Regulatory T cells sabotage anti-tumor γδ T cells by creating IL-2–deficient environments

γδ T cells represent a conserved and unique lymphocyte lineage that bridges innate and adaptive immunity (Vantourout and Hayday, 2013). Unlike conventional αβ T cells, γδ T cells are not restricted by major histocompatibility complex (MHC)–mediated antigen presentation and often respond with innate-like dynamics, thus constituting nonredundant immune effectors attractive for immunotherapeutic exploitation (Hayday et al., 2024). We and others have shown that mouse γδ T cells are functionally heterogeneous, with two main effector subsets emerging in multiple models of infection or cancer: IFNγ-producing and IL-17–producing γδ T cells (Silva-Santos et al., 2019; Ribot et al., 2021). These distinct γδ T cell subsets are developmentally programmed in the thymus (Jensen et al., 2008; Ribot et al., 2009), home to many peripheral (lymphoid and nonlymphoid) tissues (Ribot et al., 2021), and employ different mechanisms of activation upon challenge (Inácio et al., 2025). In the context of cancer, IL-17⁺ γδ T cells have been associated with tumor progression, whereas IFNγ⁺ γδ cells inhibited tumor growth in various syngeneic mouse models (Gao et al., 2003; Silva-Santos et al., 2019; Lopes et al., 2021; Reis et al., 2022). Importantly, human γδ T cells are highly biased toward IFNγ (but not IL-17) production, which underscores their potential as effectors of cancer immunotherapy (Hayday et al., 2024; Silva-Santos et al., 2019).

Despite their well-established importance in cancer immunity, the functional regulation of γδ T cells, particularly IFNγ⁺ γδ T cells, within the tumor microenvironment (TME) remains poorly understood. Previous studies have mostly focused on cell-intrinsic inhibitory pathways, particularly immune checkpoints like PD-1, LAG-3, or TIM-3 (Hoeres et al., 2019; Gao et al., 2023). Concerning extrinsic factors, neutrophils have been shown to suppress IFNγ production by human γδ T cells in vitro (Sacchi et al., 2018; Sabbione et al., 2014), but the critical in vivo immune regulators of IFNγ⁺ γδ T cells remain to be identified, which may hold the key to enhancing their performance in cancer immunotherapy.

Regulatory T (Treg) cells, namely CD4⁺ T cells expressing the master transcription factor Foxp3, are well-established suppressors of effector αβ T cell responses in cancer (Tanaka and Sakaguchi, 2017). Removal of Treg cells enhances tumor control by increasing effector αβ T cell proliferation and cytokine production and is a key target of lymphodepletion strategies prior to adoptive cell transfer (Gattinoni et al., 2005; Wrzesinski et al., 2010). These strategies function in part by eliminating “cytokine sinks”—Treg cells and other IL-2/IL-15–consuming populations—to promote homeostatic expansion and function of the adoptively transferred effector T cells. Clinical observations further indicate that in the context of chimeric antigen receptor–transduced T (CAR-T) cell therapies, poor CAR-T cell expansion and persistence correlate with elevated intratumoral Treg levels (Good et al., 2022). Although different Treg-derived anti-inflammatory molecules have been shown to suppress γδ T cells in other contexts (Okeke and Uzonna, 2019), their influence on anti-tumor γδ T cell responses has never been decisively addressed in vivo.

In this study, we fill this gap by systematically evaluating the impact of Treg depletion on tumor-infiltrating γδ T cell populations using syngeneic cancer models. We demonstrate that Treg cells selectively suppress anti-tumor IFNγ-producing (but not pro-tumor IL-17–producing) γδ T cells by limiting IL-2 availability in the TME, a cytokine required for type 1 responses. This suppression can be reversed by the administration of an IL-2Rβγ_c–selective agonist, which works not only in mice but also with human γδ T cells, notably enhancing the in vitro and in vivo functions of expanded γδ T cells used in adoptive cell therapy. Altogether, these findings uncover a novel immunoregulatory axis between Treg cells and anti-tumor γδ T cells, which can be targeted to enhance the activity of upcoming γδ T cell–based immunotherapies for cancer.

To investigate the effects of Treg depletion on γδ T cell responses to tumors, we employed a syngeneic breast cancer model involving the orthotopic implantation of the E0771 breast cancer cell line into Foxp3-diphtheria toxin (DTx) receptor (DTR) mice, which express the human DTR under control of the Foxp3 promoter. Sequential administration of DTx one week after tumor implantation, following confirmation of palpable tumors (Fig. 1 A), enabled transient and selective depletion of Treg cells across multiple tissues (Kim et al., 2007), including tumors (Fig. 1 B). In agreement with previous literature (Bos et al., 2013; Joshi et al., 2015; Simpson et al., 2013), Treg depletion led to tumor control (Fig. 1 C). While total numbers of γδ T cells remained unchanged (Fig. 1 D), Treg depletion induced a drastic phenotypic switch in tumor-infiltrating γδ T cell subsets (Fig. 1 E). Multicolor spectral flow cytometry analysis classified γδ T cells within tumors into six different clusters, revealing a pronounced expansion of those (C1–C3) presenting features of type 1 (IFNγ-producing) response in DTx-treated mice (Fig. 1, E–G). Consistently, the percentages and numbers of IFNγ⁺ γδ T cells, as well as the levels of IFNγ production by these cells (quantified as mean fluorescence intensity), were increased after Treg depletion (Fig. 1 H). By contrast, the levels of IL-17–producing γδ T cells remained virtually unchanged (Fig. 1, G and I), thus resulting in an increased IFNγ⁺/IL-17⁺ γδ T cell ratio (Fig. 1 J). Augmented Ki-67 expression and BrdU incorporation highlighted the increased proliferation of IFNγ⁺ γδ T cells underlying the expansion of this subset (Fig. 1 K). Importantly, we also observed a similar accumulation of proliferative IFNγ⁺ γδ T cells in a MC38 (colorectal) subcutaneous tumor model, correlative with an enhanced tumor control upon Treg cell depletion (Fig. S1, A–C), suggesting that this phenomenon is preserved across tumor models. The observed IFNγ⁺ γδ T cell expansion was more pronounced in tumors and tumor-draining LNs (dLNs) than in the spleen, suggesting a preferential effect in the tumor context (Fig. 1 L). By contrast, in tumor-free mice, IFNγ⁺ γδ responses were found systemically upon Treg depletion (Fig. S1 D and E). One possibility is that the TME locally retains or preferentially activates γδ T cells following Treg depletion, thereby restricting their systemic dissemination. In tumor-free mice, this constraint would be absent, allowing IFNγ⁺ γδ T cell responses to manifest systemically. However, the mechanisms underlying this phenomenon remain to be determined. Notably, the proportions of different TCRγ-specific subsets was altered in DTx-treated mice, with heightened representation of Vγ1⁺ cells in the tumors, compared with decreased Vγ4⁺ cells and unchanged Vγ1⁻Vγ4⁻ cells (Fig. 1 M). In addition to their upregulation of IFNγ expression in the absence of Treg cells, tumor-infiltrating γδ T cells also increased their expression of the cytotoxic enzyme, granzyme B (Fig. 1 N), supporting enhanced anti-tumor potential. To demonstrate a direct impact of Treg cells on γδ T cell cytotoxicity, we co-cultured them for 24 h with E0771 breast cancer cells and observed impaired tumor-cell killing by γδ T cells when in the presence of Treg cells (Fig. 1 O). These data demonstrate that Treg cells selectively suppress type 1 cytotoxic, but not type 17, γδ T cell responses to tumors.

