Therapeutic Targeting of Regulatory T Cells in Cancer

Abstract

The success of immunotherapy in oncology underscores the vital role of the immune system in cancer development. Regulatory T cells (Tregs) maintain a fine balance between autoimmunity and immune suppression. Tregs play multiple roles in the tumor microenvironment but particularly act in suppressing T cell activation. This review focuses on the detrimental and sometimes beneficial roles of Tregs in tumors, our current understanding of recruitment and stabilization of Tregs within the tumor microenvironment, and current Treg-targeted therapeutics. Research identifying subpopulations of Tregs and their respective functions and interactions within the complex networks of the tumor microenvironment will be crucial to develop the next generation of immunotherapies. Through these advances, Treg-targeted immunotherapy could have important implications for the future of oncology.

Targeting regulatory T cells in tumors

Cancer treatment underwent a revolution following advances in cancer immunotherapy and discoveries regarding the tumor microenvironment (TME). Within the TME, regulatory T cells (Tregs) are a major therapeutic target. Tregs are a small population (5%–15%) of CD4⁺ T cells with potent suppressive activity. While they limit inflammation and maintain immune homeostasis, Tregs can also promote peripheral tolerance to tumors [1,2]. Treg suppressive activity depends on the transcription factor forkhead box protein 3 (FOXP3) [2,3]. Individuals who lack FOXP3 expression develop immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome, a severe autoimmune lymphoproliferative disease that is fatal without bone marrow transplant [4]. Although Tregs are important in maintaining peripheral tolerance and preventing autoimmunity, they also suppress antitumor immunity within the TME [1,5]. This immunosuppressive effect occurs through contact-dependent mechanisms, such as direct cytolysis and dendritic cell (DC) disruption, as well as contact-independent mechanisms, such as suppressive cytokine release and metabolic dysregulation (Figure 1) [1]. Here we provide a brief overview of Treg biology within the TME, current methods that specifically target Tregs in the clinic, and a perspective on future questions and challenges.

Figure 1. Treg activation and expansion in the TME.

The TME and its high antigen burden activate and expand Tregs. Intratumoral Tregs orchestrate a complex cellular and molecular network to alter the tumor milieu. Treg contact dependent suppression mechanisms in the TME include cytolysis via perforin and granzyme A/B and transfer of cyclic AMP by membrane gap junctions. Treg contact independent suppression mechanisms include secretion of IL-10, IL-35, and TGF-β; deprivation of IL-2; conversion of extracellular ATP to adenosine by CD39; and expression of CTLA-4 and LAG3. In addition, tumor cells can recruit peripheral Tregs into the TME by chemokines such as CXCL12, CCL17, CCL22 CCL1. Furthermore, metabolic reprogramming modifies Tregs favor of the TME over effector T cells. Treg activation and expansion in the TME can be supported by multiple mechanisms: (1) IL-2, IDO and TGFβ released from tumor cells; (2) alternate metabolic pathway can be activated in intratumoral Tregs; (3) tumor necrosis factor family genes, OX40, GITR and TNFR2 as well as co-stimulatory receptor ICOS can be upregulated in effector intratumoral Tregs and (4) NRP1 can be activated to maintain their suppressive function. Created with BioRender.com.

Treg recruitment and function in solid tumors

Increased numbers of Tregs within a tumor are associated with worse prognosis across many cancers, including breast [6–9], lung [10–12], ovarian [13,14], and hepatocellular [15–19] cancers. However, not all malignancies with elevated intratumoral Tregs have poor prognosis (Table 1). Intratumoral Tregs are associated with a good prognosis in microsatellite instability-high (MSI-H) colorectal cancer (CRC), estrogen receptor-negative breast cancer, squamous cell carcinoma of the esophagus, and metastatic sites of ovarian cancer [20–28]. The divergent correlations between tumor-infiltrated Tregs and clinical outcomes may be explained by three phenomena.

Table 1.

Selected studies on prognostic implications of intratumoral Tregs

Tumor Type	Prognosis Studies
Breast	Good: IHC by FOXP3 in triple negative disease [188] Mixed: IHC by FOXP3 [189] Poor: IHC by FOXP3; FC by FOXP3, CD25, [9,190–193]
Bladder	Poor: IHC by FOXP3; FC by FOXP3, CD25 [194,195]
Cervical	Poor: IHC by FOXP3 [196,197]
Colorectal	Good: IHC by FOXP3 include MSI-H tumors [23,198] Mixed: IHCFOXP3; FC by FOXP3 [22,199,200] Poor: IHC by FOXP3; FC by FOXP3, CD25, CD127lo [201–203]
Endometrial	Poor: FC by FoxP3, CD25; IHC by FOXP3 [204,205]
Esophageal/oral	Good: IHC by FOXP3 on squamous histology [27] Poor: FC by FOXP3, CD25; IHC by FOXP3 [206,207]
Gastric	Poor: FC by FOXP3, CD25; IHC by FOXP3 [206,208]
Head and neck	Poor: FC by FOXP3, CD25 [109,209]
Hepatocellular	Poor: FC/IHC by FOXP3, CD25 [15–19]
Lung (non-small cell lung cancer)	Poor: IHC by FOXP3; FC by FOXP3, CD25 [10,12,210–213]
Melanoma	Poor: FC by FOXP3, CD25 [195,214]
Mesothelioma	Mixed: IHC by FoxP3, CD25 [215]
Ovarian	Good: IHC by FOXP3 in metastatic ovarian cancer [216] Poor: IHC by FOXP3, 1 FC by FOXP3, CD25 [13,14,40]
Pancreatic	Poor: IHC by FOXP3, 1 FC by FOXP3, CD25 [217,218]
Renal cell	Poor: FC by FOXP3, CD25 [219,220]
Vulvar	Poor: IHC by FOXP3 [221]

