Next-Generation regulatory T cell therapies: Translational considerations for inflammatory bowel disease

1. Introduction

The prevalence of inflammatory bowel diseases (IBD), Crohn’s disease (CD) and ulcerative colitis (UC), continues to rise globally. Current treatments, including immunomodulators, biologics (e.g., TNF and integrin inhibitors), and small molecules (e.g., JAK inhibitors), have improved clinical outcomes but remain limited by primary nonresponse, secondary loss of response, and adverse effects such as immunosuppression and malignancy.

T cell-based therapies, initially developed for cancer, offer precise antigen targeting through genetic modification of the T-cell receptor (TCR) or chimeric antigen receptors (CARs). CAR T cells have shown high efficacy in hematologic malignancies, including leukemia, lymphoma, and multiple myeloma, with their use now expanded to autoimmunity, including IBD. Indeed, the first reported use of CD19-targeted CAR T cells in a patient with severe, multidrug-resistant UC led to clinical, biochemical, and mucosal improvement [1].

Regulatory T cells (Tregs), with innate immunosuppressive capacity, are an attractive option for IBD and other autoimmune diseases. Early-phase trials using adoptive transfer of expanded autologous Tregs in graft-versus-host disease (GVHD) and type 1 diabetes (T1D) [2–5] have demonstrated safety and feasibility. However, progress is limited by small trial size, heterogeneity in manufacturing and dosing, potential instability, and variable persistence and efficacy. Advances in T cell engineering now enable targeted delivery and activation of Tregs for specific contexts.

Although Treg-based cellular therapies are being developed across transplantation and multiple autoimmune diseases, translation into IBD presents distinct challenges. The intestinal mucosa is characterized by continuous antigen exposure, dynamic cytokine signaling, and inflammatory pathways that can influence Treg stability and function. Therefore, successful application of engineered Treg therapies in IBD will require careful consideration of antigen specificity, tissue homing, lineage stability, and appropriate preclinical modeling systems. In this review, we examine advances in adoptive and engineered Treg strategies through the lens of their translational relevance to IBD, highlighting key biological, manufacturing, and modeling considerations necessary to advance next-generation Treg therapies for intestinal inflammation.

2. Tregs as immunosuppressive products

Tregs are essential in maintaining immune homeostasis and preventing autoimmunity [6]. The identification of CD25 as a surface marker of suppressive CD4⁺ T cells sparked interest in their therapeutic potential. In mice, depletion of CD25⁺ cells induces systemic autoimmunity, whereas reconstitution with CD4⁺CD25⁺ cells prevents autoimmune disease [7]. Tregs are also defined by the transcription factor FOXP3, which directs lineage differentiation and function. FOXP3 deficiency causes severe autoimmunity in humans, exemplified by Immune dysregulation, Polyendocrinopathy, Enteropathy, X-linked (IPEX) syndrome [8].

Treg suppression occurs via contact-dependent and contact-independent, or cytokine-mediated, mechanisms. Contact-dependent pathways involve inhibitory receptors such as cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), which binds CD80/CD86 to block CD28 co-stimulation, and PD-1, which engages PD-L1 to limit effector T cell activation, with context-dependent effects on regulatory function [9,10]. Several comprehensive reviews summarize the diverse contact-dependent and cytokine-mediated mechanisms of Treg suppression and functional regulation [11–13]. Inflammatory conditions can induce Tregs to express granzyme B and perforin, promoting apoptosis of effector T cells and antigen-presenting cells (APCs) [14]. Finally, by binding the Fas receptor on effector T cells, the Fas ligand (FasL) expressed on certain Treg subsets (ie, CCR4 + Tregs) can also induce apoptosis in effector T cells [15].

Contact-independent mechanisms rely primarily on the production of immunossupressive cytokines, including IL-10, IL-35 and TGF-β. IL-10 signals through the IL-10 receptor (IL10R) and STAT3 to support Treg suppressive function and control inflammatory responses, particularly by restraining Th17-mediated inflammation [16–19]. IL-35 suppresses effector proliferation and Th17 differentiation [20], while TGF-β inhibits effector activation and promotes infectious tolerance by inducing FOXP3 in naïve CD4⁺CD25⁻ cells [21]. Contact-independent suppression also occurs through the ectokinases CD39 and CD73, which convert extracellular ATP to adenosine, inhibiting effector T cells via their A2A receptors [22].

IL-2 is vital for Treg survival, proliferation, and function. Due to their high affinity to IL2 through CD25 (IL2 receptor), Tregs capture IL-2 even in low-cytokine environments, activating STAT5 to maintain FOXP3 and expression of CTLA-4 and CD39/CD73. This preferential usage also deprives effector T cells of IL2, limiting their expansion.

Taken together, Tregs are central to immune tolerance, and their diverse suppressive mechanisms underpin their expanding potential as candidates for cellular immunotherapy in autoimmunity, including IBD (Fig. 1).

Fig. 1. Mechanisms of Treg suppressive function.

