Report on the ‘ISCBI Workshop on the delivery of High Quality Induced Pluripotent Stem Cells (iPSC) Resources’, held February 21st-22nd 2012, at the National Institute for Biological Standards and Control, Potters Bar, UK
Earlier this year, the International Stem Cell Banking Initiative (ISCBI) held the ‘ISCBI Workshop on the delivery of High Quality iPSC Resources’ at the National Institute for Biological Standards and Control, Potters Bar, UK. ISCBI is a global network of pluripotent stem cell banks that work together to promote and facilitate ‘best practice’ in stem cell research and the delivery of pluripotent cells for clinical use (Crook et al. 2010 In Vitro Cell Dev Biol 46: 169-172). The workshop was organised by the UK Stem Cell Bank , University of Massachusetts and the British Consulate-General Boston, and brought together experts in a number of key stem cell areas including somatic cell procurement and reprogramming for iPSC derivation, as well as iPSC culture, differentiation, characterising, and banking.
The workshop started with a welcome by Glyn Stacey, Director of the UK Stem Cell Bank followed by a Keynote Session consisting of three talks on ‘The Landscape of iPSC Derivation and Use’.
Pete Coffey (University College, London, UK) spoke on ‘The Clinical Application of iPSC for Eye Disease’ and described the work of ‘The London Project to Cure Blindness’. This project focuses on age related macular degeneration (AMD) including the ‘wet’ and ‘dry’ types of this disorder. Both types lead to atrophy of the retinal pigment epithelium (RPE). Currently, ‘wet’ AMD is treated by injections of anti-angiogenic drugs such as Lucentis into the back of the eye every 6 to 8 weeks or by the translocation of peripheral RPE into the maculae. There is currently no treatment available for the dry form of the disorder.
The potential for using pluripotent stem cells to generate RPE for treating AMD has been shown using human embryonic stem cells (hESCs). Methods of graft delivery are being evaluated and include delivery on an artificial membrane to facilitate attachment of the RPE cells to Bruch’s membrane at the back of the retina. Studies using rats and pigs have demonstrated that the transplantation of RPE does improve the sight of the animals. The research is being supported by Pfizer who has agreed to manufacture the membranes used in the procedure.
The therapeutic benefit of treating AMD with hESCs extends to substantial cost savings whereby treating a single patient for wet AMD costs approximately £15,000 per year compared to £4,000 using hESC-derived RPE cells. There are some 700,000 cases of age related macular regeneration in the UK. Treatment with RPE would save the NHS around £2,000,000 annually. To date, patients treated with RPE have recovered some vision for up to 8 years, with no outright failures reported to date.
Attention is now being turned to the potential use of iPSC to produce cells for therapy as this would ameliorate the need for immunosuppression. ACT has conducted a clinical trial and has reported that hESC derived iPSC cells transplanted into patients have resulted in improved vision. These results are currently being debated by the scientific community since the results are open to interpretation.
Peter Andrews (University of Sheffield, UK) discussed ‘Culture Adaptation of ES cells’. Although focussing on hESC, Peter believed that the principles described in his talk could be applied to iPSC. He described the selection process for an hESC within a heterogenous population that pre-disposes the cell to self-renew, suggesting that nullipotency is the ultimate end-point for this type of cell. Nullipotency, a recessive trait, can also be reversed. Karyotypic changes are also observed in culture adapted cells. These changes do not appear to be random, with the most common gains comprising chromosomes 12p, 17q and X, also seen in embryonal carcinomas. The most recent ISCI studies, which looked at 125 ES lines and 11 iPSC lines, demonstrated in addition to the gains on chromosomes 12p,17q and X, changes in chromosomes 1 and 20. Changes are observed during extended passaging of cells by both bulk passaging (approximately 30%), using enzymes, as well as manual passaging (approximately 14%). The ISCI study also looked at the SNP profiles of the cell lines and demonstrated that in 20% of cells a minimal amplicon on chromosome 20 was observed which contained 3 genes including BCL2L1, a potential candidate gene for driving culture adaptation. What causes this culture adaptation? There is no difference in attachment or cell cycle time between normal and ‘adapted’ cells. However, more normal cells apoptose in culture compared to their abnormal counterparts. There is a trend to add rho kinase (ROCK) inhibitors to enzymatically treated cells in culture to prevent apoptosis, which could potentialy make the cells more susceptible to genetic changes, Peter concluded that it is inevitable that culture adaptation will arise. This adaptation may encompass both genetic and epigenetic changes, which could in turn prove problematic for the future application of these pluripotent cells. Nevertheless, the adaptations provide an opportunity to study and control stem cell fate and to model cancer progression. Caveat emptor: a normal karyotype does not preclude karyotypic changes in a cell!