FC, flow cytometry; IHC, immunohistochemistry

First, human Tregs are phenotypically and functionally heterogeneous within tumors. CD4⁺ T effector cells undergoing activation can temporarily express FOXP3 and CD25 (IL-2 receptor α-chain) without suppressive function, and thus are not traditional Tregs [29–31]. Additionally, application of immunohistochemistry commonly detects the presence of intratumoral Tregs by CD25 and FOXP3, or FOXP3 alone, which is less quantitative than flow cytometry-based studies and may require additional markers to identify functional Tregs [22,32–34]. For example, low expression of CD127 (IL-7Rα) is widely used to identify functionally active Tregs [32,33]. Furthermore, CD45RA expression marks Treg subpopulations with distinct suppressive abilities that inversely correlate with CRC patient survival [22,34]. In addition, TNFR2 (TNFRSF1B) is preferentially expressed on both mouse lymphoid tissue Tregs and human thymic Tregs [35] and is a marker for more potent suppressive Tregs in murine tumor models and human cancers [36,37]. Thus, the existence of different subpopulations of Tregs and their differential detection across studies could account for the discrepancies in reported results.

Second, findings associating Tregs with good prognosis tend to be in settings of chronic inflammation. Such is the case for CRC promoted by chronic, stromal inflammation, where the presence of Tregs suppressing chronic inflammation is beneficial [38]. Intratumoral Tregs associated with good prognosis are also consistent with the chronic inflammation associated with triple-negative breast cancer compared to hormone-receptor positive breast cancer, squamous cell carcinoma of the esophagus (smoking-induced) compared to adenocarcinoma of the esophagus, or metastatic sites of ovarian cancer compared to the primary site of ovarian cancer. Thus, the specific inflammatory state may critically and variably influence the function of Tregs in the local tumor context. Third, cancer therapies targeting tumor-infiltrating lymphocytes can affect both antitumor T cells and Tregs. While expression of FOXP3 or CD8 alone fails to show significant correlation to survival outcomes in patients with advanced HER2-positive breast cancer, a high CD8:Treg ratio is associated with significantly extended survival outcomes [39]. The tumor-infiltrating CD8:Treg ratio is widely used to predict prognosis of cancer patients and treatment responses to cancer therapies [39–42]. Thus, the prognosis of cancer patients cannot be predicted solely according to absolute cell numbers.

The TME favors the presence of Tregs via two main mechanisms: (1) release of chemokines attracting Tregs, and (2) creation of a metabolically distinct milieu that increases proliferation of Tregs over effector CD4⁺ T cells (Figure 1). Human Tregs can be recruited to the TME by specific antigens on malignant cells via TCR–MHCII interactions [43]. Tregs can also be attracted to the TME by macrophages and tumor cell-expressed chemokines, such as CXCL12, CCL17, CCL22, and CCL1. Once human Tregs leave the periphery and reside in the TME, they no longer upregulate lymphoid “homing receptors”, such as CCR7 or CD62L (L-selectin) [13,44–51]. In addition, a recent study reported a distinct metabolic pathway of intratumoral Tregs compared to effector T cells that helps Tregs maintain their suppressive activity within the metabolite-depleted and acidic TME [52]. In contrast to effector T cells and Tregs in the periphery, intratumoral Tregs use an alternate metabolic pathway that consumes lactic acid [52].

Beyond markers of chemotaxis, the TME combined with high antigen burden activate and expand Tregs. Intratumoral Tregs orchestrate a complex cellular and molecular network to alter the tumor milieu by contact-dependent mechanisms including cytolysis via perforin and granzyme A/B and transfer of cyclic AMP by membrane gap junctions, as well as contact-independent mechanisms including secretion of IL-10, IL-35, and TGF-β; deprivation of IL-2; conversion of extracellular ATP to adenosine by CD39; and expression of CTLA-4 and LAG3 (Figure 1). While there are Treg-suppressive mechanisms that are therapeutically targeted for Treg specificity within the TME (see below), there is tremendous potential in the development of novel therapeutics targeting Tregs and Treg functions within the TME.

Therapeutic implications of targeting Tregs in cancer

Suppression of antitumor immune responses by Tregs is central to tumor development and progression. Efforts to discover novel therapies targeting Tregs focus on restoring antitumor immunity while maintaining peripheral tolerance. There are seven proposed methods of targeting Tregs in human cancer: (1) depleting intratumoral Tregs, (2) halting Treg migration into tumors, (3) sensitizing intratumoral Tregs to inhibitory receptor blockade, (4) targeting co-stimulatory signals on intratumoral Tregs, (5) targeting Treg cytokine secretion, (6) altering Treg fragility, and (7) targeting Treg metabolism. Here we discuss current clinical approaches specifically targeting intratumoral Tregs.

