Abstract※
Background aims※
Few human induced pluripotent stem cell (hiPSC) lines are Good Manufacturing Practice (GMP)-compliant, limiting the clinical use of hiPSC-derived products. Here, we addressed this by establishing and validating an in-house platform to produce GMP-compliant hiPSCs that would be appropriate for producing both allogeneic and autologous hiPSC-derived products.
Methods※
Our standard research protocol for hiPSCs production was adapted and translated into a GMP-compliant platform. In addition to the generation of GMP-compliant hiPSC, the platform entails the methodology for donor recruitment, consent and screening, donor material procurement, hiPSCs manufacture, in-process control, specific QC test validation, QC testing, product release, hiPSCs storage and stability testing. For platform validation, one test run and three production runs were performed. Highest-quality lines were selected to establish master cell banks (MCBs).
Results※
Two MCBs were successfully released under GMP conditions. They demonstrated safety (sterility, negative mycoplasma, endotoxins <5.0 EU/mL and negative adventitious agents), cell identity (>75% of cells expressing markers of undifferentiated state, identical STR profile, normal karyotype in >20 metaphases), purity (negative residual vectors and no plasmid integration in the genome) and potency (expression of at least two of the three markers for each of the three germ layers). In addition, directed differentiation to somitoids (skeletal muscle precursors) and six potential clinical products from all three germ layers was achieved: pancreatic islets (endoderm), kidney organoids and cardiomyocytes (mesoderm), and keratinocytes, GABAergic interneurons and inner-ear organoids (ectoderm).
Conclusions※
We successfully developed and validated a platform for generating GMP-compliant hiPSC lines. The two MCBs released were shown to differentiate into clinical products relevant for our own and other regenerative medicine interests.
Introduction※
Human induced pluripotent stem cells (hiPSCs) are now emerging as increasingly relevant regenerative medicine products. Among those with reported success are dopaminergic cells in Parkinson disease, retinal cells for macular degeneration, beta cells for the treatment of type I diabetes and mesenchymal stromal cells in graft-versus-host disease. The clinical application of hiPSCs requires that their production meets Good Manufacturing Practice (GMP) standards to ensure quality and safety of the product.In addition to the HLA haplobank, the generation of universal hiPSC lines could provide a promising perspective for allogeneic cell therapy. Universal cell lines refer to cells that have been engineered to bypass the major histocompatibility complex mismatch. Achieving a universal hiPSC line would provide an “off-the-shelf” source for cell-based therapies reducing treatment wait time, eliminating donor search, reducing the need for immunosuppression and being a more cost-effective alternative to personalized autologous medicine.
As an academic hospital, the Leiden University Medical Center's (LUMC) main interest is the development of innovative cell and gene therapy medicinal products and especially the application of hiPSC-derived products in patient care. Our aim was to establish a platform for in-house GMP-compliant hiPSCs production using the combined expertise of LUMC's hiPSC hotel, specialized in developing research-grade hiPSC lines, and LUMC's GMP facility (Centre of Cell and Gene Therapy [CCG]), with a proven track record in cell and gene therapy production. This approach allowed the translation of a robust research protocol into a GMP production process. The translation involved, aside from using GMP-released raw materials and performing the manufacturing in a GMP-grade clean room, setting up a quality management system, training of personnel, generating a documentation system and design, validation and qualification of all necessary quality control (QC) tests for final product release. Aside from the generation of GMP-compliant hiPSC lines with demonstrated capacity for differentiation to multiple cell types relevant for cell therapy in regenerative medicine, our validated protocol could be used to produce patient-derived lines for personalized treatments. “Hospital exemption” for cell therapy in the Netherlands would create opportunities to test prospective clinical products.
Methods※
Donor selection and screening of input material※
Male donors (age 18–30 years) were recruited and signed consent forms to use their donated tissue for clinical research, pluripotent stem cell production, commercialization and whole-genome sequencing. For this project, peripheral blood was collected from each donor. A statement of “no objection” was obtained from the LUMC medical ethical committee for donor recruitment. Donors were screened according to EU directive 2006/17/EC and additional EU member state specific requirements. Dutch legislation on genetic modification considers episomal reprogramming as a genetic modification procedure requiring donor testing for the presence of pathogens (i.e., cytomegalovirus and Epstein–Barr virus) for which regulatory elements are present in the episomal vectors. Informed consent form, tests, medical examination, and inclusion decision are as described in supplementary Table 1. All data were handled confidentially and in compliance with Good Documentation Practice in research. For anonymization, donors were allocated encoded LUMC numbers, to which their tissue and data were linked; the key is safeguarded by the principal investigator.