Tregs exert their suppressive function through several mechanisms, including through inhibitory receptors (CTLA-4, PD-1), pro-apoptotic mechanisms (FasL), and contact-independent mechanisms, primarily through secretion of anti-inflammatory cytokines (IL-10, IL-35, and TGF-β) and conversion of ATP to adenosine through ectokinases CD39/CD73. Effector T cell deprivation of IL-2 occurs through the high affinity IL-2 receptor on Tregs (CD25). Figure created with Biorender.

3. Considerations in Treg engineering: implications for IBD

Polyclonal Treg therapies have shown curative effects in mouse models of IBD [23] and early-phase human trials report safety, with no significant infusion reactions or high-grade adverse events (NCT03185000) [24,25]. One case report described successful autologous Treg therapy in refractory UC with concurrent primary sclerosing cholangitis (PSC), showing endoscopic improvement, normalization of liver enzymes, and no treatment-related adverse events [24]. In pre-clinical models, efficacy improves with enhanced antigen specificity, tissue homing, and stability in inflammatory environments. Antigen-specific, engineered Tregs, outperform polyclonal Tregs in suppression and migration to sites of inflammation [26].

One approach to confer disease-relevant antigen specificity is to engineer Tregs with CARs. CARs are fusion proteins that combine an extracellular antigen-binding domain, most often an scFv, with intracellular signaling modules such as CD3ζ and costimulatory domains (CD28, OX40, or 4-1BB), enabling activation [27]. Because CARs signal independently of MHC expression, MHC matching, and exogenous costimulation, they bypass key constraints that limit engineered T-cell receptors (TCRs) [28]. CAR designs have progressed from first-generation constructs containing only CD3ζ to second- and third-generation versions with costimulatory domains that improve persistence and potency, such as CD28 or 4-1BB [29,30], and more recently to fourth-generation designs with added regulatory functions. CAR-Tregs have been developed for transplant tolerance [31–34] and are under investigation across a range of other diseases [35–37].

The first preclinical study of antigen-specific CAR-Tregs demonstrated that CARs directed Tregs to colitis lesions and enabled MHC-independent suppression of effector T cells, significantly alleviating disease [38]. Subsequent work produced flagellin-specific CAR-Tregs that localize to colonic lamina propria during barrier disruption [39]. Targeting the IL-23R, a key mediator in IBD, with CAR-Tregs has also shown promise, with studies demonstrating antigen-dependent suppression, in vivo homing to IL-23R-expressing tissues, and responsiveness to Crohn’s disease tissue derived from patient biopsy samples [40].

Optimizing CAR-Tregs requires thorough evaluation of CAR expression, antigen-specific engagement and activation, signaling, and stable genomic integration. CARs can also influence cell behavior, including cell cycle kinetics and activation thresholds, so risks such as tonic signaling, exhaustion, or uncontrolled proliferation must be carefully evaluated. These considerations are particularly critical for CAR-Tregs, as uncontrolled expansion or loss of lineage fidelity could compromise therapeutic efficacy and safety. Notably, the aforementioned IL-23R CAR-Tregs have exhibited minimal tonic signaling, maintained a regulatory phenotype, and showed antigen-dependent suppression [40].

Recombinant TCR-engineered Tregs offer physiological activation and can recognize both extracellular and intracellular antigens presented by MHC on APCs. However, HLA diversity limits generalizaed applicability. Furthermore, mispairing with endogenous TCR chains can cause off-target effects. Targeted insertion into the T-cell receptor alpha constant (TRAC) locus can replace the native TCR and reduce mispairing risk [41].

Although TCR-Tregs have not yet been extensively studied in IBD, preclinical models show efficacy in transplantation (allo-MHC class II-specific TCR-Tregs prolonging heart allograft survival [42]) and autoimmune disease (EAE with myelin-specific TCR-Tregs [43,44], T1D in NOD mice with islet-specific TCR-Tregs [45]). An emerging approach is to use TCRs from natural Tregs, which may better direct regulatory activity to inflammatory sites. Tissue-resident Tregs have a more restricted TCR repertoire than those in lymphoid organs [46], suggesting TCR specificity contributes to tissue localization and function.

A related strategy is the conversion of pathogenic effector T cells into Tregs, leveraging their endogenous antigen receptors and tissue-homing properties to redirect disease-relevant T cells toward a suppressive phenotype. Recent preclinical work demonstrates the feasibility of reprogramming autoreactive human T cells into regulatory cells while retaining their native TCR specificity [47]. Together, these approaches highlight the potential of antigen-guided regulatory strategies to enhance site-specific immune control.

4. Current development of Treg programs

The heterogeneity of human autoimmune diseases, including variability in target antigens, distinct tissue microenvironments, and differing immunological profiles, has driven the development of diverse Treg-based therapeutic strategies. Factors such as Treg sourcing, expansion methods, and engineering techniques influence the stability and efficacy of the final cell product. Additionally, external variables, including concomitant therapies like anti-CD20 agents or other immunomodulators, may impact both therapeutic effectiveness and potential adverse effects. Here, we outline the current landscape of Treg therapeutics under development for autoimmune and transplantation-related conditions (Table 1).

