Classically, CD4⁺ T_regs have been defined according to expression of FOXP3, considered a lineage-specifying transcription factor (TF), or the interleukin (IL)-2 receptor α chain (CD25). In a seminal study, Hiraoka et al. discovered that the prevalence of FOXP3⁺ T_regs increases during the progression from PanIN to advanced PDAC – at this latter stage, they constitute 35% (± 11%) of the total intratumoral CD4⁺ population (9, 15). Further, it is estimated that 54% (± 19%) of intratumoral T_regs are effector T_regs (eT_regs; CD45RA^-FOXP3^hiCD25^hi) (15). These cells express high levels of TIGIT, CTLA-4, ICOS, CD39, and HLA-DR, which are indicative of functional activation and potent immunosuppressive capacity (14, 15). This activated state has been attributed to sustained TCR stimulation, provided by the plethora of self- and quasi-self-antigens present in the inflammatory TME (43). However, a stable eT_reg phenotype is also dependent on the expression of Helios, a member of the Ikaros TF family. Indeed, intratumoral Helios⁺ T_regs exhibit significantly higher expression of FOXP3, compared to Helios^- T_regs (44).

Intratumoral eT_regs potently suppress CD8⁺ T cell-mediated immunity via the expression of co-inhibitory molecules e.g., CTLA-4, which prevents the functional maturation of dendritic cells (41); secretion of immunosuppressive cytokines e.g., IL-10, IL-35, and TGF-β; sequestration of IL-2, which hampers IL-2-dependent T cell activation; and the secretion of granzymes to lyse target CD8⁺ cells (45). In support of their immunosuppressive capacity in situ, spatial analyses reveal that 90% of T_regs reside in close proximity to a CD8⁺ T cell in the PDAC TME (15).

FOXP3⁺ T_regs exhibit extensive heterogeneity in PDAC. Strikingly, studies have discovered populations of FOXP3⁺ T_regs that, in addition to IL-10, secrete high levels of pro-inflammatory cytokines. For example, Chellappa et al. identified T_regs that co-express FOXP3 and RORγt: a factor that specifies the type-17 T-helper cell lineage (T_H17) (46). These cells retained markers associated with FOXP3⁺ T_regs e.g., CTLA-4, CD39, and ICOS, indicating an ability to robustly suppress anti-tumour immunity. However, through the simultaneous production of IL-17, these FOXP3⁺RORγt⁺ cells provide mitogenic signalling to transformed pancreatic epithelial cells, which upregulate the IL-17 receptor (47). Moreover, studies have identified populations of FOXP3^- T_reg-like cells that share expression of molecules classically associated with immunosuppressive T_reg functions (e.g., IL-10, CCR8, TIGIT, ICOS, CTLA-4) (48, 49). Barilla et al. demonstrated that the gene expression profile of one such population, termed T_r1 cells (CD49b, CD73, and AHR), was associated with decreased overall survival in PDAC patients (49). Furthermore, Whiteside et al. suggest that intratumoral T_eff cells may adopt this FOXP3^- T_reg-like phenotype following the ablation of FOXP3⁺ cells (48).

This profound heterogeneity likely explains conflicting reports regarding the overall contribution of T_regs in the pathophysiology of PDAC. One notable study reported an increased prevalence of T_regs in tumours of long-term PDAC survivors (24). Moreover, depletion of T_regs prior to the development of PanIN in KC mice has been shown to accelerate malignant progression (42). Conceivably, the use of different experimental systems, including varied methods for detecting and defining intratumoral T_regs, might accentuate specific T_reg-associated functions and thereby explain these conflicting reports. Moreover, studies have suggested that, as part of normal immune homeostasis, intratumoral T_regs accompany CD8⁺ T cell infiltrates (21, 42, 49), which may further obscure any relationship between the prevalence of intratumoral T_regs and a poor prognosis. Nevertheless, harnessing the therapeutic manipulation of T_regs will require a targeted approach, based on a detailed understanding of the heterogeneous functions ascribed to T_regs, and their spatiotemporal dynamics in the PDAC TME ( Figure 1 ). In addition, such an approach will reduce the systemic side-effects associated with T_reg-targeted immunotherapies.

