Human pancreatic islet transplantation offers a functional source of β cells for the treatment of diabetes, especially in a subset of patients with T1D who are prone to hypoglycemic unawareness and experience severe hypoglycemia despite optimal management of glycemic levels (11, 12, 19, 20). These have far-reaching benefits beyond physiological changes such as improvement in patients’ mental health, relief for caregivers, resumption of work productivity and reduced ambulance conveyance and emergency care needed. The success of the treatment was made possible due to the seminal research by Shapiro and team, who developed the Edmonton Protocol, building on previous achievements by others (for a detailed review on the history of clinical islet transplantation please see (21)). The procedure recommends transplantation of a cumulative islet mass of at least 10,000 IEQ per kilogram of the recipient’s body weight. This typically required at least two infusions from different donor material, unless insulin independence was achieved with a single infusion (11, 20). Islets are transplanted into the hepatic portal vein, the current clinical gold standard route, avoiding the need for surgery. The procedure allows the islet cells to access the circulatory system and facilitate glucose sensing and insulin release into the bloodstream, effectively restoring glycemic control in patients. Repeated islet infusions do carry the risk of procedure-related bleeding arising from elevated intraportal vein pressure and portal vein thrombosis (22). The protocol had set a standard for the infusion of an adequate islet mass combined with a glucocorticoid-free immunosuppressive regimen (e.g tacrolimus and sirolimus) (20). In recent practices, daclizumab (non-depleting monoclonal anti-interleukin-2 receptor antibody) and/or anti-thymocyte globulin is administered as pre-procedural induction immunosuppression, whereas low-dose tacrolimus (calcineurin inhibitor) in combination with mycophenolate mofetil or sirolimus is prescribed for maintenance immunosuppression (23). Sirolimus (mTOR inhibitor) has been found to be more poorly tolerated by patients with adverse side effects, hence its exclusion may result in improved longer-term outcomes. Likewise, although tacrolimus-based immunosuppression is effective against allo- and auto-immune rejection, its side effects include nephrotoxicity and diabetogenicity due to effects on the islet cells. Additionally, other anti-inflammatory agents are needed in the peri-transplant periods to counter proinflammatory cytokines and preserve islet function, such as etanercept (TNFα blocker) and anakinra (IL-1 receptor antagonist) which were found to be associated with improved clinical outcomes as compared to regimens without the use of anti-inflammatory agents (12, 24). Thus, ongoing research efforts remain important to define immunosuppressant and anti-inflammatory drug combinations with better safety profiles while remaining effective for preserving islet graft function.

Islet transplantation has been and will continue to be a life-changing therapy as it has resulted in positive outcomes for patients including insulin independence, glycemic control, freedom from SHEs and restoration of hypoglycemia awareness. These outcomes are positively correlated with graft survival and function (fasting C-peptide >0.1 nmol/L post-transplantation) and achievement of HbA1c level of <7.0% (53 mmol/mol) at least 1 year post-transplant (12, 19, 25). Furthermore, despite the long period under an immunosuppressive regimen, sustained islet function in those with sustained graft survival is possible, though the incidence of cancer appeared to be higher (12, 19, 25). In the recently FDA-approved donor-derived pancreatic islet cell therapy for T1D, known as Lantidra or donislecel (manufactured by CellTrans Inc.), the therapy is indicated for adults with T1D where exogenous administration of insulin is insufficient to maintain the HbA1c target and who experience hypoglycemia unawareness (26, 27). FDA approval was based on experiences from Phase I/II clinical trials that demonstrated graft survival in all 10 patients and insulin independence maintained in 60% of patients 5 years post-transplant, as well as in another Phase III clinical trial revealing that all 21 patients were free from hypoglycemic episodes and most maintained HbA1c levels at ≤ 6.5% at a 1-year follow up (28, 29). Importantly, no significant side effects were reported for the cell therapy except for procedural-related bleeding. The approval represents a positive step forward for T1D management in the US, though there remained controversy over the recognition of islets as drugs instead of organs that may place a limitation over patient access.

Islet transplantation has proven to be a promising curative approach for T1D patients, both saving lives and improving quality of life. However, there remain several challenges hindering its widespread clinical utility, particularly the need for chronic immunosuppressive therapy and limited donor islet availability ( Figure 1 ).

Firstly, as islet transplant patients are required to undergo intensive life-long immunosuppression to prevent graft rejection and loss of islet function, the selection of immunosuppressants used may induce side effects or autoimmunity recurrence, which will influence islet transplantation outcome. For instance, induction of immunosuppression with anti-thymocyte globulin as compared to daclizumab, and maintenance of immunosuppression with tacrolimus as compared to sirolimus, has been shown to increase risk of autoantibody recurrence in islet transplantations (30). This study highlighted the “off-target” effects of immunosuppressants, particularly how immunosuppressants influence the profile of regulatory T cells (Tregs), which are an important subset of immunomodulatory T cells responsible for promoting immune tolerance. Immunosuppressants that foster a richer Tregs environment could drive tolerance and further minimize the need for immunosuppression (31). Previous studies investigating the impact of immune-modulatory drugs on the function of Tregs showed that sirolimus has a Tregs-favoring effect as compared to tacrolimus (32). Greater clarity on the immunological mechanisms mediated by the immunosuppressants would guide future directions in preserving Tregs numbers and function for better success in balancing immunosuppression and transplant outcome.