Table 1.

Class	Product Name / Sponsor	Target Antigen or Indication	Phase/Status	Trial ID(s)	Key Notes
Autologous Polyclonal Tregs	PTG-007 (PolTREG)	Recent-onset T1D ± rituximab	Phase 1/2	ISRCTN37116985	Combination with rituximab delayed T1D progression; monotherapy less effective
	PolTREG-T1D	Stage 1/2 T1D (presymptomatic)	Recruiting	NCT06688331	Pediatric focus
	CLBS03 (Caladrius)	New-onset pediatric T1D	Phase 2	NCT02691247	Safe, no β-cell preservation
	PTG-007 (PolTREG)	MS (IV vs intrathecal)	Phase 1/2a	EudraCT database 2014–004320-22	Intrathecal delivery promising (over IV), study not powered for significance
	ONE Study	Kidney transplant	Phase 1/2a completed	NCT02371434 NCT02129881	Safe induction alternative
	TWO Trial	Kidney transplant	Phase 2b	ISRCTN11038572	Reduced maintenance immunosuppression
	GAMECHANgER-1	Highly sensitized transplant candidates	Phase 2a	ISRCTN14582152	Pre-transplant desensitization
Thymic Tregs	THYTECH	Pediatric heart transplant	Phase 1/2	NCT04924491	First-in-human safety, stable Treg frequency at 2y
Donor Allo-Reactive Tregs	ONE Study	Live donor renal transplant	Phase 1/2	NCT02244801, NCT02091232	>6y monotherapy with tacrolimus
	ARTEMIS	Living donor liver transplant	Phase 1/2	NCT02474199	Dose manufacturing challenges
Specific Tregs: CAR	QEL-001 (Quell)	HLA-A2+ liver transplant	Phase 1/2	NCT05234190	FOXP3-Phenotype Lock™, iCasp9 switch
	SBT777101 (Sonoma)	Refractory RA, hidradenitis suppurativa	Phase 1/2	NCT06201416, NCT06361836	Targets citrullinated antigens
Engineered Antigen-Specific Tregs: TCR	GNTI-122 (GentiBio)	IGRP* in T1D	Phase 1	NCT06919354	Tunable IL-2 signaling (rapamycin)
	ABA-101 (Abata)	MBP** in progressive MS	Phase 1	NCT06566261	HLA-DRB1*15:01-specific
	ABA-201 (Abata)	PPI*** peptide in T1D	IND-enabling	—	Preclinical stage
Allogeneic Products	TRX103 (Tr1X)	GVHD prevention, refractory Crohn’s	Phase 1, Phase 1/2a	NCT06462365, NCT06721962	IL-10/TGF-β producing
	Orca-T	Hematologic malignancies (HCT)	Phase 3, earlier phases	NCT05316701, NCT05507827, NCT06195891, NCT04013685	Composite HSPC + Treg + Tcon product
iPSC-Derived Tregs	Quell/Cellistic	iPSC-derived CAR-Tregs	Preclinical	—	HLA-A2 CAR platform under development

^*IGRP = islet-specific glucose-6-phosphatase catalytic subunit-related protein

^**MBP = myelin basic protein

^***PPI = preproinsulin peptide.

Autologous polyclonal Treg products, derived from the patient’s own cells, have the advantage of reduced immunogenicity and lower risk of GVHD. These products are in early clinical trials for diseases such as T1D, multiple sclerosis (MS), and solid organ transplantation. In T1D, the adoptive transfer of polyclonally expanded autologous Tregs aims to restore immune tolerance and preserve residual pancreatic β-cell function. For example, PTG-007 (PolTREG), which is generated from peripherally derived Tregs (CD4⁺CD25⁺CD127^lo), was tested in a phase 1/2 trial combining PTG-007 with the CD20 targeting monoclonal antibody, rituximab [48]. This combination showed superiority in delaying disease progression compared to controls, though PTG-007 alone did not demonstrate the same effect. Other trials, such as ongoing studies evaluating PolTreg-T1D in presymptomatic pediatric patients, continue to assess safety and efficacy (NCT06688331). Similarly, smaller studies have reported increased C-peptide levels and reduced insulin requirements following autologous Treg therapy, with favorable safety profiles [49]. Larger randomized trials, such as the Sanford Project T-Rex study, have found that while single doses of expanded autologous Tregs are safe, they did not prevent β-cell decline over one year [50]. Beyond T1D, autologous Tregs are being investigated in MS, with early phase I/IIa studies evaluating intrathecal or intravenous (IV) delivery of PTG-007, with promising safety and lesion control outcomes [51]. In transplantation, the ONE study and its follow-up, the TWO trial, are exploring autologous Tregs for promoting tolerance post-kidney transplant, aiming to reduce reliance on standard immunosuppression[52,53]. Another phase 2a trial, GAMECHANgER-1 (ISRCTN14582152), is targeting highly sensitized transplant candidates pre-transplant to improve outcomes in HLA-incompatible grafts. Collectively, these trials illustrate that autologous Treg therapies are well tolerated and safe across indications; however, robust and consistent evidence of durable clinical benefit remains to be established.