Regulatory compliance of GMP manufacturing※
Manufacturing was performed under European Union GMP regulation following authorized standardized operating procedures generated as described in EudraLex Volume 4 Part IV - GMP requirements for Advanced Therapy Medicinal Products (ATMP) . Clinical manufacturing was performed under aseptic conditions in a grade B clean room and a grade A laminar air flow cabinet.
Release of raw material for manufacture※
Raw materials used were compendial or non-compendial and released for manufacture through in-house GMP risk analysis assessment. Episomal vectors, as described by Okita et al.—were checked for integrity through restriction digest analysis and Sanger sequencing and subsequently produced in high quality ccc-classic grade (Plasmid Factory).
Isolation of peripheral blood mononuclear cells (PBMCs)※
Peripheral blood samples were collected in 10-mL ethylenediaminetetraacetic acid tubes at the LUMC and transported to the in-house GMP facility (CCG) according to GMP-compliant standardized operating procedures. Samples were diluted with an equal volume of StemSpan-ACF Erythroid Expansion Medium followed by density gradient separation using Ficoll-Paque Premium. PBMCs isolated from the Ficoll interphase were counted using the Nucleocounter NC-200 and 12-24 × 106 cells were used for erythroblast expansion. Leftover PBMCs were cryopreserved in CryoStor CS10 at a concentration of 5 × 106 cells/mL using a controlled-rate freezing system for later-stage quality testing or, if necessary, a second production run. Cryopreserved PBMCs were stored in the vapor phase of a liquid nitrogen tank (<–150°C).
Expansion of erythroblasts※
An expanded erythroid progenitor cell population was used for reprogramming. PBMCs were seeded in 12–24 wells of a six-well cell culture plate at a density of 1 × 106 cells per well in StemSpan ACF supplemented with StemSpan Erythroid Expansion Supplement 100×. Cells were cultured in a 37°C incubator with 5% CO2 and the culture medium refreshed daily. On day 2, cells in suspension were transferred to a new well to remove any adherent cells. Erythroblasts were harvested after 7–9 days when the cells exhibited visual signs of expansion.
Reprogramming of erythroblasts and colony selection※
Erythroblast cell number and viability was assessed using a Bürker chamber in a 1:1 Trypan Blue dilution. For reprogramming, two to four independent electroporations were performed per donor, using 0.5–1 × 106 live cells. Cell pellets were resuspended in P3-buffer containing 1 µg of each episomal vector. Electroporation was performed in a Nucleocuvette Vessel with 4D-Nucleofector program EO-115. Subsequently, cells were resuspended in 6 mL of prewarmed ACF supplemented with EES per cuvette and divided over six wells of a 12-well plate coated with 0.5 µg/cm2 Laminin 521 CTG. On day 2 after seeding, 1 mL of ACF plus EES was added to each well. On day 3 and day 5 after seeding, 1 mL of medium was added. From day 7 onwards, cells were refreshed with 1 mL of medium every second day until hiPSC colonies appeared. For picking, each colony was “cut” into 6–12 pieces and transferred to a Laminin-coated well of a 12-well plate. Picked colonies, labeled as passage 1, were considered clonal and treated as independent lines. All lines were coded as follows: LUMC-GMP-iPSC_DxPxLx (Dx: donor number, Px: passage number and Lx: line number).
Maintenance of hiPSC lines※
Cells were passaged (every 6–7 days based on colony morphology, size, and confluency) as 300–500 aggregates of 50–200 µm on laminin (P2) or vitronectin -coated wells of a six-well plate (P3 onwards) containing 2 mL of medium by incubating them with Versene CTS for 4 min at room temperature (RT). Medium was refreshed every 1–3 days.
Cryopreservation of hiPSC lines※
hiPSCs were cryopreserved at passages 3, 5, 7, 9, 10 and 12 for quality testing and banking. Cell aggregates equivalent to half a well of a six-well plate (P3-P7 and P12) or the equivalent of one well of a six-well plate (P9/P10) were frozen in 1 mL of CryoStor. Freezing was performed in a controlled-rate freezing system. Cryopreserved hiPSCs were stored in the vapor phase of liquid nitrogen (<–150°C).
Master cell bank (MCB) generation※
Three lines of one donor at passage P10 were selected for generation of the MCB. Five vials cryopreserved at P10 were plated onto 10 wells of six-well plates coated with vitronectin in mTeSR Plus medium supplemented with RevitaCell . Medium was refreshed the next day to avoid prolonged exposure of the cells to RevitaCell. After 6–7 days, cells were passaged to eight vitronectin-coated T175 flasks in 40 mL of medium per flask for scale-up. Cells were maintained as described previously and harvested at P12 using 18 mL of Versene CTS per flask for 4 min at RT. Cell aggregates from seven flasks were pooled and cryopreserved in CryoStor CS10 in 120 vials and the last flask was used for QC testing.
