Manufacturing clinical‐grade human induced pluripotent stem cell‐derived beta cells for diabetes treatment

Abstract The unlimited proliferative capacity of human pluripotent stem cells (hPSCs) fortifies it as one of the most attractive sources for cell therapy application in diabetes. In the past two decades, vast research efforts have been invested in developing strategies to differentiate hPSCs into clinically suitable insulin‐producing endocrine cells or functional beta cells (β cells). With the end goal being clinical translation, it is critical for hPSCs and insulin‐producing β cells to be derived, handled, stored, maintained and expanded with clinical compliance. This review focuses on the key processes and guidelines for clinical translation of human induced pluripotent stem cell (hiPSC)‐derived β cells for diabetes cell therapy. Here, we discuss the (1) key considerations of manufacturing clinical‐grade hiPSCs, (2) scale‐up and differentiation of clinical‐grade hiPSCs into β cells in clinically compliant conditions and (3) mandatory quality control and product release criteria necessitated by various regulatory bodies to approve the use of the cell‐based products.


| INTRODUCTION
Diabetes is a debilitating disease affecting millions worldwide. 1 Depending on the subtype, diabetes can be attributed to the autoimmune destruction of pancreatic beta cells (β cells) in type 1 diabetes (T1D), or dysfunction of β cells in type 2 diabetes (T2D). As such, curative treatment of diabetes may be attained by β cell replacement therapy. Apart from whole pancreas or islet transplantation, β cells may also be replaced by transplanting human pluripotent stem cell (hPSC)derived β cells into diabetes patients to achieve insulin independence.
hPSCs are one of the most attractive sources of cells for cell therapy applications due to their unlimited proliferative capacity and their ability to differentiate into lineages of the three germ layers. 2 While tremendous progress has been made in developing and refining strategies to Lay Shuen Tan and Juin Ting Chen are the co-authors. differentiate hPSCs into clinically suitable insulin-producing endocrine derivatives or functional β cells, [3][4][5][6][7] clinical application and commercialization of these functional β cells for diabetes treatment are impeded by numerous difficulties, notably the regulatory challenges surrounding the generation, release and clinical use of stem cell-derived products for cell therapy purposes. [8][9][10][11] In recent years, there have been various international efforts by key stem cell experts to harmonize the critical quality attributes (CQAs) of clinically compliant stem cells and stem cell-derived cell therapy products, [12][13][14][15][16][17] but the requirements dictated by different national and international regulatory bodies and organizations (Table 1) may still vary. [18][19][20][21][22][23][24][25][26][27][28] Therefore, the industry will need to assess which countries to operate in and engage the relevant regulatory authorities to ensure compliance in the entire cell product manufacturing process. Generally, the overarching principles that are important in the generation of clinically compliant stem cell-based products that conform to good manufacturing practice (GMP) are: writing and following of site master files and standard operating procedures (SOPs), having well-defined procurement, storage, shipment and tracking processes, incorporating proper facility design, training of staff to perform and document all laboratory and administrative processes, regular equipment maintenance and last but not least, conducting regular quality checks and compliance assessments to ensure accountability, performance and safety in all processes and end products ( Figure 1). 23,25 Here, we focus on discussing human induced pluripotent stem cell (hiPSC)-derived insulin-producing β cells as the commercial therapeutic product for diabetes treatment. While both human embryonic stem cells (hESCs) and hiPSCs have been used for cell therapy, 2 the derivation of hESCs is accompanied by the ethically controversial destruction of human embryos. As such, the derivation of hESCs is permitted only in parts of the United States (US), China and Europe (EU). 