Yoshinori Yoshida (Centre for iPSC Cell Research and Application, Kyoto, Japan) talked about the ‘Analysis of differentiation capacity from ES and iPS cells’. He discussed the potential use of pluripotent cells for cellular therapies and asked whether there was indeed variant differentiation potential between different types of pluripotent cells. Cells derived from different sources using various derivation methods may display distinct patterns in their potential to produce lineages representative of the three germ layers.
Yoshinori went on to discuss the differences between pluripotent cell lines used in cardiomyocyte differentiation. His own research suggest significant variation between the differentiability of lines. He tried to improve and standardise the differentiation conditions using a system described by Kattman, S. et al.(Cell Stem Cell (2011), vol 8, pp228-240) using the Activin A and BMP4 pathways to drive differentiation to cardiomyocytes. He was able to differentiate the majority of lines but he described one karyotypically normal clone of iPSC that would not differentiate into cardiomyocytes and would not produce teratomas.
For the purposes of banking cells and choosing clones, Yoshinori also looked at the propensity of different lines to differentiate down the neural lineage to determine ‘good’ versus ‘bad’ clones using criteria such as KLF4 and miR371-3 expression as described by Kim et al (Cell Stem Cell (2011), vol 8, pp695-706). He found these markers were not reliable indicators of neural differentiation.
Are there differences between the expression profiles of hESC and iPSC ? Building on the work of Lister, R., et al. (Nature (2011), 471: 68-73) and Ohi, Y et al.,(Nature Cell Biology (2011) vol13: pp:541–549) . an extensive study of a range of iPSC clones/lines derived from different tissue sources and a number of hESC lines revealed some differences in the methylation of theTCERG1L, C9orf64 and TRIM4 promoters. However, it appears that there is no major difference in the gene expression patterns between iPSC and hESC cell lines.
Day 2, Session 1 started with a session entitled ‘iPSC Derivation Methods’ and featured three talks on reprogramming cells using different technologies. The first talk given by Pauline Lieu (Life Technologies, USA) was entitled ‘Efficient Method to Generate Integration –free iPSCs from Different Cell Types and Novel Live Alkaline Phosphatase Substrate for PSC (pluripotent stem cell) Identification’ . Pauline discussed the need for efficient technologies to generate integration-free iPSC. She discussed the current methods used to generate these lines and the problem of low reprogramming efficiency. The use of Sendai virus, a virus that does not integrate into the nucleus, leaves no viral remnants, and is sold in a kit form as ‘Cytotune™’ by Life Technologies was described. Sendai virus has a higher reprogramming efficiency than messenger RNA and Lentivirus and can be used efficiently to reprogramme peripheral blood mononuclear cells. Life Technologies claims it takes 3-4 weeks to generate a line after a single transduction using this technology. Reprogramming using Sendai virus can be achieved in a feeder-free environment. Anti-Sendai virus antibodies can be used to determine that the virus is no longer present in the system (achieved after approximately 5 passages). The emergence of reprogrammed colonies can be tracked using a live Alkaline Phosphatase stain (Life Technologies) which does not damage the cells.
Although this method increases the efficiency of reprogramming , there is a difference observed between the reprogramming efficiences of normal versus diseased fibroblasts
Kerry Mahon (Stemgent, USA) spoke on ‘RNA reprogramming for iPSC derivation’. Kerry described the Stemgent technology using mRNA, first described by Warren et al , (Cell Stem Cell (2010); 7:618-630). Stemgent have produced a kit containing synthetic mRNA for Oct4, Sox2, KLF4, c-myc and Lin28 that are used in the reprogramming of cells by daily application of mRNA (18 days) combined with a reagent that prevents the production of interferon which would destroy the single stranded mRNA. This method of reprogramming is useful to both academia and to industry as it produces cells that have not seen a virus (integrating or non-integrating) and therefore does not require screening for integration sites or loss of virus from the cells. The reprogramming is safe, efficient, and faster than any other method to produce functional iPSC (less than one month) .The iPSC should show a closer fidelity to hESC and could potentially be used in the establishment of a QC standard. The reprogrammed cells can be produced for clinical use as the reagents are xeno-free. Reprogrammed colonies are identified using a live antibody to TRA-1-60 or TRA-1-80. The karyotype of the cells is normal and cells have been shown in vitro and in vivo to produce cells representative of all three germ layers.