Donor alloreactive Tregs represent another strategy designed to induce donor-specific immune tolerance and prevent allograft rejection in solid organ transplantation. These Tregs are peripherally derived from the transplant recipient (autologous) and expanded ex vivo in the presence of donor antigen-presenting cells (APCs), to enrich for Tregs that specifically recognize donor antigens [54]. Early clinical data from the ONE study consortium demonstrated that purified donor alloreactive Tregs administered post-renal transplant, alongside reduced immunosuppression, achieved prolonged graft survival without rejection events over at least six years [55]. Similarly, the ARTEMIS trial explored donor alloreactive Tregs in liver transplant recipients, although manufacturing challenges limited dose delivery and efficacy evaluation [56]. Nevertheless, no safety concerns were observed.

Recognizing limitations with peripherally derived Tregs, such as limited cell yield, long expansion times, and donor age-associated declines in cell quality, attention has also turned to thymic Tregs. These cells, sourced from discarded pediatric thymus tissue removed during cardiac surgeries, are less differentiated and may exhibit enhanced stability and suppressive capacity [57,58]. Early clinical experience includes a first-in-human case reporting safety and peripheral persistence of autologous thymic Tregs (thyTregs) in an infant heart transplant recipient, with ongoing phase I/II trials actively recruiting pediatric patients (THYTECH, NCT04924491)[59,60]. The practical availability of discarded thymic tissue adds value to this approach, although clinical efficacy data remain limited and further investigation is warranted to determine long-term outcomes.

The field of engineered Tregs has rapidly expanded, with products designed to improve antigen specificity and functionality. For instance, approximately 40-50% of the global population expresses HLA-A2. HLA-A2 mismatch between donor (positive) and recipient (negative) can increase rejection risk in transplantation. Autologous Tregs engineered with CARs targeting HLA-A2 have demonstrated enhanced suppressive activity and stability without altering native phenotype, showing promise in humanized mouse models [33]. TX200 (Sangamo Therapeutics) is an autologous CAR-Treg product targeting HLA-A2 for kidney transplant patients, currently undergoing clinical trials to assess safety and tolerability [61]. Similarly, QEL-001 (Quell Therapeutics) is an autologous HLA-A2–specific CAR-Treg engineered for liver transplant tolerance, featuring additional modifications including constitutive FOXP3 expression to lock in regulatory phenotype and an inducible suicide gene for safety [62]. This product is being evaluated in ongoing phase I/II trials aiming for immunosuppression withdrawal.

Beyond transplantation, engineered CAR-Tregs are advancing in autoimmune disease applications. PolTREG is developing a CAR-Treg for MS based on prior autologous polyclonal Treg experience. Sonoma Biotherapeutics is conducting clinical trials of SBT777101, an autologous CAR-Treg targeting citrullinated antigens in refractory rheumatoid arthritis and hidradenitis suppurativa, with discovery programs underway for celiac disease and IBD. GentiBio’s GNTI-122, an autologous TCR-engineered Treg product specific for islet antigens, is in early clinical development for T1D, employing advanced gene editing to stably express FOXP3 and modulate IL-2 signaling [63]. Abata Therapeutics’ ABA-101 targets myelin basic protein in MS patients carrying a high-risk HLA allele and is currently in phase 1 trials. A related product, ABA-201, is being developed for early T1D intervention.

Allogeneic Treg products, which offer potential advantages such as off-the-shelf availability, cost-effectiveness, and use of healthy donor cells, are also under active investigation. Type 1 regulatory cells (Tr1), which suppress immune responses through IL-10 and TGF-β secretion, are represented by TRX103 (Tr1X, Inc.), currently in phase I trials for graft-versus-host disease prevention post-hematopoietic stem cell transplant and for treatment-refractory Crohn’s disease (NCT06462365, NCT06721962). Orca-T combines purified hematopoietic stem and progenitor cells with Tregs and conventional T cells, aiming to balance immune reconstitution and GVHD prevention in hematologic malignancy patients undergoing transplant; it is currently in clinical trials, with phase 2 data showing improved GVHD-free survival with reduced rates of severe GVHD [64,65].

Lastly, induced pluripotent stem cells (iPSCs) represent a promising source for Treg generation due to their scalability and potential to produce large numbers of functional Tregs. FOXP3 can be transduced into iPSC-derived CD4 + T cells, yielding cells with regulatory phenotypes and suppressive function comparable to natural Tregs [66]. CAR-Tregs targeting HLA-A2 derived from iPSCs have demonstrated efficacy in vitro and in vivo in humanized GVHD models. These advances have led to product development collaborations, such as those between Quell Therapeutics and Cellistic^™, that are actively advancing iPSC-derived Treg therapies toward clinical application.

