Generation of therapeutic T cells from human induced pluripotent stem cells

Abstract

T cell exhaustion remains a critical barrier in treating chronic infections and cancer. To overcome this, our laboratory has established a platform for generating “rejuvenated” T cells from induced pluripotent stem cells (iPSCs). This review outlines our research trajectory, transitioning from the foundational autologous “T-iPSC” concept to scalable, “off-the-shelf” allogeneic strategies. We detail the establishment of clinical-grade, xeno-free manufacturing protocols and the integration of advanced gene editing—including functional enhancement through intracellular signaling modulation, immune rejection–related gene editing to minimize immunogenicity, and optimization of synthetic receptor architectures to mitigate the risk of Graft-versus-Host Disease (GvHD) while enhancing targeting efficiency. Furthermore, we describe our activity on generating regulatory T cells from iPSC for managing autoimmune disorders and GvHD. Finally, we discuss the remaining challenges and the future roadmap for translating these therapeutic T cells into clinical practice.

Introduction

Effector T cells serve as the immune system’s primary executioners, playing crucial roles not only in the elimination of pathogenic viruses and bacteria but also in the surveillance and eradication of malignant cells. The advent of cancer immunotherapy, particularly Adoptive Cell Therapy (ACT) utilizing Chimeric Antigen Receptor (CAR)-modified T cells, has fundamentally transformed the treatment landscape for hematological malignancies. By redirecting patient-derived T cells to recognize tumor antigens, these therapies have achieved unprecedented durable remission rates in refractory leukemias and lymphomas [1]. However, the current paradigm, which relies heavily on autologous T cells, faces significant biological and logistical hurdles. In patients with chronic infections or advanced cancers, T cells are often subject to persistent antigenic stimulation, driving them into a state of “exhaustion” characterized by epigenetic scarring, impaired proliferative capacity, and diminished cytotoxicity [2, 3]. Consequently, the quality of the starting cellular material is highly variable, leading to manufacturing failures or suboptimal therapeutic efficacy in a subset of patients. Furthermore, the personalized “one-patient, one-batch” manufacturing model incurs high production costs and lengthy vein-to-vein intervals, limiting accessibility for patients with rapidly progressing disease [4].

iPSC technology offers a revolutionary strategy to overcome these limitations. By reprogramming somatic cells into a pluripotent state, iPSCs provide a theoretically unlimited source of cellular material that can be differentiated into any lineage, including immune effectors [5]. A pivotal advantage of this approach in the context of immunotherapy is the concept of “rejuvenation.” Our laboratory and others have demonstrated that reprogramming exhausted antigen-specific T cells into iPSCs (T-iPSCs) effectively resets the epigenetic clock—restoring telomere length and erasing exhaustion markers—while preserving the rearranged T cell receptor (TCR) genes [6, 7]. Upon redifferentiation, these cells emerge as highly potent, “rejuvenated” T cells with superior proliferation and survival profiles compared to their parental counterparts.

While early efforts focused on autologous T-iPSC models, the field is now strategically shifting toward “off-the-shelf” allogeneic therapies to democratize patient access. This transition necessitates a convergence of developmental biology and advanced genetic engineering. Unlike primary T cells, iPSCs are amenable to extensive and precise multiplex genome editing at the clonal level. This capability allows for the systematic removal of barriers to allogeneic transplantation, such as the risk of GvHD and host-mediated rejection, through the TCR-related gene editing [8, 9] and Human Leukocyte Antigens (HLA) [10], respectively.

In this review, we outline the developmental roadmap of iPSC-derived T cell immunotherapy, tracing its evolution from the initial proof-of-concept of cellular rejuvenation [11] to the current state-of-the-art clinical manufacturing platforms. We discuss the critical optimization of differentiation protocols, specifically the transition from stromal-dependent systems to fully defined, xeno-free, and feeder-free conditions essential for industrial scale-up [12], as well as the optimization of CAR signaling and CRISPR-mediated enhancement of antitumor function [13]. We also detail our strategies for establishing allogeneic platforms, including the introduction of antigen-specific receptors into HLA-homozygous iPSCs to serve broad patient populations [9, 12].

Furthermore, we examine the integration of genetic engineering to create “stealth” cellular chassis. This includes the use of hypoimmune engineering to reduce immunogenicity [10, 14] and the implementation of novel receptor architectures—such as Mini-TCRs [8] and Synthetic TCR and Antigen Receptors (STARs) [15]—to mitigate GvHD and support T cell developmental signaling. Finally, expanding the therapeutic horizon beyond cytotoxicity, we review the development of iPSC-derived regulatory T cells (iPSC-Tregs) [16]. We highlight the potential of CAR-engineered iPSC-Tregs to suppress excessive immune responses, offering a versatile new modality for managing autoimmune disorders and preventing GvHD. We conclude by discussing the remaining challenges and future perspectives for clinical translation of iPSC derived T cell therapy.

1. Antigen-specific T cell meets iPSC technology

iPSCs are pluripotent stem cells that theoretically possess the ability to differentiate into virtually all cell types constituting the human body, as well as the capacity for unlimited self-renewal in vitro [5]. By introducing defined transcription factors into somatic cells, such as peripheral blood cells, epigenetic landscapes are reset, leading to the establishment of iPSCs while preserving the donor’s genomic sequence. When iPSCs are derived from antigen-specific T cells, the rearranged TCR loci that determine antigen specificity are retained. Based on this principle, we hypothesized that T cells re-differentiated from such T-iPSCs would re-express the same antigen-specific TCR, thereby preserving their original specificity while acquiring renewed functional capacity.