Depleting Tregs in the TME

Depleting Tregs with targeted chemotherapy

Targeted chemotherapeutic agents, such as tyrosine kinase inhibitors (TKI), can selectively impact Treg number, maintenance and function [53–55]. The TKI, sunitinib, causes a reduction of Tregs in patients’ peripheral blood contributing to efficacy in renal cell carcinoma [53,56]. In a murine hepatocellular cancer model, sunitinib inhibits tumor growth by depletion of Tregs and reduction of TGF-β and IL-10 produced by Tregs [57]. Additionally, neoadjuvant treatment with the TKI, sorafenib, significantly reduced intratumoral Tregs in patients with renal cell carcinoma [54]. Another TKI, imatinib, is found to reduce effector Tregs in patients with chronic myelogenous leukemia, especially those who have achieved complete molecular remission [58]. Data indicate that imatinib reduces Tregs in vivo and inhibits the phosphorylation of lymphocyte cell-specific protein tyrosine kinase (LCK) as an off-target effect, thereby reducing TCR signaling of Tregs [58]. Several traditional chemotherapeutic agents have selective cytotoxic activity against Tregs relative to other immune cells [53,56,59]. Particularly, use of “metronomic chemotherapy” with low-dose cyclophosphamide (CTX) has been reported to deplete Tregs (Figure 2). However, the overall benefits in various cancer types are controversial, and further clinical studies are still needed [59–61].

Figure 2. Treg targeted anti-tumor therapies.

1. Therapy targeting Treg depletion in the TME includes kinase inhibitors, low-dose cyclophosphamide and anti-CD25; 2. Blocking chemokine receptors CCR4 and CCR8 can halt Treg migration into the TME; 3. Blocking inhibitory receptors such as CTLA-4, TIGIT and LAG3 reduce Treg suppression function and induce Treg depletion. In addition, chemotherapy, radiation therapy and targeted therapies can sensitize tumors to inhibitory receptor blockade therapy; 4. Targeting co-stimulatory signals such as OX40, GITR, ICOS and TNFR2 can enable Tregs to be less suppressive and decrease intratumoral immunosuppression; 5. Blocking Treg derived inhibitory cytokines such as IL-10, IL-35 and TGF-β can dampen Treg suppressive function in the TME; 6. Blocking NRP-1 directly with anti-NRP1 can lead to Treg fragility and IFN-γ secretion; 7. Administration of metformin, IDO inhibitors and A2AR inhibitors can affect Treg metabolism in the TME, resulting in Treg instability and reduction of CTLA4 and IL-10.

Depleting Tregs by targeting IL-2 and CD25

Tregs express high levels of the IL2 receptor, CD25, but produce only a small amount of IL-2, which can lead to local IL-2 deprivation impacting effector T cell function (Figure 1). Targeted therapy with monoclonal antibodies against CD25 aims to deplete intratumoral Tregs within the TME [62,63]. Preclinical studies in mice showed inhibition of tumor growth and improved survival with administration of anti-CD25 antibody [64,65]. Denileukin was approved in 1999 by the US Food and Drug Administration (FDA) for T cell lymphoma and demonstrates mild efficacy in renal cell carcinoma but no effect in advanced melanoma patients [62,63]. Similar to denileukin, daclizumab in combination with an antigen-specific DC vaccine demonstrates mild efficacy in glioblastoma patients, yet only a limited number of patients pretreated with daclizumab developed efficient responses [64,65]. However, a drawback of CD25-directed global intervention is that while it achieves transient Treg depletion, it also depletes CD25⁺ effector T cells that upregulate CD25 following activation [65]. Thus, a non-IL-2-blocking anti-CD25 antibody (RG6292) was recently proposed to specifically deplete Tregs without affecting IL-2 receptor signaling on effector T cells [66]. Data indicate that RG6292 binds to the opposite side of the IL-2–CD25 interaction site, which leads to depletion of Tregs through antibody-dependent cell-mediated cytotoxicity (ADCC) without affecting activated effector CD4⁺ and CD8⁺ T cells, thus promoting antitumor activities [66]. A phase I trial of RG6292 in patients with advanced solid tumors is currently ongoing (NCT04158583ⁱ) (Table 2).

Table 2.

Selected studies on human Treg-targeted immunotherapies

Immunotherapy	Treg Mechanism	Animal Studies	Human Cell Studies	Clinical Trials	Moderate/Severe Adverse Events
Low-dose CTX	Depletion		[222]	[59–61]	Local allergic reaction
Anti-CD25 (daclizumab)	Depletion	[223]	[224]	[64,225]	Local allergic reaction
Sunitinib	Depletion	[57]	[53]	[56,90,226]	Thrombocytopenia, fatigue, asthenia and neutropenia
Sorafenib	Depletion		[54]	[227]	Diarrhea, hand-foot skin reaction, and fatigue
Imatinib	Depletion	[58]	[58]	[228]	Stevens-Johnson syndrome and hepatotoxicity
Denileukin	Depletion	[229]	[62,63]	[64,230]	Anaphylaxis
Anti-CD25 (RG6292)	Depletion	[66]	[66]	NCT04158583 i
Anti-CCR4 (mogamulizumab)	Halting migration	[69]	[69]	[70,72]	Stevens–Johnson syndrome and type I diabetes mellitus.
Anti-CCR-4 (FLX475)	Halting migration			NCT03674567ii, NCT04894994iii NCT04768686iv
Afucosylated anti–CTLA-4	Depletion	[92]	[231]	NCT03110107 viii
Anti-TIGIT (Ociperlimab)	Inhibitory receptor	[97]	[97]	NCT04047862 xi
GITR agonist (TRX518)	Co-stimulation	[118]	[118]	[118]	Fatigue, nausea, cough
OX40 agonist (OX86)	Co-stimulation	[123]		[124]	Fatigue, nausea and decreased appetite
Anti-ICOS (MEDI-570)	Co-stimulation			[133]	Anemia, hypophosphatemia and thrombocytopenia
ICOS agonist (JTX-2011)	Co-stimulation	[134]	[134]	NCT04319224 xiv NCT02904226 xv
Anti-TNFR2 (BI-1808)	Co-stimulation			NCT04752826 xviii
Afucosylated anti-GARP (DS-1055a)	Cytokine secretion	[155]		NCT04419532 xix
Anti-NRP-1 (Fc(AAG)-TPP11)	Inducing fragility	[170,174]	[174]	[232]	Proteinuria, cytopenia