Across sourcing and manufacturing techniques, early-phase trials have established the safety of Treg-based cell therapy, while efficacy results remain variable. The field is evolving toward greater antigen specificity, phenotype stabilization, and scalable manufacturing, with engineered platforms poised to broaden applicability.

5. Characterizing identity and stability of Treg products

A comprehensive, universally accepted marker for Treg stability is lacking. As the master transcription factor of Tregs, FOXP3 is consistently used as a proxy for lineage stability due to its central role in maintaining Treg identity and function [6,67,68]. FOXP3 expression correlates with sustained Treg function both in vitro and in vivo [69,70]. Strategies to enhance FOXP3 expression include ectopic expression in CAR-Tregs [71], promoting demethylation of epigenetic FOXP3 regulatory regions [72–74], and treatment with metabolic inhibitors [75]. However, FOXP3 induction alone is insufficient to guarantee a stable Treg phenotype [74,76], and loss of FOXP3 does not always equate to total loss of suppressive function or transcriptional identity [77,78]. Co-expression of FOXP3 with the transcription factor Helios has been linked to enhanced Treg stability and suppressive capacity [79,80], though Helios deficiency alone does not significantly impair function [81], underscoring the need for more reliable markers of stability and fitness (Table 2).

Table 2.

Marker/Pathway	Role	Advantage/Implication
FOXP3	Master transcription factor	Surrogate for Treg identity/stability
Helios	Transcription factor	Enhances stability but non-essential
CD25/CD127	Surface markers	Define Tregs
CD45RA	Surface marker	Associated to epigenetically stable FOXP3 and Treg stability
HLA-DR, CD39, CD27	Functional/stability markers	Predict efficacy and suppressive ability
4-1BB+ / CD40L−	Functional/stability marker combination	Enriches for highly suppressive Tregs and distinguishes them from activated effector T cells
TIGIT+ / CD40L−	Functional/stability marker combination	Associated with enhanced suppressive capacity/improved functional stability
mTOR signaling	Metabolic regulation	Balances proliferation and stability
Wnt/β-catenin	Epigenetic regulation	Can destabilize under inflammation
Notch1 signaling	Transcriptional regulation	Destabilizes via Th1 skewing
IL-2/IL-6 signaling	Cytokine-mediated signaling	IL-2 promotes stability, IL-6 promotes instability

Surface markers are essential for monitoring Treg efficacy and isolating functionally stable populations. CD25 is a well-characterized marker of activated Tregs [82,83], while CD127 is typically down-regulated [83]. Additional markers such as HLA-DR+ [84], CD39+ [85], CD27+ [86], and CD226− [87] have been associated with enhanced suppressive capacity and lineage stability. Other phenotypic features linked to functional Treg subsets include low expression of CD40 ligand (CD40L) and expression of TIGIT and 4-1BB (CD137), which have been associated with highly suppressive or activated Treg populations [88,89]. Importantly, environmental factors such as inflammation can disrupt lineage stability [90–92]. Tregs expressing TNFRSF members CD27 and OX40 demonstrate reduced lineage plasticity [93], with CAR-Tregs engineered to express FOXP3-driven OX40-Ligand showing enhanced suppressive function and reduced proinflammatory gene expression [94]. Epigenetic regulators such as UHRF1 promote FOXP3-independent transcriptional stability. Inhibition of acetyl-CoA carboxylase 1 (ACC1) supports suppressive epigenetic and metabolic states, limiting conversion to proinflammatory Th17 and Th1 phenotypes [75].

Recent advances using single-cell technologies have uncovered novel, context-specific markers of Treg stability. For instance, CYP1A1 was identified as a metabolic regulator preventing Treg conversion to Th17-like cells upon IL-6 exposure [95]. Disease-specific Tregs in spondylarthritis and Crohn’s disease exhibit a TNFα gene signature [96,97], and Tregs in atherosclerosis can express Th1-like signatures with reduced suppressive capacity [98].

In contrast to T effector cells that switch to aerobic glycolysis upon activation, Tregs upregulate oxidative phosphorylation besides glycolysis to sustain their suppressive functions [99]. The PI3K-AKT-mTOR pathway regulates this metabolic switch, balancing proliferation and stability [99]. While mTOR activity is essential for Treg activation and function [100], excessive mTORC1 signaling drives HIF-1α-mediated glycolysis and Th17 differentiation, reducing FOXP3 and suppressive capacity [101–103]. Consequently, tightly controlled mTOR signaling is critical for optimal Treg fitness. Rapamycin, an mTOR inhibitor, enhances Treg function and improves outcomes in animal models [104–106], suggesting potential translational benefit. Seahorse metabolic flux assays could further inform in vivo Treg stability.