QC testing※
Sample preparation for QC tests※
To assess sterility, mycoplasma and endotoxin levels, the culture supernatant from all wells or flasks was collected and pooled. Matrix verification was performed to ensure the validity of the assays. For cell viability, the undifferentiated state, trilineage differentiation, residual vector analysis, STR-profiling, whole-genome/exome sequencing and adventitious agents testing, hiPSCs were dissociated into single cells by incubation with 18 mL of Versene CTS at 37°C for 10 min. For karyotyping, a small sample was taken of cell aggregates from the pooled pellet obtained after MCB harvest. Table 1 shows QC tests performed, material used and acceptance criteria.
Cell and colony morphology※
The cells were examined under a Leica stereomicroscope at every step of the manufacturing process to confirm typical morphology and to monitor the degree of spontaneous differentiation. Before passaging, colony and cell morphology were checked and imaged using an inverted microscope.
Cell count and viability※
Cell numbers and viability after harvest was assessed using the Nucleocounter according to European Pharmacopoeia (Ph.Eur.) 2.7.29 Nucleated cell count and viability.
Sterility testing※
Sterility testing was performed at the LUMC's accredited microbiology laboratory according to Ph.Eur. 2.6.27, Microbial Examination of cell-based Preparations (aerobic and anaerobic microbial contamination) procedures using BacTec flasks.
Mycoplasma testing※
Mycoplasma tests were performed at the LUMC's accredited microbiology lab with PCR in compliance with Ph.Eur. 2.6.7., Mycoplasmas.
Endotoxin testing※
Endotoxin testing was performed at the LUMC's accredited Clinical Pharmaceutical Lab using the LAL test and complies with Ph.Eur. 2.6.14, Bacterial Endotoxins.
Adventitious agents testing (AAT)※
AAT was performed using the Twist Comprehensive Viral Research Panel. The assay was performed by and validated for sensitivity and reproducibility by GenomeScan BV (Leiden, Netherlands). The validation process was supervised and reviewed by the QC ATMP team of the LUMC GMP facility (CCG).
Residual episomal vector analysis※
gDNA was extracted from a maximum of 5 × 106 hiPSCs at the indicated passage numbers using a DNeasy Blood & Tissue Kit according to the manufacturer's protocol. The presence of residual reprogramming vectors was tested with a quantitative polymerase chain reaction (qPCR)-based assay using primers annealing to the gene encoding for EBNA-1 present in each of the reprogramming vectors and primers annealing to hexokinase 2 (HK2) as a reference for DNA input. qPCR was performed using SYBR Green Master Mix , reverse and forward primers , and 120 ng of genomic DNA on the Quant Studio for 35 cycles. A negative PCR signal (no residual vector detected) was set to a CT value of 35. The assay was validated and qualified in-house.
Analysis of markers of the undifferentiated state※
Dissociated cells were fixed and permeabilized using BD Kit according to the manufacturer's recommendations. Permeabilized cells were incubated at 4°C for 30 min with the following antibodies at the respective dilutions: . Cells were washed with buffer and FACS buffer , resuspended in 150 µL of FACS buffer and subjected to spectral flow cytometry using the Cytek Aurora. The assay was validated and qualified in house.
Short tandem repeats (STR) profiling※
STR profiling of primary PBMC and hiPSCs was performed by the accredited department of Clinical Genetics of the LUMC. The STR profile of the hiPSCs was compared with the input PBMC to ensure authenticity of the hiPSC lines.
Karyotype analysis by G-banding and KaryoStat※
Undifferentiated hiPSC cultures of 5–6 days were treated with solution for 1 h at 37°C to arrest cells in metaphase and processed further by the LUMC's accredited Diagnostic Genome Analysis laboratory for G-banding. At least 20 metaphases were examined for cytogenetic aberrations. Karyotypes were described according to the International System for Human Cytogenetic Nomenclature. Karyotype analysis was verified using the KaryoStat+ assay performed by Thermo Fisher Scientific on hiPSCs harvested at P10 and ≥P26.
Whole-genome sequencing (WGS) and whole-exome sequencing (WES)※
To assess possible integration of reprogramming vectors, WGS was performed using the NovaSeq6000 platform at P9/10. WES was performed using the NovaSeq6000 platform and the Agilent SureSelectXT human all exon v7 capture library on MCB samples at P12 and P26. Both assays were outsourced to the accredited company GenomeScan.