Since Treg cell depletion also induces type 1 responses in αβ T cells (Fig. S2, A–C), we next aimed to evaluate the specific contribution of IFNγ⁺ γδ T cells to the observed tumor control. As Vγ1⁺ cells were the main IFNγ-producing γδ T cell subset expanded upon Treg elimination (Fig. 1 M), we administered (intraperitoneally) an anti-Vγ1 TCR monoclonal antibody (αVγ1 mAb) together with the DTx treatment to selectively deplete Vγ1⁺ γδ T cells (Zheng et al., 2017) (Fig. 2 A). Notably, depletion of Vγ1⁺ cells partially reversed the tumor control achieved by Treg depletion, resulting in increased tumor growth, which demonstrated a nonredundant role for Vγ1⁺ cells in the anti-tumor response elicited by Treg ablation (Fig. 2 B). As expected, αVγ1 mAb treatment did not significantly affect the frequency or extent of Treg depletion (Fig. 2 C) but effectively reduced the proportion of intratumoral Vγ1⁺ cells by approximately two-thirds (Fig. 2 D). Importantly, this intervention reduced the frequency of IFNγ⁺ γδ T cells in DTx-treated mice to levels observed in control animals (Fig. 2 E), thus supporting Vγ1⁺ cells as the critical IFNγ⁺ γδ T cells in this context. Most interestingly, transient Vγ1⁺ cell depletion also resulted in reduced frequencies of IFNγ-producing CD4⁺ and CD8⁺ αβ T cells, supporting a model in which Vγ1⁺ γδ T cells orchestrate IFNγ-based anti-tumor immunity unleashed following Treg depletion (Fig. 2 F), consistent with the established γδ T cell ability to modulate broader immune responses (Arias-Badia et al., 2024). While CD8⁺ T cells also contribute to tumor control following Treg depletion (Fig. S2, D–F), CD8 depletion did not affect γδ T cell responses (Fig. S2 G), indicating that IFNγ⁺ γδ T cells act nonredundantly and upstream in this anti-tumor immune cascade. Whether this effect is mediated directly by γδ T cell–derived IFNγ remains to be determined, as formally addressing this question would require conditional ablation of IFNγ specifically in γδ T cells. These data position IFNγ-producing Vγ1⁺ γδ T cells as effectors of anti-tumor immunity upon Treg depletion therapy.

To investigate how Treg cells suppress anti-tumor γδ T cells, we co-cultured sorted γδ T cells with Treg cells in the presence of αCD3/αCD28 stimulation under defined conditions targeting key immunosuppressive pathways employed by Treg cells (Tanaka and Sakaguchi, 2017). We found the presence of Treg cells directly inhibited the proliferation of IFNγ⁺ γδ T cells. However, neither blockade of Treg-derived IL-10 or IL-35 nor inhibition of adenosine production restored IFNγ⁺ γδ T cells proliferation (Fig. S3 A). In line with this, in vivo IL-10 neutralization neither enhanced IFNγ⁺ γδ T cell responses nor improved tumor control (Fig. S3, B–D). We cannot rule out a contribution of IL-10 to Treg-mediated suppression in other contexts, as the effect of this cytokine is highly variable across tissue environments. In the experimental system employed, our findings suggest that Treg-mediated suppression of IFNγ⁺ γδ T cells is independent from IL-10, IL-35, or adenosine. On the other hand, the increased expression of IL-2 receptor α (IL-2Rα/CD25) on IFNγ⁺ γδ T cell clusters in tumors lacking Treg cells (Fig. 1 G) could reflect enhanced IL-2 signaling. This observation is particularly relevant because a well-established mechanism of Treg suppression of both αβ T cells and NK cells is IL-2 deprivation (Chinen et al., 2016). Treg cells constitutively express the high-affinity trimeric IL-2 receptor, composed of IL-2Rα (CD25), IL-2Rβ/IL-15Rβ (CD122), and the common γ chain (IL-2Rγ/CD132). In contrast, effector αβ T cells lack CD25 expression at steady state, resulting in lower receptor affinity for IL-2. Consequently, Treg cells effectively outcompete effector cells for limited IL-2 availability, thereby constraining their activation and function (Kastenmuller et al., 2011). In agreement with previous reports (Corpuz et al., 2017; Shibata et al., 2008), we found that peripheral IFNγ-committed γδ T cells at steady-state express higher levels of CD122 compared with IL-17–committed subsets, which in turn show a relative preference for CD25 expression, although not to the level observed in Treg cells. Co-expression of CD25 and CD122 remained largely restricted to Treg cells (Fig. S4 A). Importantly, IL-2 receptor expression was not significantly altered following 3-h PMA/ionomycin stimulation, allowing reliable assessment of cytokine-producing effector subsets (Fig. S4 B).