Canonical Wnt signaling via β-catenin dysregulates the Treg/Th17 balance, particularly in inflammatory contexts [107,108]. In murine models, β-catenin expression in Tregs promotes pro-inflammatory phenotypes and colitis [108,109]. TCF-1 interaction with FOXP3 limits proinflammatory gene expression; however, β-catenin stabilization alters epigenetic landscapes and enhances proinflammatory transcription, resulting in lineage plasticity [110,111]. In humans, MS patient Tregs with elevated β-catenin activation show IFN-γ/IL-10 imbalance via the β-catenin–SGK1–Foxo1/3 axis [112].

Notch1 signaling negatively regulates Treg function. Canonical and non-canonical Notch1 activation destabilizes Tregs via Th1 programming and CNS2 methylation, while Notch1 abrogation increases peripheral Treg frequency [113]. Notch1 blockade promotes long-term tolerance in cardiac transplant models and suppresses rheumatoid arthritis progression [114,115]. Screening for low activated β-catenin and Notch1, with high TCF-1, may identify Tregs resistant to inflammatory conversion.

Tregs depend on exogenous IL-2 for sustained FOXP3, CD25, and CTLA-4 expression [116,117]. IL-2/IL-2R/STAT-5 signaling maintains Treg identity, yet infused Tregs typically diminish within three months due to limited IL-2 [5]. Low-dose IL-2 can improve persistence but may also activate effector T cells [118]. Orthogonal IL-2/IL-2R constructs targeted to the FOXP3 locus enhance Treg function in murine T1D models [119]. In contrast, IL-6/STAT-3 signaling can destabilize Tregs by antagonizing FOXP3 expression and promoting proinflammatory transcriptional programs [120]. IL-6R and STAT-3 expression may thus serve as indicators of Treg stability.Future standardized approaches to data generation and analysis will enable development of comprehensive panels to monitor Treg stability in clinical contexts.

6. Manufacturing considerations

Manufacturing Tregs presents unique challenges distinct from conventional adoptive or CAR-T cell therapies. While CAR-T manufacturing protocols are well established, no consensus exists for Treg production. Key hurdles include low Treg yield from leukapheresis, isolation difficulties, and maintaining phenotypic stability and functional integrity throughout ex vivo expansion.

The initial cellular material profoundly impacts product quality. Most trials start with autologous or allogeneic leukapheresis collections, but Tregs are rare and require enrichment and expansion. The Treg subtype used (natural/thymic, peripheral/induced) carries inherent advantages and limitations. Advances in immuno-magnetic and flow sorting have improved starting population purity. Notably, CD45RA+ Tregs show greater phenotypic stability and FOXP3 expression and expand efficiently without rapamycin [121,122]. Marek et al. demonstrated better expansion kinetics in bulk Tregs compared to naïve and memory subsets cultured separately [123]. Defining critical quality attributes of starting material is vital to enhance clinical efficacy.

After isolation, ex vivo culture and expansion aim to reach therapeutic cell doses. Preservation of Treg phenotype during culture is crucial, as prolonged expansion reduces suppressive function [123]. Activation mimicking TCR co-stimulation may cause epigenetic destabilization, shifting toward effector phenotypes with altered function [124]. High proliferation rates may impair efficacy; notably, lower-fold expansion in Tregs from T1D patients correlated with better clinical outcomes [50]. Combinations like all-trans-retinoic acid improve function in Crohn’s disease Tregs [121]. In addition, recent work demonstrates that metabolic conditioning with lactic acid during ex vivo expansion can enhance Treg viability, purity, and suppressive function, including in CAR-engineered products [125]. Genetic engineering offers further avenues, with FOXP3 overexpression enhancing stability and suppressive function without compromising cytokine secretion or metabolic fitness [71]. Despite advances, lack of standardized culture protocols remains a critical barrier.

Effector T cell suppression assays and phenotypic characterizations are commonly used, yet clinical correlations remain inconsistent. Emerging phenotypic markers such as TNFR2 [126], Helios [127], and memory markers CD45RA and CCR7 may better link phenotype to function. Engineered Tregs with antigen specificity may enable more robust potency assays compared to polyclonal products.

Three critical technical challenges persist in Treg manufacturing: inconsistent starting material quality, absence of standardized culture protocols, and lack of validated potency assays. Addressing these issues should be a central focus in advancing the development of more effective Treg products. Genetic engineering and synthetic biology, as successfully applied in CAR-T cancer therapies, offer promising solutions to these hurdles [128].

7. In vivo models for characterizing Treg cell products

The evaluation of Treg-based immunotherapies in vivo requires the presence of a functional human immune system. While several murine models of colitis, such as adoptive T cell transfer from B6 to Rag-knockout B6 mice [129,130], the Muc2 knockout [131] and IL-10 knockout models [132], and TCR-mutant mouse models [133], have advanced our understanding of colitis pathogenesis and therapeutic strategies, translating these findings into clinical applications remains challenging due to species-specific differences in immune development, function, and response to treatment. To bridge this translational gap, researchers have developed various human immune system (HIS) mouse models that incorporate human immune components into immunodeficient mice. These models differ primarily in the source of human immune cells, either peripheral blood mononuclear cells (PBMCs) or hematopoietic stem cells (HSCs), as well as the thymic tissue used, including the native murine thymus, human fetal thymus, or fetal pig thymus. Distinct HIS mouse models have been established to suit different experimental needs.

