Cost-Effective Master Cell Bank Validation of Multiple Clinical-Grade Human Pluripotent Stem Cell Lines From a Single Donor

This study addressed protocols for testing cell lines for human viral pathogens. It is highly unlikely that multiple cell lines derived in parallel from tissue samples taken from one donor would have different endogenous viral pathogen profiles. Master Cell Banks of sibling lines were pooled for tiered validation; all results were negative. This cost-effective strategy could be applied for validation of Master Cell Banks of multiple clinical-grade iPS cell lines from a single donor.

Abstract

Standardization guidelines for human pluripotent stem cells are still very broadly defined, despite ongoing clinical trials in the U.S., U.K., and Japan. The requirements for validation of human embryonic (hESCs) and induced pluripotent stem cells (iPSCs) in general follow the regulations for other clinically compliant biologics already in place but without addressing key differences between cell types or final products. In order to realize the full potential of stem cell therapy, validation criteria, methodology, and, most importantly, strategy, should address the shortfalls and efficiency of current approaches; without this, hESC- and, especially, iPSC-based therapy will not be able to compete with other technologies in a cost-efficient way. We addressed the protocols for testing cell lines for human viral pathogens and propose a novel strategy that would significantly reduce costs. It is highly unlikely that the multiple cell lines derived in parallel from a tissue sample taken from one donor would have different profiles of endogenous viral pathogens; we therefore argue that samples from the Master Cell Banks of sibling lines could be safely pooled for validation. We illustrate this approach with tiered validation of two sibling clinical-grade hESC lines, KCL033 and KCL034 (stage 1, sterility; stage 2, specific human pathogens; and stage 3, nonspecific human pathogens). The results of all tests were negative. This cost-effective strategy could also be applied for validation of Master Cell Banks of multiple clinical-grade iPSC lines derived from a single donor.

Introduction

Optimism that human embryonic stem cells (hESCs) would provide a virtually unlimited source of selected cell types for future cell therapies and drug screening and development has resulted in considerable progress in stem cell biology since the first hESCs were derived in 1998 [1]. The initial difficulties in obtaining embryos for stem cell research has largely been overcome, especially in the U.K., where a regulatory route map to facilitate clinical research application has been produced relatively quickly [2, 3]. By the end of 2009, more than 1,000 hESC lines had been derived worldwide; however, only a few had been thoroughly characterized [4, 5]. With the discovery of induced pluripotent stem cells (iPSCs) in 2006 [6], interest in hESCs started to wane.

An Investigational New Drug application for the first clinical trial of hESC-based therapy received initial Food and Drug Administration clearance in January 2009. The company, Geron Corp. (Menlo Park, CA, http://www.geron.com), generated a Master Bank from hESC line H1, which was derived as a research grade line in the presence of animal products [1]. The inner cell mass was isolated by immunosurgery with a rabbit antiserum and plated on mitotically inactivated mouse embryonic fibroblasts in a medium supplemented with 20% fetal bovine serum. The Master Bank used for production of hESC-derived oligodendrocyte progenitors GRNOPC1 had to be “re-derived” under current good manufacturing practice (cGMP) conditions and subjected to validation, not only for the presence of human viral pathogens but also for the presence of animal pathogens, significantly increasing the costs.

Similarly, the hESC line MA-09 used in clinical trials for Stargardt’s macular dystrophy and age-related macular degeneration was also derived as a research grade line in the presence of animal components [7, 8] and had to undergo a similar process and costly validation for the presence of animal pathogens to be used for cellular therapy.