Halting Treg migration

Tregs infiltrate into the TME via a combination of chemokine and chemokine receptor interactions, such as CCL22–CCR4 [13] and CCL1-CCR8 [67,68], providing an additional approach to target Tregs. CCR4 is expressed on tumor Tregs in patients with ovarian cancer, and in vitro data further confirmed that a neutralizing monoclonal antibody to CCL22 significantly blocks ascites-induced Treg migration [13]. Moreover, CCR4 is highly expressed on Tregs from melanoma patients compared to peripheral blood and healthy individuals, specifically on intratumoral suppressive Tregs [69]. Anti-CCR4 mAb treatment selectively depletes effector Tregs and augments effector T cells both in vitro and in vivo murine models [69]. Mogamulizumab, a defucosylated anti-CCR4 mAb targeting malignant T cells expressing high levels of CCR4, was approved in 2016 by the European Medicines Agency and in 2018 by the US FDA for treatment of relapsed/refractory mycosis fungoides and Sézary syndrome cutaneous T-cell lymphoma [70]. A completed phase I trial of mogamulizumab demonstrated a well-tolerated safety profile with encouraging overall response rates among patients with cutaneous T-cell lymphoma: 47.1% in Sézary syndrome and 28.6% in mycosis fungoides [71]. Mogamulizumab has been evaluated in phase I clinical trials for patients with advanced or metastatic solid tumors (including non-small-cell lung cancer, ovarian cancer, and pancreatic cancer) in combination with nivolumab [72]. However, grade 3/4 treatment-related adverse events were observed in patients treated by mogamulizumab and nivolumab, which might be due to the depletion of effector Tregs in both the periphery and the TME [72] (Table 2). Furthermore, a murine study with CCR4 blockade demonstrated a reduction of Treg migration into murine tumors, an increase of both CD4⁺ T cell:Treg and CD8⁺ T cell:Treg ratios, and promising antitumor efficacy in combination with anti-CTLA4 antibody in tumor models with low baseline levels of CCR4 ligands [73]. Based on these preclinical findings, another CCR4 antagonist, FLX475, is being investigated alone or in combination with checkpoint blockade in clinical studies for safety and antitumor activity in patients with advanced cancer (NCT03674567ⁱⁱ, NCT04894994ⁱⁱⁱ and NCT04768686^iv).

Upregulation of CCR8 in tumor-infiltrating Tregs has been reported in various cancers in comparison to normal tissues and tumor-infiltrating effector T cells [74–76]. In various human tumors CCR8 expression marks a more phenotypically activated or differentiated Treg subpopulation as indicated by the expression of CD25, CTLA-4, CD39 and IL-10 [68,74,76]. Recent studies using murine models further show that Treg-depleting anti-CCR8 antibodies significantly decrease tumor Treg frequency, evoke anti-tumor responses and suppress tumor growth. Interestingly this was not achieved with function-blocking CCR8 antibody [77,78]. This was further supported from data from in a murine study that indicated that CCR8 was dispensable for Treg migration into the TME [79]. An anti-CCR8 mAb (JTX-1811) targeting Tregs by ADCC depletion is currently in a phase I trial for patients with advanced solid tumors in combination with anti-PD-1 (NCT05007782^v). Another phase I trial of anti-CCR8 antibody S-531011 in combination with pembrolizumab in patients with advanced or metastatic solid tumors is currently ongoing (NCT05101070^vi). Together, these preclinical and clinical studies show that targeting chemokine receptors on Tregs is a promising therapeutic approach to selectively deplete intratumoral Tregs. In addition to the high expression of CCR4 and CCR8 on Tregs in human tumors, recent studies report CCR4 and CCR8 expression on effector T cells and Tregs from peripheral blood and skin [80–82]. Given the promising safety and efficacy profiles of CCR4- and CCR8-targeted therapies in clinical studies, further investigation on the off-target effects are warranted.

Sensitizing intratumoral Tregs to inhibitory receptor blockade

Inhibitory receptors or immune checkpoints dampen effector T cell responses to prevent immunopathology and autoimmunity, yet they also can be dysregulated by tumors as an immune resistance mechanism that limits antitumor responses [83]. Thus, administration of checkpoint blockade restores the functions of exhausted intratumoral CD4⁺ and CD8⁺ effector T cells and are becoming standard-of-care for several tumor types. However, inhibitory receptors are also expressed on Tregs and play important roles in Treg function [84,85]. Understanding the effects of checkpoint blockade on Tregs is warranted to elucidate the unresponsiveness of checkpoint inhibitors in cancer patients and the mechanisms of toxicity.