To validate this hypothesis, we first established iPSCs from human CD8⁺ T cell clones specific for human immunodeficiency virus (HIV) antigens and demonstrated that the original TCR gene sequences were faithfully retained in these iPSCs. Subsequently, through stepwise coculture with two types of mouse stromal cells in the presence of lineage-directing cytokines, these iPSCs were successfully differentiated into the T cell lineage, yielding CD3⁺TCR⁺CD8⁺ T cells. These differentiated T cells (iPSC-T cells) preserved the same antigen specificity as the original T cells and exhibited significantly enhanced cytokine production and proliferative capacity compared with their parental counterparts (Fig. 1A) [6]. Thus, T cells derived from T cell–derived iPSCs (T-iPSCs) were expected to (i) retain antigen specificity, (ii) be generated in large numbers, and (iii) display a “rejuvenated” phenotype relative to the original T cells, potentially conferring superior therapeutic efficacy.

Fig. 1.

Platform establishment for T cell rejuvenation and lineage optimization. A The T-iPSC-T rejuvenation loop. Antigen-specific T cells are reprogrammed into induced pluripotent stem cells (T-iPSCs), preserving the rearranged T cell receptor (TCR) sequence. Subsequent redifferentiation regenerates antigen-specific T cells with retained antigen specificity. B Directed differentiation toward adaptive CD8αβ T cells. Conventional PHA/PBMC-based stimulation favors CD8αα homodimer expression, associated with innate-like features. An optimized protocol incorporating TCR stimulation and dexamethasone at the CD4⁺CD8⁺ double-positive stage promotes physiological CD8αβ heterodimer expression and adaptive cytotoxic lineage commitment. C Monoclonality enforcement through RAG2 ablation. Expression of RAG enzymes during differentiation can induce secondary TCRα rearrangements. CRISPR/Cas9-mediated knockout of RAG2 prevents secondary recombination, ensuring stable monoclonal TCR expression

Following this initial work, multiple research groups further advanced the field of T cell regeneration using T-iPSCs. However, it also became evident that regenerated T cells possess properties distinct from primary CD8 T cells [17]. While cytotoxic T lymphocytes normally express the CD8αβ heterodimer, early iPSC-T cells generated by our group and others predominantly expressed the CD8αα homodimer, a feature more characteristic of innate lymphoid cells (ILCs) (Fig. 1B) [17]. To generate regenerated CD8αβ T cells that more closely resemble endogenous antigen-specific T cells, we further refined the differentiation protocol. In vivo, T cells undergo selection and maturation in the thymus, where only thymocytes receiving appropriate TCR signaling through major histocompatibility complex (MHC) class I molecules differentiate into CD8αβ T cells. Based on this principle, we reasoned that providing thymus-mimicking TCR stimulation to CD4⁺CD8⁺ double-positive (DP) cells could promote differentiation into CD8αβ T cells. In addition, glucocorticoids are known to modulate TCR signal strength in thymocytes and contribute to terminal T cell maturation. We found that the addition of the glucocorticoid dexamethasone (DEX) during TCR stimulation of DP cells resulted in efficient induction of CD8αβ T cells (Fig. 1B) [9].

However, in CD8αβ T cells derived from T-iPSCs using modified differentiation method [9], recombination-activating genes (RAGs), which mediate TCR gene rearrangement, were found expressed at the DP stage, similar to physiological T cell development. Consequently, additional rearrangements occurred at the endogenous TCRα locus, leading to partial loss of antigen specificity. This loss not only markedly diminishes therapeutic efficacy but also raises the risk of generating newly induced, potentially cross-reactive CD8αβ T cell clones, necessitating a fundamental solution. We therefore focused on the observation that expression of RAG2 was upregulated in DP cells derived from T-iPSCs. Using CRISPR/Cas9 genome editing, we generated RAG2-knockout T-iPSCs. These RAG2-deficient T-iPSCs successfully differentiated into CD8αβ T cells without undergoing additional TCR rearrangements thereby kept the original TCR antigen specificity (Fig. 1C) [9].

2. Establishing clinical-grade manufacturing and functional enhancement of CAR-iPSC-T cells

The T cell differentiation methods described above relied on mouse stromal cells (e.g., 10T1/2 and OP9-DL1) as supporting cells [7, 9, 17, 18]. However, for therapeutic applications, viral safety represents a critical aspect of risk management. Because regenerative medical products, classified as cell-based medicinal products, cannot be terminally sterilized, stringent control of viral contamination throughout raw materials and manufacturing processes is essential. To address this issue, we pursued the development of methods to generate functional CD8αβ killer T cells without the use of animal-derived components, which are potential sources of viral contamination.