Chronic immunosuppression has also been associated with other detrimental side effects, such as β cell toxicity, kidney toxicity, higher risk of cancer and opportunistic infections as the protective function of the patient’s immune response is jeopardized. Reported symptoms experienced by patients include anaemia, nausea, fatigue, diarrhoea, and abdominal pain, though the incidents varied amongst patients and depended on the number of islet infusions and length of follow-up (33, 34). Common adverse effects occurred in between 20% to 90% of patients from initial infusion up to 1 year following final infusion, that are mostly related to the infusion procedure and immunosuppressive regimen administered (34). In the event of a life-threatening infection or cancer requiring discontinuation of immunosuppressive medications, there was eventual loss of islet cell function and resumption of insulin dependence (34). Additionally, patients who are contraindicated for immunosuppression, such as those with relevant drug allergies or who are highly susceptible to acute or chronic infections, are not eligible for islet transplant. T1D patients often possess pre-existing renal impairment due to longstanding diabetes and their renal dysfunction may be exacerbated by tacrolimus which can result in calcineurin-induced nephrotoxicity (35, 36). In addition to nephrotoxicity, tacrolimus is further associated with gastrointestinal side effects leading to episodic diarrhoea. Sirolimus has also been linked to several side effects including mouth ulcerations, neutropenia, dyslipidemia, small bowel ulceration, peripheral edema, and the development of ovarian cysts in females (37). The lifetime risk of lymphoma is estimated to be 1-2% in transplant recipients undergoing long-term immunosuppression, with the most common malignancies being non-melanomatous skin cancers (38). At supratherapeutic levels, tacrolimus and sirolimus have also been associated with human islet toxicity caused by increased amyloid deposition and disrupted insulin granule formation, though the detrimental effects on the β cells may be reversible upon withdrawal of drug treatment (39). For all these reasons, patients must be screened for endogenous infections or pre-existing medical conditions that can be aggravated following immunosuppressant therapy, and the risks weighed alongside the benefits (non-recurring severe hypoglycemia and achieved target HbA1c) to patients. Hence, islet transplantation has only been considered for T1D patients complicated by IAH and SHEs, when other lines of treatment have failed to prevent life-threatening SHEs, placing a significant limitation on the large pool of diabetes patients who could benefit from islet cell replacement therapy.

A second major factor limiting the accessibility of islet transplantation is the lack of sufficient cadaveric donor islets to meet the global demands for human islet cell replacement. More than 2,000 patients have received allogeneic islet transplantation globally since year 2000 (10). This is only a small fraction of the millions of individuals who have been afflicted with brittle T1D, not to mention an even larger number of patients with insulin-deficient T2D for which such a treatment is currently not an option. A recent report published by National Health Service (NHS) England stated that the waiting time for islet transplant was 631 days and the number of patients on the active islet transplant list in the UK is 29 by the end of September 2021 (40). In addition to the lack of suitable deceased donor pancreases, the quality of islet products is highly variable depending on the circumstances under which the donor organ is obtained, donor characteristics, the complex islet isolation and culturing process and the preservation conditions before transplant (41). The method of isolating islets by enzymatic digestion and mechanical separation can lead to potential damage of the endocrine cells. Majority of islet transplant recipients receive islets from multiple donors (2 to 4) as up to 60% of islet mass is lost within the first few days following islet infusion. While transplantation of larger islet mass (>11,000 IEQ/kg of recipient weight) over multiple islet infusions contributes to a larger mass of surviving β cells, this limits the number of patients that can receive the islet allografts (42). Efforts have been made across several islet isolation facilities to harmonize the donor selection criteria, manufacturing procedures, and lot release attributes, but this remains a huge undertaking to be controlled at all phases and implemented at a larger scale (43, 44).

Xenogeneic islets, in particular porcine islets, have been explored as an alternative primary cell source to supplement the supply of primary human islets for transplantation. Porcine islets are more readily available and possess functional characteristics that make them a suitable substitute for human islets. They have weaker immunogenicity and porcine insulin is structurally similar to human insulin (with one amino acid difference that is alanine in pigs and threonine in humans) (45). Major hurdles need to be overcome for xenografts to be a feasible alternative in the clinic. These include physiological incompatibility and immunological reaction to non-human donor tissues that trigger both innate and adaptive barriers of the immune system, resulting in rejection. In addition, xenotransplantation presents the potential risk of zoonosis and porcine endogenous retrovirus (PERVs). As such, previous studies have evaluated the feasibility of xenotransplantation of porcine islets into non-human primates (NHPs). One study showed that an anti-CD40 (2C10R4) monoclonal antibody-based immunosuppressive regimen together with tacrolimus was effective in circumventing graft rejection and prolonging porcine islet graft survival in diabetic rhesus monkeys, with median survival (serum porcine C-peptide concentration of >0.15 ng/mL) of 60 days. All monkeys also received anti-thymocyte globulin, cobra venom factor (CVF), adalimumab, and sirolimus (46). In another study, a newly engineered anti-CD40L-specific monoclonal antibody AT-1501 was tested in a cynomolgus macaque model that had undergone intrahepatic islet allotransplantation. The study showed that AT-1501 enabled long-term graft survival with higher C-peptide levels detected compared with conventional immunosuppression (47). AT-1501 was modified to minimize risk of thromboembolic complications that were previously reported for CD40L-based therapies in clinical trials, and therefore appears to be a promising and safe agent for further testing. Another strategy to enhance graft survivability is to utilize genetically-modified pigs with alterations in expression of known xeno-antigens, and modification of the complement and coagulation systems to improve immunological compatibility between pigs and NHPs (48). In one example, cardiac xenografts from genetically-modified pigs with alpha 1-3 galactosyltransferase gene knockout, expression of human complement regulatory protein CD46 and human thrombomodulin, were transplanted into baboons (49). The pre-transplant immunomodulatory induction regimen included anti-thymocyte globulin and 2C10R4 antibody, followed by maintenance with intensively-dosed 2C10R4 antibody and mycophenolate mofetil (49). This combination of genetic modifications and immunosuppressive regimen resulted in sustained survival of the xenografts with median of 298 days up to the longest of 945 days observed (49).

