Skip to main content



Executive Summary:
With such a long and arduous path to develop new drug or regenerative medicine therapies, improved discovery and pre-clinical tools could help select more appropriate candidates to succeed in the clinical trial process. A paradigm shift to use more human relevant data earlier in the discovery process would address the problem of poorly predictive animal models.
Stem cells, particularly hIPSC could be potentially be used to produce large stocks of cells of any type, including recapitulation of donor diseases, a feat that would be very difficult or impossible through other means. Some key issues to overcome for their use include developing suitable differentiation protocols and keeping stocks growing in an undifferentiated state
3D culture and organ-on-a-chip technology also have the potential to aid in more predictive models and can be developed together with cell lines to create more complex systems that could provide co-culture of two or more cell types and a greater range of physical and chemical ques to give a model with a more organotypic phenotype.
The TissueGen project has sought to combine the disciplines of hIPSC derived disease models with the 3D cell culture and organ-on-a-chip models to create a human relevant disease model with a focus on the liver amenable to small molecule drug discovery and regenerative medicine therapies.
The partners used the experience of; Takara to provide stem cell hepatocytes from healthy donors. University of Cambridge to provide stem cell hepatocytes and cholangiocytes from diseased donors. CN Bio Innovations to provide their LiverChip® organ-on-a-chip perfused cell culture platform to facilitate and assess the ability of the cells to recapitulate various inherited diseases. Transtissue to develop functionalised scaffolds to be incorporate into the LiverChip®. Micronit to work together with CN Bio Innovations to develop higher throughput versions of the LiverChip® platform.
The project has succeeded in generating a bank of hESC and hIPSC from both healthy and diseased (inherited metabolic disorders) as well as improving protocols for the differentiation of the cells to hepatocyte that exhibit reproducible functional CYP and transporter expressions.
The function of hESC derived hepatocytes was shown to be significantly enhanced through 3D culture in the LiverChip® over the same lots in 2D culture and (numerous) functionalised scaffolds were incorporated to test their effect in the system. In addition, the throughput of the 12 well LiverChip® device was augmented with the subsequent development of 24 and 36 well devices that were shown to support extended culture with comparable microtissue formation and function.
A 3D disease models were then developed using hIPSC hepatocytes and cholangiocytes to model glycogen storage disease, OCT deficiency, Alagille syndrome and Cystic Fibrosis. The Cystic Fibrosis model was shown to recapitulate key aspects of the disease in vitro and thus provides a unique system to model the biliary disease induced by the condition.
The project has produced commercially available stem cell derived hepatocytes as well as a commercially available higher throughput LiverChip® device. The range of 3D disease models can help identify and modulate new targets in the disease pathways. The healthy cell models themselves can be further developed to be a useful tool in toxicological testing.