This hypothesis is fortified by the discovery that apoptotic T_regs, defined by increased expression of Ki67 and cleaved caspase-3, exert immunosuppressive effects in the oxidative TME. They release large quantities of ATP, which is converted into adenosine via CD39 and CD73 – ectoenzymes that are expressed by T_regs and remain catalytically active after cell-death (50). Through the A_2A receptor, accumulating extracellular adenosine inhibits T_eff cell proliferation; induces immunosuppressive dendritic cells; and stabilises surviving T_regs (51). Thus, CD39 and CD73 expression correlate with a poor prognosis in patients with various solid tumours (52, 53). Importantly, this paracrine signalling pathway is likely to be operating in human PDAC, as intratumoral T_regs express high levels of CD39.

Allison and colleagues originally attributed the anti-tumour activity of anti-CTLA-4 monoclonal antibodies (mAbs) to the reinvigoration of dysfunctional T_eff cells (56). However, accumulating evidence suggests that anti-CTLA-4 mAbs can preferentially deplete CTLA-4^hi T_regs in vivo by antibody-dependent cellular cytotoxicity (ADCC) (57–60). Thus, in spite of the failure of prior clinical trials (4, 61), this novel mechanistic insight provides a rationale for the continued development of anti-CTLA-4 mAbs to treat PDAC. Clearly, however, this will necessitate re-engineering of existing anti-CTLA-4 mAbs; specifically, the fragment crystallisable (F_c) domain to enhance affinity for activatory F_cγ receptors and decrease affinity for inhibitory receptors, thereby promoting ADCC. This approach can be optimised with consideration of the relative abundance and distribution of specific F_cγRs on local effector cells; indeed, the engineering of anti-CTLA-4 mAbs in this manner has been shown to increase therapeutic activity in tumour-bearing mice (59, 62). Therefore, it is important that studies have identified intratumoral populations of F_cγRIIIA (CD16)-expressing natural killer and myeloid cells in human PDAC (14–16). Moreover, Agenus recently initiated a phase I/II trial to investigate botensilimab – an F_c-engineered anti-CTLA-4 mAb with enhanced affinity for F_cγRIIIA – in metastatic PDAC patients (NCT05630183).

Further testament to the widespread interest in strategies to selectively deplete intratumoral T_regs, there is renewed attention on the development of anti-CD25 mAbs. For example, Solomon et al. developed an anti-CD25 mAb (RG6292) that selectively depletes CD25^hi T_regs, whilst preserving CD25-STAT5 signalling required for T_eff cell-mediated anti-tumour immunity (63). Indeed, a phase I trial of RG6292, conducted in patients with advanced/metastatic solid tumours, indicated a manageable safety profile and preliminary anti-tumour activity (64). However, multi-omic analysis of patient-derived tumour samples obtained during treatment with RG6292 is required to confirm this proposed mechanism of action in vivo.

Since the discovery of CTLA-4 and PD-1, studies have identified a plethora of immune checkpoints – both inhibitory (e.g., TIGIT, LAG-3, TIM-3) and co-stimulatory (e.g., ICOS, OX40, GITR, 4–1BB) – that might be exploited therapeutically to augment anti-tumour immunity. In PDAC, TIGIT and ICOS are expressed at high levels on intratumoral eT_regs (14, 15). TIGIT is also expressed, albeit at lower levels, by dysfunctional T_eff cells, whereas ICOS is induced upon the activation of intratumoral T_eff cells (14, 15, 27). Therefore, anti-TIGIT and agonistic ICOS mAbs might have a dual mechanism of action: the re-invigoration of dysfunctional T_eff cells and selective depletion of activated T_regs (65). However, achieving the optimal balance between these mechanisms will require F_c engineering to effectively engage specific F_c receptors (66).