In further attempts to reduce immune rejection after xenogeneic islet transplantation, porcine islets may be encapsulated in a protective layer to avoid immune cell recognition. In one study, neonatal porcine islets were encapsulated in a stable and permeable alginate gel and enclosed in a biocompatible, immunoprotective membrane, and transplanted in the abdominal cavities of immunocompetent diabetic mice. Islet xenograft survival, rapid lowering of blood glucose and long-term glycemic control for >200 days was achieved without any immunosuppressants (50). Furthermore, the devices were shown to retain their integrity after they were retrieved and re-transplanted in new immunocompetent diabetic mice. In a clinical study, alginate-based encapsulation of neonatal porcine islets were transplanted into the peritoneal cavity of eight T1D patients without immunosuppression at up to 20,000 IEQ/kg body weight over two separate transplantations (51). The procedure was shown to be safe with no PERVs infection detected. Some fibrosis of the microcapsules were observed post-transplant, however long-term efficacy was shown with HbA1c <7% over more than 600 days and significant reduction of serious unaware hypoglycemia (51). More encapsulation studies involving porcine islets are discussed in a later section and in Table 1 , which lend support to the clinical benefit provided by porcine islet xenotransplantation in T1D patients.

Besides the need to address the genetic and molecular discrepancies between human recipients and xeno-organs, other challenges to note include psychosocial and ethical barriers, tension from religious beliefs, concerns for animal welfare and the use of animals for research. Nonetheless, the careful use of existing or novel immunosuppressive therapies, development of genetically-modified pigs to obtain porcine islets with better immune tolerance, and use of encapsulation to provide immune protection (to be discussed in greater detail in section 3.1) make it possible for porcine islets to be considered as another safe, functional and readily available source of primary cells for T1D patients. This will help to overcome the ongoing shortage of donor human islets.

To protect from allograft rejection and recurrence of autoimmunity, cell encapsulation is a common strategy that can provide a physical barrier to shield the islets from immune cell recognition and attack. Specifically, encapsulation helps to mediate IBMIR which is caused by islet contact with the blood and is highly detrimental to cell survivability. An ideal encapsulation device or material should be biocompatible but not biodegradable, made of a semi-permeable material to allow entry of nutrients and oxygen, and enable release of hormones and metabolic by-products into the bloodstream. Such a strategy can be based on microencapsulation or macroencapsulation ( Table 1 ). Microencapsulation is a method in which islet cell clusters are individually encapsulated in spherical capsules typically 300 μm to 600 μm in diameter (67), and commonly in alginate-based hydrogels. In contrast, macroencapsulation is based on devices greater than 100 mm with capacity to house a larger mass of islet cells within the membrane. Micro-capsules have an optimal surface-to-volume ratio compared to macro-capsules that require vascularization and sufficient oxygenation to improve islet survival and function. Biomaterials used for microencapsulation are often made of natural polymers such as alginate (55), agarose (68), chitosan (52) but also synthetic polymers such as polyethylene glycol (PEG) that form a hydrogel (50, 54, 56, 69) ( Table 1 ). Microencapsulation encompasses a semi-permeable membrane that has demonstrated some success in providing immune protection and mechanical stability in mice (50, 52), NHP (55) models and human studies (51, 53). However, microencapsulation requires more complex and individualized fabrication processes (70), as opposed to macroencapsulation devices that may be easier to manufacture, are more easily retrievable after implantation, and are more favourable for commercialization. Several devices that have been developed include Theracyte™ from TheraCyte Inc., βAir from BetaO₂ Technologies, the Cell Pouch System from Sernova, and PEC-Encap (VC-01) and PEC-Direct (VC-02) from ViaCyte (now acquired by Vertex Pharmaceuticals) ( Table 1 ).

The Theracyte™ planar macroencapsulation device consists of an inner hydrogel semi-permeable membrane layer for immune protection and an outer polytetrafluoroethlene membrane layer for neovascularization (71). Rodent islets encapsulated within the Theracyte™ device demonstrated survival and functionality in an immunized rat model for at least 6 months post-transplantation (59). Porcine islets transplanted subcutaneously within Theracyte™ also survived and was able to reverse diabetes up to 8 and 16 weeks in cynomolgus monkeys and NOD mice respectively (72). Another device known as the βAir bioartificial pancreas (BAP) has been tested clinically and consists of two compartments for islets, an oxygenated chamber that maintains physiological oxygen pressure, covered with a porous polytetrafluoroethylene (PTFE) membrane impregnated with alginate to provide the immunoprotective barrier (73–75). Stable graft function and insulin secretion were observed in NHP models of diabetes (73) and human patients (75) who received βAir containing porcine islets and human islets respectively, both in the absence of immunosuppressants, though complete insulin independence was not achieved. In one Phase I clinical study, βAir containing allogeneic human islets were subcutaneously implanted in T1D patients (clinicaltrials.gov: NCT02064309) (63). Although the transplanted islets survived 6 months post-transplantation, limited functionality was observed based on minute levels of circulating C-peptide with no impact on glycemic control in the patients (63). Additionally, fibrosis or inflammation were observed on the surface of the chamber. Another promising technology is the proprietary Cell Pouch system developed by Sernova, which is an implantable device that provides a vascularized tissue matrix for cells in addition to local microencapsulation of cells in polymer spheres. The Cell Pouch is undergoing testing in T1D patients in an ongoing Phase I/II clinical trial (clinicaltrials.gov: NCT03513939). In an interim update, patients with islet transplants in the 8-channel Cell Pouches were found to achieve insulin independence for as long as 3 years as a result of both functional islet grafts in the Cell Pouches supplemented by a modest intraportal islet transplant top-up through the portal vein (66). Additionally, a second version of the Cell Pouch with higher capacity is being evaluated and early patient data so far revealed persistent serum C-peptide levels detected from a single islet transplant in the 10-channel Cell Pouch (66). The company announced a collaboration with Evotec to test out human induced pluripotent stem cell (hiPSC)-derived islet cells in the Cell Pouch system in future clinical trials (66).