Project Context and Objectives:
The path to developing a new drug or regenerative medicine therapy is long and large numbers of potential new therapies fail in clinical trials due to a lack of efficacy or unexpected toxicity. This suggests improved discovery and pre-clinical tools are required to more appropriately select novel compounds or regenerative medicine therapies through the discovery pipeline. One key issue with current technologies is the reliance on the use of animal models which are in some cases poorly predictive of human response. In some areas there are no suitable animal models or ethical concerns make the use of animal e.g. primates unacceptable. A paradigm shift is required to bring more human relevant data earlier in the drug or regenerative medicine discovery process.
Stem cells and particularly human induced pluripotent stem cells (hIPSC) with appropriate differentiation protocols offer the opportunity to produce large stocks of cells of any type and in the case of hIPSC recapitulate donor diseases. The ability of cells differentiated from hIPSC of diseased donors to recapitulate the disease phenotype, in particular genetically inherited disorders allows the generation of disease models which would be difficult/impossible through other means. It is envisaged stem cells will play an important role in the paradigm shift required in the drug discovery process. Two significant issues are i) to date developing differentiation protocols which allow stem cells to differentiate to fully mature phenotypes has proved a challenge ii) to keep undifferentiated stocks of stem cells growth on feeder cells is usually required. The former is particularly the case with differentiation to hepatocytes, the parenchymal cell of the liver. Most stem cell derived hepatocytes have a foetal phenotype with loss of the stem cell markers, production of liver specific secreted proteins and expression of important drug metabolising enzymes (CYP450) but at levels lower than those found in fully mature primary human hepatocytes. Improving differentiation protocols and implementing new technologies which drive stem cell derived cells to more mature phenotypes remains a key challenge in the field.
A second area which has held great promise for improvements in drug discovery is 3D cell culture and organ-on-a-chip technologies. It is widely acknowledged the culture of cells in 2D monolayers on hard plastic surfaces is unsatisfactory for certain cell types, causing changes in or loss of phenotype. There has been much interest in shifting from 2D culture to 3D culture either as unsupported cells aggregates (spheroids) or in scaffolds supported configuration. The appropriate format depending to some degree on the cell type being cultured. This has proved fruitful and there are numerous examples in which the culture of cells in 3D has improved phenotype or extended the useful duration of culture. In recent years it has become more widely appreciated that in a number of cases simply taking the same cells from 2D to 3D is insufficient. More complex systems are required which may involve co-culture of two or more cell types and provision of a greater range of physical and chemical ques to achieve a more organotypic phenotype. Here organ-on-a-chip technologies offer a solution which combines 3D cell culture, typically co-culture, together with fluid flow and optionally mechanical stimulation to achieve an improved phenotype. Organ-on-a-chip systems are inherently more complex than simple 2D cultures but have the potential to offer levels of information more similar to animal models.
The TissueGen project has sought to combine three emerging and potentially transformative technologies, namely hIPSC derived disease models, 3D cell culture and organ-on-a-chip to create human relevant disease models with a focus on the liver amenable to small molecule drug discovery and regenerative medicine therapies. The team brings together expertise in the fields of stem cell production, development and optimisation of differentiation protocols, 3D cell culture, scaffold technologies and the design, development and commercialisation of organ-on-a-chip technology. The project has five major objectives:
• Generate a bank of hESC and hIPSC from healthy donors and donors with a number of inherited metabolic disorders (IMD) of the liver, together with protocols to keep cells in an undifferentiated form.
• Development of improved protocols for the differentiation of human embryonic stem cell (hESC) and hIPSC to hepatocytes
• Optimisation of scaffold technologies for 3D cell culture of stem cell derived hepatocytes
• Design and development of medium/high throughput organ-on-a-chip technologies to enable scalable testing of stem cell derived hepatocytes
• Combine the above aspects to produce human relevant 3D liver disease models and provide proof of concept data for the utility of these models for drug discovery and regenerative medicine applications.
Stem cells from healthy (Takara, CEL) and diseased (University of Cambridge, UCAM) donors will be differentiated to hepatocyte like cells using defined and where possible xeno-free protocols. UCAM will also develop a modified differentiation process to allow stem cells to be differentiated to cholangiocytes, the cells which line bile ducts in the liver. At various stages of the differentiation process cells will be seeded into 3D and cultured in CN Bio Innovations (ZXL) LiverChip® organ-on-a-chip perfused cell culture platform. The functional performance of the cells will be assessed as will their ability to recapitulate various IMDs. Through the project functionalised scaffolds able to guide differentiation will be developed (TransTissue (TRA) and incorporated into the LiverChip® platform. Micronit (MIC) will work together with CN Bio to develop higher throughput versions of the LiverChip® platform offering improved function and lower number of cells per well. Taken together the key scientific question the project seeks to address is: by carefully defining differentiation protocols and utilising advanced cell culture techniques (3D, organ-on-a-chip) can stem cell derived hepatocytes be driven to a more mature phenotype.

Project Results:
Development of hESC and hIPSC Cell Banks and protocols for 2D differentiation and culture.
The ability to maintain stem cells in an undifferentiated state, store these cells and then to forward differentiate those cells in a controlled xeno-free process to cells of mature phenotype remains a significant challenge. The generation of rigid hepatocyte protocols with comparable functionality to fresh human hepatocytes is of crucial importance if hESC/hIPS cells are to be used as a source to eventually complement or replace existing models based on today’s techniques in the industry. As part of this development, CEL undertook a number of important steps:
1) CEL developed a very robust protocol to generate hepatocytes in any 2D format. This is of uttermost importance to be able to develop future novel systems.

2) CEL has also developed a method to freeze and thaw hepatocytes that keeps their functionality similar to the fresh hepatocytes generated.