Ipilimumab, specific for CTLA-4, was one of the first approved checkpoint blockade therapies, initially for advanced melanoma in 2011 and then subsequently in the adjuvant setting [86]. Ipilimumab was originally thought to reinvigorate dysfunctional T cells, although murine tumor models suggest that its antitumor response may depend on depletion of Tregs in the TME mediated through the Fcγ receptor and ADCC [87–90]. Data indicate that both human IgG1 and IgG2 isotypes of anti-CTLA-4 promote depletion of intratumoral Tregs by FcγR-dependent cytotoxicity [90]. However, Treg depletion appears to have differential efficiency depending on the context, such as tumor mutational burden and human FcγR polymorphism [90]. In another study, both ipilimumab and tremelimumab did not deplete Tregs in human tumors but instead increased infiltration of intratumoral CD4⁺ and CD8⁺ T cells [91]. Thus, increasing the IgG1 Fc binding affinity to improve Treg depletion by anti-CTLA-4 is being further explored. Given that ADCC function is modulated by N-linked glycosylation in the Fc region of IgG antibodies, an afucosylated IgG1 can significantly improve ADCC by the absence of core fucose on the Fc N-glycan [92]. In a preclinical study associated to a phase I/IIa trial, the novel nonfucosylated anti-CTLA-4 antibody, BMS-986218 demonstrated increased intratumoral effector T cells, enhanced tumor rejection and increased depletion of intratumoral Tregs in mice, while minimal changes in Treg or CD8⁺ T cell levels in the spleen were detected^vii,viii. To date, first-in-human phase I/IIa trial of BMS-986218 in combination with nivolumab in advanced solid tumors showed acceptable safety profiles and improved effects of CTLA-4 blockade (NCT03110107^ix,x). T cell Ig and ITIM domain (TIGIT) is upregulated on CD4⁺ and CD8⁺ T cells upon activation as well as on Tregs [93,94]. TIGIT is expressed on the majority of FOXP3⁺ intratumoral Tregs and a high percentage of intratumoral CD8⁺ T cells in patients with hepatocellular carcinoma [95]. In a preclinical study, TIGIT is found to be significantly expressed on Tregs and CD8⁺ T cells in the TME compared to FOXP3⁻CD4⁺ T cells in murine tumor models, and the function of TIGIT in Tregs seems to dominate over its function on CD8⁺ T cells [96]. Effector T cell responses can be dampened directly by TIGIT-mediated, T cell-intrinsic inhibitory function and indirectly by prevention of DC maturation [93]. In contrast, TIGIT defines a subset of Tregs with an enhanced suppressive phenotype [94]. Preclinical studies using in vitro and murine tumor models further illustrate that TIGIT⁺ Tregs possess highly suppressive activities and play the major role of TIGIT-mediated CD8⁺ T cell suppression, which can be countered by co-blockade of TIGIT/PD-1, but not PD-1 blockade only, emphasizing the potential of modulating intratumoral Tregs by targeting TIGIT [95,96]. Ociperlimab (BGB-A1217) is an anti-TIGIT antibody that have shown Treg depletion in preclinical studies [97]. Several anti-TIGIT antibodies EOS-448, SEA-TGT and Ociperlimab are currently in clinical trials against advanced cancers (NCT04047862^xi, NCT04335253^xii and NCT04254107^xiii).

The roles of PD-1 in Tregs remain less understood compared to effector T cells. Human glioblastoma intratumoral Tregs expressing high PD-1 show exhausted phenotypes and dampened suppression activity [98]. However, PD-1 signaling may also promote Treg suppression and development of induced Tregs [99,100]. The outcomes of these findings may be dependent on the disease context, as patient response rates to PD-1 blockade vary based upon tumor subtype, histology, smoking status, PD-L1 status, and antigen burden [101–103]. The model system could influence the results as well. Thus, combining anti-PD-1 blockade with other immunotherapies and the development of bispecific antibodies targeting PD-1 and other inhibitory receptors simultaneously may overcome resistance to single anti-PD-1 treatment and provide more effective antitumor responses [104–107].

Given the prevalence and efficacy of checkpoint blockade therapy, new studies are evaluating mechanisms of sensitizing the TME to inhibitory receptor blockade. Chemotherapies such as oxaliplatin, carboplatin, cisplatin, doxorubicin, and cyclophosphamide demonstrated differential sensitivities towards intratumoral Tregs and CD8⁺ T cells when checkpoint blockade therapy in lung cancer mouse models [108]. Human head and neck squamous cell carcinoma studies also identified upregulation of membrane CTLA-4 on Tregs after targeted therapy with cetuximab [109]. Together, additional targeting on Tregs with tumor debulking by chemotherapy might provide potential to treat patients with limited efficacy on immune monotherapies. A variety of autoimmune phenomena, such as pneumonitis, colitis, myositis, and even reports of Guillain-Barre syndrome, are seen with checkpoint blockade treatments, which are accompanied by Treg dysfunction or deficiency. It remains uncertain whether checkpoint blockade causes these adverse events via modulation of effector T cell function or Treg dysregulation. Further investigation of the effects of checkpoint blockade therapy on Tregs is warranted to elucidate these mechanisms of toxicity (Table 2).

Targeting co-stimulatory receptors on Tregs

Co-stimulatory receptors are critical for Treg development and function [110,111]. Intratumoral Tregs express multiple tumor necrosis factor super receptor family (TNFRSF) genes and are associated with a late differentiated state found in various cancer types [112–114]. Furthermore, glucocorticoid-induced TNF receptor family-related protein (GITR) is significantly higher in tumor-infiltrating Tregs compared to CD4⁺ effector T cells in patients with non–small cell lung carcinoma [115]. A murine study demonstrated that a GITR agonist, DTA-1, reduced intratumoral Tregs by >50% and displayed minimal effects on peripheral Tregs [116]. Several GITR agonists have entered clinical phase I/IIa studies against advanced solid tumors [117–119]. A Fc-dysfunctional aglycosylated GITR agonist, TRX518, demonstrated induction of effector Treg cell death by promoting Treg hyper-activation and loss of Treg stability in a phase I clinical study [118]. While TRX518 showed Treg depletion in tumors and peripheral blood in patients with colorectal cancer and bladder cancer, the treatment failed to produce a clinical response [118]. One explanation is that effector T cells accumulated at the tumor site have established an exhausted phenotype, and depletion of Tregs in the TME alone cannot reinvigorate dysfunctional T cell functions. Studies with GITR agonists alone or in combination with anti-PD1 blockade are well tolerated in patients with advanced solid tumors, with the combinatorial therapy exhibiting modest clinical benefit [119,120].