As with the Cytotune™kit, the reprogramming efficiency is lower with diseased rather than normal fibroblasts. However, cells can be reprogrammed on Matrigel™ under feeder free conditions. In addition the frequency of application of mRNA for reprogramming can potentially be reduced. A new kit is being developed to incoporate these improvements and make iPSC derivation less laborious and more efficient.
Kwang-Soo Kim (McLean Hospital, Harvard Medical School, USA) spoke on ‘Protein-based reprogramming for iPSC derivation’. He began by discussing the problems with cell therapies for Parkinson’s disease where technically challenging and ethically contentious foetal cells are used. The use of iPSC in this setting could circumvent these problems. The production of dopamergic neurons for therapy could be monitored by the expression of Pitx3 gene which is considered to be the ‘gold standard’ marker of this cell type (Chung S. et al (PNAS (2011) vol 108 pp 9703-9708). However, the majority of human iPS cells are generated by retroviral or lentiviral vectors, resulting in multiple chromosomal integration and residual reprogramming genes. The remaining transgenes and chromosomal disruptions could be harmful and may cause genetic, molecular, and cellular dysfunctions such as tumor formation.
Kwang-Soo called for the establishment of non-genome integrating and/or gene-less reprogramming methods. He favored the generation of safer iPSCs by direct protein delivery without any virus or DNA vectors and recommended Streptolysin O-induced membrane permeability
or protein delivery using CPP fusion. The work of Chang was discussed (Chang, M-Y et al. PLOS (2010)5(3): e9838) who generated iPSC from rat neural cells and fibroblasts usng Streptolysin O.
He went on to describe work using cell penetrating peptide (CPP) or protein transduction domain (PTD): short peptides capable of overcoming the cell membrane barrier. This approach was originally described in 1988 by Frankel and Pabo (Frankel, A et al.Cell(1988); 55:1189-1193; and Science (1988);240:70-73) where the HIV-TAT protein with a short basic segment at amino acid 48-60 was able to penetrate the cell membrane and activate and activate HIV specific genes. Other naturally occurring peptides with a high percentage of basic amino acids are also able to cross the cell membrane. These are known as CPP (Ziegler, A. et al (2005) Biochemistry; 44:138-148 and El-Sayed, A et al. (2009) AAPS J.; 11: 13–22 ). This type of direct protein transduction technology would enable the delivery proteins without the risks associated with viral systems. Initial experiments using human newborn fibroblasts have revealed that that CPP can successfully reprogramme these cells to iPSC that exhibit similar morphology, cell growth characteristics, and cell surface and genomic markers to those seen in hESC (Kim,D. et al (2009) Cell Stem Cell; 4:472-476). These cells also produce all three germ layers in vitro and in vivo.
Just how similar or indeed how different are hESC and iPSC? There are a number of publications that indicate hESC are more efficient at producing certain cell types than iPSC (Hu, B-Y et al (2010) PNAS; 197:4335-40 and Feng, Q et al (2010) Stem Cells;28 :704-712). Viral reprogramming may not be the best way to reprogramme cells. Rhee (Rhee, Y-H et al (2011 ) J. Clin.Invest. ;121 : 2326-2335) has demonstrated that human iPSC generated by different methods can efficiently differentiate to neural progenitor cells and dopamergic neurons that can be used to successfully treat rats. Residual expression of reprogramming factors was observed in differentiated iPSCs produced using lentiviral delivery systems. Human neural precursor cells derived from lentiviral/retroviral-iPSCs were not highly expandable and exhibited early senescence wheras those from cells reprogrammed using CPP could be expanded without senescence. What causes these problems in the iPSC? Genomic instability, copy number variation, aberrant epigenomic reprogramming? Reprogramming using protein based methods has been shown to produce cells that better resemble hESC than those generated by viral methods. Unpublised data by Park (2011) using hepatocytes reveals that using protein based delivery of reprogramming factors produces cells with significantly better genome integrity compared to virus-based iPSCs. Therefore, the use of protein based delivery systems provides an ideal source of human iPSCs for drug discovery, disease modeling, and cells that may be suitable for clinical translation. However, low efficiency of iPSC generation potentially caused by the production of heterogeneous reprogramming proteins and the low efficiency of protein delivery still needs to be addressed for this to be a viable delivery method.