3) An extensive study to compare hepatocytes in 2D format derived from hiPS and hES cell lines, to frozen human primary hepatocytes at RNA, characterisation by immunostainings and measure hepatocyte activity levels of certain detoxification enzymes (so called CYP analyses) has been performed. This is of importance to better understand the functionality of the hepatocytes and understand what can be done for further maturation of the cells. Exemplified data in Figure 1.

4) CEL has undertaken an extensive screen of >500 compounds and combinations thereof, to try to increase the CYP activity in the hepatocytes.

5) When hIPS/hESC cells are differentiated into Definitive Endoderm (DE, an early progenitor stage of cells to become hepatocytes), -hepatoblast-hepatocytes, the surface as well as the seeding densities are crucial to generate a robust system. For establishment of hepatocytes in 3D bioreactor formats it is necessary to establish several different delivery options. CEL is continuously evaluating this process, and for the progress within TissueGen, CEL has tested/evaluating;
a. Delivery of fresh hepatocytes in attached formats
b. Delivery of hepatocytes in suspension
c. Delivery of frozen hepatocytes.
d. Delivery of frozen DE cells for further differentiation of end user. Exemplified data in figure 4.
e. Delivery of spheroids differentiated to “hepatospheres”

6) The Batch to batch reproducibility of the developed protocols has been extensively tested with data in Figure 2.
7) To better be able to study the capacity of cells to bind to different scaffolds within programme, CEL has used a GFP cell line that constantly expresses Actin-GFP. This certainly facilitates and enhances the progress when studying reattachment of cells to new surfaces. Exemplified data in Figure 3.

Figure 1- Comparative study of CEL hepatocytes (hES-hep = hepatocytes derived from a hESC cell line & hIPS-hep = hepatocytes derived from a hIPS cell line) vs human primary hepatocytes 4h and 48h after thawing. The study is represented by immunostaining images.

Figure 2- The five batches presented in figure 4 were harvested for qPCR analyses 6 and 11 days post thaw. The small batch-to -batch variations as well as the stability over time for several CYPs were evaluated and is presented in the figure.