Downstream analysis of WGS and WES data※
For single nucleotide polymorphism/indel calling, and structural variant analyses, the Illumina DRAGEN output was used on donor-derived PBMCs and hiPSC lines. Variants in the exon coding region were double checked using WES data generated from hiPSC lines at P12 (MCB), and P26. The workflow, based on GATK-compliant algorithms, entails quality control and a read preprocessing stage in which adapter sequences, low-quality regions (phred < 20) and poly G trimming was performed.
Analysis of differentiation capacity※
Trilineage differentiation was performed using the StemDiff Trilineage Differentiation Kit according to the manufacturer's recommendation. Subsequently, cells were fixed with 2% paraformaldehyde. Differentiation capacity was confirmed by immunocytochemistry (ICC) using lineage-associated markers.
Stability testing※
To monitor the effect of long-term storage, a stability study protocol was established in which the post-thaw sterility, viability, identity, and potency of the cryopreserved product is and will be assessed at 6 months, 1 year, 2 years, 5 years and 10 years.
Differentiation to prospective clinical products※
SC-islets (hiPSC-derived pancreatic islets)※
hiPSC lines were switched to expansion culture on vitronectin-coated plates in Essential 8 medium and passaged using Versene. Differentiation towards pancreatic islets was conducted using a fully clinically-compliant protocol in three-dimensional suspension.
Kidney organoids※
hiPSCs were switched to Essential 8 medium and differentiated using previously described protocols.
Cardiomyocytes※
hiPSCs were differentiated into cardiomyocytes and analyzed by FACS and ICC as described previously by van den Brink et al.
Keratinocytes※
hiPSCs were differentiated to keratinocytes as previously described by Ruiz-Torres et al.
Neural lineage※
hiPSCs were first differentiated for 16 days into medial ganglionic eminence (MGE) progenitors and then matured into GABAergic interneurons for 42 days.
Inner-ear organoids※
hiPSC lines were differentiated toward inner ear organoids following the protocol described by van der Valk et al.
Somitoids※
Somitoids were generated according to the protocol described by Sanki-Matsumiya, et al.
Results and Discussion※
Critical steps involving donor selection, screening of input material, translation of the research protocol to a GMP-compliant production process and QC testing are described herein. For platform validation, a minimum of three production runs was required. Hence, three donors were recruited for GMP production up to cryopreservation and QC testing of the seed lot. GMP manufactured seed lots passing all controls, including in-process monitoring, in-process control and QC tests, were qualified person (QP)-released. Production of three MCBs was required including their full QC release testing. MCBs passing all QC release testing were GMP released.
Donor selection and screening of input material※
Male donors were chosen to avoid possible erosion of X chromosome inactivation described for hiPSC lines from female donors. Tissue from three donors consenting to research use was used under standard laboratory conditions to translate the research reprogramming protocol to a GMP-compliant process. Ten lines were first generated from each donor and used for validating hiPSC-specific release tests. The next run was a test in the GMP facility, using PBMC from a bone marrow donor who consented to research use of surplus donated material. For the three GMP production runs, donors were recruited from November 2021 through March 2022. These donors consented to the use of their material for clinical research, stem cell production, commercialization and whole-genome sequencing. All donors passed the selection criteria and medical examination outlined in supplementary Table 1; these tests were performed on the day of blood collection. To exclude potentially missed infectious agents during the initial material collection, donors consented to a second blood collection 6 months after their first donation for repeated testing. Medical examinations were performed by an internist. Donated material was released for ATMP production by a qualified hematologist and a QC ATMP officer.
Translation of the research protocol to a GMP protocol※
PBMCs were used as input material for erythroblast expansion and subsequent reprogramming. The in-house reprogramming protocol for generating research grade hiPSCs underwent a number of changes to translate the process to GMP. First, we replaced research-grade erythroid expansion medium containing bovine serum albumin with an animal component–free erythroid expansion medium. Second, instead of pCXLE-OCT3/4-shRNA-p53, which contains an shRNA against p53, we used pCXLE-Oct3/4 plasmid without the p53 knockdown. The rationale for omitting p53 knockdown was related to potential safety issues in patients, where an intact p53 (tumor suppressor) pathway would be advisable. ReproTeSR was replaced by mTeSR-Plus as reprogramming medium because its composition is known. All media were free of antibiotics. In addition, (undefined, animal-derived) Matrigel was replaced by laminin 521 CTG (first three passages) and vitronectin-CTS (later passages). During reprogramming laminin supported superior colony outgrowth and up to 40% more colonies formed. However, on vitronectin, hiPSC colony morphology was more homogenous which facilitated elimination of spontaneous differentiation and better culture maintenance at later passages. These modifications to the research protocol did reduce erythroblast expansion and reprogramming efficiency. Compared with Okita and Yoshida, reprogramming using our GMP protocol was 100–1000 times less efficient. We therefore scaled up to 24 × 106 PBMCs for erythroblast expansion and performing multiple electroporations per donor in parallel to increase the number of colonies obtained. We noted that erythroblast expansion rate and reprogramming efficiency were highly donor dependent, varying from 3 to 300 colonies per donor. Details of erythroblast expansion and reprogramming efficiency are shown in supplementary Table 2.