Session 2 entitled ‘Differentiating iPSCs and hESCs : Characterisation and Quality Control’ comprised three technical talks, the first given by Paul Sartipy (Cellectis, Sweden) on ‘Cardiomyocytes derived from human pluripotent stem cells’ . Peter outlined the applications for human cardiomyocytes . These included studies of cardiomyocyte function, modesl for normal and abnormal cardiac development, safety pharmacology, target identification and validation, functional genomics, in vitro disease modeling, cell replacement therapy and customized screening assays. He went on to discuss the models available for efficacy and safety testing during pre-clinical drug discovery . These models include favorable stem cell lines, primary cardiomyocytes, engineered cardiac tissue, explanted hearts and cardiac tissue, and large and small animal models. Cardiomyocytes can be derived from hPSC in a number of different ways (Burridge et al. (2012) Cell Stem Cell ;10 :16-28). However, interline variability in cardiac differentiation can arise due to genetic background, derivation/reprogramming method, culture conditions, passage number, source cell type for iPSC, levels of expression of endogenous growth factors and/or receptors, epigenetic status and kinetics of differentiation, with the optimization of exogenous and endogenous signaling ultimately determining the efficiency of cardiac differentiation. There are a plethora of different factors involved in the differentiation of hPSC into cardiomyocytes (Burridge et al. (2012) Cell Stem cell; 10 :16-28). The molecular and functional characteristics of hPSC derived cardiomyocytes are well defined and include the expression of cardiac markers and ion-channels, the display of ventricular-, atrial-, and nodal-like action potentials, functional blocking of ion-channels and respone to pharmacological stimuli. However, the developmental status of hPSC- cardiomyocytes are in some aspects immature and display a fetal phenotype:
The fetal phenotype of hPC-cardiomyocytes may reduce their general applicability for in vitro drug testing, although this might benefit cell transplantation. Potential limitations aside, they may be useful for evaluating cardiotoxicity (Sartipy, P and Björquist, P (2011) Stem Cells;29:744-748 and Mandenius C., et al (2011) J. Appl. Toxicol.; 31:191-205) including drug induced arrhythmia and myocardial injury. Importantly, action potential recordings in hES-cardiomyocytes demonstrate the presence of cardiac phenotypes (Jonnson, M et al (2010) Stem Cell Res;4:189-200), and can be selected for in vitro testing on the basis of electrophysiological readouts (beat frequency <50bmp, APD90>300ms) and pharmacological response (response to cisapride). In addition, consistent with ion channels in the heart enabling the heart to beat in rhythm, blocking of the inward potassium channel in hESC-cardiomyocytes causes a response analogous to ‘early after depolarizations’ and subsequent ‘torsades de pointes’,tachycardia and other arrhythmias.
Nalos L (2012) B J Pharm; 165:467-478 made a comparison of the inhibition of the rapid activating delayed rectifier potassium channel (IKr ), using blocking drugs in five models of pro-arrhythmia. This study demonstrated that the selective response obtained from the hES-cardiomyocyte model was comparable to in vivo animal models.The hESC-cardiomyocytes also lend themselves to analysis by micro electrode array (MEA) which allows changes in field potential to be measured. MEAs can be used as a surrogate way to measure arrhythmias and Q-T prolongation. In this system signal retrieval is non invasive and interpretation is similar to an ECG. Recently clusters of hESC-cardiomyocytes have been used to examine field potential duration to hERG channel inhibitors and the results showed a difference between responsive and non-responsive clusters (Yamazaki, K et al. (2012) Toxicol In Vitro;26:335-42). This establishes an assessment system with potential to influence the QT interval, using pharmacologically selected clusters.