Figure 3- In attempt to study the rebinding potential of either spheroids differentiated to hepatocytes (A-C) or hepatoblasts (D) to a MESH filter (PGA 20mmx30mmx1.1mm pre-coated with collagen 10ug/ml solution, MESH provided from partner TransTissue), a monoclonally derived cell line constitutively expressing Actin-GFP was used. Cells were grown as spheroids and differentiated to hepatocytes (A-C), or differentiated in a 2D format to hepatoblasts (D) and re-plated to the MESH filter. Seven days post re-seeding, the filters (C, D) were studied in a fluorescence microscope. Quite unexpectedly the spheroids had fallen through the MESH, while a high concentration of single cells seeded at hepatoblast stage had successfully attached to the MESH filter. B) Represents a successful rebinding of spheroids to collagen surface when reseeded to a 2D format.
Functionalisation of Scaffolds
The ability to design and functionalise scaffolds is of key importance for 3D cell culture. To prove the utility of functionalised scaffolds polyglycolic acid (PGA) scaffolds functionalised with bioactive molecules from platelet rich plasma (PRP) were developed and the culture of cells on these scaffolds assessed. In 2012 the composition of platelet-rich plasma (PRP) particularly with regard to putative bioactive chondrogenic growth factors (GF) was analyzed. PRP was obtained from healthy blood donors by apheresis using an automated blood collection system. The concentration of platelets was 0.6-1.3x1010/ml. Leucocytes were less than 0.3x104/ml. PRP activation was performed by repeated freeze/thaw cycles.
GF composition of a pool of 6 PRP preparations was determined using Protein Antibody Membrane Arrays covering 507 GF, signaling molecules and receptors. To verify the chondrogenic GF variability in PRP, Growth Factor Antibody Membrane Arrays covering 26 GF and 15 GF receptors were applied to 6 individual PRP preparations. Selected GF involved in chondrogenic differentiation were quantified by Enzyme-Linked Immunosorbent Assay (ELISA).
Results showed that 417 out of 507 possible detectable proteins were present in the PRP pool, including 35 intracellular proteins, 44 transmembrane proteins, 108 receptor proteins, 69 extracellular proteins, 36 interleukins, 16 metalloproteinases and their inhibitors, 33 chemokines and 76 GF.
In individual PRP preparations GF content was analyzed by quantification of spot intensity. The highest spot intensities were found for the epidermal growth factor (EGF; 119.2-152.7 mean 140.5) platelet-derived growth factor-AA (PDGF-AA; 52.9-72.8 mean 60.9) PDGF-AB (92.3-140.4 mean 114.0) and PDGF-BB (97.0-142.4 mean 120.1). Differences in the presence of GF between the individual PRP preparations were found for heparin-binding epidermal growth factor (HB-EGF; 0-5.4 mean 1.3) neurotrophin-3 (NT-3; 0.0-4.2 mean 1.9) NT-4 (0.0-5.7 mean 2.9) TGF-alpha (0.0-13.8 mean 2.3) TGF beta 1 (0.0-3.8 ,mean 1.3) and TGF-beta 2 (0.0-4.3 mean 2.7). Furthermore, BMP-2, 4 and 7, CTGF, FGF-2, FGF-18, growth and differentiation factor 5 (GDF-5), TGF-beta 2, TGF-beta 3 and insulin-like growth factor-1 were detectable in all individual PRP preparations.
Quantification of selected GF by ELISA showed an average of 0.31ng/ml bone morphogenic protein-2 (BMP-2), 0.50ng/ml connective tissue-factor (CTGF), 0.76ng/ml fibroblast growth factor-2 (FGF-2) and 0.59ng/ml transforming growth factor-beta3 (TGF-beta 3) (Krüger et al. 2013).
In a second part an efficient method for combination of growth - and differentiation factors with three-dimensional matrices made of polyglycolic acid and hyaluronan (PGA-hyaluronan) was established. As a model substance TGF-beta 3 was chosen to test optimal loading and preservation of protein stability in the matrix by lyophilisation, which was verified by demonstration of biological activity after factor release.
Results showed that a matrix loaded with initially 200ng/ml TGF-beta 3 led to a release of 150 ng/ml TGF-beta 3 in the first 24h followed by a release of additionally 40 ng/ml after 14 days in vitro. The recovery rate of TGF-beta 3 after a 14 day release was 95%. The use of 200ng/ml lyophilized TGF-beta 3 as a single dose in a high density pellet culture of human expanded mesenchymal progenitor cells (MPC) resulted in the induction of the gene expression of aggrecan, collagen type II and cartilage oligomeric matrix protein in these cells, which was verified by real-time RT-PCR on RNA level two days after supplementation. On protein level a proteoglycan- rich matrix could be detected histochemically 4 days after supplementation (Krüger et al. 2014).
The third part of the project describes the establishment of an implant consisting of a combination of PRP with the PGA-hyaluronan scaffold. It could be demonstrated that the use of 5 % PRP as an chondrogenic inductor for the differentiation of MPC in PGA-hyaluronan scaffolds over a period of
21 days was accompanied by an increase of the gene expression of typical chondrogenic markers and a deposition of cartilage-like matrix proteins. In contrast to TGFß-3 mediated chondrogenesis, type X collagen, a marker usually found in hypertrophic cartilage was decreased, if 5% PRP was added to the cultures.
Furthermore, a qualitative analysis of presence or absence of growth and differentiation factors in human platelet rich plasma (PRP) (n=6) was performed. The chondrogenic differentiation factors TGF-beta 1, 2 and 3, as well as Insulin like growth factor binding protein, fibroblast growth factor and epidermal growth factor as typical proliferation factors could be detected. Additionally, the presence of stem cell factor receptor was observed (Krüger et al. 2013). Also, the biological activity after freeze drying process of PRP could be demonstrated in a migration assay. For assessment the freeze-dried PRP-loaded PGA-hyaluronan scaffolds were reconstituted in phosphate-buffered saline and centrifuged. The resulting eluate was tested on its migratory effect on human mesenchymal progenitor cells in a modified Boyden chamber assay. No differences in the migratory effect between fresh or reconstituted freeze-dried PRP on MPC was detected (unpublished data).
MPC as well as human multipotent stromal cells (MSC) isolated from bone marrow aspirates maintain their chondrogenic potential in PGA-hyaluronan scaffolds as shown by induction with 10ng/ml TGFß-3 over a period of 21 days. Additionally, the presence of stem cell factor receptor could be detected by flow cytometric analyses (Patrascu et al. 2013).
These results showed that lyophilisation of PRP has no negative impact on its migratory and chondrogenic effect on MPC and MSC. MPC and MSC have the potential to undergo chondrogenic differentiation in PGA-hyaluronan scaffolds. Three dimensional culture of MPC and MSC can be used as an assay system to identify possible new inductors in single dose or under continuous addition of the substances.
Culture of stem cell derived hepatocytes in LiverChip®
At the start of the project a 12 well LiverChip® system was available. This consisted of 12 fluidically isolated bioreactors, each containing a scaffold designed for the culture of primary hepatocytes. The system is designed to mimic the liver sinusoid. Perfusion for each bioreactor is driven by a pneumatically actuated pump embedded into the plate. Cell culture media is perfused directly through the scaffold containing the cells ensuring cells are adequately supplied with oxygen. Figure 4c shows a schematic of the LiverChip® system Stem cell derived hepatocytes are seeded into the LiverChip® as single cells and over a number of days undergo morphogenesis to form 3D microtissue structures within the scaffold. In some experiments for comparison cells were cultured in 2D or primary human hepatocytes (PHH) were used to benchmark the performance of the stem cell derived hepatocytes. Figure 4a shows the formation of microtissues within the scaffold. The morphology of the microtissues formed from PHH and stem cell derived hepatocytes in the system is similar.