After receiving activation signals 1 and 2, autocrine IL-2 induces STAT-5 phosphorylation and subsequent CD25 expression in effector T cells to gain affinity for IL-2 and boost expansion and effector function (Shouse et al., 2024). This mechanism is particularly important for IFNγ⁺ γδ T cells, whose effector function is promoted by IL-2 signaling, and less so for IL-17–producing subsets (Ribot et al., 2012). In tumor-bearing mice, IFNγ⁺ γδ T cells exhibited higher levels of CD25 within tumors than in the periphery; however, they were still far from those found on Treg cells. Strikingly, Treg depletion upregulated CD25 expression on IFNγ⁺ γδ T cells in the tumors and tumor dLNs (but not in the spleen) (Fig. 3 A), where an increased frequency of IFNγ⁺ γδ T cells was also observed (Fig. 1, H and L), suggesting that enhanced IL-2 signaling accompanies the local expansion of these effector cells in the absence of Treg cells. IL-2R upregulation was observed across Vγ-based γδ T cell subtypes in DTx-treated mice (Fig. S4 C). By contrast, the levels of the IL-2R on IL-17⁺ γδ T cells remained unchanged across the different tissues, including tumors, following Treg depletion (Fig. S4 D). These data suggest that the presence of Treg cells may constrain IL-2 signaling in IFNγ⁺ (but not IL-17⁺) γδ T cells in tumors and tumor dLNs.

To directly assess Treg interference with IL-2 signaling on γδ T cells, we co-cultured sorted Treg cells and γδ T cells and exposed them to a 10-min pulse of either recombinant IL-2 or the IL-2Rβγ_c–selective agonist neoleukin-2/15 (Neo2/15), which activates IL-2R signaling independently of CD25 (Silva et al., 2019). In the absence of Treg cells, both IL-2 and Neo2/15 elicited robust and comparable STAT5 phosphorylation in IFNγ-committed (Vγ1⁺) γδ T cells, demonstrating that Neo2/15 can act directly on this subset. Notably, the presence of Treg cells selectively suppressed IL-2–driven STAT5 activation, but allowed for Neo2/15-induced signaling in Vγ1⁺ γδ T cells (Fig. 3 B and Fig. S5 A). Of note, STAT5 phosphorylation levels in Treg cells remained unchanged regardless of the presence of γδ T cells (Fig. S5 B). These findings support the notion that IL-2 preferentially binds to the high-affinity trimeric IL-2 receptor on Treg cells, whereas Neo2/15 signaling on γδ T cells circumvents this advantage.

In extended 3-day co-cultures with αCD3/αCD28 stimulation (that induces IL-2 production by IFNγ⁺ γδ T cells [Ribot et al., 2012]), we confirmed that IFNγ⁺ γδ T cells are directly suppressed by Treg cells (Fig. 3, C and D). Importantly, supplementation with either exogenous IL-2 or Neo2/15 reversed Treg-mediated suppression of both proliferation and IFNγ production by IFNγ⁺ γδ T cells (Fig. 3, C and D). Neutralization of endogenous IL-2 recapitulated the suppressive effects of Treg cells, and the addition of Treg cells did not further inhibit IFNγ⁺ γδ T cell responses under these conditions in vitro, supporting IL-2 deprivation as the primary mechanism of suppression (Fig. 3, C and D). By contrast, the proliferation of IL-17⁺ γδ T cells, which was already modest under these stimulation conditions, remained largely unaffected by either Treg cells or IL-2/Neo2/15 supplementation (Fig. 3 E), in agreement with this subset being less dependent on IL-2 signaling (Shibata et al., 2008; Ribot et al., 2012). Furthermore, IL-2 deprivation impaired the cytotoxic function of γδ T cells, as supplementation with IL-2 or Neo2/15 restored their ability to kill E0771 tumor cells, overcoming Treg-mediated suppression in in vitro cytotoxicity assays (Fig. 3 F). These in vitro data clearly suggest that IL-2 deprivation is a key Treg-based mechanism of suppression of anti-tumor type 1 cytotoxic γδ T cells.

To determine whether elevated IL-2 availability in the absence of Treg cells drives the in vivo expansion of anti-tumor γδ T cells, we devised a strategy to neutralize IL-2 concomitantly with Treg depletion in E0771 tumor-bearing mice. Beginning on the day of DTx administration, mice were treated daily with equal amounts of two mAbs: one targeting the CD25-binding site of IL-2 (clone S4B6-1) and the other targeting the CD122-binding site (clone JES6-1A12) (Fig. 4 A). This dual blockade ensures effective IL-2 neutralization and prevents the formation of IL-2/anti-IL-2 immune complexes, which paradoxically would enhance receptor signaling by stabilizing the cytokine through its unbound domain (Boyman et al., 2006; McNally et al., 2011; McKinstry et al., 2014). Importantly, this strategy neutralizes IL-2 without directly impacting its receptor. Notably, IL-2 neutralization partially restored tumor growth in Treg-depleted mice, with no changes in Treg-sufficient controls (Fig. 4 B). While the frequencies of tumor-infiltrating Treg cells and total γδ T cells remained minimally affected by transient IL-2 neutralization (Fig. 4, C and D), CD25 upregulation on γδ T cells following Treg-depletion was abrogated upon anti-IL-2 administration, supporting attenuated IL-2 signaling in this population (Fig. 4 E). Accordingly, IL-2 neutralization impaired both αβ T cell responses (Fig. S5 C) and the effector function of γδ T cells, markedly reducing both IFNγ and granzyme B production (Fig. 4 F), as well as the proliferation of IFNγ⁺ γδ T cells (Fig. 4 G). Therefore, enhanced IL-2 availability following Treg depletion unequivocally contributes to the activation and expansion of type 1 cytotoxic γδ T cells within the TME.