GMP-compliant hiPSCs production※
Individual colonies were picked between day 14 and 24 after electroporation. To avoid cross-contamination of colonies, we only selected those that were isolated from other colonies and picked only one colony per well. One colony was placed in one well of an independent 12-well plate, designated passage 1 and assigned a unique line number for identification (DxPxLx). Each of these lines was considered an individual product, meaning that from passage 1 onwards different lines were never handled simultaneously. From passage 2 onwards, colonies were maintained in six-well plates and passaged as aggregates once every 6–7 days as soon as they had centers that were dense and phase-bright compared to their edges. Every second passage, several wells of each line were cryopreserved. Selection of lines during the production process was based on visual monitoring of spontaneous differentiation, colony morphology and episomal plasmid retention.
Selection of lines and production of MCB※
The three pre-seed lots of D2 were selected for validation of MCB production based on donor characteristics. Blood type incompatibility can cause acute transplant rejection. Therefore, donor D2 with AB0 type 0+ was preferred to D4 with AB0 type B+. Scale-up to allow cryopreservation of 120 MCB vials was achieved by seeding 5 pre-seed lot vials in 10 wells of 6-well plates. MCB production D2MCBL01B was successful after the first attempt (D2MCBL01) failed due to infection. Next to all QC tests performed on the pre-seed lots, additional QC tests for endotoxin, AAT and WES/WGS were performed on the MCBs. For AAT, a next-generation sequencing–based method was used. We used the highly sensitive comprehensive viral research panel developed by Twist Biosciences, capable of screening for >3000 different viruses, including >15 000 viral strains. This method is able to detect viruses that are missed using conventional AAT much faster, does not require animal testing, and is more cost-efficient compared to conventional AAT. While its suitability is still under debate, reports from workshops organized with regulatory agencies predict authorization of NGS-based AAT in the near future. No viral contamination was observed in the MCBs tested.
Differentiation to prospective clinical products※
To test the capacity for differentiation to potential clinical products, hiPSC lines D2MCBL01B and D2MCBL03 underwent directed differentiation into various lineages. SC-islets generated from both lines (n = 3 for each line) generated on average 50.6 ± 12.0% SC β cells (C-peptide–expressing cells), 11.0 ± 7.0% SC α cells (glucagon-expressing cells) and 8.1 ± 6.8% polyhormonal cells (positive for both C-peptide and glucagon), validated by ICC (Figure 3A). Also, dithizone was used to selectively stain the SC-islet clusters red. Overall, SC islets produced from both lines exhibit a cytoarchitecture and composition that is similar to previously published SC islets and primary donor islets.
Successful differentiation was achieved for both GMP-released hiPSC lines into somitoids and six potential clinical products from all three germ layers: SC-islets (endoderm), kidney organoids and cardiomyocytes (mesoderm), and keratinocytes, GABAergic interneurons, and inner ear organoids (ectoderm).
Conclusions※
We successfully developed and validated a GMP-compliant production platform for in-house hiPSCs production. This resulted in five pre-seed lots that passed all QC-testing and two MCBs that were QP-released and can be used as starting products for allogeneic clinical products. Furthermore, the platform as such will allow the production of patient-specific hiPSCs and their derivative cell therapy products for autologous therapies when the time arises. The establishment of this validated platform represents a critical milestone in the implementation of hiPSC-based regenerative medicine within our own university hospital and likely that of others. Moreover, the potential of the GMP-released MCBs to differentiate into putative clinical products highlights their suitability for therapeutic applications.
It is important to acknowledge that the methodology and materials we used here fall partly under patents of third parties and therefore the released products would have licensing restrictions when used. It will be up to users to negotiate licenses as necessary should they request use of these lines. Furthermore, clinical products from allogenic hiPSCs would require immune suppression to counteract transplant rejection. In some cases, the immunogenicity of the organ for transplantation (like the eye and brain) may only require minimal or short periods of immune suppression. However, for highly immunogenic tissues (like the kidney and beta cells), immune evasive hiPSCs lines would be an alternative for off-the-shelf allogeneic cell therapies.