The first clinical-grade hESC lines were derived at a cGMP facility in Brisbane, Australia, from frozen embryos donated at Sydney IVF Ltd., under sponsorship of the Singapore-based company ES Cell International Pte. Ltd. (ESI Bio, BioTime, Inc., Alameda, CA, http://www.biotimeinc.com) [9]. To promote the lines, the company was willing to make them easily accessible in a research grade version through the A*STAR Singapore Stem Cell Consortium (SSCC). The intention was that the Stem Cell Bank within the SSCC would expand and distribute the cells under non-cGMP conditions for minimal reimbursement. Despite international efforts and multimillion investments, they did not gain the popularity of the H1 and H9 hESC lines derived by Thomson et al. [1] and the cells were never used in clinical trials. In May 2010, the California-based BioTime acquired ESI, including a bank of six clinical-grade hESC lines. Five of them the company characterized further [10]. BioTime has also recently acquired Geron’s hESC portfolio; there might, therefore, be long-term plans for using these cell lines in the clinic.

In the U.K., the Medical Research Council systematically invested in derivation of clinical-grade hESC lines in five centers across the country. Currently, more than 30 such lines have been derived. At King’s College London, in our purpose-built cGMP laboratory, we derived, under animal product-free conditions [11–14], eight clinical-grade lines: KCL031–KCL034 and KCL037–KCL040. All eight clinical-grade lines will be available through the U.K. Stem Cell Bank, which will conduct further validation of these lines before releasing them. The lines are also under consideration for placement on the NIH hESC registry [15], which will make them eligible for use in NIH-supported research.

In the next step, following cGMP, we adapted four clinical-grade hESC lines, KCL033, KCL034, KCL037, and KCL040, to feeder-free conditions using methods described previously [12]. However, owing to the high costs of running a cGMP facility and Master Cell Bank validation, we were not able to proceed with all the lines. Of the four lines, KCL040 and KCL037 had haplotypes that matched ∼0.016% and ∼0.019% of the population, respectively, and were therefore potentially more clinically useful than the other two, which only matched ∼0.004% [13]. However, the lines KCL033 and KCL034 were sibling lines, meaning that they were derived from embryos donated by the same couple. In addition, these embryos had been generated and cryopreserved as supernumerary in a single in vitro fertilization (IVF) cycle. Furthermore, the embryos had been thawed at the same time, and the lines were generated in parallel under the same conditions using the same reagents. The embryos were thus coming from a single source and were undergoing all procedures in parallel until initial stock cryopreservation. We, therefore, reasoned that if one line had been positive for any human pathogens, it was highly unlikely that the other line would have a different status. Testing of only one cell line should therefore theoretically be sufficient to exclude or confirm the presence of pathogens in all sibling lines; however, regulations require that all lines should be tested. We therefore considered the strategy of pooling samples from such Master Cell Banks, thereby reducing costs without compromising the detection of human viral pathogens.

Hence, KCL033 and KCL034 were taken further, and Master Cell Banks were generated and validated in vitro for the presence of human viral pathogens. All validation studies were conducted by SGS Vitrology (Glasgow, U.K., http://www.sgs.com), in compliance with the principles of GMP as set out in Directive 2003/94/EC for medicinal products for human use [16] and 91/412/EEC for veterinary medicinal products [17]. In the present study, we describe the strategy, methodology, and outcomes of these in vitro validation studies.

Materials and Methods

Human Samples

The present work was ethically approved (U.K. National Health Service Research Ethics Committee Reference: 06/Q0702/90). The experiments were performed under licenses from the U.K. Human Fertilization and Embryology Authority (license no. R0133) and the U.K. Human Tissue Authority (license no. 22621).

Cell Culture

The methods have been previously described in detail [11–14].

Sterility Testing

Five vials each of KCL033 and KCL034 from the respective Master Cell Banks were pooled and tested for sterility. The presence of microorganisms was assessed after 14 days of incubation at the appropriate temperature. Tryptone soya broth (TSB) was used for the growth of aerobic bacteria and fungi at 20°C–25°C. Fluid thioglycolate (THIO) medium was used for the growth of both anaerobic and aerobic bacteria at 30°C–35°C.