The PBMC-HIS model, or the xenogeneic GVHD model, is widely used for rapidly evaluating human T cell–mediated immune responses in vivo [31,134–136]. This model involves IV injection of human PBMCs, with or without Treg depletion, into immunodeficient NSG (NOD-scid IL2Rγnull) mice, which lack functional murine T, B, and natural killer (NK) cells. Sublethal irradiation (1 Gy) prior to PBMC transfer is necessary for optimal engraftment and GVHD induction. Because human T cells from PBMCs are not tolerized to murine antigens, they recognize mouse tissues as foreign, leading to robust activation and expansion. Within a few weeks, human T cells typically comprise over 90 % of immune cells in peripheral blood and tissues, causing systemic inflammation and multi-organ tissue damage that recapitulate several hallmarks of clinical GVHD. The major strengths of this model are its speed, simplicity, cost-effectiveness, and scalability, making it a valuable platform to assess the suppressive function, persistence, and in vivo trafficking of human Tregs, including CAR-engineered Tregs. However, the model has notable limitations: the limited presence of human APCs and poor interaction between murine APCs and human T cells restrict the study of Treg-APC interactions and antigen-specific tolerance. The rapid onset of GVHD and associated morbidity create a short experimental window that limits long-term tolerance studies. Additionally, the absence of structured lymphoid tissues in NSG mice further constrains investigations into lymphoid-driven immune regulation [137,138]. Despite these limits, the model has been adapted to explore specific features of human Tregs, such as trafficking and antigen-specific activity. For instance, HLA-A2–directed CAR Tregs have been shown to selectively migrate to and persist within HLA-A2–positive human skin grafts transplanted onto the dorsal flank of engrafted mice, highlighting the model’s utility in evaluating engineered Treg behavior. Similarly, antigen-specific CAR Tregs have demonstrated trafficking to and persistence within human islet grafts in murine models, supporting their capacity for antigen-dependent localization and tissue engraftment in relevant transplant settings[139–141]. Typical experimental endpoints include monitoring clinical signs of GVHD, such as weight loss and skin lesions, histopathological analysis of affected organs (liver, intestine, skin, lung), flow cytometric assessment of human immune cell infiltration and activation, and cytokine profiling and evaluation of Treg stability or plasticity. In summary, the PBMC-HIS mouse model provides a rapid, and flexible system for assessing the function of human Tregs in a xenogeneic inflammatory environment, though is less suited for studying thymic development, antigen-specific selection, or long-term immune tolerance due to its limited APC representation and short lifespan.

In contrast, HSC-based HIS mouse models, such as the HU/HU or BLT (Bone marrow–Liver–Thymus) model, offer a more physiologically relevant system with robust multilineage human immune reconstitution [137,138,142]. This model is generated by transplanting human HSCs, typically derived from fetal liver or cord blood, together with a fragment of human fetal thymus under the kidney capsule of immunodeficient NSG mice. To ensure human T cell development occurs exclusively within the human thymic graft, the native murine thymus is surgically removed. After approximately 10 weeks, the thymic graft matures producing human T cells that develop within an organized microenvironment resembling native human thymus, including cortical and medullary regions and Hassall’s corpuscles. This setting supports physiological human thymopoiesis, generating diverse T cell subsets such as conventional CD4⁺ and CD8⁺ T cells and FOXP3⁺ Tregs, alongside a human APC compartment. The TCR repertoire develops similarly to humans, with high diversity at the double-positive thymocyte stage that narrows through positive and negative selection, culminating in mature CD4⁺, CD8⁺, and FOXP3⁺ Treg lineages [143,144]. Tregs arising in this model express canonical markers such as FOXP3, Helios, and CTLA-4, and demonstrate potent suppressive function both in vitro and in vivo [145]. The presence of a human thymic microenvironment combined with human and murine APCs enables the development of T cells broadly tolerant to both human and mouse antigens. However, tolerance is not absolute; by 25 to 40 weeks post-humanization, mice often develop a spontaneous autoimmune-like syndrome with weight loss, alopecia, and lymphocytic infiltration in multiple organs, including the intestine, skin, liver, and lungs. Colitis is a prominent feature of this late-onset inflammatory pathology. Although this autoimmune syndrome limits long-term studies of immune homeostasis, it provides a valuable platform for testing the in vivo efficacy of human Tregs, including engineered products, in preventing or treating colitis and systemic autoimmunity. This model thus offers a comprehensive and physiologically relevant system for investigating human T cell development, tolerance, and immune regulation.