In another Phase I/II clinical trial by ViaCyte, human embryonic stem cell (hESC)-derived pancreatic endoderm cells (PECs) were encapsulated in a cell-impermeable device designed to be immunoprotective against recipient immune systems (clinicaltrials.gov: NCT02239354). The macroencapsulation devices containing cells (also known as VC-01 or “PEC-Encap”) were implanted subcutaneously in T1D patients in the absence of immunosuppression (62, 76), and evaluated for efficacy, tolerability, and safety. Formation of neovasculature was observed in the grafts and the PECs were able to mature in vivo into insulin-expressing β cells, as shown by immunohistochemical staining for pancreatic islet cell markers (NKX6.1, insulin, and glucagon) (62). VC-01 was found to be safe, well-tolerated and immunoprotective with evidence of prolonged cell survival up to 24 months. However, some inconsistency of cell survival was observed amongst subjects due to varying foreign body responses in the host (62, 76). Furthermore, no evidence of insulin secretion was found due to chronic damage to islets resulting from device fibrosis (62). This suggested that the macroencapsulation device, although well-tolerated in recipients, resulted in poor long-term engraftment and diminished efficacy due to poor oxygenation and nutrient supply to the transplanted cells. In efforts to mitigate cell loss due to device fibrosis, a subsequent version of the combination product (also known as VC-02 or “PEC-Direct”) was developed to include engineered portals in the device to enable direct capillary permeability and facilitate better vascularization to the implanted cells (clinicaltrials.gov: NCT03163511) (76). However, this was a non-immunoprotective device and patients still required immunosuppression to limit allo- and autoimmune responses. This time, patients exhibited meal-stimulated C-peptide secretion following maturation of the PECs in vivo and achieved the target blood glucose range for longer periods (26 weeks) as compared to VC-01 (76). Subjects in which substantial cell engraftment were observed after evaluating the explants were shown to have higher meal-responsive C-peptide levels during the follow-up period as compared to those with poor cell engraftment (64). While VC-02 is generally safe and well-tolerated, the side effects of immunosuppression accounted for majority of adverse events (AEs). Another study utilizing the same VC-02 device but with an optimized membrane perforation increased the initial implanted cell dose (14 x 10⁶ cells per kg body weight) such that it is within the range of that used for intrahepatic primary islet transplants (6-18x 10⁶ cells per kg body weight) (20, 77, 78). After 6 months post-transplantation, only 3 of 10 patients achieved C-peptide levels ≥0.1 nmol/L with reduced insulin dependence, and the detectable β cell mass in the retrieved implants was found to be less than 5% of the initial cell mass, indicating high cell loss and limited efficacy from the device-delivered PECs (78). This could be due to insufficient vascularization in the devices to support the metabolically functional β cell mass. Further optimization remains needed to increase the efficacy of the encapsulated PECs to be comparable to that of conventional human primary islet transplantation. These outcomes suggest that the macroencapsulation devices not only need to prevent entry of immune cells, but also facilitate (even encourage) vascularization to enable better cell survival and maturation into functional β cells and reduce infiltration of fibroblasts into the devices.

In another recent effort to evaluate macroencapsulated stem cell-derived islets in the absence of immunosuppressants, Vertex Pharmaceuticals, who has an ongoing first-in-human Phase I/II clinical trial (clinicaltrials.gov: NCT04786262) for their allogeneic stem cell-derived, fully differentiated islet cells (79), revealed the development of their second cell therapy program (clinicaltrials.gov: NCT05791201) investigating the islet cell product encapsulated in a “channel array” device and implanted subcutaneously (65). Though the design of the device used in the trial has not been disclosed, based on publicly-available patent filing information from the company, such a device would have a thickness of at least 300μm, an average pore size ranging from 5 nm to 2500 nm and comprising 1 x 10⁶ to 1 x 10⁹ (PCT/US2018/053665). The proprietary design also showed deformation of the membrane to a formed configuration instead of a flat configuration, with channels that enable vascularization in and around the device. The pore size had to be fine-tuned to ensure long-term structural integrity while allowing release of insulin and restricting leakage of cells out of the device (PCT/US2018/053665 and PCT/US2018/037637). It remains to be seen whether the device allows sufficient nutrient and oxygen supply, as well as provide immune tolerance in the absence of immunosuppression.

Across the numerous efforts from academic labs and commercial companies to develop and test macroencapsulation devices for islet cell transplant ( Table 1 ), prevention of immune attack is found to be achievable, but a balance needs to be struck to achieve other outcomes including better vascularization and oxygenation, maximising transplanted β cell mass, preservation of β cell viability and function following implantation, and reduction of foreign body reactivity. These devices also require additional unique considerations related to manufacturing and regulatory oversight as medical devices for use in a clinical setting, evaluation of the biocompatibility of the materials, and selection of transplantation site and protocol given the larger size of the devices.