The experiment controls were as follows. Uninoculated TSB and THIO in duplicate served as the negative control (NC) and when inoculated with 1 ml of sterile demineralized water served as the operator/assay technique control. TSB inoculated with 10–100 colony forming units of Bacillus subtilis, Aspergillus brasiliensis, and Candida albicans and THIO inoculated with Clostriduim sporogenes, Pseudomonas aeruginosa, and Staphylococcus aureus served as the positive controls (PCs). Environmental monitoring was performed in the sample isolator (settle plates) and before and after manipulation (contact plates).

Sterility testing was performed in accordance with the current requirements of the European Pharmacopoeia, Section 2.6.1 Sterility, U.S. Pharmacopeia <71> Sterility Tests, and International Conference on Harmonisation Topic Q5D guidelines.

Mycoplasma Testing

One vial each of KCL033 and KCL034 from the respective Master Cell Banks were pooled and increased to 5 ml volume by adding sterile Dulbecco’s modified Eagle’s medium. The pooled sample was sonicated on ice for 2 minutes, increased to a final volume of 15 ml, and tested for Mycoplasma without additional treatment.

Vero indicator cell cultures were inoculated with the test material and cultured for 3 days. Then, the cells were mechanically transferred into the chamber slide and cultured for an additional 3 days. At day 6, the cells were stained with DNA-binding Hoechst 33258 fluorescent dye for Mycoplasma detection.

Test material was also inoculated into broth known to have nutritive properties for Mycoplasma species and either cultured further in broth or plated on agar plates, which were then observed microscopically for typical Mycoplasma colony morphology.

Mycoplasma testing was performed in accordance with the current requirements of the European Pharmacopoeia, Section 2.6.7, Mycoplasmas.

Fluorescent-Product Enhanced Reverse Transcriptase Assay for Retroviral Activity

Spent medium from the KCL033 and KCL034 cultures, 8.7 ml each, was pooled and centrifuged at 11,000g for 10 minutes at 4°C. Then, the supernatant was filtered through a 0.45-µm filter. The sample was spiked with 3.0 × 10⁴ retroviral particles (amphotropic murine leukemia virus [A-MLV]) per milliliter, and retroviral particles were purified by ultracentrifugation at 100,000g for a minimum of 60 minutes at 4°C. To release reverse transcriptase (RT) activity from the retroviral particles, the pellets were resuspended in disruption buffer. Recovery of retroviral particles in the absence of the test sample was assessed by ultracentrifugation of the same number of A-MLV in culture medium. Aliquots (5 µl) of processed test sample and retrovirus particle recovery control were added to the RT reaction mix containing all reagents necessary for reverse transcription (dNTP, MgCl₂, reverse primer, RNase inhibitor) and Brome mosaic virus RNA as a template. The mix also contained activated calf thymus (aCT) DNA to reduce false-positive signals arising from DNA polymerases. To assess the level of DNA polymerase activity, the sample was also analyzed in the absence of aCT DNA. Synthesized cDNA was subsequently amplified using TaqMan (Life Technologies, Carlsbad, CA, http://www.lifetechnologies.com) real-time polymerase chain reaction (PCR). The data were collected and analyzed using an Applied BioSystems 7900HT Fast Real Time PCR System (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com). The negative controls included PCR no-template control (PNTC), NC, and medium only control. The positive controls included PC and spike positive control (SPC). SPC consisted of retroviral particles at or near the limit of detection spiked into the test sample before processing. The cycling parameters were as follows: for reverse transcription: 30 minutes at 37°C (reverse transcription) and 5 minutes at 95°C (reverse transcription inactivation), and 5 minutes at 25°C (reaction cooling); for c-DNA PCR: 2 minutes at 50°C (uracil-DNA glycosylase [UNG] activation) and 10 minutes at 95°C (AmpliTaq Gold activation; Life Technologies), followed by 40 cycles of 15 seconds at 95°C (denaturation) and 1 minute at 60°C (annealing/extension).

The methods used were compliant with the current edition of the European Pharmacopoeia, 2.6.21, Nucleic Acid Amplification Techniques.