To aim at therapeutic application by overcoming Treg-mediated suppression without going through Treg depletion, and after showing that the IL-2Rβγ_c agonist Neo2/15 directly activates type 1 cytotoxic γδ T cells, we treated mice daily with this compound starting on day 7 after tumor inoculation (Fig. 5 A). As expected, Neo2/15 treatment significantly reduced tumor growth (Fig. 5 B). Consistent with previous reports (Silva et al., 2019), Neo2/15 had minimal impact on Treg cell frequency (Fig. 5 C). Although the overall frequency of tumor-infiltrating γδ T cells remained largely unchanged (Fig. 5 D), these cells upregulated CD25 expression following Neo2/15 administration (Fig. 5 E), coinciding with a marked shift toward a type 1 cytotoxic phenotype characterized by increased IFNγ production and granzyme B expression (Fig. 5 F). Importantly, the activation of IFNγ⁺ γδ T cells was substantially responsible for the therapeutic effect of Neo2/15, as depletion of IFNγ-committed Vγ1⁺ γδ T cells significantly impaired Neo2/15-driven tumor control (Fig. 5, G and H). In conclusion, Neo2/15 treatment bypasses Treg-mediated IL-2 sequestration and promotes anti-tumor murine γδ T cell responses contributing to tumor control.

To determine whether the Treg suppression mechanism found in mice is conserved in humans, we isolated Treg cells (sorted as CD4⁺CD25⁺CD127⁻) and γδ T cells (sorted as CD4⁻TCRVδ1⁻TCRVδ2⁺, corresponding to the Vδ2 subset that predominates in circulation) from human peripheral blood mononuclear cells (PBMCs) and co-cultured them in the presence or absence of recombinant human IL-2 or Neo2/15. Consistent with our findings with mouse cells, human Treg cells prevented CD25 upregulation and inhibited both the proliferation and IFNγ production of Vδ2⁺ γδ T cells (Fig. 6 A and Fig. S5 D). Notably, this suppression was abolished by adding exogenous IL-2 or Neo2/15, which were both able to stimulate potent Vδ2⁺ γδ T cell proliferation and effector functions even in the presence of Treg cells (Fig. 6 A and Fig. S5 D).

Finally, we hypothesized that Treg cells could represent a hurdle to emerging adoptive γδ T cell therapies. Our laboratory previously developed Delta One T (DOT) cells, a Vδ1-enriched γδ T cell therapy product generated upon expansion of blood-derived human γδ T cells, and demonstrated its efficacy in xenograft models of leukemias and colorectal cancer (Almeida et al., 2016; Di Lorenzo et al., 2019; Blanco-Domínguez et al., 2025; Mensurado et al., 2024). We now found that Treg cells are able to suppress the activation and proliferation of DOT cells in vitro and that supplementation with IL-2 or Neo2/15 fully restored their activity (Fig. 6 B). Moreover, the cytotoxic function of DOT cells against MDA-MB-231 triple-negative breast cancer cells was impaired following co-culture with Treg cells, while stimulation of the IL-2 receptor via Neo2/15 bypassed Treg-mediated suppression (Fig. 6 C).

To assess whether IL-2Rβγ_c engagement could enhance the therapeutic activity of DOT cells in vivo, we orthotopically implanted MDA-MB-231 breast cancer cells into the mammary gland of NSG mice engineered to produce basal levels of human IL-15. Mice were treated intravenously with DOT cells, with or without concomitant intraperitoneal administration of Neo2/15 (Fig. 6 D). While DOT cells alone achieved partial tumor control, co-administration of Neo2/15 significantly enhanced their anti-tumor efficacy (Fig. 6 E).

Collectively, these findings indicate that human Treg cells can suppress the activity of both endogenous (Vδ2) and expanded (DOT) γδ T cells and that promoting IL-2Rβ signaling can boost γδ T cell/DOT cell performance in adoptive cell therapy.

γδ T cells are increasingly recognized as important effectors in anti-tumor immunity, particularly through their capacity to produce abundant IFNγ and exert potent cytotoxicity in complementary modes to αβ T cells or NK cells (Hayday et al., 2024). In this study, we unraveled a novel immunoregulatory axis between Treg cells and anti-tumor γδ T cells, which can be targeted via IL-2Rβγ_c signaling to enhance the efficacy of γδ T cell–based cancer immunotherapy.

First focusing on a syngeneic mouse model of breast cancer, we found that induced Treg depletion unleashed the proliferation of IFNγ-committed (mostly Vγ1⁺) cells in the tumor bed. Importantly, Treg cell depletion did not impact γδ T cell subsets biased toward IL-17 production. We associated this differential impact on γδ T cell subsets to the corresponding expression of distinct IL-2R chains, which determine binding affinity, and functional dependence on IL-2 signaling. Nevertheless, we acknowledge that the E0771 and MC38 models used in the study are not particularly well-suited to study IL-17⁺ γδ T cells, since they are biased toward an IFN-γ response derived from activated Vγ1⁺ cells. In addition, a limitation of the E0771 breast cancer model is that it does not allow discrimination between γδ T cell responses originating from the mammary gland epithelium versus those arising from the surrounding fat pad, thus the precise tissue source of the proliferating IFNγ⁺ γδ T cells was not determined.

We demonstrate that Treg cells suppress tumor-associated IFNγ⁺ γδ T cells by limiting their access to IL-2 that is critical for their proliferation and effector functions. Although Treg cells cannot produce IL-2 themselves, they depend on it for their survival and function, and their constitutive expression of the high-affinity IL-2 receptor incorporating the α chain (CD25) allows them to efficiently sequester this cytokine, thereby limiting its availability to effector T cells and suppressing their activation (Pandiyan et al., 2007). A genetic approach to uncouple IL-2 signaling from Treg function by expressing a constitutively active STAT5b in IL-2R–deficient Treg cells revealed that IL-2 limitation is necessary to restrain CD8⁺ T cell expansion (Chinen et al., 2016). In CD4⁺ T cells, this mechanism limits IFNγ and granzyme B expression via T-bet and Blimp-1, respectively (Śledzińska et al., 2020). On the other hand, CD8⁺ T cell expansion was promoted by a mAb specific for IL-2 administered in combination with recombinant IL-2, leading to an increased biological activity via the formation of stabilizing complexes (Boyman et al., 2006). Moreover, Treg depletion with a CD25 antibody that preserves IL-2 bioactivity led to augmented CD8⁺ T cell responses and enhanced anti–PD-1 mAb therapeutic activity (Galvez-Cancino et al., 2025).