In patients with colorectal cancer, an elevated expression of OX40 Tregs is found in tumor-infiltrating Tregs and circulating Tregs [121]. A preclinical study using in vitro and in vivo models showed that negative modulation of OX40 affected FOXP3 expression and suppressive function in Tregs [122,123]. Furthermore, a combination of OX40 agonist (OX86) with CTX shows promising tumor reduction in a B16 murine melanoma model, inducing profound Treg depletion in the TME and influx of effector CD8⁺ T cells. OX86 preferentially binds to Tregs after CTX administration, and the combination treatment causes Treg hyperactivation and apoptosis, resulting in Treg depletion at the tumor site [123]. Additionally, a completed phase I trial of OX40 agonist (PF-04518600) shows acceptable safety and enhanced memory T cell proliferation, although no data were reported regarding any impact on Tregs [124].

Co-stimulatory receptor ICOS is highly expressed on intratumoral Tregs in multiple cancer types, and an increased population of ICOS⁺ Tregs is associated with poor prognosis [125–131]. For example, Tregs exhibit detrimental roles in follicular lymphoma, and data suggest an accumulation of Tregs in follicular lymphoma tissues with preferential expression of ICOS [132]. These ICOS⁺ Tregs display activated phenotypes and suppress effector T cells and follicular lymphoma B cells, and ICOS antagonist treatment disrupts enrichment of activated Tregs [132]. The ICOS antagonistic mAb MEDI-570 has been evaluated in a phase I trial in patients with relapsed or refractory peripheral T cell lymphoma follicular variant or angioimmunoblastic T cell lymphoma (Table 2). The treatment is well-tolerated and reduces circulating ICOS⁺CD4⁺T cells [133]. JTX-2011 (vopratelimab) is an ICOS agonist that decreases the frequency of intratumoral Tregs without affecting peripheral Treg frequency in a murine model [134]. The combination of JTX-2011 with Ipilimumab and Nivolumab is under clinical evaluation (Table 2). Given the role of ICOS in Treg proliferation and immunosuppression, further human studies on the effects of an ICOS-specific agonist on intratumoral Tregs are warranted [132,135].

TNFR2 (TNFRSF1B) is preferentially expressed on human FOXP3⁺ Tregs and TNFR2-expressing Tregs exhibit a more potent suppressive phenotype in human cancers [136–139]. APX601 is a potent TNFR2 antagonist that can overcome immunosuppression in the TME. Data in a preclinical murine study showed APX601 significantly inhibits Treg suppressive function by blocking TNFR2:TNFα interaction and depleting TNFR2-expressing Tregs and tumor-associated myeloid cells^xvi. Another antagonistic anti-human TNFR2 antibody, AN3025, demonstrated inhibition of induced Treg suppression and reduction of Treg frequency in a murine tumor model [140]. The potent anti-tumor efficacy of APX601 and AN3025 in preclinical studies support the clinical development of TNFR2 antagonistic antibodies. Further analyses of their safety and toxicity in human cancers are warranted. In addition, TNFR2 antagonist BI-1808, which has similar mechanism of action, is well tolerated in cynomolgous monkeys and is being evaluated in a phase I/IIa clinical trial in treatment of advanced malignancies (NCT04752826^xvii,xviii).

Targeting Treg cytokine secretion

Tregs are a key source of inhibitory cytokines, such as TGF-β, IL-35, and IL-10, which are important contact-independent mechanisms of Treg immunosuppression [1,141–148]. Infiltration of FOXP3⁺ Tregs is associated with poor prognosis of patients with pancreatic ductal adenocarcinoma (PDAC) and positively correlates with TGF-β expression in PDAC samples [149]. In addition, in a murine tumor model, TGF-β produced by tumor cells enhanced the expansion of Tregs and blocking TGF-β signaling inhibited tumor growth through a Treg dependent fashion [150]. However, systemic depletion of TGF-β may cause severe toxicity due to its pleiotropic roles [151]. Glycoprotein-A repetitions predominant (GARP) functions as a carrier and surface receptor for latent TGF-β1 and is highly expressed on the surface of human activated Tregs and platelets [152]. GARP promotes TGF-β secretion and Treg suppressive function in murine models [152–154]. DS-1055a, an afucosylated anti-GARP antibody, effectively depletes Treg in tumors and subsequently activates effector T cells, which leads to tumor clearance in a humanized murine model [155]. A phase I clinical trial is currently ongoing to evaluate the safety and tolerability of DS-1055a in patients with metastatic solid cancer (NCT04419532^xix). Furthermore, integrin αVβ8 (ITGβ8) is essential to bind and activate TGF-β from GARP/latent TGF-β1 complexes [156]. Interestingly, of all the ITGβ8 expressing tumor-infiltrating T cells in various human cancers, 65–70% are FOXP3⁺ T cells and consequently expression of ITGβ8 in tumor-infiltrating T cells is associated with poor survival [143]. Blocking integrin αVβ8 expression resulted in reduced Tregs suppression of pathogenic T cells during active inflammation in murine models [157]. Furthermore, neutralizing anti-ITGβ8 antibody significantly increases cytotoxic activity of CD8⁺ T cells in human melanoma and breast cancer tumors in ex vivo culture experiments, highlighting the potential of this Treg-targeted anti-cancer immunotherapy [143].