Real-Time PCR

For detection of the enteroviruses, hepatitis A virus (HAV), and hepatitis C virus (HCV), RNA was extracted using RNA easy kit (Qiagen, Venlo, The Netherlands, http://www.qiagen.com) from 3.0 × 10⁶ cells in 200 µl of phosphate-buffered saline (PBS) per column. Eluted RNA from multiple columns was pooled together. During the RNA extraction procedure, the sample was spiked with cucumber mosaic virus (CrMV) interference control RNA from an unrelated species. From the pooled RNA, 3 µl (67 ng/µl) was used per replicate in testing for pathogens in real-time PCR assays. This is equivalent to approximately 1.2 × 10⁵ cells per reaction assuming 100% recovery from the extraction process.

For detection of sexually transmitted disease (STD) pathogens and all other viruses, DNA was extracted using the QIAamp DNA mini-kit (Qiagen) from either 3.0 × 10⁶ cells (for detection of viruses) or 3.6 × 10⁶ cells (for bacterial pathogen detection) in 200 µl of PBS per column. Eluted DNA from multiple columns was pooled together. During the DNA extraction procedure, the sample was spiked with CrMV interference control DNA from an unrelated species. From the pooled DNA, 3 µl (280 ng/µl for viruses, 238.3 ng/µl for STD pathogens) was used per replicate in testing for pathogens in real-time PCR assays. This is equivalent to approximately 1.2 × 10⁵ cells for viruses and 1.0 × 10⁵ cells for STD pathogens.

Extracted nucleic acids were tested in real-time PCRs containing either CrMV interference control specific primers and probe or pathogen target specific primers and probe. The accumulation of the PCR products was monitored and data collected using an Applied Biosystems 7900HT Fast Real Time PCR System. Real-time PCR controls included PNTC, NC, PC, postspike control, exogenous internal positive control, extraction no-template control, and extraction recovery and contamination control. The cycling parameters were as follows.

For the HAV, HCV, and enteroviruses, we used initial steps of 10 minutes at 50°C (reverse transcription) and 10 minutes at 95°C (reverse transcription inactivation), followed by 40 cycles of 15 seconds at 95°C (denaturation) and 1 minute at 60°C (annealing/extension).

For HIV-1, HIV-2, human T-lymphotropic virus (HTLV) type 1 and 2, hepatitis B virus (HBV), human herpesvirus (HHV)-6, HHV-7, HHV-8, human cytomegalovirus (hCMV), Epstein-Barr virus (EBV; also known as HHV-4), B19 virus, Simian virus 40, human polyomavirus JC, human polyomavirus BK, and Treponema pallidum, we used 2 minutes at 50°C (UNG activation) and 10 minutes at 95°C (AmpliTaq Gold activation), followed by 40 cycles of 15 seconds at 95°C (denaturation) and 1 minute at 60°C (annealing/extension). For Chlamydia species and Neisseria gonorrhoeae, the number of cycles was increased to 50.

The methods used were compliant with the current edition of the European Pharmacopoeia, 2.6.21, Nucleic Acid Amplification Techniques.

In Vitro Assay for Detection of Virus Contaminants Using Indicator Cell Lines

The cell lines used in this assay were normal human fetal lung fibroblasts MRC-5, African green monkey Cercopithecus aethiops kidney Vero C1008 (Vero 76, clone E6, Vero E6), and human cervical carcinoma HeLa. The control viruses used in the assay were herpes simplex virus type 1, human adenovirus type 5, and influenza A virus.