Our data now extend this “IL-2-sink” mechanism to γδ T cells, positioning IL-2 competition as a key regulatory mechanism against IFNγ⁺ γδ T cells. Previous work demonstrated that autocrine IL-2 production is essential for mouse γδ T cell survival and proliferation (Ribot et al., 2012) and that stimulation with IL-2–induced IFNγ and cytolytic activity in human γδ thymocytes (Ribot et al., 2014). In line with this, we now show that increased IL-2 availability promotes robust expansion of IFNγ⁺ γδ T cells, reinforcing the notion that IL-2 is a critical limiting factor for their activation and anti-tumor function. We note, however, that Treg depletion also induces strong αβ T cell activation. In this regard, concomitant depletion of Treg cells and CD8⁺ T cells did not impair the accumulation of IFNγ⁺ γδ T cells, indicating that CD8⁺ T cells are dispensable for this response, although a contribution from CD4⁺ T cells, as a major source of IL-2, remains probable. Interestingly, the depletion of the main IFNγ⁺ γδ T cell subset, Vγ1⁺ cells, resulted in reduced frequencies of IFNγ-producing CD4⁺ and CD8⁺ αβ T cells, consistent with our previous findings in a malaria infection model (Ribot et al., 2019) and with the established concept that γδ T cells modulate broader immune responses (Vantourout and Hayday, 2013; Arias-Badia et al., 2024).

Recent advances in cytokine engineering offer promising tools to circumvent Treg-mediated suppression. For example, an IL-2 variant (AB248) selectively targeted to CD8⁺ T cells was recently developed, enabling robust effector T cell responses with limited Treg or NK cell activation (Moynihan et al., 2024). Another example of cytokine engineering is the development of the Neo 2/15 agonist used in this study, which takes advantage of the differential signaling properties of IL-2 and IL-15. While IL-2 signals through the trimeric high-affinity receptor (IL-2Rαβγ_c), primarily expressed by Treg cells and activated T cells, IL-15 signals through the shared IL-2Rβγ_c complex and is presented in trans by IL-15Rα–expressing cells (Tamzalit et al., 2014). Our data with Neo2/15, which binds to the IL-2Rβγ_c heterodimer and elicits downstream cell signaling independent of IL-2Rα and IL-15Rα, clearly showed its ability to circumvent Treg-mediated inhibition of both mouse and human γδ T cells, thus providing the proof of concept for its use in combination with upcoming γδ T cell–based therapies, including DOT cells. These are expanded in response to IL-2 or IL-15, and the latter is essential for their enhanced cytotoxic capacity (Almeida et al., 2016) and for in vivo persistence (Sánchez Martínez et al., 2022). Although other lymphocyte subsets may also respond to Neo2/15, our in vitro and in vivo data firmly establish mouse and human γδ T cells as direct targets of this engineered cytokine. Thus, the co-administration of Neo2/15 with adoptive DOT cell therapy may enhance clinical efficacy, in line with our preclinical data obtained using the MDA-MB-231 triple-negative breast cancer xenograft model. We acknowledge, however, that this xenograft model lacks Treg cells, whose modulatory effects on γδ T cell responses were therefore only assessed in vitro, representing a limitation of our study. Interestingly, CAR-NK cells engineered to secrete Neo2/15 enhanced their cytotoxicity and persistence in different solid tumor models, suggesting that similar armoring strategies on DOT (or other γδ T) cell therapies may also bypass TME immunosuppression (Luo et al., 2025).

Beyond IL-2 competition, Treg cells may suppress γδ T cells through additional pathways, as suggested by studies in other murine and human disease contexts. Treg-derived mediators such as CTLA-4, IL-10, IL-35, and adenosine have been shown to selectively suppress IL-17⁺ γδ T cells in models of psoriasis (Lee et al., 2023; Sivasami et al., 2023), colitis (Park et al., 2010; Yurchenko et al., 2011), lung inflammation (Faustino et al., 2020; Xu et al., 2023), and myocardial infarction (Blanco-Domínguez et al., 2022), while Glucocorticoid-Induced TNFR-Related protein (GITR) was shown to impact both IL-17⁺ and IFNγ⁺ γδ T cells in the context of malaria infection (Gonçalves-Sousa et al., 2010). With regard to human γδ T cells, they were inhibited in vitro by Treg-derived IL-10 and TGF-β in tuberculosis and hepatocellular carcinoma (Yi et al., 2013; Li and Wu, 2008). Our in vivo data confirm an overtly dominant role of IL-2 deprivation specifically preventing anti-tumor γδ T cell responses in our study, although these alternative mechanisms might also play a role in other cancer settings.

Collectively, our findings identify a previously underappreciated axis of regulation in which Treg cells constrain anti-tumor γδ T cells through IL-2 cytokine competition. This adds to the growing body of evidence that the suppressive capacity of Treg cells extends beyond conventional T cells and positions IL-2 bioavailability as a key modulator of γδ T cell function in cancer. Therapeutic strategies that deplete or functionally disable Treg cells, or that selectively deliver cytokine signaling to effector cells, such as the Neo2/15 agonist explored here, may thus be the key to unlocking the full potential of γδ T cells in cancer immunotherapy.

C57BL/6J Foxp3-DTR (B6.129 [Cg]-Foxp3tm3 [DTR/GFP] Ayr/J) (Foxp3-DTR [C57BL/6 background] [Kim et al., 2007]) were obtained from The Jackson Laboratory and are backcrossed for at least eight generations to C57BL/6NTac mice, and NOD.Cg-Prkdcscid Il2rgtm1Wjl Tg(IL15)1Sz/SzJ (NSG-Tg[Hu-IL15] [NOD/ShiLtJ background]) were obtained from The Jackson Laboratory, and C57BL/6J FOXP3-hCD2/IL-17A-GFP (CD57BL/6 background) mice were bred in house from C57BL/6J FOXP3-hCD2 mice, kindly provided by Prof. Shohei Hori (University of Tokyo, Japan), and C57BL/6J IL-17A-GFP mice were obtained from Biocytogen. Mice between 6 and 20 wk old were used in all experiments. All the in vivo experiments were performed in female mice due to the biological relevance of breast cancer. Control mice used throughout the study were of the same strain (C57BL/6J Foxp3-DTR) but were not injected with DTx. Both male and female mice were used for ex vivo co-culture experiments. Mice were maintained under a 14-h/10-h light/dark cycle, with access to food and water ad libitum and controlled temperature and humidity conditions. All mice were kept at Gulbenkian Institute for Molecular Medicine (GIMM)’s rodent facility in a specific pathogen–free environment.