IL-10 contributes to Treg stability and phenotype by sustaining expression of FOXP3 and TGF-β [158,159]. IL-10 is now recognized not only as the most potent anti-inflammatory cytokine but also for its ability to enable cancer immune surveillance and tumor rejection [145,160]. Studies in mouse tumor models demonstrate the suppressive effects of Treg-derived IL-35, as treatment with IL-35-blocking mAb or Treg-restricted deletion of the IL-35:p35 heterodimer protein EBI3 leads to decreased tumor growth [148,161]. These studies highlight a contribution of IL-35-induced Tregs (iTR35) and the identity of the IL-35 receptor [162,163]. In patients with non-small cell lung cancer, an elevated number of IL-35⁺FOXP3⁺ Tregs is found in tumors compared to normal tissue controls, also positively correlating with PD-1-expressing cells [164]. Lastly, distinct Treg subpopulations can express IL-10 or IL-35, but collectively both cytokines are required to drive inhibitory receptor expression on tumor-infiltrating T cells in a BLIMP1-dependent manner [145,148]. While future human studies on modulations of IL-35⁺ Tregs are still warranted, IL-35 may serve as a novel viable therapeutic target in patients undergoing tumor immunotherapy.

Altering Treg fragility

The importance of the phosphatidylinositol-3-kinases (PI3K) pathway, which mediates signaling downstream of the TCR, in modulating Treg maintenance and stability has been studied in murine models [165,166]. Inhibition of the PI3Kδ isoform significantly reduced phosphorylation of Akt in Tregs but had no effect on effector CD4⁺ T cells, which was further confirmed using mutant mice with PI3Kδ inactive kinase [55]. PI3Kα and PI3Kβ render PI3Kδ redundant in effector CD4⁺ T cells but not in Tregs, enabling selective Treg depletion by PI3Kδ inhibitor in the TME and subsequent inhibition of tumor growth in a murine tumor model [55]. In addition, inactivation of PI3K p110δ in Tregs unleashes the anti-tumor activity of CD8⁺ T cells and results in inhibition of tumor growth and metastasis in multiple murine models [167]. A phase I trial is currently evaluating the efficacy and safety of the PI3Kδ inhibitor, INCB050465, combined with pembrolizumab in patients with advanced solid tumors, including melanoma, head and neck cancer and non-small cell lung cancer (NCT02646748^xx). The PI3Kδ-specific inhibitor, idelalisib, is approved to treat human B cell malignancies, such as relapsed/refractory chronic lymphocytic leukemia (CLL). However, immune-mediated adverse effects (irAEs) such as hepatotoxicity, enterocolitis and pneumonitis have been reported in idelalisib-treated patients, which could be explained by the systematic effects of PI3Kδ inhibitor on Treg survival or maintenance [168]. A phase II trial has been conducted to assess the effects of the PI3Kδ inhibitor, AMG319, in patients with head and neck cancer [169]. Unfortunately, 12 out of 21 patients discontinued treatment with AMG319 due to irAEs. In addition, an experimental study in murine models revealed a systemic reduction of Treg frequency in mice treated with the PI3Kδ inhibitor. Single-cell RNA sequencing analysis further revealed a loss of tissue-resident colonic ST2⁺ Tregs with expansion of pathogenic T helper 17 (Th17) cells and type 17 CD8⁺ T (Tc17) cells, which resulted in colitis in PI3Kδ inhibitor-treated mice [169]. These studies highlight the issue of modulating Tregs in non-malignant tissues by anti-tumor immunotherapies, which may cause severe irAEs. This highlights the importance of developing Treg-targeted therapies that are highly selective for intratumoral Tregs. Further preclinical and translational research is warranted to prove not only the effectiveness of Treg-targeted therapies but also their selectivity for intratumoral Tregs without affecting peripheral Tregs and effector T cell populations. It is also possible that alternate dosing strategies could be beneficial. Indeed, a recent study has shown that intermittent dosing can still enhance anti-tumor immunity while limiting irAEs [169].

The neuropilin-1 (NRP1)–semaphorin-4A (SEMA4A) axis stabilizes intratumoral Tregs and helps maintain their suppressive function [170] (Figure 1). NRP1 expression on Tregs can also promote DC interaction and facilitate intratumoral migration via VEGF binding to NRP1 [171,172]. NRP1⁺ Tregs are enriched in multiple human cancers compared to non-cancerous tissues. Infiltration of intratumoral NRP1⁺ Tregs is associated with poor prognosis in head and neck squamous cell carcinoma [168]. Importantly, loss of NRP1 does not cause instability or loss of FOXP3 expression or Treg identity, but rather loss of function, a phenotype referred to as Treg fragility [173]. In preclinical murine models, induction of Treg fragility results in tumor clearance while still maintaining peripheral tolerance. Further analysis has identified IFN-γ as the primary driver of Treg fragility [174].