First, we modified an embryoid body (EB)-based differentiation protocol that enabled the generation of hematopoietic progenitor cells without the use of mouse stromal cells [19]. Subsequently, T cell differentiation from hematopoietic progenitors was achieved using immobilized Delta-like 4 (DL4). The addition of factors known to play important roles during thymocyte development—such as stromal cell–derived factor 1α (SDF-1α) and p38 MAPK inhibitors—to the culture medium resulted in a marked increase in CD4⁺CD8⁺ DP T cell yield [12]. Finally, feeder-free expansion conditions for the generated CD8αβ T cells were optimized without the use of peripheral blood mononuclear cell (PBMC) feeder cells by supplementation with cytokines including IL-15, IL-7, IL-12, IL-18, and IL-21 [20] (Fig. 2A).

Fig. 2.

Clinical-grade manufacturing and functional optimization of iPSC-derived CAR T cells. A Transition to feeder-free, xeno-free manufacturing. Stromal-dependent systems are replaced by fully defined platforms incorporating embryoid body–mediated hematopoietic induction, DL4-based T-lineage commitment, and cytokine-driven maturation, enabling scalable production of CD8αβ T cells. B Optimization of CAR signaling for lineage fidelity. CD28-based CARs induce strong tonic signaling that suppresses BCL11B expression and impairs T-lineage specification. In contrast, 4-1BB-based CARs maintain controlled signaling and preserve physiological T cell differentiation. C CRISPR-mediated enhancement of antitumor signaling. Deletion of DGKα and DGKζ relieves inhibitory constraints on ERK activation, restoring robust downstream signaling and improving antitumor efficacy in solid tumor models

To evaluate whether this clinical grade differentiation method derived CD8αβ killer T cells could serve as starting material for CAR T cell production, we introduced a CD19-specific CAR gene together with a membrane-bound IL-15/IL-15Rα complex, which is known to promote T cell survival in vivo. When these iCAR T cells were administered to immunodeficient mice previously engrafted with the acute lymphoblastic leukemia cell line Nalm-6, treated animals exhibited a marked extension of survival [12].

We extended our investigation to the application of the CAR-iPSC platform for treating solid tumors, which currently lack effective therapeutic options. To establish a robust foundation for this approach, we first optimized the CAR design to ensure compatibility with iPSC-to-T cell differentiation. We observed that first-generation and CD28-based second-generation CARs (e.g., 1928z) triggered non-specific phosphorylation of CD3z during the early stages of differentiation [13]. This antigen-independent tonic signaling may suppress the expression of BCL11B, a critical transcription factor for T cell commitment, thereby impairing differentiation into CD4⁺CD8⁺ double-positive (DP) cells and skewing development toward an NK-like phenotype [21, 22]. Based on these findings, we established that prioritizing constructs with milder signaling profiles—specifically those incorporating 4-1BB co-stimulatory domains—is a prerequisite for successfully generating T-lineage cells from iPSCs (Fig. 2B). With this CAR design in place, we sought to further potentiate iCAR-T cells against refractory solid tumors by editing key regulatory genes. Our detailed functional analysis revealed that, compared to peripheral blood-derived primary T cells, iPSC-T cells exhibited attenuated activation signaling, characterized by reduced ERK phosphorylation [13]. To overcome this signaling bottleneck, we targeted diacylglycerol kinases (DGKs), specifically isoforms α and z, which act as negative regulators of TCR signaling [13, 23, 24]. By employing CRISPR/Cas9 to knock out DGK α and DGK z in iPSCs, we generated engineered iCAR-T cells from DGK-DKO iPSCs which showed restored signaling potency. In a subcutaneous xenograft mouse model, these DGK-DKO iCAR-T cells demonstrated superior efficacy, yielding significantly improved tumor control and prolonged survival compared to both conventional iCAR-T cells and unmodified primary CAR-T cells [13] (Fig. 2C).

3. The autologous TIL-to-iPSC-T cell therapy

Compared with embryonic stem (ES) cells, iPSCs pose fewer ethical concerns, as their derivation does not require the destruction of human embryos. Moreover, when iPSCs are generated from a patient’s own somatic cells, they can be used for autologous transplantation with reduced risk of immune rejection due to HLA matching. The original concept of regenerative T cell therapy was therefore based on autologous transplantation, in which iPSCs are generated from a patient’s tumor-specific T cells, differentiated into T cells, and subsequently returned to the same patient.

To test this concept, we isolated tumor-infiltrating lymphocytes (TILs) from colorectal cancer patient samples and co-cultured them with autologous colorectal cancer spheroids [11]. Tumor-reactive T cells were identified and enriched based on CD107a or CD137 expression (Fig. 3A). These selected tumor-specific T cells were reprogrammed into iPSCs using Sendai virus vectors with optimized protocols. The resulting TIL-derived iPSCs (TIL-iPSCs) were successfully differentiated into T cells (TIL-iPS-T cells) (Fig. 3B), which retained intrinsic T cell functions and tumor specificity while exhibiting enhanced proliferative capacity and increased cytotoxic activity. Moreover, compared with primary TILs, TIL-iPS-T cells displayed a less differentiated phenotype and prolonged persistence in Patient-Derived Xenograft (PDX) mouse model (Fig. 3C), highlighting their potential as a cellular product for autologous adoptive cancer immunotherapy (Fig. 3D).

Fig. 3.