Figure 4- a) Schematic of the LiverChip® scaffold, b) LiverChip® 12 well system, c) Cross Sectional view of a single LiverChip® bioreactor, d) Microtissues formed within the scaffold in LiverChip®

The function of stem cell derived hepatocytes was significantly enhanced through 3D culture in the LiverChip® organ-on-a-chip system. Figure 5 shows a comparison of the same lot of stem cell derived hepatocytes cultured in 2D and LiverChip®. The cells in LiverChip® produce more albumin, a marker of mature cells and less alpha-fetoprotein (AFP) a marker of foetal like cells than in 2D. The gene expression of important drug metabolising CYP450 enzymes is between 50 and 125 folder higher for cells cultured in LiverChip®. When the performance of stem cell derived hepatocytes is compared to PHH in LiverChip® a similar morphology of cells is noted and the levels of cell death are comparable (Figure 6). For this particular lot of cryopreserved stem cell derived hepatocytes the albumin production level on a per cell basis was lower than PHH, but was maintained throughout the culture period.

Figure 5- Functional comparison of iPSC-derived hepatocyte like cells in 2D and in ZXLs LiverChip® platform. a) Cells could be reattached to collagen coated plates or scaffolds. b) Excretion of maturity marker proteins albumin and c) AFP were equivalent in LiverChip® and 2D culture but gene expression analysis demonstrated greater expression of mature hepatocyte marker genes. Data is mean +/- SD of 3 wells for albumin and AFP secretion. Gene expression compared using the ΔΔCt method using GAPDH as a stable housekeeping gene

Figure 6- Culture of iPSC-derived hepatocytes and PHH in MICs bioreactor platform A) Light micrographs of collagen coated scaffolds after 11 days of culture. B) LDH production during culture. C) Excretion of albumin during culture measured by ELISA. Data is mean +/- SD of 3 wells.

High Throughput Organ-on-a-Chip
CN Bio Innovations together with Micronit have developed 24 and 36 wells versions of the LiverChip® platform. In developing these scale down versions special consideration was given to ensuring the 3D microtissues remained sufficiently oxygenated. Oxygen probes were used to measure the dissolved oxygen concentration in the media before and after the cells, from this and given a known flow rate the oxygen consumption can be calculated (Figure 7). To investigate the effects of changing oxygen tension LiverChip® cultures were run under reduced oxygen tension in a hypoxic incubator. LiverChip® perfusion culture plates were installed in a hypoxic incubator set at 10% oxygen, circa 100µM saturation concentration. Cells cultured under these conditions experienced significantly lower oxygen tensions throughout the culture period (Figure 8) when compared to culture under 21%, i.e. ambient oxygen. At both conditions microtissue formation and albumin secretion were acceptable, indicating primary human hepatocytes are tolerant to a wide range of oxygen tensions during culture and seeding. To determine if changes to the design of the LiverChip® could be used to change the oxygen tension experienced by the 3D microtissues, versions of the bioreactors with reduced oxygen transfer area were produced. Figure 9 shows a comparison between a regular 12 well LiverChip® and a LiverChip® with only 10% of the available oxygen transfer area. The oxygen tension experienced by the cells is significantly reduced in the low oxygen transfer bioreactor, however this did not have a detrimental effect on microtissue formation or function.