KCL033 and KCL034 in frozen culture medium (2 × 10⁶ cells per milliliter) were thawed, pooled together, and centrifuged at low speed. The supernatant (SN) was removed and saved, and the pellet was resuspended in a small volume of SN and subjected to two cycles of freezing and thawing to facilitate release of cell-associated viral particles. Following low-speed centrifugation, the SN was removed and pooled with the original SN. The pooled SN was diluted with appropriate maintenance medium at a ratio of 1:4 and used for inoculation of the indicator cell lines. Subconfluent monolayers of indicator cell lines in T80 flasks were inoculated with the test item or with relevant controls in duplicate (2.5 ml inoculum per flask). MRC-5 cultures were fed on days 7, 14, and 21, VeroC1008 on day 14, and HeLa on days 11, 18, and 25 after inoculation. The HeLa cells were also subcultured on days 7, 14, and 21. The cultures were monitored for 28 days for cell health and evidence of a virus-induced cytopathic effect (CPE). On day 28 after inoculation, the test item cultures and relevant controls were tested for hemadsorption with guinea pig erythrocytes, one half at 2°C–8°C and the other half at 20°C–25°C.

Transmission Electron Microscopy

Cell pellets containing 1 × 10⁷ cells fixed in glutaraldehyde solution were postfixed in OsO₄, stained en bloc in uranyl acetate, dehydrated in an ethanol series and propylene oxide before being infiltrated to Araldite resin (Huntsman Corp., Salt Lake City, UT, http://www.huntsman.com), and polymerized. Semithin sections (1 µm) were mounted on glass slides, stained with toluidine blue, and examined by light microscopy for general quality of fixation, gross cell morphology, and the presence of mitotic, dead, or dying cells. Ultrathin sections (0.1 µm) were mounted on electron microscope grids and stained with uranyl acetate and lead citrate solutions. A minimum of 100 median cell profiles of the tested cell culture were examined for the presence of viruses, virus-like particles, and other extraneous agents.

Results

Stage 1: Sterility and Mycoplasma Testing

Validation of the Master Cell Banks was divided into three stages. In order to proceed to the next stage, the samples had to pass each validation test.

The first stage was to test for any contamination that might have arisen from our handling, despite the continuous monitoring during sample processing in our cGMP-compliant facility. Pooled samples from the KCL033 and KCL034 Master Cell Banks were therefore tested first for sterility, to detect the presence of bacteria, fungi, and Mycoplasma species and to ensure that no inhibitory materials were present that would mask the detection of pathogens.

The growth of bacteria or fungi was not detected in the test article after 14 days of incubation. Spiking the test material with either bacteria or fungi resulted in growth at a similar rate as the positive control, indicating that no inhibitory factors were present in the tested sample (Table 1).

Table 1.

Sterility testing

Mycoplasma validation was performed in the indicator cell line and liquid (broth) and solid (agar) media known to have nutritive properties for Mycoplasma species. Growth was not detected, and no inhibitory factors were present in the tested sample (Table 2A–2C).

Table 2A.

Mycoplasma testing: indicator cell cuture test

Table 2C.

Mycoplasma testing: culture test on agar

Table 2B.

Mycoplasma testing: culture test in broth

Stage 2: Detection of Specific Human Pathogens

Having confirmed in stage 1 that our handling did not introduce exogenous pathogens, in the second stage, we tested for the presence of human viral pathogens that might have been inherited from the donor tissue. Two different types of assay were performed: fluorescent-product enhanced reverse transcriptase (F-PERT) for detection of reverse transcriptase, which would indicate retroviral activity, and a series of real-time PCRs for detection of various common human viruses, including T. pallidum, Chlamydia species, and N. gonorrhoeae.

Regulatory guidance recommends the use of reverse transcriptase activity packaged into extracellular retroviral particles as a marker for the potential presence of retroviral contamination. The assays are designed to detect conversion of RNA template to cDNA when retroviral reverse transcriptase is present in the test sample [18–21]. PERT assays are routinely required for cell banks or any product originating from mammalian and avian cell substrates manufactured in a process that does not contain endogenous retroviral activity. The F-PERT assay for KCL033 and KCL034 pooled sample was negative to a sensitivity of approximately 1,000 retroviral particles per reaction or 2,600 particles per milliliter of pooled sample (Table 3).