Autologous regeneration of tumor-specific T cells from iPSCs. A Functional enrichment of tumor-reactive T cells. Tumor-infiltrating lymphocytes (TILs) are isolated and enriched based on activation markers following coculture with autologous tumor spheroids. B Reprogramming and redifferentiation of TIL-derived iPSCs. Selected TILs are reprogrammed using non-integrating vectors while preserving the rearranged TCR sequence. Redifferentiation generates a renewable source of patient-specific T cells. C Rejuvenation-associated functional restoration. Compared with exhausted original primary TILs, regenerated T cells display enhanced proliferative capacity, and potent antigen-specific cytotoxicity in vitro and prolonged persistence in vivo PDX model. D Autologous therapeutic reinfusion. Rejuvenated T cells are reinfused into the original patient, ensuring complete HLA matching and eliminating the risk of immune rejection

4. The leap to allogeneic “ready-to-use” therapies

For autologous regenerative T cell therapy, multiple steps—including the cloning of antigen-specific T cells, establishment of iPSCs, and redifferentiation into T cells—are required, each of which is associated with substantial cost and time, representing a major obstacle to clinical implementation. In contrast, allogeneic transplantation eliminates the need for patient-specific cell preparation, thereby reducing costs. Moreover, if cells are cryopreserved in a ready-to-use state, treatment can be initiated immediately when needed.

One of the major concerns in allogeneic transplantation is immune rejection. To address this issue, the Center for iPS Cell Research and Application (CiRA) and the iPS Cell Research Foundation (CiRA Foundation) at Kyoto University have been advancing a project to establish a stock of iPSCs derived from HLA-homozygous donors, which are expected to minimize immune rejection and enable allogeneic transplantation [25–27]. Building on previously established cell production methods, we applied these techniques to HLA-homozygous iPSCs to generate allogeneic immune cells (Fig. 4A). Specifically, a TCR gene specific for the Wilms tumor 1 (WT1) antigen, which is expressed in a wide range of tumors, was introduced into HLA-homozygous iPSCs, followed by differentiation into CD8αβ T cells. As a result, nearly all generated CD8αβ T cells expressed the WT1-specific TCR. In immunodeficient mouse models engrafted with WT1-expressing human tumor cell lines, these cells exhibited significant suppression of tumor progression and effective control of metastasis [9].

Fig. 4.

Strategies for “Off-the-Shelf” allogeneic iPSC-T cell therapy. A HLA-homozygous iPSC banking to reduce alloimmune mismatch. Clinical-grade iPSC lines derived from HLA-homozygous donors provide partial population coverage with reduced immunogenic disparity. B Multilayered immune evasion engineering. CRISPR-mediated deletion of B2M and CIITA eliminates HLA class I and II expression to avoid host T cell recognition. NK cell responses are controlled through HLA-E knock-in or selective HLA-C retention, PVR deletion, and ablation of ICAM-1 and CD58 to disrupt immune synapse formation. C Pre-installation of antigen specificity in hypoimmunogenic iPSCs. Defined antigen receptors are introduced at the iPSC stage to generate standardized therapeutic T cell products. Mini-TCR constructs provide CD3-mediated developmental signaling without antigen specificity, creating an inert chassis for safe receptor engineering. STAR receptors couple antibody specificity to TCR constant domains, enabling T cell development with enhanced antigen sensitivity

Subsequently, to evaluate the clinical applicability of immune cells derived from HLA-homozygous iPSCs, we also conducted a representative non-clinical study using induced innate lymphoid cells (ILCs)/NK cells. A third-generation chimeric antigen receptor (CAR) containing a single-chain variable fragment (scFv) specific for the cancer-associated antigen glypican-3 (GPC3) was introduced into the most frequently occurring HLA-homozygous iPSC line in the Japanese population. ILC/NK cells derived from these iPSCs (GPC3-CAR-iPS-ILCs) were then evaluated in preclinical studies. Administration of GPC3-CAR-iPS-ILCs into immunodeficient mice revealed no evidence of acute systemic toxicity or tumorigenicity arising from residual iPSCs. Furthermore, in a peritoneal dissemination model established by intraperitoneal inoculation of GPC3-expressing ovarian tumor cells into NSG mice, repeated local administration of GPC3-CAR-iPS-ILCs resulted in a significant extension of survival compared with untreated controls [28].

Based on these preclinical outcomes, a phase I clinical trial has been conducted to assess the safety (including immunogenicity and cytokine release syndrome), pharmacokinetics, and, where possible, the efficacy of intraperitoneally administered GPC3-CAR-iPS-ILCs against peritoneally disseminated tumors. This physician-initiated clinical trial has been ongoing since 2021, involving a collaborative clinical team led by the National Cancer Center Hospital East, a manufacturing team comprising CiRA Foundation, our laboratory and Shinobi Therapeutics.

Although the use of HLA-homozygous iPSCs represents one strategy to reduce immunogenicity in allogeneic immune cell therapy, the current HLA-homozygous iPSC stock maintained by the CiRA Foundation is estimated to cover only approximately 40% of the Japanese population (Fig. 4A). To expand population coverage, multiple approaches have therefore been explored to reduce immunogenicity through genetic modification of HLA molecules in iPSCs (Fig. 4B) [29–33].