Figure 7- Schematic of the on-line, non-invasive oxygen sensors (Lucid Scientific Inc., U.S.) used to monitor the oxygen concentration in the culture medium before and after the 3D microtissues. The red line indicated oxygen tension below the scaffold. The blue line represents oxygen tension above the scaffold after passing over the cells.

Figure 8- Comparison of oxygen tensions, tissue formation and function of hepatocyte microtissues in conventional and hypoxic incubators. Albumin was quantified on day 3 of culture in a minimum of 3 wells per condition (data is mean +/- Standard deviation). Standard = standard incubator throughout. Standard/hypoxic = seeded in standard incubator and removed to hypoxic after 24h. Hypoxic = seeded and maintained in hypoxic incubator.

Figure 9- Oxygen concentrations measured in single wells of a regular LiverChip® or a modified plate with a surface channel having only 18% of the oxygen transfer area.

A 24 and 36 well high throughput LiverChip® were developed using different materials of construction and construction methods. MIC developed the 24 well unit which in its final iteration was produced using polystyrene (or derivative materials) and thermally bonded (Figure 10). The 36 well unit was developed by ZXL using polysulfone and assembled using screws (Figure 11). Both units were tested for the culture of stem cell derived hepatocytes and found to be able to support extended culture (Final figure in pervious section and Figure 12).

Figure 10- Photograph of a thermally bonded 24-well bioreactor. The size of the device is 128 x 85 x 7 mm3.

Figure 11- Computer Aided Design (CAD) drawings of the LC36. (a) A Polysulfone top plate which contains 36 individual channels and an acrylic bottom plate which provides the system with pressurised air. (b) Components that fit directly into the culture wells. A filter, scaffold and retaining ring are placed inside an insert accordingly. The assembled component is then pushed down into a culture well.

Figure 12- Functional comparison of iPS-derived hepatocytes cultured in LiverChip® and ZXLs LC36 bioreactor platform. A) Light micrographs of collagen coated scaffolds after 7 days of culture. B) LDH production during culture. C) Excretion of albumin during culture measured by ELISA. Data is mean +/- SD of 3 wells.
3D Disease Models
The production of disease models of IMD was a key output for the project. Stem cell derived hepatocytes (HLC), stem cell derived cholangiocytes (CLC) and co-cultures of the two cell types were used to produce disease models in a number of 3D format. The disease models were for glycogen storage disease (GSD), OCT deficiency (OCTD), Alagille syndrome (AGLS) and cystic fibrosis (CF). Genetic GSD is associated with an in born error in the ability to metabolise glycogen, Figure 13 shows the accumulation of glycogen in cells from a donor with GSD, as compared to a healthy control. Glycogen is stained purple in the images. Similar results were achieved for OCT deficiency, a disease associated with the urea cycle. The modelling of AGLS a disorder leading to abnormalities of the bile ducts was more complex and required not only differentiation of hIPSC to cholangiocyte like cells but also inhibition of the NOTCH signalling pathway. Figure 14 shows hIPSC from an AGLS patient differentiated in 3D to cholangiocyte like cells expressing the cholangiocyte markers SOX9 and CK19.
Cystic Fibrosis (CF) is the most common lethal genetic disorder in the Caucasian population. The disease is caused by mutation in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR) protein. CFTR is s a multi-domain cAMP-regulated chloride channel expressed in polarized epithelia lining many tissues which is required to regulate the components of sweat, digestive juices, and mucus. The most common mutation causing CF is an absence of phenylalanine at position 508 (DF508 CFTR) which disrupts the folding pathway of CFTR in the endoplasmic reticulum and ultimately block the expression of CFTR at the cellular membrane. The precise mechanism by which CFTR dysfunction leads to the phenotypic disease remains unclear. However, the absence of CFTR strongly affect the activity of several organs including the lungs but also in the biliary system. CFTR gene mutations in cholangiocytes result in reduced intra-luminal chloride secretion, increased bile viscosity, and focal biliary cirrhosis secondary to bile plugs occluding the intrahepatic bile ducts. The hIPSCs were generated from skin fibroblasts of a patient homozygous for the most common CF mutation ΔF508 (CF- hIPSC) (Fibroblasts obtained from the biodepository Coriell) and then differentiated into CLCs using our 3D culture system based on Matrigel. CF-hIPSC derived CLCs (or CF-CLCs) expressed markers (Figure 15a) and displayed functionality characteristic of biliary epithelial cells (Figure 15b). Transcription of the CFTR gene was confirmed using QPCR while immunofluorescence analyses detected minimal CFTR protein expression (Figure 15 c,d). Finally, we used the fluorescent chloride indicator N-(6-methoxyquinolyl)acetoethyl ester (MQAE) to monitor intraluminal chloride concentration, which is physiologically regulated through CFTR. Wild type (WT) CLC organoids appropriately modified intraluminal chloride in response to media with varying chloride concentrations while no change was observed in CF-CLCs (Figure 15 e,f) thereby confirming the absence of functional CFRT in these cells. Overall, these results demonstrate that CF-CLCs recapitulate key aspects of CF in vitro and thus provide a unique system to model the biliary disease induced by CF.