All experiments with mice were approved by the Animal Welfare Body of the institute (Órgão Responsável pelo Bem-Estar Animal (ORBEA)-GIMM) and by the national competent authority, Direção Geral de Alimentação e Veterinária, according to national and European regulations.

E0771 breast cancer cell line was purchased from Tebubio. Cells were maintained in DMEM (Life Technologies) with glutamine, sodium pyruvate, 10% FCS (Life Technologies), and 1% penicillin-streptomycin at 37°C and 5% CO₂. For inducing orthotopic breast tumors, either C57BL/6J Foxp3-DTR or wild-type mice were injected with 1 × 10⁶ viable E0771 cells in 100 μl of 1:1 PBS:Matrigel solution (Corning Matrigel Matrix) in the fat pad of the fourth left nipple, directly under the nipple. One week after tumor implantation, tumors were measured using a digital caliper, and mice were equally distributed in experimental groups. To deplete Treg cells, C57BL/6J Foxp3-DTR mice received intraperitoneal injections of 1.5 μg DTx in 100 μl of PBS at days 7, 9, and 11 after tumor inoculation, followed by a last injection of 0.75 μg at day 13/14. Animals were monitored daily or every 2 days from day 7, and their weight and tumor volume were assessed to guarantee the well-being of the animal. Tumor volumes were estimated following the formula width² × length/2 (Faustino-Rocha et al., 2013). Mice were sacrificed on day 15 after the tumor implantation. When indicated, either Vγ1⁺ T cells or CD8 T cells were depleted by intraperitoneal injection of either 100 µg αVγ1.1/Cr4 (clone 2.11; BioXcell) or αCD8β (clone Ly3.2; BioXcell) mAbs, respectively, in 100 μl PBS on days 7, 9, 11, and 14. In vivo IL-10 neutralization was induced with the injection of 200 µg of anti-IL-10 mAb (clone JES5-2A5; BioXcell) (or isotype control) intraperitoneally at days −1, 0, 1, and 3 after tumor inoculation and intra-tumoral at days 7, 10, and 13. In vivo IL-2 neutralization was induced by daily intraperitoneal injections of anti-IL-2 mAbs (S6-1A12 and S4B6 clones, 100 µg each, both from BioXcell) in 100 μl PBS. IL-2Rβ agonism was achieved by daily intraperitoneal injections of 10 µg Neo2/15 (Silva et al., 2019) in 50 μl PBS. For in vivo proliferation assays, the animals were injected intraperitoneally with 1.5 mg of BrdU (Sigma-Aldrich) diluted in PBS on days 7, 9, 11, and 13. Additionally, BrdU was also incorporated in the drinking water of these animals (0.8 mg/ml) since day 7 of tumor injection.

MC38 cell line was purchased from Kerafast. Cells were maintained in DMEM with glutamine, nonessential amino acids, 10% FCS (Life Technologies), and 1% penicillin-streptomycin at 37°C and 5% CO₂. For tumor inoculation, C57BL/6J Foxp3-DTR female mice were injected with 2 × 106 viable MC38 cells in 100 μl of 1:1 PBS:Matrigel solution (Corning Matrigel Matrix) in the right flank subcutaneously. Six days after tumor implantation, tumors were measured using a digital caliper, and mice were equally distributed in experimental groups. To deplete Treg cells, C57BL/6J Foxp3-DTR mice received intraperitoneal injections of 1.5 μg DTx in 100 μl of PBS at days 6, 8, and 10 after tumor inoculation, followed by a last injection of 0.75 μg at day 12. Animals were monitored daily or every 2 days from day 7, and their weight and tumor volume were assessed to guarantee the well-being of the animal. Tumor volumes were estimated following the formula width ² × length/2 (Faustino-Rocha et al., 2013). Mice were sacrificed on day 13 after the tumor implantation.

MDA-MB-231 cancer cell line was purchased from ATCC. Cells were maintained in DMEM (Life Technologies) with glutamine, sodium pyruvate, 10% FCS (Life Technologies), and 1% penicillin-streptomycin at 37°C and 5% CO₂. For inducing orthotopic breast tumors, NOD.Cg-Prkdcscid Il2rgtm1Wjl Tg(IL15)1Sz/SzJ (Jackson Laboratories) were injected with 1 × 10⁶ viable MDA-MB-213 cells in 100 μl of 1:1 PBS:Matrigel solution (Corning Matrigel Matrix) in the fat pad of the fourth left nipple, directly under the nipple. One week after tumor implantation, tumors were measured using a digital caliper, and mice were equally distributed in experimental groups. Mice received 5–10 × 10⁶ viable DOT cells intravenously at the time points indicated in the figure and daily intraperitoneal injections of 10 µg Neo2/15 (Silva et al., 2019) in 50 μl PBS. Tumor volumes were estimated following the formula width² × length/2

For analysis of tumor infiltrates, the tumor was harvested and cut into small pieces. 100 μg/ml of DNAse I (Roche), 1 mg/ml of collagenase (I) (Roche), and 0.4 mg/ml of collagenase (IV) were used to digest the tumor for 30 min at 37°C. After that time, the digested sample was filtered through a 100-μm cell strainer. For the analysis of the spleen and dLN, the organs were mashed and filtered through a 40-μm cell strainer. Cells were counted and transferred to a V-bottom 96-well plate where they were stimulated at a maximum density of 4 × 10⁶ cells/well in 200 μl of complete RPMI medium with PMA (50 ng/ml, Ref: 16561-29-8; Sigma-Aldrich) and ionomycin (1 μg/ml, Ref: 56092-81-0; Sigma-Aldrich) in the presence of Brefeldin A (BFA; 10 μg/ml, Ref: 20350-15-6; Sigma-Aldrich) and Monensin (Invitrogen eBioscience Monensin Solution [1000X] Ref: 00-4505-51) for 3–4 h at 37°C, 5% CO₂. The cells were then washed in FACS buffer (79.6% H₂O; 10% PBS 1 M; 10% heat-inactivated FCS; 0.4% ultrapure EDTA). For surface staining, cells were incubated with anti-mouse CD16/CD32 (1:200, Ref: 14-0161-82; Thermo Fisher Scientific) concomitantly with the surface marker staining (antibody clones in Table 1) that was performed by adding the appropriate antibody mix in FACS buffer (20 min, 4°C). LIVE/DEAD Fixable Near-IR Dead Cell Stain (Invitrogen Ref: L10119) dye was also added in this step. For intracellular cytokine staining, cells were then fixed and permeabilized using eBioscience Foxp3/Transcription Factor Fixation/Permeabilization, following the manufacturer’s instructions, and then incubated for 1 h with mAbs to detect cytokine production. Antibodies were purchased from eBioscience, BioLegend, and Invitrogen as summarized in Table 1. For BrdU staining, fluorescein isothiocyanate BrdU Flow Kit (Ref: 559619; BD Pharmingen) was used following the manufacturer’s instructions. Flow cytometry data acquisition was performed using the BD LSRFortessa X-20 (2010; BD Biosciences) or FACSymphony A5 SE (2023; BD Biosciences) cell analyzers and the FACSDiva software (version 6.2, BD Biosciences). Data were analyzed using the FlowJo software (version 10.8.1, BD Biosciences). Clustering of γδ T cells based on spectral flow cytometry data was performed with the FlowSOM plugin for FlowJo (version 4.1.0) considering the expression following markers: Vγ1, Vγ4, CD25, CD122, CD4, CD8, T-bet, RORγt, granzyme B, IFNγ, IL-17A, and Ki-67.