Although pure populations of cardiomyocytes can be derived from hPSC with relatively high yield, the cardiomyocyte differentiation efficency differs betweeen hESC and hiPSC lines. To reitetate, hPSC derived cardiomyocytes display many critical functional properties of human cardiomyocytes, but in some aspects, hPSC cardiomyocytes display a foetal phenotype. hPSC derived cardiomyocytes are useful tools for drug testing. They are especially useful for predicting the effects of IKr blockers . However, standardization of assays/cells is required and more research is needed to achieve the “adult” cardiomyocyte phenotype and to generate preparations of pure ventricular-, atrial-, and nodal-like CMs
David Hay (University of Edinburgh, UK) spoke on ‘The generation of metabolically active and predictive hepatocytes from pluripotent stem cells’ . In his talk David described how hepatic endoderm can be produced from both hESC and iPSC by directed differentiation via definitive endoderm ( Terrace J. et al (2007) Stem Cells and Development;16:771-778. Hay D. et al. (2008) PNAS ;105:12301-6; Hay D. et al (2008) Stem Cells.;26:894-902, Terrace J et al (2009) Experimental Cell Res.; 315:2141-2153 Terrace, J et al (2010)Experimental Cell Res.; 316: 1637-1647; Sullivan G (2010) Hepatology;51:329-35 ). These cells express genes characteristic of bona fide hepatic endoderm, able to be used for novel liver models, metabolic disease models, the assessment of novel medicines and the discovery of biomarkers. The cells are potentially amenable to high-throughput automated screening since the process of differentiation is highly reproducible. The differentiation process can be divided into a number of well defined stages leading to differentiation to hepatocyte-like cells in 14 days. However maintaining stable differentiated cells is proving to be problematic using currently available substrates including matrigel™. David described the screening of polymer libraries to identify cell culture substrates capable of increasing the phenotypic stability of both primary and hPSC derived ‘hepatocytes’ (Hay et al (2011) Stem Cell Res; 6:92-102). Libraries of polymer microarrays were produced and screened for the attachment of stem cell derived hepatic endoderm and long-term fuction using high content screening. From the screening, six polymers produced hits and these were subjected to detailed investigation. One polymer (code number 134) displayed superior hepatic maturation and function over the other candidates and was selected for further characterisation alongside matrigel, a commonly used biological substrate. Unlike on matrigel where differentiated cells displayed phenotypic instability and a reduction of phenobarbital-induced CYP3A and albumin levels with time, polymer 134 displayed a stable morphology, higher levels of inducible CYP3A and better albumin levels. Other parameters such as integrin expression were also studied revealing differences between adhesion substrates.
These cells were also shown to be a valuable for predicting drug and metabolite toxicity.
David concluded by describing the work of the company Fibromed, a company working towards the provision of a genotyped and phenotyped biobank of cells that can be used to produce hPCs as starting materials for the generation of hepatocytes that can be utilised as discovery technologies to build novel products.
Simone Haupt (Life and Brain, Germany) discussed ‘Human pluripotent stem cells-a source for neurological disease modelling’. iPSCs can be used to generate a wide range of disease models such as schizophrenia, spinal muscular atrophy, Retts syndrome, Parkinsons etc. However, protocols to differentiate cells into neural cells are not standardised and do not provide stable intermediate cell populations. This has a knock-on effect resulting in limited scalability and banking, poorly defined culture conditions, impure populations, protracted differentiation , low yield, lengthy and complex differentiation procedures and batch-to-batch variations. Indeed the situation is not helped when the hPSC source material shows differences in the potential to differentiate (Osafune, K. et al. (2008) Nat. Biotech.; 26:313-315, Hu B-Y et al. (2010) PNAS; 107 :4335- 4340). How can we work towards standardisation? Bock et al (2011, Cell;144:439-52) recently published a reference map of iPSC and hESC variation. By assessing the epigenetic and transcriptional similarity of ES and iPS cells the authors predicted the differentiation efficiency of individual cell lines.
A combination of assays (gene expression and DNA methylation) yields a scorecard for quick and comprehensive characterization of pluripotent cell lines. Recently a well characterized set of iPSC cells have been described (Boulting, G. et al. (2011) Nat Biotechnol.;29:279-86). The set comprises 16 lines that have been assessed for their pluripotentiality and ability to terminally differentiate. Therefore, the lines provide a robust resource for the study of the basic biology of PSCs.
A couple of stable hPC-neuronal intermediates have been described and can be used for neuronal modeling: A rosette-type of cell (Koch, P. et al. (2009)PNAS;106: 3225-3230) and a primitive neuronal precursor (Li, W. et al (2011) PNAS; 108:8299-8304). These intermediates enable the cells to be captured in a self-renewing state for expansion. The cells could then be directed to efficiently generate defined neuronal and glial cells. The rosette type of neural cells are a homogeneous neural stem cell population that can self-renew (> 100 Passages) and this renewal is independent of genetic background (hESC). They are also stable neural epithelial-like stem cells (Lt-NES) that display a tripotent differentiation potential (neuronal, glial and oligodendrocytes), form functional synapses and are amenable to regional patterning.