Tiragolumab (IgG1κ anti-TIGIT) has demonstrated tolerability and preliminary anti-tumour activity in patients with advanced solid tumours (67, 68). Consequently, two early-stage trials are investigating anti-TIGIT mAbs, incorporated into combinatorial regimens, for the treatment of metastatic PDAC (NCT03193190, NCT05419479). By contrast, a phase I/II trial, investigating vopratelimab (IgG1κ agonistic ICOS) for the treatment of advanced solid tumours, including three PDAC patients, reported limited efficacy (69). However, on-treatment emergence of ICOS^hi CD4⁺ T_eff cells was associated with therapeutic responses, suggesting that vopratelimab might indeed re-invigorate dysfunctional T_eff cells in patients through ICOS activation. More generally, this illustrates that multi-omic analyses of on-treatment patient-derived samples during clinical trials may further advance our understanding of the PDAC immune landscape.

The development of strategies for selectively drugging T_regs has been the subject of considerable research. One potential target is Helios; in PDAC patients, Helios⁺ T_regs are significantly enriched in the TME (70). Moreover, T_reg-intrinsic deletion of Helios has been shown to enhance anti-tumour immunity in tumour-bearing mice (71). Interestingly, Helios-deficient T_regs acquire a T_eff phenotype including the production of pro-inflammatory cytokines (e.g., IFN-γ), which is attributed to downregulation of FOXP3 and de-repression of T_H1/T_H2 lineage determinants (43). In the absence of the stabilising influence of Helios, it appears that the inflammatory TME promotes the trans-differentiation of T_regs into activated T_eff cells. Intriguingly, this novel T_eff population is equipped with an inherently self-reactive TCR repertoire, which might be expected to direct a potent immune response against ‘quasi-self’ tumour antigens.

Transcription factors are traditionally considered difficult to drug. However, several recent studies have described small-molecules that selectively enhance the proteasomal degradation of Helios (72, 73). Future in vivo studies must determine whether these small-molecules can selectively destabilise activated intratumoral eT_regs; one clinical trial is currently evaluating this approach in advanced solid tumours (NCT03891953).

The origin of intratumoral FOXP3⁺ T_regs is unclear – they may differentiate locally from T_eff cells or be recruited from the circulation. For the latter, targeting chemokine signalling axes (e.g., CCL2-CCR4; CCL5-CCR5) that can recruit T_regs into the PDAC TME is of interest. However, this strategy has proved disappointing thus far; clinical trials investigating mogamulizumab (IgG1 anti-CCR4) reported off-target depletion of T_H2/T_H17 cells, reflecting heterogeneous expression of CCR4 (74, 75).

It is notable, therefore, that intratumoral eT_regs uniquely express CCR8 (76). However, functional blockade of CCR8 does not affect T_reg recruitment; they acquire CCR8 expression in the TME, perhaps suggesting that this axis mediates retention of intratumoral T_regs (77). Nevertheless, CCR8 constitutes a target for the selective depletion of intratumoral eT_regs in PDAC. Pre-clinical studies have demonstrated that anti-CCR8 mAbs profoundly suppress tumour growth in tumour-bearing mice (76, 78). Further, this response coincided with the expansion of intratumoral CD4⁺ T_eff cells and the preservation of systemic T_reg populations, which may mitigate the risk of autoimmune-related adverse events. Currently, eight early-stage trials are investigating anti-CCR8 mAbs for the treatment of advanced solid tumours (NCT04895709, NCT06131398, NCT05635643, NCT05537740, NCT05007782, NCT05518045, NCT05101070, NCT05935098).

Apoptotic T_regs convert ATP to adenosine, an immunosuppressive metabolite, via ectoenzymes that remain catalytically active after cell-death. This raises the paradoxical possibility that the therapeutic depletion of T_regs might not limit T_reg-cell-mediated immunosuppression. This discovery provided impetus to the development of immunotherapies that target the adenosinergic pathway: CD39, CD73, and the A_2A/A_2B receptors. It is hoped that these therapies will synergise with T_reg-targeted approaches, or other immunotherapeutic modalities, to induce potent anti-tumour immunity. To date, however, attempts to target this pathway with anti-CD73 mAbs have demonstrated no clinical benefit for PDAC patients; a phase-II trial investigating the combination of anti-CD73, anti-PD-L1, and chemotherapy revealed comparable efficacy to chemotherapy alone (79).