Figure 13- Hepatocytes like cells generated from GSD hIPSCs lines accumulate glycogen when compared to cells generated from control hIPSC lines.

Figure 14- AGLS-hIPSCs can differentiate into cholangiocytes spheroid expressing biliary markers such as CK19 and Sox9.

Figure 15- Modelling CF using hIPSCs derived CLCs grown in 3D based matrigel culture system. (a) Q-PCR analyses showing that CF-hIPSCs can be differentiated into CLCs. (b) ELISA showing that CF-CLCs can secrete gamma GT. (c) Expression of CFTR in CF-CLCs. (d) Immunostaining analyses showing the absence of CFTR in CF-CLCs. (e) MQAE assay showing that CF-CLCs can only transport chloride when grown in the presence of VX809

Potential Impact:
The TissueGen project has produced a number of important findings with broad potential scientific impact:
Definition of robust, protocol based xeno-free culture and differentiation protocols for hIPSC (and hESC) derived hepatocytes and cholangiocytes.
The need for robust feeder free stem cell maintenance and differentiation protocols has been well recognised for a number of years. The need to use feeder cell layers to maintain stem cells is undesirable as it i) introduces flask to flask and batch to batch variability, ii) represents an uncontrolled aspect within the biology making the process ill defined, iii) adds cost as multiple cell lines must be maintained. A second important issue, particularly if the stem cells, or any derived products are ultimately to be used for regenerative medicine applications is the presence of animal derived products in maintenance or differentiation media. Animal products are less well defined than chemical products and bring with them the risk of zoonotic transmission of pathogens.
The protocols developed in the TissueGen project by CEL seek to address be chemically defined and xeno-free. CEL has incorporated these new protocols into manufacturing and now makes available products, such as stem cell derived hepatocytes which are generated using TissueGen based protocols. This will allow pharmaceutical companies and academics access to these well-defined cells for their research, generating revenues and profits for CEL and consequent taxes within the EU economic area.
Development of medium throughput organ-on-a-chip bioreactors for the 3D culture of primary and stem cell derived liver cells
Liver cell culture has traditionally been considered difficult owing to the rapid dedifferentiation of primary cells cultured in monolayer 2D formats. Whilst in recent years this problem has been to some degree overcome by improved culture protocols and a regular supply of high quality cryopreserved cells there remains a need for i) longer lasting primary cultures, ii) cultures derived from non-finite sources, e.g. stem cell derived hepatocytes which more closely recapitulate the phenotype of primary cells. Organ-on-a-chip or more broadly bioreactor type technologies have offered advancements in this area but many bioreactors have very low throughput with only one or two reactors per unit. Whilst this makes them suitable for interrogating basic biological problems, this makes them inherently unsuitable to the production of disease models which had form a part of the screening and discovery cascade. Thus the need for higher throughput organ-on-a-chip bioreactors which are open well and based on standard footprints utilised by the pharmaceutical and regenerative medicine sectors is clear.
The 24 and 36 well organ-on-a-chip liver bioreactor units developed as part of the TissueGen project represent amongst the highest throughput units currently available. The designs are based on a standard microtiter plate footprint and they are of open well design, cell culture media can be changed simply by pipette, making them compatible with standard cell culture techniques and robotic liquid handling. The flow within the organ-on-a-chip bioreactors has been carefully characterised and is reproducible from unit to unit ensuring precise control of the cellular microenvironment, a facet which clearly assists in the reproducibility of primary cell culture and stem cell differentiation. Many microfluidic organ-on-a-chip cell culture devices make use of polydimethylsulfoxide (PDMS) a polymer well suited to prototyping and cell culture owing to its low cost, easy mouldability and high oxygen permeability. The devices designed in TissueGen have specifically sort to exclude this material owing to a number of undesirable properties namely: i) high sorption of small molecule drugs, particularly lipophilic compounds ii) leaking of unbound polymer into the cell culture media which may have detrimental effects on cells. All TissueGen units are free from this material and in the case of the 24 well unit designed by MIC the unit is predominantly composed of polystyrene the industry standard for cell culture or a closely derivative material.