Lt-NES cells derived from different sources exhibit similar characteristics and can be used as a tool to model neurological disease (Falk, A. et al. (2012) PLOS ONE;7:e29597). The cells reliably differentiate into functional neurons and glia. Lt-NES cells also respond to morphogenes by altering regional marker expression and neurotransmitter phenotype. Genetic profiling reveals that both hESC and iPSC derived Lt-NES express genes seen in neural progenitors and the neural stem cell niche. In addition, pluripotent cell lines cluster together and are distinct from the Lt-NES cell lines. Lt-NES inter-profile clustering is also independent of the cell of origin (hESC or hiPSC). These cells also exhibit a distinct ventral anterior hindbrain identity. Morphologically Lt-NES appear to resemble human foetal neural stem cells. However, Lt-NES and foetal NS are distinct neural stem cell types representing different developmental stages .
Lt-NES-derived neurons can be used for studying human disease such as Alzheimer’s Disease (AD) (Koch ,P. (2012) Am. J. Pathol.;180:2404-2416) as they can be used to address patho-physiological changes in familial monogenic forms of AD and might provide a valubale platform for the development of pharmaceutical compounds. They have also been used to study Spinocerebellar ataxia type 3 (MDJ) (Koch, P. et al. (2011) Nature; 480;543-546).
Simone concluded her presentation by submitting that hPSC are a good source for neurological disease modeling. LT-NES provide a standardized source of human neurons that are independent of genetic background (hESC line or hiPSC line). They are capable of self-renewing and therefore scalable. LT-NES exhibit a stable differentiation potential along with regional identity. They are also amenable to genetic modifications. Importantly, these cells are cryopreservable (able to survive repeated freeze-thawing) and when thawed, can be successfully differentiated into neurons. Thus, these cells lend themselves to automated plating in a multiwell format and could provide a robust source of cells for pharmacological screening.
Lt-NES are human cells for studying human diseases. They can be used as a as platform to express candidate genes/mutants in a physiological cellular context . When used in reprogramming these cells can facilitate the study of pathogenic processes since hiPSC-derived Lt-NES reflect genotype and expression levels found in disease-relevant somatic cells. However, the ‘one size fits all’ approach is inappropriate for neural disease modelling and repair . Tight temporal control over patterning is seen in hPSC cell-derived neural precursor cells. Moreover, at early stages cells are fully competent to respond to extrinsic developmental cues (Zeng, H. et al 2010 PLOS ONE;7:e11853). Run–through protocols can be manipulated accordingly to derive cells that express gene profiles which reflect distinct rostrocaudal and dorsoventral neural identities. Therefore, we need to design bespoke derivation methods tailored to the specific pathology and application. These protocols should be adaptable to industrialized methods (GMP, robustness, scalability, costs) and provide well characterized cell types (publications/applications).
The third session was titled ‘Establishling iPSC lines from donors and preliminary characterisation’, with the first talk given by Ludovic Vallier (University of Cambridge, UK) on ‘Human Induced Pluripotent Stem Cells: Old challenges and new opportunities. Ludovic discussed the human iPSC facility in Cambridge, UK, which was established in 2009, promotes and facilitates the use of PSCs by clinicians and academics for disease modelling and cell therapies. The focus is on cardiac, neuronal, blood and metabolic disorders. Over the last 2 years more than 400 lines have been generated from 70 patients with a 90% success rate. Cells for reprogramming are sourced from skin biopsies. The dermal fibroblasts generated are reprogrammed using the 4 Yamanaka factors delivered using lentiviruses or retroviruses. From biopsy to the characterisation of an hiPSC takes 2 to 3 months. Studies using blood derived endothelial progenitor cells indicate this cell type can be efficiently reprogrammed making it suitable for high throughput generation of iPSCs. The EPSCs can be reprogrammed using Sendai virus which does not integrate into the host genome. Genome-wide analysis of iPSCs and the parental cells from which they were derived has been carried out in collaboration with the Sanger centre and this work has demonstrated that the profiles of the iPSCs are almost identical to the parental lines. Cells were differentiated into neuroectoderm by inhibiting activin and BMP signalling and endodermal differentiation was achieved using a combination of activin, BMP-4 and FGF-2.