As one such approach, we generated iPSCs in which β2-microglobulin (B2M) and CIITA, molecules essential for the expression of HLA class I and class II, respectively, were genetically ablated. CD8αβ killer T cells differentiated from these iPSCs were shown to evade immune responses mediated by allogeneic CD8⁺ and CD4⁺ T cells derived from healthy donors. Furthermore, by introducing a single chain HLA-E molecule—an inhibitory ligand for NK cells—and simultaneously knocking out the activating ligand PVR, we demonstrated that regenerated CD8αβ killer T cells could also evade immune responses mediated by resting allogeneic NK cells. These engineered cells were therefore designated as “first-generation hypoimmunogenic” iPSC-T cells.

Subsequent introduction of a CD20-specific chimeric antigen receptor (CD20-CAR) into these stealth cells further demonstrated their therapeutic potential. In immunodeficient mouse models engrafted with third-party peripheral blood lymphocytes (including T cells, B cells, and NK cells) as well as CD20-positive B lymphoblastoid cell lines (B-LCLs) generated from the same donor, administration of CD20-CAR–expressing stealth regenerated CD8αβ killer T cells resulted in significantly prolonged in vivo persistence and effective suppression of tumor growth compared with non-genetically edited iPSC-derived T cells. These findings suggested that stealth CD8αβ killer T cells could potentially substantially enhance the versatility of allogeneic T cell therapy platforms [10].

We showed that NK evasion strategy for “first-generation hypoimmunogenic” iPSC-T cells was effective primarily under short-term coculture conditions with resting NK cells [10]. However, when coculture duration was extended or when NK cells were activated by inflammatory cytokines like IL-15, first-generation hypoimmunogenic iPSC-derived T cells were still subject to robust NK cell–mediated cytotoxicity [14]. These observations highlighted the need for a more broadly effective NK evasion strategy.

To address this limitation, we focused on the initial adhesion step of NK cell–target cell recognition, hypothesizing that genetic ablation of adhesion molecules on iPSC-derived T cells could prevent the early physical interactions required for immune synapse formation. Based on this concept, we developed a strategy to mechanically dismantle the immunological synapse (IS) required for cytotoxic execution.

Specifically, by genetically ablating the key adhesion molecules ICAM-1 (CD54) and LFA-3 (CD58) in iPSCs, the resulting T cell progeny were rendered incapable of supporting stable conjugation with NK cells mediated through LFA-1–ICAM-1 and CD2–LFA-3 interactions. This “slippery target” phenotype effectively aborted lytic granule polarization and immune synapse maturation, thereby conferring broad-spectrum resistance to diverse allogeneic NK donors, even under highly inflammatory conditions such as exposure to IL-15 [14].

5. Antigen specific construct for iPSC derived T cell platform

We have extensively investigated diverse strategies for generating antigen-specific T cells, ranging from the use of T-iPSCs to iPSCs engineered with antigen-specific TCRs or CARs introduced at either the pluripotent (Fig. 4C) or differentiated stages. However, a fundamental bottleneck across these protocols is the obligate requirement for a structural TCR complex. This is because the physiological maturation of T cells—specifically the transition from CD4⁺CD8⁺ DP progenitors to single-positive cells—strictly depends on developmental signaling transmitted through the TCR-CD3 complex, which we recapitulate in vitro via agonist CD3 stimulation. In the context of allogeneic transplantation, however, reliance on these full-length TCRs introduces a significant risk of GvHD. To overcome this, we developed a novel “Mini-TCR” strategy by engineering synthetic receptors that retain the TCR constant domains while deleting the variable regions. We demonstrated that these Mini-TCRs successfully recruit the endogenous CD3 complex to the cell surface and enable iPSC-derived DP cells to respond to anti-CD3 (OKT3) stimulation, thereby driving functional maturation. This innovation establishes a highly flexible and efficient platform: by effectively decoupling developmental signaling from antigen specificity, it allows for the modular introduction of therapeutic CARs or TCRs at either the iPSC or T cell stage, creating a versatile “blank chassis” for off-the-shelf immunotherapies (Fig. 4C).

Furthermore, we leveraged the Synthetic TCR and Antigen Receptor (STAR) [34–36] architecture to transcend the fundamental limitations of conventional CARs—specifically, their inability to drive T cell developmental signaling and their restriction to high-density surface antigens (Fig. 4C). By fusing a single-chain variable fragment (scFv) derived from a TCR-mimic antibody (which recognizes MHC-peptide complexes) directly to the TCR constant domains, the STAR construct preserves the physiological capacity to recruit and assemble with the endogenous CD3 complex.

Using this platform, we engineered a WT1-specific STAR and demonstrated that it successfully provides the requisite signaling to drive the differentiation of iPSCs into phenotypic, adaptive-type CD8αβ T cells. Crucially, because these cells rely on the STAR for development, they do not express endogenous TCRs, thereby mitigating the risk of GvHD. Functional validation via in vitro co-culture assays with WT1-peptide-pulsed or WT1-expressing targets confirmed that STAR-T cells can effectively recognize intracellular antigens presented on MHC molecules. Notably, the STAR platform exhibited superior antigen sensitivity compared to traditional CAR constructs. These findings establish a robust Proof-of-Concept (POC) for the application of “off-the-shelf” iPSC-T cells in targeting refractory malignancies characterized by low-density or intracellular antigens [15].