The units developed in the TissueGen project are near commercial prototypes and it is expected within the next 6-12 months ZXL will make the 36 well version available as a product or for service/research collaborations. This will allow the wider community access to the developments made through the TissueGen project, generate revenues and profits for ZXL/CN Bio and consequent taxes within the EU economic area.
A range of 3D disease models for IMD utilising hIPSC derived hepatocytes and cholangiocytes
The disease models generated through the use of hIPSC are difficult to replicate through other technologies making them a unique resource for those studying IMD. Together with study of the basic biology of the diseases the models are amenable to use as a screening tool in the pharmaceutical industry. As the differentiated cells used in the models are derived from a bank of hIPSC there is essentially an unlimited supply of cells making large screening campaigns against multi-million compounds decks feasible.
UCAM has licenced a number of the differentiation protocols and disease models developed through TissueGen to a spin out company (Definigen, Cambridge, UK). Definigen is making hIPSc derived hepatocytes and other cells available either fresh or cryopreserved to academic and pharmaceutical companies, generating revenues for the company, employment and tax revenues within the European economic area.
Improved in vitro models should lead to societal benefit in bringing forward more rapidly efficacious new therapies with improved safety and tolerability profiles for IMDs. There are currently over 150,000 people afflicted by IMD worldwide of with one in every 5000 live births in Europe suffering from an applicable condition. Many IMDs are life limiting resulting in severe symptoms and reduced life expectancy. Through an ability to more fully understand the basic biology and pathology of these diseases using hIPSC derived models new targets and modes of action will likely be identified. Therapies can then be developed against these targets which may relate to the aberrant gene itself or to downstream pathways which whilst not curing the disease may lessen symptoms and improve length or quality of life.
A spill over benefit of developing liver disease models is that the comparator models developed with healthy donors may find utility as tools for the in vitro assessment of the toxicity of small molecules and cell based therapies. Regulatory toxicology still relies heavily on animals as these are immune competent models capable of capturing the complex interactions between organs. There has been an increasing push to look for improved in vitro liver models able to provide complimentary information to animal models. Hepatocytes and hepatocyte co-cultures with cholangiocytes derived from hIPSC may prove to be a useful tool for toxicology investigation. Indeed it may also be of interest to compare the toxicity of new agents in the healthy and diseased models to compare how these different patient populations will react.
Building Relationships
Through the project both ideas and materials were exchanged between academic laboratories and companies within the project. Frequent discussions were held by telephone and email to plan the project, discuss issues and brainstorm solutions, this built relationships between researchers in the UK, Netherlands, Germany and Sweden.
MIC and ZXL are in discussions regarding commercialisation of the bonded 24 well bioreactor developed by MIC and biologically validated at ZXL during the TissueGen project. ZXL is a leading supplier of organ-on-a-chip solutions, particularly with relation to liver and the MIC bioreactor with some further development may be an attractive product.
Further Research
Through working relationships built as part of the TissueGen grant CN Bio Innovations (ZXL) and Definigen the UCAM spin off company have applied and been successful in securing further grant funding to continue to develop 3D models utilising hIPSC derived hepatocytes. In 2015 they were awarded an Innovate UK grant to develop a 3D non-alcoholic steatohepatitis (NASH) model using hIPSC derived hepatocytes and to incorporate this new model into a multi-organ platform with hIPSC pancreatic cells. The use of a multi-organ platform enables the interplay of the liver and pancreas to be explored and the importance of organ interaction in the disease pathology to be studied.
The results of project have been disseminated through a variety of mechanisms:

• Peer –reviewed scientific publications
• News and updates on the public project website
• News features on company websites
• Mini-workshop at European organ-on-a-chip conference, Cambridge UK April 16

List of Websites:
David Hughes
CN Bio Innovations Ltd
Biopark, Welwyn Garden City, U.K.