However, it was demonstrated that the variability in differentiation capabilities of cell lines derived from different individuals is on a par with the variability seen between lines derived from the same patient. Both iPSC and hESC can be differentiated using the same protocols but there is still a need to standardise the methods to ensure reproducibility.
Ludovic concluded his talk by highlighting a number of challenges that need to be addressed in the production of iPSC . These include the optimisation of the reprogramming factors used, the cell type (this influences the quality of the iPSC), derivation conditions, delivery of reprogramming factors, scale-up and characterisation. With respect to characterisation, this will probably move from phenotypic analysis and qPCR to genome wide studies (exome, histone and methylome analysis). Another important consideration in this area is the ethical governace of the human material used to generate the cell lines. Appropriate consent should be in place for the downstream application of these cells. Indeed it might require further regulatory input to assure that all the ethical issues are addressed.
The second talk given by David Kahler (New York Stem Cell Foundation, USA) detailed the ‘Derivation and Characterisation of iPSC lines in the NYSCF Personalised Medicine Bank’
This bank is not actively banking at present but is still setting up. The foundation is privately funded and focusses on high risk/low return projects that would not be funded from other sources .
The bank employ standard protocols where skin biopsies and fibroblasts are the starting materials and reprogramming will likely use Sendai virus. The cells will be from healthy and diseased subjects that are over the age of 18 where informed consent has been obtained and a medical history taken. A personalised medicine approach will necessitate high throughput screening for cell line characterisation.
David has a background in flow cytometry and this could be put to great use in selecting early reprogrammed cells by sorting on day 7 with the markers CD13, SSEA4 and Tra-1-60 and replating the cells on inactivated feeders. This would clean up the cultures and enrich for iPSCs.
Once the hiPScs have been generated they will be characterised using a panel of markers and assessed for their ability to differentiate into cells of the three germ layers.
David finished his talk by emphasising the challenges that the NYSCF personalised medicine bank faces to fulfil its remit. There will be contraints on the time and labour (skilled technicians) for this resource intensive activity.
The afternoon session started with a Moderated discussion session on ‘Critical points for iPSC banking for clinical use’ led by Glyn Stacey. From this discussion it was established that early evaluation of disease needs to be addressed: to establish how many patients to capture the range of genotypes presented in the disease of interest and to determine what is the likely penetration of the genetic trait in an in vitro model.
A second point of discussion related to the choice of tissue. Recent work in Yamanaka’s laboratory indicated that a broad range of tissues generate iPSCs equally well, although no data has been presented on their relative epigenetic states.
Thirdly, the reprogramming method may influence research data. Lengthy deivation times may lead to higher levels of genetic change/mutation, but there is a balance to be struck since clones that appear later and survive may be more stable . Viral transduction methods are effective and produce lines able to replicate for at least 50-100 passages. However, the quality of cells may be influenced by the reprogramming technique, methodological variations and the stability of the donor cells. Low efficieny methods could also be selective for certain genotypes. These variables might induce greater differences between individual cell lines than the genotype or defect. Further studies are necessary.
Intellectual property (IP) is also a serious constraint. While the situation is relatively clear with use of Yamanka factors for reprogramming, there may be constraints on other techniques. Also, the patent landscape is becoming more rather an less complex with uncertainty arising from some patent claims previously rejected being appealed. Plans for banking which do not address IP issues risk stalling regardless of whether activities are for commercial or academic purposes. A model where central funding secures ownership of lines and IP within a project would potentially address this issue. In addition, use of model material transfer agreements (MTAs) was discussed.
There are many issues associated with the derivation and development of cell lines deemed to be of fundamental importance to research, cinical product development and commercial exploitation. Firstly, care should be taken to ensure ethical procurement and application, including donor consent for actual use, ie. the potential to commercialise the cell lines, export cell lines to other countries, use for a wide range of purposes etc. Another point to consider is how many “clones” should be isolated from each donor to ensure a representative clone can be identified (ideally this number should be between 5 and 10). It might be possible to preserve large numbers by simple preservation methods in a multiwell tray format, as has been achieved for primary hepatocytes. Thirdly, cell line characterisation must encompass two key and distinct features: self-renewal capacity and pluripotency. Self-renewal is different to pluripotentiality and there is a need for data and possibly new markers to distinguish between these two functions. Early screening of the lines should be performed to give an indication of pluripotency, ideally using a straightforward test that can be used in research laboratories on a regular basis. The generation of small quality controlled pre-master banks would give adequate material for these early selection procedures.