6. Clinical translation of iPSC derived CD8 T cells

To accelerate the clinical translation of iPSC-derived immunotherapies, we established Shinobi Therapeutics, integrating our proprietary differentiation technology—which reliably regenerates highly potent, cytotoxic CD8αβ T cells—with an advanced hypo-immune strategy developed in collaboration with our U.S. partners [32]. Technologically, the Shinobi platform utilizes mini-TCR engineered iPSCs to drive efficient T cell differentiation. To ensure persistence, the resulting cells function as “stealth” agents via a multi-layered design: we combine B2M/CIITA knockout (KO) to evade T cell recognition with the expression of a SIRPα engager and truncated CD64. This architecture effectively evades rejection mediated by CD4 and CD8 T cells, NK cells, macrophages, and antibodies, thereby enabling safer, scalable, and re-doseable allogeneic therapies. Furthermore, these cells are armored with genetic enhancements to withstand and deliver molecular payloads within the hostile, immunosuppressive tumor microenvironment. Leveraging this “universal” platform, our lead candidate, NJA-001, is engineered to express a GPC3-specific TCR for the precise targeting of solid tumors. We are currently advancing NJA-001 through final preclinical validation, with clinical trials scheduled to initiate in 2026 for patients with GPC3-positive malignancies, including Hepatocellular Carcinoma (HCC), Colorectal Cancer (CRC), and Non-Small Cell Lung Cancer (NSCLC).

7. Human iPSC-derived CD4⁺ regulatory T cell–like cells

While most of our studies have focused on generating cytotoxic CD8⁺ T cells from human iPSCs, the plasticity of the iPSC platform also potentially enables the derivation of CD4⁺ T cell subsets with distinct immunological functions. Among these subsets, regulatory T cells (Tregs) play a central role in maintaining immune tolerance and suppressing pathological immune responses, including GvHD following allogeneic hematopoietic stem cell transplantation. Although adoptive primary Treg therapy has shown promising clinical results [37], its broader application has been hampered by the difficulty of expanding stable, functionally suppressive Tregs to therapeutically relevant numbers while preserving their regulatory phenotype. These limitations raise the possibility that iPSC technology could provide a scalable and programmable platform to overcome current manufacturing and stability challenges in Treg therapy.

Recent advances have demonstrated that CD4⁺ T cells can be efficiently derived from human iPSCs using three-dimensional artificial thymic organoid (ATO) culture systems [38]. However, the majority of iPSC-derived CD4⁺ T cells generated under conventional conditions resemble conventional helper T cells (Tconv) and are prone to producing inflammatory cytokines such as IFN-γ and IL-4, rather than acquiring a regulatory phenotype [39, 40]. To overcome this limitation, we developed a strategy to induce a stable regulatory T cell–like state in iPSC-derived CD4⁺ T cells through targeted modulation of differentiation and epigenetic programs.

By treating iPSC-derived CD4⁺ T cells with a defined combination of four reagents—AS2863619 [41], MR2-1 [42], rapamycin [43], and transforming growth factor-β (TGF-β) [44], collectively termed AMRT—robust induction of FOXP3, the master transcription factor governing Treg identity and function, was achieved (Fig. 5A). Importantly, AMRT-treated iPSC-derived CD4⁺ T cells (iCD4⁺ Treg-like cells) exhibited key hallmarks of bona fide regulatory T cells. These included sustained FOXP3 expression, absence of inflammatory cytokine secretion upon activation, and the acquisition of Treg-specific DNA demethylation patterns, indicating epigenetic stabilization of the regulatory phenotype rather than transient FOXP3 induction.

Fig. 5.

Generation and therapeutic application of HLA-A2 CAR-engineered, iPSC-derived regulatory T cells for GvHD suppression. A Induction of regulatory T cells using chemical programming. Human iPSCs are differentiated into CD4⁺ single-positive T cells using the ATO system. These cells are then treated with the “AMRT” protocol (IL-2, AS2863619, MR2-1, rapamycin, and TGF-β), inducing stable FOXP3 expression and TSDR demethylation consistent with regulatory lineage commitment. B Antigen-specific CAR engineering of iTregs. Transduction with an HLA-A2–specific CAR confers targeted immunosuppressive activity. C Targeted suppression of GvHD a xenograft model. In an immunodeficient mouse model challenged with HLA-A2⁺ human PBMCs, adoptive transfer of CAR-engineered iTregs attenuates effector expansion, reduces inflammatory pathology, and prolongs survival in xenogeneic GvHD model

Functionally, iCD4⁺ Treg-like cells demonstrated potent immunosuppressive capacity in vitro. In mixed lymphocyte reaction assays, AMRT-treated iCD4⁺ Treg-like cells suppressed the proliferation of CD8⁺ cytotoxic T lymphocytes (CTLs) in a dose-dependent manner. These suppressive effects were comparable to those observed with naturally occurring CD25⁺CD127⁻ Tregs isolated from peripheral blood, supporting the functional equivalence of iPSC-derived Treg-like cells.