Spleen and LNs (superficial cervical, brachial, axillary, and inguinal) were harvested from reporter Foxp3-hCD2/IL-17A-GFP mice and mashed onto 40-μm cell strainers to generate cell suspensions. RBCs were lysed using RBC Lysis Buffer (BioLegend) and then incubated with anti-TER119, anti-CD11b, and anti-CD19 for 20 min at 4 °C. Subsequently, cells were incubated with anti-biotin microbeads (Miltenyi Biotec) and negatively enriched using LS columns (Miltenyi Biotec) following the manufacturer’s instructions. Negative fractions were stained with viability dye and fluorescence-conjugated antibodies against mouse TCRγδ (clone eBioGL3; eBioscience), mouse CD3e (clone 145-2C11; eBioscience), and human CD2 (clone RPA-2.10; Invitrogen) for 20 additional minutes on ice. Then, γδ T cells were sorted as alive CD3⁺TCRγδ⁺, whereas Treg cells were sorted as alive CD3⁺hCD2⁺ cells. After sorting, γδ T cells were stained with CellTrace Violet (Thermo Fisher Scientific) at a concentration of 2.5 µM in PBS for 15 min in the dark at room temperature, after which cells were washed and cultured. γδ T cells and Treg cells were cultured for 3 days in a 96-well round-bottom plate coated with anti-CD3 (2.5 μg/ml, clone 145-2C11; BioLegend) together or separated in complete RPMI with IL-7 (20 ng/ml, PeproTech), anti-CD28 (2.5 μg/ml, clone 37.51; Invitrogen), and in the presence or absence of recombinant murine IL-2 (10 ng/ml, PeproTech) or Neo2/15 (10 ng/ml [Silva et al., 2019]). Cells were cultured for 72 and 3 h before they were incubated with PMA (50 ng/ml, Sigma-Aldrich) and ionomycin (1 μg/ml, Sigma-Aldrich) in the presence of BFA (10 μg/ml, Sigma-Aldrich) and Monensin (Invitrogen eBioscience Monensin Solution [1000X]). Following incubation, cells were washed and analyzed by flow cytometry. For all culturing steps, the complete RPMI refers to RPMI (Gibco) supplemented with 10% FCS, 1% penicillin-streptomycin, 10% HEPES, 10% Sodium Pyruvate, 10% non-essential amino acids solution, 0.1% Gentamycin, and 0.1% 2-Mercaptoethanol (all from Gibco). For phosphorylated STAT-5 stainings, sorted γδ T cells and Treg cells were re-stained with surface markers and cultured together or separated for 10 min at 37° and 5% CO₂ with/without 10 ng/ml recombinant IL-2 (PeproTech) or Neo2/15 (Silva et al., 2019). Then, the medium was removed, and cells were fixed for 30 min and intranuclearly stained with PE-conjugated anti-STAT5 pY694 (clone 47/Stat5(pY694); BD Biosciences) for 30 additional minutes using the BD Phosflow protocol (BD Biosciences).

Buffy coats from healthy volunteers were obtained under the agreement (15.12.2003) between GIMM and Instituto Português do Sangue e da Transplantação. These procedures are authorized by the respective authorities, namely the Autoridade para os Serviços de Sangue e Transplantação and the Comissão Nacional de Proteção de Dados, and follow the applicable national and European regulations, including EU Directives 2004/33/EC, 2005/61/EC, and 2005/62/EC, and the General Data Protection Regulation (EU) 2016/679. All experimental work involving human cells complies with all ethical principles and with European Directive 2004/23/EC and Portuguese Law No. 12/2009, which establish standards for the quality and safety of human cells used in research.

Human PBMCs from healthy donors were stained for 20 min on ice with viability dye and fluorochrome-conjugated antibodies against CD4, TCRVδ1 (REA173, Miltenyi), TCRVδ2, CD25, and CD127 (clones and references listed in Table 1). Treg cells were sorted as alive CD4⁺CD25⁺CD127⁻ cells, whereas Vδ2 cells were sorted as CD4⁻TCRVδ1⁻TCRVδ2⁺ cells from the same PBMC sample.

In addition, we used DOT cells expanded as previously described. In brief, αβ T cell–depleted were seeded in G-REX vessels (Wilson Wolf Manufacturing). The culture medium used was OpTmizer-CTS, enriched with 2.5% heat-inactivated human plasma (LifeSciences), 2 mM L-glutamine (Thermo Fisher Scientific), and penicillin/streptomycin at concentrations of 50 U/ml and 50 µg/ml, respectively (Thermo Fisher Scientific). On day 0, the following recombinant, animal-free human cytokines were added to the culture: IL-4 (100 ng/ml), IFNγ (70 ng/ml), IL-21 (7 ng/ml), IL-1β (15 ng/ml), and a soluble anti-CD3 mAb (140 ng/ml, clone OKT-3; BioLegend), all sourced from PeproTech unless otherwise noted. On day 7, cultures were further supplemented with additional anti-CD3 (clone OKT-3, 1 µg/ml), IL-21 (13 ng/ml), and IL-15 (70 ng/ml). On day 11, cultures received fresh medium containing anti-CD3 (1 µg/ml) and IL-15 (100 ng/ml). Cells were maintained for 14 days at 37°C in a 5% CO₂ atmosphere. DOT cells were collected at the end of the culture period and used either fresh for in vitro applications or cryopreserved for in vivo use. For cryopreservation, cells were resuspended in CryoStor cell cryopreservation media (Merck) and then stored in liquid nitrogen.