Characterisation could extend to methylation studies giving broad epigenetic profiles to reveal differences between clones of the same line. However, very detailed testing could make the supply of cells by banks too expensive. However, it should be noted that the cost of genetic testing is decreasing.
Banks should ask for information on the iPSC isolation methods, so that details such as lengthy isolation procedures (which are more likely to induce genetic changes) can be reported
Banks should evaluate generic features of iPSCs such as those established by ESTOOLS (see appendix 1). Fundamental characteristics need to be met for a cell line to be accepted as an iPSC (the group proposed that the cells need to satisfy 10 as a minimum).
The need to define an iPSC clone was also considered as well as use of standardised nomenclature and reporting criteria (eg Luong et al (2011) Cell Stem Cell;8:357-9. ,) for PSC lines. It was noted that the point of sampling at many cell banks was post-thaw from a sample of the bank (not prior to freezing) and a European Medicines Agency representative emphasised the need to demonstrate equivalence of the pre- and post- cryopreservation cultures of the cell lines. Testing for mycoplama and sterility was also recommended since mycoplasma infection can have a devastating effect on the cell lines.
With respect to differentiation iPSC and hESC generally perform similary. Nonetheless, there is a need for stable reference cell lines such as those described by Bock (Bock, C (2011) Cell; 144:439-452) There is also a requirement for good positive and negative control materials for genomic and phenotypic studies as well as standard reference materials for benchmarking studies. Currently differentiation protocols are long, complex and expensive, being unsuitable for industry. There is a requirement to refine these protocols and if possible make them more robust and reproducible with better yields of pure cell populations.
Although hPSC demonstrate useful functional attributes, both iPSC and hESC are still producing only foetal phenotypes. This needs to be addressed to see if we can generate cells with the relevant functional and morphological characteristics of adult cell types.
In summary, the meeting was informative and addressed many contemporary and relevant questions. By bringing together eminent scientists, with industry, funding bodies and regulators it was able to address many contemporary and important questions and highlighted the challenges ahead for this exciting dynamic field where best practice is essential.
Minimal Criteria for the classification of putative iPS cells for further study
(as suggested by ESTOOLS)
- Stable ES cell like morphology and growth pattern
- Expanded in culture as established line for > 10 passages
- Viable frozen stocks
- Human ES cell surface antigen profile: Expression of SSEA3, SSEA4, TRA-1-60/TRA-1-81, L-ALP (TRA-2-54 or TRA-2-49) – quantitated by flow cytometry
- Express key endogenous pluripotency-associated genes: Oct4, Nanog, Sox2, Rex, TDGF, assessed by:
- immunostaining/western blot
- Neural differentiation in vitro – immunostaining for TuJ1 and GFAP
- Primary evidence of pluripotency in embryoid body or other in vitro differentiation assays by qRT-PCR for lineage markers
- Transgenes down-regulated
- Diploid karyotype
Note: These minimal criteria should be met before putative iPS cells are entered into further study, unless a strong case can be made that one or other criterion should not exclude the cells from specific experiments. For example, it should be noted that, although no SSEA3(-) or SSEA4(-) human ES cells have yet been identified, rare polymorphisms in the human population indicate that SSEA3(-) or SSEA4(-) human ES or iPS cells might be encountered.
Advanced characteristics that should be assessed for putative iPS cells
- Array CGH or SNP analysis of genetic integrity
- DNA fingerprint confirming identity with somatic cell of origin
- Teratoma formation
- Detailed evidence of differentiation in vitro to three germ layers with functional markers
- Copy number of transgenes with evidence of silencing; or evidence of transgene deletion or non-incorporation
- Gene expression profile – quantitative assessment by TLDA
- Methylation status of Oct4 and Nanog promoters
- X chromosome activation/inactivation status for female cells
- Comprehensive transcriptomic analysis by microarray or high throughput cDNA sequencing (for selected lines)
Note: ESTOOLS can provide central resourcing for some aspects of advance phenotyping when all minimal criteria have been documented provided the cell line is made freely available to other ESTOOLS partners for research use.