Since the use of hiPSC technologies became widespread, hiPSCs have proven to be highly versatile and amenable to genetic manipulation. The generation of functional hiPSC-derived islets (SC-islets) has also made significant headway in recent years, making it possible for regenerative medicine to be part of a not-so-distant future in diabetes therapy. The journey of developing SC-islets in the lab to be as close to their primary human islet counterparts as possible, and the promise of using these cells as off-the-shelf therapy for islet cell replacement, have been extensively discussed in other recent reviews (80–82). Previously, promising preliminary clinical results had been released from Vertex Pharmaceuticals on their ongoing first-in-human Phase I/II clinical trial of lead candidate VX-880 (clinicaltrials.gov: NCT04786262), which is a hiPSC-derived, fully differentiated islet cell product administered to T1D patients in a similar fashion as primary human islets, in the presence of intensive chronic immunosuppressive therapy (83). Six patients with a history of undetected insulin secretion tolerated the therapy well, demonstrated islet cell engraftment with production of endogenous glucose-stimulated insulin and had improved glycemic control. Patients that were followed up at the 1-year mark also displayed successful elimination of SHEs and reduction in HbA1c <7.0% (83). This was a landmark shift from ViaCyte’s PEC grafts, which required cellular maturation in vivo into functional glucose-sensing and insulin-secreting β cells, a process that cannot be monitored and qualified before transplant.

In combining SC-islet differentiation protocols and genome engineering techniques (84–89), several novel approaches have been employed to generate human islet cells that are protected from immune rejection, potentially eliminating, or reducing the need for systemic immune suppression and/or encapsulation. These immune evasive strategies typically work by either artificially elevating immune suppressive proteins (e.g. immune checkpoint manipulation) or removing receptors important for immune cell recognition on the cell surface ( Figure 3 ). An essential component of innate and adaptive immune responses is the major histocompatibility complex (MHC) class I and II molecules which serve to present foreign antigens to the cell surface for recognition by the host immune system. In humans, these MHC molecules, also known as human leukocyte antigens (HLA), are highly polymorphic with almost 10,000 alleles. Immune rejection of hiPSC-derived cells or tissues from an allogeneic donor are mediated through these MHC molecules, limiting the survivability and therapeutic potential of the transplanted cells. Immunological mechanisms governing allograft rejection occurs in two stages: (1) non-specific innate responses predominate in the early phase, and (2) antigen-specific adaptive responses by antigen presenting cells (APCs) and dendritic cells (DCs) that result in recognition of donor antigens by host T cells (90). Both innate and adaptive immunity contribute to acute or chronic graft rejection.

To prevent innate immune rejection and further suppress adaptive immune responses, various groups have developed genetically engineered hypoimmunogenic hiPSCs or hESCs through modification of selected HLA genes and other immunomodulatory factors, and evaluated the ability of these pluripotent stem cells and their derivatives to escape immune recognition (91–93). Table 2 provides a summary of key studies that developed hypoimmunogenic cells and evaluated the ability of the cells to evade the host immune response in both in vitro and in vivo assays.

The potential application of hypoimmunogenic hPSC-derived islet cells for cell replacement therapy has been demonstrated in several pre-clinical studies in immunocompetent diabetic animal models (85–89). Simultaneous deletion of beta-2-microglobulin (B2M), a component of MHC class I-encoded HLA-A/B/C molecules, and MHC class II transactivator CIITA in hiPSCs (93) and primary human islets (85) resulted in ablation of cytotoxic CD8⁺ and CD4⁺ helper T cell responses. When evaluated for immune tolerance in vitro, B2M and CIITA knockout hESC- and hiPSC-derivatives were co-cultured with T cells, NK cells and PBMCs and were found to be resistant to T cell, NK cell, and complement-dependent cytotoxicity and macrophage engulfment (86–88, 93, 94) ( Figure 3 ). The expression of the non-classical MHC molecules, HLA-E and HLA-G, were also found to contribute to establishing an immunosuppressive microenvironment by binding the inhibitory NK cell receptors CD94/NKG2A and facilitating the escape of human tumors from the host immune response (95). A previous study showed that HLA-A/B/C knockout hESCs and their differentiated CD45+ cells and RPE cells that overexpress HLA-E are resistant to CD8+ T cell cytotoxicity and NK cell-mediated lysis in both in vitro and in vivo models (92).

Another immunomodulatory effector is CD47 which acts as an anti-phagocytic ligand to inhibit activation of the innate immune system (96). CRISPR/Cas9-mediated knockout of B2M and CIITA together with lentiviral transduction-based overexpression of CD47 in mouse and human iPSCs were effective in generating hypoimmunogenic derivatives that did not trigger NK cell activation in vitro (91). Loss of immunogenicity was also recapitulated in vivo, as observed by significant improvement in graft survival post-injection of mouse iPSC- and hiPSC-derived smooth muscle cells (SMCs), endothelial cells (ECs) and cardiomyocytes (CMs) into the right thigh muscle of immunocompetent C57BL/6 mice (91). This approach was replicated using hiPSC-derived SMCs, ECs and CMs that were transplanted into humanized NSG-SGM3 mice. Humanized mice have been widely used as a pre-clinical in vivo model that recapitulates the human context, in this case the human immune system. NSG-SGM3 mice used in the study supports the stable engraftment of human myeloid lineages, regulatory T cell populations and production of hIL-15, thereby promoting the development and/or function of human NK cells. B2M^-/-/CIITA ^-/-/CD47⁺ hiPSC-derivatives showed sustained graft survival for more than 50 days, whereas unedited WT derivatives were rejected within 14 days (91).

A similar strategy was also validated in hypoimmune B2M^-/-/CIITA ^-/- and CD47-overexpressing primary human islets (85). WT and hypoimmune human islets injected intramuscularly in humanized immunocompetent mice were monitored using bioluminescence imaging. WT islets were fully rejected within 7 to 10 days, exerted no beneficial effect on glucose homeostasis and no detectable C-peptide secretion after 29 days. In contrast, mice injected with hypoimmune islets showed allograft survival and achieved glycemic control, indicating that the function of allogeneic hypoimmune islets was sustained and confirming the ability of B2M^-/-/CIITA ^-/-/CD47⁺ to modulate immunogenicity and escape immune attack (85). More recently, Hu et al. also reported the successful rescue of an immunocompetent, diabetic cynomolgus monkey with allogeneic, hypoimmune B2M^-/-/CIITA ^-/- and CD47-overexpressing primary rhesus macaque islets (89). C-peptide remained detectable in the monkey serum and insulin independence was achieved without immunosuppression for up to 6 months (89). These results show that hypoimmune islets can be protected from immune rejection while maintaining graft function in vivo.