Table 3.

F-PERT assay

Screening of embryo donors for STDs, mandatory in the U.S. and most European Union (EU) countries, is not required in the U.K. However, because the U.K., as a member of the EU, follows the EU Tissue and Cells Directive [22–24], we validated our samples for the presence of STD pathogens, in addition to the list of specific human viral pathogens. Real-time PCR, commonly used to detect human viruses in biological materials [20, 21], confirmed that KCL033 and KCL034 were negative for all viruses tested and for T. pallidum, Chlamydia species, and N. gonorrhoeae (Table 4). Negative results were expected for some of the viruses included in the panel (i.e., HIV-1 and -2, HBV, HCV, HTLV-1 and -2, hCMV, EBV), either because the embryo donors were tested before IVF treatment, or because we had previously performed the test with a non-cGMP service provider at low cost to minimize unnecessary spending.

Table 4.

Real-time PCR detection of specific human pathogens

Stage 3: Detection of Nonspecific Viral and Other Adventitious Contaminants

The assays in stage 2 covered only specific types of human pathogens. In the third stage, we therefore performed two additional tests for the detection of nonspecific viral contaminants: the 28-day in vitro assay and transmission electron microscopy (TEM).

The 28-day in vitro assay was designed to meet the general requirements of the tests for extraneous agents in cell culture as described in European Pharmacopoeia, chapters 5.2.3 (“Cell Substrates for Production of Vaccines for Human Use”) and 2.6.16 (“Tests for Extraneous Agents in Viral Vaccines for Human Use”), with extension to 28 days to increase the sensitivity of the assay for slow growing viruses, as suggested by the Food and Drug Administration and later accepted by the World Health Organization. MRC-5, Vero C1008, and HeLa cell cultures were inoculated with the pooled samples from the KCL033 and KCL034 Master Cell Banks and monitored for 28 days. None of the cultures inoculated with the test material showed signs of a virus-induced cytopathic effect or hemadsorption with guinea-pig red blood cells. The KCL033 and KCL034 samples were therefore considered negative for the presence of cytopathic or hemadsorbing adventitious virus contaminants (Table 5A–5C).

Table 5A.

Controls in the assay with indicator cell lines

Table 5C.

HAD assay results

Table 5B.

Virus-induced cytopathic effect assay results

Regulatory agencies recommend the use of TEM on harvested cells to ascertain the presence of viruses and adventitious agents. TEM can reveal the size, structure, and localization of viral, fungal, bacterial, or Mycoplasma contaminants in the sample. Obviously, for this test, KCL033 and KCL034 could not be pooled, because we would not be able to be sure how many cells from each cell line were analyzed in a mixed sample. We tested only KCL034 in the TEM assay and no viruses, virus-like particles, or extraneous agents, including Mycoplasma, yeast, fungi, or bacteria were detected in at least 100 random median cell profiles examined (Fig. 1).

Figure 1.

Typical KCL034 transmission electron microscopy cell profiles. The cells were generally adherent, with intercellular connections and a normal range of organelles and nuclei, with sparse heterochromatin and prominent nucleoli. Abbreviations: des, desmosome; f, fibrils; tj, tight junction.

We have not tested for extraneous agents that might be present without causing cytopathic or other apparent effects in the cell culture. Such tests are done in vivo, commonly involving embryonated eggs and suckling and adult mice; they generally cost >$12,000 per sample. Multiple cell lines derived in parallel from a tissue sample from a single donor could also be pooled for these tests, as described above for our two clinical-grade hESC lines.

Discussion

Once human pluripotent stem cell (hPSC) therapy is approved and available, it will have to compete with other technologies over a range of criteria, including costs. Even if the technology is at least as beneficial as established alternatives, the high production costs could become a major obstacle in successful dissemination of the therapy and return of investments for developers and investors. Nevertheless, it is essential for hPSC-based therapy that quality control should be very rigorous for each step, from isolation of the stem cells through to the final differentiated product.