The critical role of NRP1 in modulating Treg stability and suppressive function makes it an attractive therapeutic target. While current clinical development of antitumor therapy targeting NRP1 has been focused on its regulation of tumor angiogenesis through the NRP1–VEGFR2 axis [175,176], a recent study showed that the NRP1 antagonist Fc(AAG)-TPP11 selectively inhibits intratumoral NRP1⁺ Treg function and stability and enhances antitumor activity in the TME, with no impact on Tregs in peripheral tissues [177]. In murine models, Fc(AAG)-TPP11 inhibits Treg function by inducing Treg fragility, with lower expression of FOXP3 and enhanced production of IFN-γ. Increased numbers of intratumoral CD8⁺ T cells and enhanced cytotoxic function were also detected [177]. Another small-molecule NRP1 antagonist (EG01377) reduces production of TGF-β from Tregs in the presence of glioma-conditioned media [178]. Furthermore, a phase I trial evaluating the anti-NRP1 mAb ASP1948 is ongoing to investigate its tolerability and safety as a single agent and in combination with nivolumab (NCT03565445^xxi). Although the clinical significance and the potential of targeting Treg fragility through NRP1 remains under investigation, its correlation with clinical outcome makes this therapeutic approach promising.

Targeting Treg metabolism

Metformin is a well-known prescribed drug for type 2 diabetes but has recently garnered considerable attention for its antitumor effects [179,180]. Previous studies suggest that intratumoral Tregs have distinct metabolic programs and maintain their suppressive function by lactate uptake, while effector T cells depend on glycolysis and compete with tumor cells in the TME [52,181]. Data from in vitro studies show that metformin treatment causes activation of mTORC1 and metabolic reprogramming of Tregs toward glycolysis, which results in reduction of suppressive function and induced apoptosis of Tregs [180]. Administration of metformin demonstrates a reduction of Treg frequency in a murine model but not in peripheral lymphoid tissues, implying a potential for a novel therapeutic method of targeting Treg metabolism in the TME by metformin [180]. In contrast, patients with multiple sclerosis treated with metformin had a significant decrease in brain lesions and an increase in Treg frequency and suppressive function. This discrepancy highlights the need for more studies [182]. Furthermore, indoleamine 2,3-dioxygenase (IDO) and adenosine treatment within the TME can promote Treg proliferation and suppression. Metformin, IDO inhibitors, and A2AR inhibitors have been tested in clinical trials against multiple cancer types and support the novel therapeutic potential of these inhibitors of Treg metabolism without altering effector T cell function [183–185].

Concluding remarks and future perspectives

Tregs are crucial regulators of immune homeostasis, tissue repair, autoimmune diseases, and cancer. Several Treg-targeted therapies are currently under clinical investigation alone or in combination with other modalities, such as immune checkpoint blockade, targeted vaccine therapy, radiotherapy, and chemotherapy (Figure 2). Recent progress of Treg-targeted therapies highlights the importance of future studies centered on Treg biology and suppressive functions within the TME to develop effective cancer treatments. Potential targets include IL-35, which is an important Treg suppressive mechanism in tumor models [148,162]. Furthermore, TIGIT marks a Treg subpopulation that is highly suppressive against Th1, Th17, and CD8⁺ T cell function [94,96]. Preclinical research has found a unique metabolic program of Tregs compared to effector T cells in the TME [52]. While effector T cell functions are hampered by the low-glucose, hypoxic, and acidic environment within tumors, intratumoral Tregs maintain their suppressive functions by activating alternative metabolic pathways for lactate uptake, which provides opportunities to develop novel Treg-targeted interventions [52]. Multiple data suggest Tregs express CD39 to convert extracellular ATP to adenosine and suppress effector T cells in tumor models [1,186,187], yet further investigations, including human studies, are warranted and would provide additional therapeutic combinations to improve treatment efficacy against cancers.

In summary, while there has been some success targeting Tregs in the clinic, many gaps remain in our understanding of Treg biology within the context of the TME (see Outstanding Questions). Human Tregs are phenotypically and functionally heterogeneous within the TME, suggesting the possibility of additional mechanisms controlling Treg maintenance, metabolism, and suppressive functions within the TME. Molecular signaling within the immune system consists of extremely complex networks. However, multi-omics approaches combining novel computational tools may be beneficial to help understand gene regulatory networks that modulate Treg function as well as identify additional Treg biomarkers that define functionally suppressive Treg subpopulations within the TME. These analyses could lead to novel therapeutics targeting Treg subpopulations critical to tumor escape and tumor progression without affecting immune homeostasis across various cancer types.

Outstanding Questions Box:

1. Are there other suppressive mechanisms of Tregs as yet undiscovered?

2. Are there other markers that can identify effector Treg function within the TME?

3. What other molecular mechanisms drive Treg migration into the TME?

4. Are the transcriptional networks controlling intratumoral Treg function and survival distinct from effector T cells?

5. Can we achieve Treg-targeted therapies that are tumor-specific without affecting tissue Tregs?

6. How does the TME in different cancer types affect Treg biology?

7. How do Tregs interact with other cell populations in the TME, such as myeloid-derived suppressor cells, tumor-associated macrophages, and epithelial cells, and modulate their maintenance and function reciprocally?

Highlights:

• Current regulatory T cell-targeted therapies exhibit clinical benefit, highlighting the importance of targeting regulatory T cells in cancer.

• Understanding regulatory T cell biology, recruitment, and stabilization within the tumor microenvironment may provide enhanced cancer therapeutic opportunities to target regulatory T cells without perturbing effector T cell function or immune homeostasis.

• Multi-omics approaches combining novel computational tools may help define the complex gene regulatory networks that modulate regulatory T cell function and identify biomarkers that define functionally suppressive regulatory T cell subpopulations within the tumor microenvironment.