Programmed death-ligand (PD-L1) is an immune checkpoint protein that has also been in the spotlight as it plays a role in suppression of adaptive immune response by inducing a co-inhibitory signal in activated T cells and promoting T cell apoptosis (97). Gerace et al. reported genetically engineered B2M-deficient hESCs with PD-L1 overexpression in addition to HLA-E overexpression. The authors found that in response to PBMC injection, WT SC-islets transplanted under the kidney capsule of diabetic humanized NSG-double knockout (hu-NSG-DKO) mice were destroyed within 2 weeks due to PBMC-mediated cytotoxicity, whereas graft rejection was delayed when B2M^-/- SC-islets were transplanted. At 7 weeks post-PBMC injection, B2M^-/- SC-islets gave rise to positive glucose-stimulated insulin secretion (GSIS) outcomes and were able to reverse diabetes in mice whereas graft function was lost in mice transplanted with WT SC-islets (86). These results suggested that removal of B2M could delay the rejection of the SC-islets, though it is possible that the grafts may eventually be completely rejected in longer term studies. The authors however showed that overexpression of PD-L1 in the B2M^-/- SC-islets did not protect the cells from xeno-rejection, and that overexpression of HLA-E did not provide additional protective benefit against NK cell cytotoxicity in their model. Instead, they found that SC-islets engineered to secrete tolerogenic cytokines such as IL-10 and TGF-β are protected against xeno-rejection likely due to recruitment of Tregs to induce a tolerogenic environment. The engineered SC-islets could reverse diabetes in NOD mice up to 8 weeks post-transplantation. Another group however showed that PD-L1 overexpression could create an immune-evasive microenvironment for SC-islets transplanted in immunocompetent diabetic mice, by restricting T cell activation and delaying graft rejection (88). While both SC-islets with and without PD-L1 overexpression had similar efficacy in restoring glycemic control in diabetic mice within a few days, the functionality of the islets lacking PD-L1 was quickly lost. On the other hand, islet cells overexpressing PD-L1 provided sustained blood glucose homeostasis, with human C-peptide levels correlating with glycemic control for more than 50 days (88).

While most of the reported work on the development of hypoimmune cells have been within pre-clinical settings, new efforts are now emerging to evaluate the cells in clinical studies. CRISPR Therapeutics (previously in conjunction with ViaCyte) is conducting first-in-human Phase I clinical trials with an investigational, allogeneic, gene-edited, hypoimmune stem cell-derived PECs for T1D (clinicaltrials.gov: NCT05210530, NCT05565248). The cells are also encapsulated in a device to be implanted in patients without immunosuppressive therapy. Vertex Pharmaceuticals also announced that it will license CRISPR Therapeutics’ gene-editing technology to add value to their ongoing efforts in the clinical development of iPSC-derived islet cell therapy for T1D (98). Although details of the partnership were not disclosed, the collaboration is likely to explore the development of hypoimmune, fully-differentiated iPSC-derived islet cells for transplantation into T1D patients without immunosuppression (with or without encapsulation). These studies aim to establish whether generation of universal hypoimmunogenic hPSCs differentiated into insulin-producing islets could provide long-term survival due to evasion of immune-mediated detection and killing. Positive outcomes from the trials will mean maximising the efficacy of the transplanted islet cell mass and providing a longer term, immunosuppression-free curative therapy for allogeneic recipients.

Despite the attractiveness of genome-edited hypoimmune cells as a cell source for allogeneic cell therapy, the long-term safety and efficacy remains to be ascertained as most studies are currently conducted in vitro or in animal models. As hypoimmune cells can escape immune detection, this raises concerns on cell malignancy, especially for hPSC derivatives which may give rise to tumour formation in the presence of any residual hPSCs or incompletely differentiated cells in the graft. Furthermore, CRISPR-based genome editing may induce unintended off-target genomic mutations that may contribute to aberrant gene expression that may contribute to malignancy (99–101). Therefore, tumorigenicity tests as well as evaluation of the genomic and epigenomic stability of modified cell lines remain essential to qualify any hPSC-based cell product for clinical applications. New generations of gene-editing tools have also been developed to improve on the design of nucleases, repair templates and analysis of potential off-target editing to reduce tumorigenicity risk and unintended outcomes (102–104). For instance, expression systems containing suicide gene constructs can potentially eliminate any tumorigenic cells that arise, to safeguard against tumour formation in hPSC-based cell therapies (105). A recent study showed that a combination of immune-cloaked mouse ESCs in which several immunomodulatory transgenes are being expressed, coupled with a genomically integrated FailSafe suicide transgene system, was able to generate various ESC-derived tissues that possess immune privilege. Allogeneic cells transplanted in these ‘artificially-created’, immune-privileged sites could be protected from rejection for months (106). The FailSafe system is a patented technology which creates a transcriptional link between the suicide herpes simplex virus thymidine kinase gene (HSV-TK) and a cell division gene (CDK1) to enable killing of any undesired dividing cells using a pro-drug treatment (106–108). Such safeguards help to improve the safety profile of cell therapy products, particularly those that are engineered to be more immune tolerant. For similar safety reasons, having the cells implanted within an encapsulation device also facilitates easy removal of the cells in case there is a need for the graft to be excised.