Then, sorted Vδ2 cells and expanded DOT cells were cultured alone or in the presence of sorted Treg cells for 3–4 days in a 96-well round-bottom plate coated with anti-CD3 (2.5 μg/ml, clone OKT3; BioLegend) in complete RPMI with anti-CD28 (2.5 μg/ml, clone CD28.2; Invitrogen) in the presence or absence of recombinant human IL-2 (PeproTech) or Neo2/15 (Silva et al., 2019) at 10 ng/ml. Following co-culture, Treg cells were depleted by staining with APC-conjugated anti-CD4 (clone RPA-T4; eBioscience), followed by magnetic separation using anti-APC microbeads and MS columns (Miltenyi). The remaining DOT cells were subsequently co-cultured with CellTrace Violet–labeled MDA-MB-231 breast cancer cells in complete RPMI for 3 h at a 5:1 effector-to-target ratio. Tumor cell death was assessed by annexin V staining (Alexa Fluor 647; BioLegend).

For statistical analysis, the normality of data distributions was evaluated using the Shapiro–Wilk test. When data followed a normal distribution, two-group comparisons were performed with an unpaired Student’s t test (or a paired t test for matched samples), while comparisons across >2 groups were assessed using one-way ANOVA followed by Tukey’s multiple comparisons test. For non-normally distributed data, two-group comparisons were conducted using the Mann–Whitney U test, and multiple-group comparisons were analyzed with the Kruskal–Wallis test followed by Dunn’s multiple comparisons test. In kinetic experiments or when analyzing two pairs of groups, a two-way ANOVA with Šidák’s post hoc multiple comparisons test was applied. Data analysis was performed using GraphPad Prism 9 software (GraphPad Software Inc.). The applied tests are indicated in the figure legends, where data are expressed as mean ± standard deviation, and P < 0.05 was considered significant. Results are presented as P values * = P < 0.05, ** = P < 0.01, *** = P < 0.001, and **** = P < 0.0001.

Fig. S1 is related to Fig. 1 and contains additional data on IFNγ⁺ γδ T cell expansion upon Treg cell depletion in the MC38 colon cancer model, as well as in tumor-free mice, including proliferation (Ki-67) and tissue distribution analyses. Fig. S2 is related to Figs. 1 and 2 and shows the systemic increase of IFNγ-producing αβ T cells following Treg depletion, as well as functional data supporting the contribution of CD8 T cells to tumor control. Fig. S3 is related to Fig. 3 and includes experiments assessing alternative suppressive pathways, demonstrating that blockade of IL-10, IL-35, or adenosine signaling does not impair Treg-mediated suppression of IFNγ⁺ γδ T cells. Fig. S4 is related to Fig. 3 and provides detailed characterization of IL-2 receptor subunit expression (CD25 and CD122) across γδ T cell subsets and Treg cells in steady-state and tumor-bearing conditions. Fig. S5 is related to Figs. 3, 4, and 5 and contains additional data on IL-2 signaling, including STAT5 phosphorylation, the impact of IL-2 neutralization on αβ T cell responses and presents donor-specific data demonstrating the suppressive effect of human Treg cells on γδ T cell proliferation and effector function across independent donors.

Data supporting the findings of this study are available from the corresponding authors on reasonable request.

We thank Natacha Gonçalves-Sousa (GIMM, Lisbon, Portugal) and the technical support of the Rodent Facility and the Flow Cytometry Platform of the GIMM, funded by FLxFlow-HDSpect-LISBOA2030-FEDER-01318900.

This work was funded by “la Caixa” Foundation under the project code [LCF/PR/HR24/00929] (to B. Silva-Santos). We also acknowledge support from EMBO Postdoctoral Fellowship (233-2023 to R. Blanco-Domínguez), Fundação para a Ciência e Tecnologia, Ministério da Ciência, Tecnologia e Ensino Superior, Portugal (2021.01953.CEECIND, https://doi.org/10.54499/2021.01953.CEECIND/CP1673/CT0007 to S. Mensurado), and Takeda Development Center Americas, Inc. (to B. Silva-Santos, R. Blanco-Domínguez, L. Barros, B. Henriques-Alves, M. Carreira, and S. Mensurado). The authors have no additional financial interests. This work was supported by GIMM-CARE (funded by the European Union under grant agreement No. 101060102. GIMM-CARE is co-funded by the Portuguese Government, the Foundation for Science and Technology (FCT), ARICA – Investimentos, Participações e Gestão, Jerónimo Martins, the Gulbenkian Institute for Molecular Medicine, and CAML – Lisbon Academic Medical Centre) [https://doi.org/10.3030/101060102], and by national funds through FCT under the Associate Laboratory programme (LA/P/0082/2020) [https://doi.org/10.54499/LA/P/0082/2020], and under the R&D Unit funding programme (UID/06357/2025) [https://doi.org/10.54499/UID/06357/2025].

Author contributions: Rafael Blanco-Domínguez: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, visualization, and writing—original draft, review, and editing. André Miguel Vaz-Pinto: formal analysis, investigation, and visualization. Leandro Barros: investigation and writing—review and editing. Noella Lopes: conceptualization, investigation, and methodology. Beatriz Henriques-Alves: investigation and writing—review and editing. Mariana Carreira: investigation and writing—review and editing. Carlos Labão-Almeida: resources and writing—review and editing. Julie C. Ribot: conceptualization, methodology, and writing—review and editing. Gonçalo J.L. Bernardes: resources. Bruno Silva-Santos: conceptualization, funding acquisition, supervision, and writing—original draft, review, and editing. Sofia Mensurado: conceptualization, investigation, methodology, supervision, and writing—original draft, review, and editing.