29 In contrast, hiPSCs are a more universal resource for clinical and commercial purposes as they do not face the same ethical issues that hESCs face, and can also be prepared from various somatic cell types of choice. 30 In this review, we analyse the suitability of various somatic cell reprogramming strategies to generate hiPSCs destined for clinical translation, discuss the importance of xeno-free culture systems and summarize the mandatory CQAs of clinically compliant hiPSCs.
We also provide insights into the methods and technical complexities involved in the scale-up of clinical-grade hiPSCs before differentiating them into insulin-producing β cells-a process that is technically challenging but vital for the success of large-scale commercial cell manufacturing. 31 Furthermore, we offer suggestions on key parameters for quality control (QC) testing during specific stages in the differentiation process and provide a summary of the current efforts on transitioning to serum-free and/or xeno-free differentiation culture medium to rule out the possibility of animalderived infections in future transplant patients. Last but not least, we present a broad overview of the guidelines governing the release of hiPSC-derived β cells for clinical applications in diabetes treatment, detailing the extensive testing required by international regulatory bodies to ascertain product identity, viability, sterility, safety and potency before product release.  Takahashi and Yamanaka. 32 However, the use of retroviruses causes undesired permanent integration of viral vector transgene and backbone into the genome, thus raising concerns pertaining to the risk of unintended insertional mutagenesis. 32 Since then, safer reprogramming alternatives to generate clinicalgrade hiPSCs have been developed. Amongst these, the use of the Sendai virus (SeV), episomal vectors and synthetic mRNA are some of the most efficient reprogramming methods and will be discussed in detail below. A comparison of reprogramming approaches with SeV, episomal vectors and synthetic mRNA is also summarized in Table 2.
As SeV is a single-stranded negative-sense non-integrative RNA virus that does not replicate through the DNA phase, SeV reprogramming strategies thus harbour no risk of genome integration. Its use in reprogramming was first reported in 2009 by Fusaki et al., 50 and many studies have demonstrated high efficiency in reprogramming multiple somatic cell types into hiPSCs. 37,44,50 However, it is to be noted that the use of virus in SeV-based reprogramming complicates the regulatory authority approval process, and there is a need to dilute and eventually rid hiPSCs of residual SeV via passaging. 37,51 Based on Macarthur et al., 45 complete vector clearance can be achieved without complications within 10 passages after SeV reprogramming.
In contrast, the use of plasmid DNA in episomal reprogramming renders ease of assimilating this method into clinically compliant processes as compared to SeV reprogramming. 37,51 While episomal reprogramming efficiency is poorer than that of SeV, several groups have developed techniques that improved the efficiency of this method of reprogramming. 46,49,53,54 It is notable that one drawback of the episomal method is that there remains a low possibility of episomal vector integration into the genome. 37,51 Out of the three methods, the use of synthetic mRNA is possibly the safest for clinical translation as it does not utilize virus or plasmid DNA that carries an inherent risk of genome integration. 52 However, its main limitation is its poorer efficiency in reprogramming some non-invasive cell sources such as blood and keratinocytes (Table 2). Furthermore, this method is laborious as mRNAs must be fed to the cells daily until colony emergence. 37,52 Therefore, this method is mainly recommended when skin fibroblasts are used as the starting somatic cell type.