Another strategy for transplanted grafts to be potentially shielded from the immune system involves the use of naturally immune privileged stem cell sources. Several tissues in the body are evolutionarily adapted to be protected from inflammatory immune responses, including extra-embryonic tissues such as the amnion, placenta and umbilical cord. These tissues possess immune privileges so that maternal tolerance toward fetal cells may be maintained. Specifically, these tissues contain stem cells such as umbilical cord lining mesenchymal stromal cells (CL-MSCs), amniotic MSCs and placenta-derived trophoblast stem cells (TSCs). These cells also represent valuable cell sources for the generation of hiPSCs. It is postulated that hiPSCs derived from immune privileged cells may retain some of the same genetic signatures and epigenetic memory. Whether the differentiated cells from these hiPSCs also maintain their ‘privileged’ status however remains to be tested in a cell type-specific manner. Accumulating evidence suggests that the immunogenicity of hiPSC-derived cells are cell type-dependent, as different cell types exhibit different immunomodulatory mechanisms. Retaining at least some extent of the immune privileges of the original tissue stem cells may help in resisting immune destruction in the event of allogeneic transplantation.

MSCs are multipotent stem cells with high proliferative capacity, low immunogenicity and immune modulation properties due to the expression of tolerogenic factors. Successful reprogramming of placental amniotic membrane MSCs and amniotic fluid stem cells at high efficiency has previously been shown (109–111). The hiPSCs retained the immunomodulatory signatures of the MSCs, such as absence of expression of MHC class I and II proteins, and expression of HLA-G and CD59 (109). Umbilical cord lining epithelial cells (CL-ECs) are another population of cells that do not express MHC class II molecules and co-stimulatory molecules, and express non-classical HLA-E and -G, that function to suppress maternal T cell and NK cell responses (112). Therefore they not only have low immunogenicity but also possess some immunosuppressive capacity (113, 114). In a recent study, CL-ECs differentiated into retinal pigment epithelial (RPE) cells and transplanted into mice and monkey models were found to elicit reduced pro-inflammatory responses and immune cell infiltration compared to transplanted RPE cells differentiated from skin-derived hiPSCs (115). There have been few published reports, if any, exploring differences in immune tolerance of islet cells derived from hiPSCs reprogrammed from different cell sources, representing a gap that warrants further study.

Besides using hiPSCs as the starting source of cells for differentiation into insulin-secreting islet cells, MSCs have also been directly differentiated into islet-like cells by genetic manipulation (116) or step-wise induction using specific medium and small molecules in vitro (117–125). Umbilical cord-derived MSCs (UC-MSCs) are attractive as a starting material as they can be obtained through pain-free and non-invasive methods, are available in abundance, and have high proliferation and differentiation capacity (118). Previous studies demonstrated that UC-MSCs do not induce allogeneic PBMC immune responses and can suppress the function of mature dendritic cells in vitro (113). UC-MSC-derived islet-like cell clusters also retained their immune privileged properties in vivo and were capable of regulating glucose homeostasis (118). Primitive stromal cells isolated from the umbilical cord Wharton’s jelly have also been differentiated directly into insulin-secreting islet-like cell clusters that express beta cell markers C-peptide and PDX1, and higher levels of secreted insulin compared to bone marrow-derived MSCs (126). Such an approach is however less widely adopted than hiPSC-based differentiation into islet cells, in part due to the lack of reproducibility in the functionality of the β-like cells derived using these methods, and the fact that MSCs experience replicative senescence, which will limit the ability to continuously generate differentiated cells at larger scale. Another immune privileged cell type of note are TSCs, which are a unique population of stem cells derived from the placenta that are fetal in origin, and that form the interface between the fetus and mother throughout pregnancy (127). They are immune tolerant as they have little to no expression of the classical MHC molecules and may also be differentiated into the different germ layers. Therefore, TSCs are another understudied cell source for allogeneic cell therapy, and attempts to derive islet cells directly from TSCs have yet to be reported.

In conclusion, the use of immune privileged stem cells as alternative cell sources for hiPSC generation or for direct differentiation into islet cells could be another strategy to eliminate or reduce the intensity of immunosuppressive therapy without the need for genome editing. Nonetheless, whether the differentiated cells retain their immune privilege, and to what extent, would be crucial to ascertain in a cell type-specific manner.

It is known that the predominant mediator of allograft rejection is HLA mismatch triggering T cell-mediated rejection. HLA molecules found on the surface of most cells have an important role in enabling the immune system to recognise “self” versus “non-self” antigens (128). The MHC system in humans consists of the classical MHC class Ia (HLA-A, -B, -C), non-classical MHC class Ib (such as HLA-E, -F, -G) and MHC class II (HLA-DR, -DQ, -DM, -DP) molecules that are involved in antigen presentation to CD8⁺ T cells (129), natural killer cells (NK cells) (130), and CD4⁺ T cells (131). HLA matching is not currently a criterion for primary human islet transplantation, however retrospective studies have showed that matching at selected loci, particularly HLA-DR and HLA-B, could improve long term islet allograft survival (132, 133), which would further lead to more prolonged insulin independence in patients. Another study involving follow-up of pancreas transplants purported that the number of HLA-DR and HLA-B matches correlated with a reduction in acute graft rejection, though there was no evidence to suggest a similar correlation with graft or patient survival rate (134). In other studies featuring other tissue transplantations, HLA matching has been shown to result in reduced allogeneic immunogenicity, increased graft survival, and therefore potential reduction in the intensity of immunosuppression required (135–138). Although HLA typing to match donor and recipient antigens at selected loci is clinically feasible, incorporating HLA haplotype matching in primary islet transplantation remains challenging due to existing pressure from limited cadaveric donors and the need for islet infusions from multiple donors. Thus, an avenue that may be explored relates to the formation of a repository of hiPSCs carrying different HLA haplotypes that may be matched to many recipients within a given population. Such a repository would be made up of clinical-grade HLA homozygous hiPSCs derived from carefully selected donors with homozygous HLA types to enable HLA matching when the need for an allogeneic transplantation arises (139). Based on prior experiences from cord blood and kidney grafting studies, HLA-A, -B, and -DR have been indicated as the most important HLA loci to match for long-term graft survival, with or without immunosuppressive drugs (140–142).