To enhance antigen specificity and therapeutic precision, an HLA-A2–specific chimeric antigen receptor (CAR) was introduced into iCD4⁺ Treg-like cells (Fig. 5B). CAR expression did not compromise FOXP3 expression or suppressive capacity in vitro. Instead, CAR-modified iCD4⁺ Treg-like cells retained robust inhibitory effects on CTL proliferation, consistent with the concept that antigen-directed Treg activity can improve efficacy while minimizing global immunosuppression [16].

The therapeutic potential of these cells was further evaluated in vivo using a xenogeneic GvHD mouse model. Immunodeficient NSG mice were engrafted with HLA-A2–positive human peripheral blood mononuclear cells (PBMCs) to induce GvHD and subsequently treated with either CAR-modified iCD4⁺ Treg-like cells or control populations (Fig. 5C). Remarkably, only CAR-expressing regulatory populations—namely CAR-engineered CD25⁺CD127⁻ natural Tregs and CAR-engineered iCD4⁺ Treg-like cells—significantly suppressed body weight loss and prolonged survival (Fig. 5C). Notably, CAR-modified iCD4⁺ Treg-like cells demonstrated GvHD suppression comparable to that achieved with primary human Tregs, highlighting their functional equivalence in vivo [16].

8. Summary and future perspectives

In this review, we have outlined the evolutionary trajectory of iPSC-derived T cell immunotherapy, primarily detailing the developmental roadmap established within our laboratory—from the fundamental proof-of-concept of cellular “rejuvenation” to the establishment of clinically viable, off-the-shelf therapeutic platforms. We introduced the transition from autologous T-iPSC models to allogeneic strategies, highlighting the critical role of feeder-free differentiation protocols in ensuring manufacturing scalability. Furthermore, we discussed how the integration of gene-editing technologies—specifically the decoupling of developmental signaling from antigen specificity via Mini-TCR and STAR architectures, alongside hypoimmune engineering—aims to mitigate the dual challenges of GvHD and host rejection. However, the long-term functional stability and safety of these engineered cellular platforms need to be carefully validated in complex in vivo settings.

Expanding the therapeutic horizon beyond CD8 T cell generation, we also discussed our activity on the generation of iPSC-derived regulatory T cells. By utilizing defined chemical induction strategies (the AMRT cocktail), we have made significant strides in producing functional Tregs capable of suppressing excessive immune responses. However, their lineage stability upon exposure to intense inflammatory conditions needs to be rigorously evaluated. Future development will focus the establishment of robust, clinical-grade, xeno-free manufacturing protocols aimed at maintaining lineage stability within highly inflammatory environments. As a precedent, early research successfully utilized Foxp3 overexpression and Notch ligation to generate functional Tregs from Treg derived murine iPSCs [45]. Building on this logic, we believe that by fine-tuning these developmental signals alongside our induction protocols, we can generate a more stable and robust human iTreg population. Furthermore, it is critical to construct appropriate, high-fidelity preclinical models to rigorously evaluate the functional stability and therapeutic efficacy of these cells prior to human application.

Despite some progress we discussed above, the transition of iPSC-derived T cells from experimental tools to clinically viable therapeutics requires a holistic optimization strategy. This “Next-Generation Roadmap” focuses on three critical pillars:

• (i)Precision manufacturing: From “Elite” clones to scalable 3D systems. The foundation of reproducible therapy lies in the quality of the starting iPSC material. Many existing iPSC lines exhibit clonal heterogeneity, differentiation bias, or latent genomic instability due to residual epigenetic memory or spontaneous mutations [46]. Future efforts must prioritize the identification and banking of “elite” iPSC clones [47] with superior hematopoietic [48] and lymphoid differentiation capacity. This may be achieved through predictive molecular screening using single-cell transcriptomic, epigenomic, and methylomic signatures to pre-select clones with definitive hematopoietic potential.

  At the differentiation level, protocols should more faithfully recapitulate definitive hematopoiesis while avoiding primitive waves that compromise T lineage competency. Controlled generation of high-quality hematopoietic progenitor cells (HPCs) [49] and T cell progenitors (ProTs) [50] with robust engraftment and differentiation potential will be essential.

  To meet global demand, the field must transition from labor-intensive 2D co-culture systems to automated, closed 3D bioreactor platforms compatible with GMP standards. Fully defined, feeder-free, and xeno-free systems should support large-scale expansion while preserving a stem-memory–like phenotype critical for long-term persistence and functional durability.

• (ii)Synthetic biology integration to enhance function and immune compatibility. Beyond manufacturing scalability, the next step is to make iPSC-derived T cells truly functional in real disease settings. It is not sufficient for these cells to simply recognize antigen. They must survive, migrate, persist, and remain functional within complex and often suppressive in vivo environments.

  From a functional standpoint, future modifications should focus on improving cytokine responsiveness and chemokine-guided migration. For example, fine-tuning chemokine receptor expression could enhance tissue homing, while reinforcing intracellular signaling pathways may improve in vivo persistence. Metabolic fitness is another critical factor. T cells derived from iPSCs must maintain mitochondrial integrity and avoid premature exhaustion, particularly within the tumor microenvironment. Intrinsic signaling rewiring or controlled modulation of checkpoint pathways may represent a more stable and potentially safer strategy.