| Somatic cell types suitable for reprogramming
Blood cells are generally prioritized over skin fibroblasts as the starting somatic cell type for reprogramming due to the ease of accessibility. [55][56][57] T A B L E 1 Non-exhaustive list of regulatory authorities and stem cell organizations involved in stem cell therapies Furthermore, skin fibroblasts harbour somatic mutation risks due to outward exposure to environmental mutagens such as sunlight. 58 Other non-invasive cell sources such as exfoliated renal epithelial cells from urine and keratinocytes from hair can also be considered if desired. 59,60 Ultimately, the choice of reprogramming method and starting somatic cell type will be based on the industry's preference when aligned with the regulatory requirements of their site of operation.

| Xeno-free culture conditions
The use of xeno-free culture reduces the risk of immune reactions and zoonotic infections associated with the use of animal-derived reagents.
The manufacturing process should thus be designed in which the entire pipeline is done in chemically defined xeno-free culture systems. In addition, all cell culture components should be defined chemically for F I G U R E 1 Workflow to generate clinically compliant stem cell-based products with good manufacturing practice (GMP). First, planning of the correct facility design and processes is enabled by putting together a multidisciplinary team of stem cell biologists, process engineers and skilled laboratory managers. Standard operating procedures (SOPs) need to be devised for both administrative procedures such as procurement and shipping of raw materials, reagents and equipment and laboratory procedures such as stem cell maintenance, protocol for differentiating stem cells to end-stage cell products, operating bioreactor systems and performing flow cytometry for cell characterization. After the planning phase and setting up of all GMP facilities and processes, staff must be trained on all relevant SOPs before proceeding with the manufacturing process. Trained staff will be required to execute the SOPs, document all their activities and observations in logbooks and record all quality control data generated. To ensure quality performance, routine equipment maintenance and on-site audit checks by regulators on current processes, previous batch records, staff practices and hygiene will need to be conducted. Processes will need to be reviewed and improved if necessary. This figure is created with BioRender.com standardization when culturing hiPSCs, for documentation purposes and for preventing issues arising from lot-to-lot variability in undefined components such as fetal bovine serum (FBS). [61][62][63][64][65] Currently, xeno-free conditions have been successfully incorporated into SeV, episomal and synthetic mRNA reprogramming processes by several groups. [44][45][46][47][48] A multitude of xeno-free stem cell culture media and chemically defined feeder-free extracellular matrix proteins, such as vitronectin and laminin, are now commercially available to replace feeder cells and Matrigel in the surface coating of hiPSC culture dishes. These should all be integrated into the production process. [44][45][46][47][48]61,62,[64][65][66][67][68][69][70][71][72][73][74] Finally, the hiPSC passaging reagent should also be carefully chosen based on efficiency, reliability and reproducibility. 62 To ascertain genomic identity/stability, the clearance of residual reprogramming vector and/or virus in seed and master cell banks must be proven by appropriate methods. 77,78 Karyotyping analysis by Giemsa banding (G-banding) should also be performed to confirm with 95% probability that the hiPSCs do not carry chromosomal abnormalities and where possible, whole genome or exome sequencing can also be performed. 61,78 To distinguish between hiPSC lines and to avoid cross-contamination, short tandem repeat (STR) fingerprinting analysis should be performed with minimally eight core STR loci according to International Cell Line Authentication Committee guidelines, 19 while minimally 15 loci are required if hiPSCs are destined for autologous cell therapy. 79 In Lonza's cell line authentication process, as many as 16 loci are usually analysed for at least 80% match. 80 To assess pluripotency, flow cytometry analysis is a robust quantitative method to determine pluripotency marker expression in hiPSC lines. A combination of surface pluripotency markers (e.g., TRA-1-60) and intracellular pluripotency markers (e.g., OCT3/4, SOX2, NANOG) should be selected for flow cytometry, and >70% pluripotency marker expression can be set as the minimum standard for QC testing. 75 Last but not least, pluripotency tests such as the teratoma assay or embryoid body-based three germ layer differentiation will also be ideal to demonstrate pluripotency of the hiPSCs. 77 • High efficiency-blood, 33 urine, 38 hair keratinocyte 39,40 • Low efficiency-skin fibroblast 37 • High efficiency-skin fibroblast, 37 urine 41 • Low efficiency-blood, not efficient but possible with blood-derived endothelial progenitor cells, 42 hair keratinocytes 43 Examples of xeno-free methods described • Churko et al. 44 and Macarthur et al. 45 • Chen et al. 46 • Warren et al. 47,48 Reprogramming agent clearance • Within 10 passages 45 • Within 11-20 passages 49 • Immediately Ease of assimilation into clinical processes differentiation into insulin-producing β-cells, as it is difficult to scale up during the differentiation process. About 5000-10,000 islet equivalent (IEQ) per kilogram of the recipient's body weight is required to improve metabolic control of blood glucose levels. 81,82 As 1000 β cells are estimated to be present in an IEQ, 83 close to 1 billion hiPSCderived β cells will be required for each diabetes transplant patient.
Similarly, a therapeutic dose of more than 1 billion cells per patient is also commonly estimated by pharmaceutical manufacturers in the context of large-scale commercial manufacturing for allogeneic cell therapy. 31 Hence, to cater to the demand for hiPSC-derived β cells, hiPSCs, which are the starting material for the differentiation process, will need to be readily expanded. This therefore underlines the need for optimized culture conditions and standardized vessels for robust manufacturing of cell-based products and guarantee the production of sufficient hiPSC-derived β cells for cell therapy. To that end, various strategies pertaining to the inoculation methods and feeding strategies for two-dimensional (2D) static culture system and three-dimensional (3D) suspension-based conditions, as well as the vessel choices for hiPSC expansion ( Figure 2) are described in more detail below.
T A B L E 3 Summary of relevant in-process and final product testing during hiPSC-derived β-cell manufacturing  [100][101][102] However, such methods prove cumbersome in bioprocessing as additional re-encapsulation and de-encapsulation steps will be required for routine hiPSC passaging. 103 In addition, nutrient diffusion and monitoring of cell growth in the capsule may also be limited by the physical properties of encapsulation.