Several groups across different countries have embarked on efforts to derive repositories of HLA-homozygous hiPSCs that capture the high frequency HLA haplotype backgrounds (most typically for HLA-A, -B and -DRB1) in their population. The Center for iPS Cell research and Application (CiRA) of Kyoto University runs an iPSC Stock Project that aims to establish an HLA-homozygous iPSC haplobank for most of the Japanese population (143, 144). They recently reported a clinical-grade iPSC haplobank consisting of 27 iPSC lines from seven HLA-homozygous donors that could cover 40% of the Japanese population (145). Generation of HLA-homozygous iPSC lines for coverage of other geographical populations and ethnic groups have also been shown in Korea, where ten of the most frequent HLA-homozygous lines can match 41% of the population (146, 147), in United Kingdom (139), Spain (148), and China (149). The feasibility of such an endeavour has also been explored in Brazil, where it is estimated that 3.8 million people have to be screened to obtain 559 triple HLA-homozygous cell lines covering 95% of the population (150). In Finland (151), the top ten most frequent haplotypes homozygous for HLA-A to -DQB1 were compatible with 49.5% of the population. In Australia (152), haplotyping frequencies could be estimated from existing national blood banks or cord blood banks. A probabilistic model developed by Gourraud et al. evaluated multiple ancestry backgrounds and estimated that construction of a hiPSC bank representing 20 of the most frequent HLA haplotypes in each of the European American and African American populations would require screening of 26,000 and 110,000 donors respectively (153). This would match with over 50% of the European American and 22% of the African American populations respectively (153). Population-specific hiPSC haplobanks may also be deployed for other populations, especially closely related ones compared to ethnically-diverse populations. For such a biobank to be used for clinical applications, the cell banks have to be manufactured in compliance with Good Manufacturing Practice (GMP) and be qualified as a clinical grade cell bank, which is is highly costly, resource-intensive and not a trivial undertaking. The efficiency of such an effort depends on successfully identifying the desired HLA-homozygous haplotypes in an opportunistic manner within a screened population (potentially a prohibitively large one), unless a population-wide genomic data or blood bank typing data exists that allows for donor recall by genotype. In addition, data have shown that HLA-matching alone may be insufficient for successful allogeneic engraftment, and immunosuppression may still be required (137, 154). Furthermore, tolerability of HLA-matched hiPSCs may be dependent on other non-MHC factors such as the method of reprogramming, the cell type being transplanted (different cell types have different levels of immunogenicity), and the site of transplantation (whether the site is immunologically privileged) (155).

In addition to universal and ‘super donor’ hiPSC banks that we have discussed so far, personalized donor-specific hiPSC banks can also be generated for autologous use. As protocols to reprogramme easily accessible somatic cells (such as blood) are now standardized, it is possible to generate hiPSCs successfully for many individuals, at scale. There are increasingly more solutions being developed for large scale, high throughput generation of hiPSCs (156–159). These support the potential for autologous transplantation using patients’ own iPSCs as starting material to reduce immune-mediated graft rejection and therefore eliminate or reduce the need for immunosuppressive therapy. Guha et al. demonstrated that syngeneic mouse iPSCs differentiated into the three embryonic germ layers had little to no immunogenicity when transplanted into the subcapsular renal space of preclinical models (160). Morizane et al. also demonstrated that autologous transplantation of iPSC-derived dopaminergic neurons in the brains of NHPs elicited minimal immune response (161). These studies evaluated immunogenicity in different transplant sites, and transplantation in immune-privileged sites such as the central nervous system, brain and eyes do not generally trigger an immune response as compared to other sites, hence the site of transplantation must be considered even in an autologous setting. There are also several other considerations unique to hiPSC-derived cell therapy. The immunogenicity of autologous grafts remains to be validated as potential causes of immunogenicity include immaturity of the hiPSC derivatives, the reprogramming process and extended period of culturing and passaging of hiPSCs leading to genetic and epigenetic changes, and other off-target effects when gene correction is done on the donor cell line (162). To support the notion that stem cell derivatives exhibit variable immunogenic properties, Zhao et al. showed that autologous iPSC-derived SMCs are highly immunogenic to the immune system due to the dysregulated expression of immunogenic proteins, whereas iPSC-derived RPE cells are not immunogenic (163). Therefore, hiPSC-derived cells may not retain their immune privileged properties upon differentiation, resulting in immune attack, possibly due to differing levels of expression and activity of immunomodulatory proteins during the cellular differentiation process (163). Long-term (>4 months) evaluation of graft function and immune responses need to be considered when translating pre-clinical findings to the human context. The creation of donor cell banks for patients in need would be costly and time-consuming, given that reprogramming and qualification of the cell bank, followed by subsequent differentiation into islet cells, could easily take at least a few months. This is in contrast with other currently approved autologous therapies such as chimeric antigen receptor (CAR)-T cell immunotherapy which takes 2 to 3 weeks from apheresis to cell infusion. Nonetheless, autologous hiPSC-based cell therapy remains a useful platform for evaluating the safety and efficacy of regenerative medicine treatments for disease without many of the concerns that allogenic transplantations pose (164).