  At the same time, immune compatibility gene-editing strategies need to be guided by a mechanistic understanding of how these cells are rejected in vivo. In some contexts, selective HLA modulation may be sufficient; in others, attenuation of NK cell–activating signals or disruption of immune synapse–associated adhesion molecules may be required. Importantly, the clonal and highly editable nature of iPSCs enables multiplex genetic modifications at the master cell bank stage, thereby ensuring uniform immune characteristics across all differentiated products.

  Overall, the integration of synthetic biology into the iPSC platform should not aim merely at increasing cytotoxicity (iPSC-CTL) or enhancing suppressive function (iPSC-Treg). Instead, it should enable the generation of programmable T cells that are tissue-adapted, resistant to exhaustion, and rationally optimized to coexist with the host immune system. Whether iPSC-derived T cells remain sophisticated experimental systems or evolve into broadly applicable therapeutic platforms will ultimately depend on achieving this level of design precision.

• (iii)Addressing safety concerns. As with all pluripotent stem cell–derived therapies, safety remains paramount. The risk of tumorigenicity due to residual undifferentiated cells or insertional mutagenesis must be rigorously minimized through stringent purification, genomic surveillance (e.g., WGS-based integrity checks), and defined differentiation endpoints. In addition, genetically encoded safety switches should be considered a core design element rather than an optional add-on. Inducible suicide systems such as iCasp9 or HSV-TK–based approaches provide a controllable “off-switch,” allowing selective elimination of infused cells in the event of severe toxicity or unexpected immune complications [51, 52]. These fail-safe mechanisms provide a necessary “off-switch,” allowing clinicians to selectively eliminate the grafted cells in the event of adverse events or unexpected toxicity, thereby ensuring the overall safety profile of these next-generation living drugs.

  Taken together, although this review has primarily outlined the developmental trajectory established within our own research program, the progress of the field has been shaped by the substantial contributions of many investigators working in parallel across disciplines and regions. The continued advancement of iPSC-derived T cell therapies will require sustained, multidisciplinary collaboration within the global scientific community. Through such collective effort, these engineered cellular platforms may ultimately fulfill their promise as broadly accessible treatments for human disease.

 

Abbreviations

ATO

Artificial Thymic Organoid

B2M

Beta-2 Microglobulin

BCL11B

B-Cell Lymphoma/Leukemia 11B

CAR

Chimeric Antigen Receptor

CD

Cluster of Differentiation (e.g., CD3, CD4, CD8, CD19, CD28)

CIITA

Class II Major Histocompatibility Complex Transactivator

CRC

Colorectal cancer

CRISPR

Clustered Regularly Interspaced Short Palindromic Repeats

Cas9

CRISPR-associated protein 9

CTL

Cytotoxic T lymphocyte

DEX

Dexamethasone

DGK

Diacylglycerol Kinase

DL4

Delta-Like 4

DP

Double positive (CD4+ CD8+)

EB

Embryoid body

ERK

Extracellular Signal-Regulated Kinase

ES

Embryonic stem (cells)

FOXP3

Forkhead Box P3

GMP

Good manufacturing practice

GPC3

Glypican-3

GvHD

Graft-versus-host disease

HCC

Hepatocellular carcinoma

HIV

Human immunodeficiency virus

HLA

Human Leukocyte Antigen

ICAM-1

Intercellular Adhesion Molecule 1 (CD54)

iCART

IPSC-derived Chimeric Antigen Receptor T cells

iPSC

Induced pluripotent stem cell

IL

Interleukin (e.g., IL-2, IL-7, IL-15)

ILC

Innate lymphoid cell

LFA-3

Lymphocyte Function-Associated Antigen 3 (CD58)

MAPK

Mitogen-Activated Protein Kinase

mbIL15

Membrane-bound IL-15

MHC

Major histocompatibility complex

NK

Natural killer (cell)

NSCLC

Non-small cell lung cancer

NSG

NOD scid gamma (immunodeficient mouse strain)

PBMC

Peripheral blood mononuclear cell

pCAR-T

Peripheral blood-derived CAR-T cell

PVR

Poliovirus Receptor (CD155)

RAG

Recombination-Activating Gene

scFv

Single-chain variable fragment

SDF-1α

Stromal Cell-Derived Factor 1 alpha

SIRPα

Signal Regulatory Protein Alpha

SIRPα-E

Signal Regulatory Protein Alpha Engager

STAR

Synthetic T cell Receptor and Antigen Receptor

T-iPSC

T cell-derived induced pluripotent stem cell

Tconv

Conventional T cell

TCR

T Cell receptor

TGF-β

Transforming Growth Factor Beta

TIL

Tumor-infiltrating lymphocyte

Treg

Regulatory T cell

WGS

Whole-genome sequencing

WT1

Wilms Tumor 1

Authors’ contributions

Bo Wang drafted the manuscript and prepared the figures. Shin Kaneko provided overall guidance on the structure and content of the manuscript and reviewed and approved the final version. All authors have read and approved the final manuscript.

Funding

Not applicable.

Data availability

No datasets were generated or analyzed during the current study.

Declarations

Competing interests

Shin Kaneko is a founder, shareholder, and director of Shinobi Therapeutics Inc. S.K. has also received research funding from Takeda Pharmaceutical Co., Ltd., Astellas Pharma Inc., Panasonic Corporation, Mitsui-soko Co., Ltd., KOTAI Biotechnologies, and Shinobi Therapeutics Inc.