| Feeding strategies
Unlike 2D planar systems where limited culture size will require frequent medium replacement (repeated batch feeding), 3D vessels with larger volume capacity and capability for automated processing can allow for fed-batch systems and perfusion systems to be implemented. In fed-batch systems, nutrient supplements can be added to prevent growth inhibition of hiPSCs. In contrast, spent medium is continuously removed while fresh medium is simultaneously added in perfusion systems. Overall, perfusion feeding is more advantageous as it allows for more homogeneous culture conditions and perfusion feeding has in fact been previously documented to lead to 47% higher expansion yield than batch-fed cultures. 104 In addition, as a closed-loop system, perfusion feeding strategies also minimize the risk of contamination and thus, reduce the number of in-process sterility tests. 105 However, the overall operational complexity and costs associated with perfusion systems make it prohibitive for widespread adoption. 106

| 2D versus 3D culture vessels for cell culture and differentiation
Despite the incompatibilities of 2D planar culture with some of the afore-mentioned inoculation strategies, there are some merits to this method. 2D static culture system has been routinely employed in laboratories for stem cell maintenance as it is simple to implement and is more cost-effective than its 3D suspension-based counterpart. Culturing hiPSCs as a monolayer, rather than aggregates in suspension, also allows the cells to be more evenly exposed to nutrients. In the 2D culture system, multi-layered cell stacks and factories can be utilized for the expansion of hiPSCs through a scale out process, where capacity increases linearly with the number of cell stacks added. 107,108 However, as cell growth will still be limited by the surface area of culture plates, 3D culture system is a more viable alternative for hiPSC expansion. Moreover, variability in conditions between cell stacks of the multi-layered cell plates have also been documented. 108  Despite the afore-mentioned progress, one key limitation of current differentiation methods is that the entire process, or at least part of it, is not serum-free and/or xeno-free. This may lead to translational issues which we will address in the next subsection.

