine 0.03 L-phenylalanine 0.066 L-serine 0.095 L-threonine 0.095 L-tryptophan 0.016 L-tyrosine disodium salt (dihydrate) 0.10379 L-valine 0.094 Vitamins Choline chloride 0.004 Folic Acid 0.004 myo-inositol 0.0072 Niacinamide 0.004 D-pantothenic acid hemicalcium salt 0.004 Pyridoxine monohydrochloride 0.004 Riboﬂavin 0.0004 Thiamine monohydrochloride 0.004 Other D-glucose 4.5 Phenol red monosodium salt 0.0159 Sodium pyruvate 0.11

70 Table A.3: Complete formulation for MCDB 131. From www.thermoﬁsher.com/pt/en/home/technical- resources/media-formulation.85.html

Component mg/L Component mg/L Inorganic Salts Ammonium molybdate (tetrahydrate) 0.0037 Ammonium metavandanate 0.00060 Calcium chloride (dihydrate) 235.0 Copper(II) sulfate (pentahydrate) 0.0012 Iron(II) sulfate (heptahydrate) 0.283 Magnesium sulfate (heptahydrate) 2464.0 Manganese (II) sulfate (monohydrate) 0.00020 Nickel (II) chloride (hexahydrate) 0.000071 Potassium chloride 298.0 Selenious acid 0.0038 Sodium bicarbonate 1176.0 Sodium chloride 6430.0 Sodium metasilicate (nonahydrate) 2.8 Disodium phosphate 134.0 Zinc(II) sulphate (heptahydrate) 0.00030 Amino Acids Glycine 2.3 L-alanine 2.7 L-arginine monohydrochloride 63.2 L-asparagine (monohydrate) 15.0 L-aspartic acid 13.3 L-cysteine dihydrochloride (monohydrate) 35.0 L-glutamic acid 4.4 L-histidine monohydrochloride (monohydrate) 42.0 L-isoleucine 66.0 L-leucine 131.0 L-lysine monohydrochloride 182.0 L-methionine 15.0 L-phenylalanine 33.0 L-proline 11.5 L-serine 32.0 L-threonine 12.0 L-tryptophan 4.1 L-tyrosine 18.1 L-valine 117.0 Vitamins Biotin 0.0073 Choline chloride 14.0 D-calcium pantothenate 12.0 Folinic acid calcium salt 0.6 Niacinamide 6.1 Pyridoxine monohydrochloride 2.1 Riboﬂavin 0.0038 Thiamine monohydrochloride 3.4 Vitamin B12 0.0136 i-inositol 7.2 Other Adenine 0.135 D-glucose 1000 Lipoic acid 0.0021 Phenol red 12.4 Putrescine dihydrochloride 0.0002 Sodium pyruvate 110.0 Thymidine 0.024

71 72 B Vessel Design

73 Figure B.1: Representation of the complete vessel without the front panel.

Figure B.2: Representation of the complete vessel without the lid and cut through the center of the front support columns. In this image it is possible to see how the acrylic legs fit with the lower steel weight and screw and how these in turn fit in the floor indentations.

74 Figure B.3: Technical drawing of the upper shelf.

Figure B.4: Technical drawing of the lower shelf.