In order to decrease the costs of quality control while maintaining rigorous monitoring, we considered the possibility of mixing multiple lines derived in parallel from tissue of a single donor, before testing for the presence of various human viruses and other adventitious agents. The risks associated with this approach seem very low, because (a) the tissue has been processed under identical environmental conditions, exposed to identical cell culture media and reagents, cryopreserved and stored in the same vacuum flask, and (b) the possibility of introducing any human viruses from the controlled environment of a cGMP-compliant laboratory is rather remote. Therefore, the only source of contamination would be the donor tissue itself. Although possible, it is unlikely that a part of tissue from which one cell line is derived would be positive for some virus or other pathogen and the other part would not. Such a risk might be somewhat higher in the case of embryos harvested during a single IVF cycle than it would be for multiple iPS cell clones obtained from reprogramming of a single batch of donor cells.

Pooling the lines derived from different donors could be considered as potentially providing additional cost savings. However, an important consideration in the testing of pooled lines using PCR methods is the determination of sensitivity. For nucleic acid testing methods, the European Pharmacopoiea 2.6.21 states a requirement of sensitivity at the 95% confidence level. For example, the assay sensitivity for the detection of T. pallidum is 10 copies per reaction at 95% confidence, the minimum number of copies detected in 95% of the reactions. Thus, we can follow that the reaction containing 100,000 cells (or 714.9 ng of DNA, based on the estimate that 7 pg of DNA corresponds to 1 genome equivalent) will detect 10 copies of the pathogen with a sensitivity of 95%, regardless of the number of lines pooled. Therefore, if testing 10 cell lines with 10,000 cells from each cell line in the reaction, all 10 pooled cell lines would need to have 1 copy of T. pallidum to give a low-level positive result of 10 copies per reaction at the assay detection limit.

It has been estimated that in the U.S., CMV infects 50%–80% of the adult population [25] and nearly 95% of adults aged 35–40 years have been infected with EBV [26]. When considering the cost of deriving hPSC lines, it would therefore seem vitally important to prescreen potential donors for common viral pathogens such as HIV, HBV, HCV, HTLV, CMV, and EBV, along with screening for other health risks such as hereditary diseases or cancer. This approach would diminish the risk of investing in lines that would not be safe for therapeutic use and might reduce the pathogen burden sufficiently to enable larger scale pooling. This strategy would be impractical for hESC lines owing to donor confidentiality but should be easily applicable when recruiting donors for clinical-grade iPSC lines. Alternatively, screening of biopsied samples in non-GMP facilities before cell line derivation would exclude contaminated material and reduce the overall derivation costs, although the tests would have to be repeated once the cell lines were derived to satisfy regulations.

Conclusion

The requirements for validation of hPSC lines are even more rigorous than for other biological material for human therapy, including adult or fetal stem cells. The paramount issue is to prove the safety of these cells. However, the costs associated with validation of the products could render hESC- and iPSC-based therapy prohibitively expensive. Using two clinical-grade hESC lines, KCL033 and KCL034, derived in parallel from sibling embryos harvested in a single IVF cycle, we demonstrated as a proof-of-principle that multiple cell lines can be safely tested together and thus reduce the validation costs significantly without compromising safety and/or efficiency. This pooling strategy would be especially cost-effective in the case of clinical-grade iPSC lines, because three lines are generally derived from each donor. For example, a clinical iPSC bank from 100 donors with 3 lines from each donor, such as the Clinical iPSC Bank of Kyoto, could save on validation $18 million using our approach of pooling the samples from a single donor.

Author Contributions

L.D., A.P., and C.M.: experimental work; S.C., N.B., and A.L.: conception and design, manuscript writing. Y.K.: conception and design, financial support; C.O. and D.I.: conception and design, financial support, manuscript writing.

Disclosure of Potential Conflicts of Interest

N.B. and A.L. have compensated employment with SGS Vitrology.