| Current serum-free and/or xeno-free standards incorporated into β-cell differentiation methods
As an important source of various complex growth factors and molecules, FBS is often included as an essential nutrient supplement in culture media supporting the differentiation into pancreatic cells. 4,5,115,116,120,125,126 While most regulatory bodies do not explicitly ban the use of FBS in the manufacturing process of stem cellderived cell products, the undefined nature of FBS with batch-tobatch variation may increase the complexity of QC and safety testing required. 127 Combined with the risks of disease transmission and ethical issues associated with its zoonotic origin, there is a need to move away from FBS towards other serum-free and xeno-free defined alternatives for hiPSC differentiation. 127 A popular serum-free defined alternative adopted by various groups to obtain hiPSC-derived β cells is bovine serum albumin (BSA), as it is an abundant protein present in serum, even though it is not strictly considered xeno-free, being a bovine-derived protein. 3,6,7,114,128 The use of supplements such as B-27, which are commercially available in both serum-free and/or xeno-free versions, have also been reportedly added during β-cell differentiation, though these protocols did not completely eliminate FBS use. 5,115,128 This partial serum-free replacement approach has also been utilized in Pagliuca et al., 114 where differentiation media in the first 20 days of differentiation (S1, S2, S3, S5 media) is supplemented with fatty-acid free BSA instead of FBS. However, the basal CMRL-1066 medium for S6 media used in the final 15 days of differentiation, involving committing endocrine progenitors to the functional insulin-producing β-cell fate, is still supplemented with 10% FBS. 114 Separately, there have also been some efforts in recent years to replace S6 media, such as the use of an enriched serum-free media designed to ensure a serum-free differentiation process. 87 Separately, Rezania et al. and several other groups were reportedly able to completely replace FBS with BSA supplementation in their β-cell differentiation protocol. 3,7,126,128 While the complete elimination of FBS supplementation is a huge step towards serum-free and xeno-free differentiation conditions, it is notable that the basal MCDB131 medium commonly adopted for β-cell differentiation still contains low amounts of dialyzed FBS. 129 Until a completely serumfree and/or xeno-free GMP compliant β-cell differentiation process is devised, it is currently recommended for manufacturers to adopt minimal serum use in their β-cell differentiation procedure to lower the risks of spreading zoonotic disease and infections post-administration.
Thus, should low amounts of serum components such as FBS need to be used for β-cell differentiation, safety and sterility testing in endstage cell products should be rigorously implemented before clinical applications.
Eventually, to completely eliminate animal-derived components in differentiation media, human serum albumin may be used as a replacement for BSA. 130 Furthermore, apart from transitioning to differentiation media with minimal serum supplements, other plausible strategies such as using human recombinant growth factors or small molecules devoid of any animal-derived component can be used.
However, the change has to be thoughtfully considered, selected, tested by manufacturers and vetted and approved by regulatory bodies before incorporating into the finalized hiPSC-derived β-cell differentiation procedure. 131 Finally, in relation to purity, it will also be necessary to demonstrate that cytokines and growth factors used during the differentiation process are not present in the final product. 132 Endotoxin testing will also be required. While an acceptable range of endotoxin levels has not been set for cell therapy products, a maximum endotoxin level of 0.25 EU/ml has been established for water in injection products and may be extrapolated for transplantable hiPSC-derived β cells. 133  Ultimately, it is recommended for manufacturers to weigh the pros and cons of each method and choose the most suitable method aligned to industry standards. Even with the steps to remove undifferentiated hPSCs, after each batch of hiPSC-derived β cell is produced, testing for any tumourigenicity due to residual hiPSCs is still required. Details on product release testing and criteria based on international regulations will be discussed in the next section under the safety subsection.

| QC and characterization based on identity during β-cell differentiation
To reduce heterogeneity, increase cell survival and maximize differentiation outcomes when clinical-grade hiPSCs are differentiated in 3D suspension cultures as cell clusters (the current preferred choice for generating insulin-producing β cells as mentioned in the previous section), hiPSCs with a high nuclear-to-cytoplasmic ratio should be used ( Figure 3A). One QC strategy is to monitor and standardize the starting hiPSC clump sizes. 88,114,125 Single cell-based dissociation method as above-mentioned can be utilized to increase homogeneity.
An example of uniform hiPSC clump sizes at 200 μm in diameter is shown in Figure 3B and the morphology of clumps during the differentiation process is shown in Figure 3C. Suggestions of both mandatory and optional QC parameters during the pancreatic differentiation process are summarized in Table 3. 3,6,[114][115][116]128,144,145 It is notable that while some protocols reportedly generate up to 90% PDX1 + NKX6.1 + double positive pancreatic progenitor cells, 144,146 the bar is currently calibrated at 60% at the afore-mentioned stage in Table 3 based on multiple considerations such as the variability in differentiation efficiencies between different hPSC lines. 3,144 As β-cell differentiation methods are optimized to become more efficient and less cell-line dependent over time, it is prudent for manufacturers to anticipate increasingly stringent QC parameters for stage-specific markers as the field advances.

| PRODUCT RELEASE TESTING AND CRITERIA BASED ON INTERNATIONAL REGULATIONS
Testing of the final cell therapy product is required to demonstrate its identity, purity, sterility, viability, safety and potency. While QC methods and release testing to ensure proper identity and purity of hiPSC-derived β cells have been addressed earlier, the appropriate assays to document sterility, viability, cell count, safety and potency of hiPSC-derived β cells are not the same. Here, we detail the various tests necessary to meet the remaining criteria for product release.