Other HSC-based models include the MU/HU and SW/HU models, which serve as alternatives when human fetal thymus is unavailable. In the MU/HU model, human fetal or cord blood-derived HSCs are introduced into NSG mice that retain their native murine thymus. However, this murine thymus is underdeveloped and lacks the organized cortical and medullary architecture necessary for optimal human T cell selection, resulting in less physiologically relevant T cell repertoires [137,142,146]. The SW/HU model, which employs a fetal pig thymus graft, better supports human thymopoiesis, though MHC mismatch between the thymus and peripheral APCs remains a limitation [147]. Despite these challenges, both models can generate human T cells and APCs and eventually develop immune activation resembling autoimmunity, including colitis, making them useful for evaluating Treg-based therapies targeting systemic inflammation and tissue-specific immune responses.

This late-onset pathology is considered autoimmune rather than a classical xenogeneic graft-versus-host response. Although murine dendritic cells can migrate to the thymus and contribute to negative selection and promote tolerance to many mouse hematopoietic antigens, incomplete presentation of tissue-restricted antigens within the human or porcine thymic graft results in imperfect central tolerance[146]. As a result, autoreactive human T cells can emerge and target peripheral tissues, ultimately leading to multi-organ autoimmunity, including colitis.

The Personalized Immune (PI) HIS mouse model utilizes postnatal human HSCs, typically from adult bone marrow or G-CSF-mobilized peripheral blood, combined with variable thymic sources such as human fetal thymus, fetal pig thymus, or native murine thymus [137,148]. This model replicates patient-specific immune profiles, including genetic predispositions to autoimmune diseases, by transferring HSCs from individuals with conditions such as IBD. This enables in vivo study of how patient-specific risk alleles influence immune development, particularly mechanisms related to Treg dysfunction and plasticity. While valuable for personalized therapeutic evaluation, PI HIS mice face limitations due to the lower efficiency of adult HSCs in generating a diverse and functional immune repertoire compared to fetal-derived HSCs, with reduced thymopoiesis and T cell output [149,150]. Advances in HSC rejuvenation may enhance the robustness of this model.

While HSC-based HIS models typically do not develop colitis spontaneously in the early weeks post-humanization, immune dysregulation and autoimmune-like pathology, including colitis, commonly emerge at later time points (25–40 weeks) [142]. To accelerate disease onset and enable more controlled study windows, chemical or immunological triggers are often used. Trinitrobenzene sulfonic acid (TNBS) administered rectally is the most commonly used chemical agent, inducing T cell-mediated colitis characterized by epithelial damage, inflammatory cell infiltration, and cytokine production that mirrors Th1/Th17-driven intestinal inflammation. TNBS-triggered colitis in HIS mice has been used to study the contribution of human immune cells to intestinal inflammation and offers a controlled platform for evaluating anti-inflammatory therapies, including engineered or antigen-specific Tregs [151–154]. Other chemical triggers such as dextran sulfate sodium (DSS) and oxazolone, which induce epithelial barrier disruption and Th2-like colitis respectively, have also been used in mouse and HIS models, though their applicability in HIS mice remains less characterized [155–158]. Additionally, immunological stimuli such as systemic anti-CD3 antibody injection provoke polyclonal T cell activation and gut inflammation, modeling acute immune activation within humanized immune systems [159]. Overall, combining HSC-based HIS mouse models with defined chemical or immunological triggers offers a practical and clinically relevant approach to studying human immune responses during the early stages of colitis, facilitating mechanistic dissection and therapeutic testing before spontaneous pathology onset.

8. Future directions

We have reviewed the current understanding of Treg identification and function and highlighted emerging therapeutic applications, with particular emphasis on the translational challenges relevant to IBD. In IBD, adoptively transferred Tregs must function within a chronically inflamed intestinal microenvironment characterized by epithelial barrier dysfunction, sustained innate immune activation, and cytokine milieus rich in IL-6, TNFα, and IL-23, signals known to influence Treg stability and plasticity. Thus, major challenges to be addressed include achieving durable tissue distribution and persistence within the intestinal lamina propria, maintaining lineage stability under pro-inflammatory conditions, and ensuring antigen-specific potency in vivo.

Our group and others are invested in the identification of targeting strategies for engineered cells in a site-specific fashion, including both TCR- and CAR-based techniques, which may enable focused suppression at sites of intestinal inflammation while limiting systemic immunosuppression. Understanding how environmental signals shape Treg function remains critical, particularly in IBD where inflammatory cues can reprogram FOXP3⁺ cells toward effector-like phenotypes. Finally, and perhaps most importantly, widespread application and democratization of Treg therapeutics will depend upon improving scalability and reducing cost of goods. Enhancements in manufacturing technologies such as closed systems and robotics represent important steps forward.

Looking ahead, in vivo engineering approaches and intermittent, on-demand autologous Treg therapeutics may offer transformative potential for chronic autoimmune diseases such as IBD. The discovery of Tregs, recently recognized by the Nobel Prize in Physiology or Medicine, underscores how fundamental immunologic insights can translate into innovative therapeutic strategies and highlights the promise of engineered Treg approaches in reshaping treatment paradigms for intestinal inflammation.