| Sterility
As generating β cells from hiPSCs requires multiple manipulations and as they are cultivated for a relatively long duration in vitro, sterility testing of the culture environment and in-process products should be incorporated into the acceptance criteria for release. Since the final hiPSC-derived β cell product cannot be terminally sterilized, sterility testing of the final product is ever more essential to ensure that it is free of adventitious agents prior to administration into recipients. With reference to regulatory policies from countries such as US, EU, Japan and China, 147-151 mandatory tests for virus, bacteria and mycoplasma contamination should be regularly administered during the process of derivation of hiPSC-derived β cells and on the final product. A summary of the sterility tests to be implemented at various derivation steps for hiPSC-derived β cells is provided in Table 3, and the methods for bacterial and mycoplasma testing will also be further discussed below.
In accordance to The International Pharmacopoeia that was harmonized with the US Pharmacopoeia, European Pharmacopoeia and Japanese Pharmacopoeia, a 14-day sterility test and a 28-day mycoplasma test should be administered to certify that the product is free from bacterial and mycoplasma contamination, respectively. 152 Unfortunately, longterm in vitro cultivation of stem cell-derived β cells has been shown to result in decreased functionality. 87 Hence, due to the probable short shelflife of hiPSC-derived β cells, these conventional testing methods with long read-out durations are less feasible for final product release testing.
It is noted that the US Food and Drug Administration (FDA) and European Medicines Agency (EMA) also permit product administration before final product sterility test results are obtained if justifiable QC can be conducted. 147,148 While the EMA did not specify the sterility assurance required, the FDA outlined it as (1)

| Safety
The safety concerns of hiPSC-derived cell therapy products largely pertain to its genetic stability in long-term culture and residual tumorigenic potential. hiPSCs destined for large-scale cell therapy product in genetic abnormalities. 162,163 With the potential for these mutations to confer a growth advantage, it is important to regularly assess the genomic stability of hiPSC-derived products during the manufacturing process and before product release (Table 3). If aberrations are detected, the final cell product can only be administered when the aberrations are documented to be functionally insignificant. 164 In addition, as hiPSCs can form teratomas, 165  Furthermore, it is not practical to conduct in vivo tumourigenicity studies for batch release due to the long duration required. Therefore, to address the safety concerns of hiPSC-derived β cells, the use of in vitro assays to examine the propensity of tumour formation for each lot release is recommended by various regulatory bodies. 151,167,170 For instance, telomerase repeated amplification protocol can be conducted to test for telomerase activity, which should be completely absent in hiPSC-derived β cells. 171 With reference to FDA and PMDA, tests focused on detecting residual hiPSCs should also be conducted to properly assess the tumorigenic risk in hiPSC-derived β cells. 134 Simple and sensitive methods of detection through quantitative flow cytometry of hiPSC markers as aforementioned for hiPSC characterization can be carried out. 172

| Potency
The choice of potency assay(s) is subjective and dependent on product characteristics. Hence, current global regulations do not dictate a specific type of potency assay to be used for cell therapy products. In accordance with FDA guidelines, the potency assays used should be validated and should be able to directly measure the cell product's activity relevant to its mode of action with accuracy, precision and robustness. 173 Here, we propose a myriad of in vitro and in vivo assays that can be used to demonstrate the functionality of β cellssecretion of insulin in response to high blood glucose levels to maintain euglycemia.
As the goal of hiPSC-derived β-cell therapy is to mitigate the scar- In addition, an examination of the calcium signalling involved in insulin secretion will be beneficial to detail the physiological efficacy of β cells in vitro. When glucose levels increase in vivo, uptake of glucose is mediated through glucose transporters, and glucose is then catabolized in the aerobic respiration pathway to generate adenosine triphosphate (ATP). 174 In response, ATP-sensitive K + (K ATP ) channels close and initiate a wave of membrane depolarization, which allows for calcium influx through voltage-dependent calcium channels and the eventual insulin exocytosis. 174  loss of viability and functionality post-thaw. 148 However, this can be difficult to achieve due to the 3D architecture of hiPSC-derived β-cell 'organoids'. During the freezing process, layers of cells within the organoid structure with different intracellular water potential are exposed to varying temperatures. This culminates in the formation of damaging intracellular ice crystals. 182,183 Hence, cryopreserved islets/ β-cell organoids generally show a decrease in viability and functionality post-thaw. 184 The lack of optimized cryopreservation methods therefore hinders an on-demand availability of hiPSC-derived β cells.
To that end, research efforts have been geared towards the development of less damaging cryopreservation processes, including dissociation of hiPSC-derived β cells to single cells for freezing before reaggregation post-thawing, 185 encapsulation of β-cell organoids with cryoprotective hydrogel, 186 and vitrification. 187 However, these methods of cryopreservation remain relatively untested, with a lack of comprehensive data on the viability and functionality of these hiPSCderived β cells. Hence, with the current state-of-the-art, hiPSCderived β cells are likely to be made on demand, with an optimized supply chain procedure in place for near-immediate transplantation.
The second challenge relates to the immunological rejection that obstructs successful allogenic transplantation. Although immunosuppression drugs can be taken, these medications are life-long and potentially toxic, which diminishes the benefits of hiPSC-derived β-cell therapy. 188,189 In lieu of that, encapsulation strategies are increasingly being explored to protect the cells from the host immune system and to prevent immune reactions against the encapsulated β cells. [190][191][192][193] While traditional implantable devices are typically constructed with materials such as silicon and titanium, concerns with nutrient diffusion have pushed researchers to adopt other capsule materials such as alginate and polyethylene glycol. 192,193 However, as exemplified by ViaCyte's clinical trial, foreign body reactions may still occur, leading to fibrosis around the devices and affecting the viability of the encapsulated cells. 191 In addition, besides functionality issues, other regulatory concerns regarding biocompatibility, sterility and functionality of these accompanying medical devices will need to be addressed during clinical trials and in product release testing. Depending on the site of transplantation, increasingly stringent regulations may also be imposed. Hence, efforts geared towards the derivation of hypoimmunogenic hiPSC lines for eventual directed differentiation, through deletion of HLA proteins or overexpression of PDL1-CTLA4Ig molecules which can modulate Tcell activation, may deliver greater promise for the transplantation of hiPSC-derived β cells. 194,195 In addition, given the novelty and complexity of cell therapy products for clinical treatment, regulatory authorities may also be hesitant to define the potency testing(s) required and implement rules for QC checkpoints. Here, we have attempted to outline suggested potency and QC testing based on current standards. However, these proposed tests will still be subjected to regulatory oversight. Without additional clear guidelines, bench-to-clinic translation may be hindered. To facilitate the commercialisation of regenerative medicine products, collaborations with expert panels to establish guidelines for preclinical and clinical testing will be helpful. Harmonization of standards will also be useful to facilitate a wider adoption of hiPSC-based cell therapy.
Last but not least, for manufacturers focused on cell therapy, the considerations for personnel involved may be different from that of an academic setting, where diverse interdisciplinary collaboration will be required early in the development of the product (Figure 1). For instance, the expertise of process engineers will be needed for developing scale-up manufacturing and optimization of encapsulation devices whereas stem cell biologists will be needed to scrutinize cell culture reagents and growth factors used in the hiPSC culture, expansion, and differentiation processes. In addition, individuals with the legal and regulatory expertise will also be required to facilitate regulatory compliance with the jurisdiction of various countries. If encapsulation devices are to be incorporated into the final product, under EU regulations, a person responsible for regulatory compliance will also need to be appointed. 196 Close partnership with clinicians and hospitals will certainly need to be established for widespread data collection on patient safety to conduct appropriate risk management and monitoring. Hence, other than a regulatory framework that needs to be established within the company, careful planning of the required various job roles should be carried out as early as possible.
In conclusion, it is now possible to generate clinically compliant hPSC-derived β cells with higher yield and better functionality in vitro.
Further harmonization of the regulation of hPSC-based cell products and even optimisation of cryopreservation methods should be made to pave the way for its upcoming clinical translation and commercialisation. For industries focused on hPSC-based cell therapy, the establishment of a complete team that can address all the considerations listed above will be necessary in order to progress towards clinical trials. By surmounting these obstacles, it is highly probable that hiPSC-derived β cells have a chance of being a curative treatment for diabetes patients.

ACKNOWLEDGEMENTS
The authors thank members of the Teo Laboratory for contributing com-

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.