Regulatory T cells (Tregs) play important roles in maintaining immune homeostasis and preventing autoimmunity in normal physiological conditions by suppressing self-reactive T cells. However, under pathophysiological conditions such as cancer, their suppressive functions promote the growth and progress of disease by hindering antitumor immune responses (1). Earlier identified as CD4⁺CD25⁺ T lymphocytes (2), the majority of Tregs were later characterized by expression of Forkhead box protein 3 (FoxP3), a master regulatory transcriptional factor (3), which is critical for the development and suppressive functions of Tregs (4). Tregs are heterogeneous in their development and can be grouped into two major subsets; natural or thymic-derived Tregs (tTregs) and peripheral-induced Tregs (pTregs). However, specific markers, which could reliably distinguish between the two subsets, are elusive. Helios, a member of the Ikaros transcription factor family, was proposed as a marker to distinguish tTregs from pTregs (5), but this work was later questioned by several other studies (6–8). Helios, however, is shown to be a marker for T cell activation and proliferation (9), and its expression is essential for the stable Treg inhibitory activity (10). We have recently shown that Helios is a marker of activated Tregs expressing immunosuppressive molecules GARP/latency-associated peptide (LAP) (11) and that co-expression of Helios and FoxP3 with GARP/LAP can be used to identify expanded Treg subset in cancer patients (12).

The immunosuppressive mechanisms of Tregs are not fully understood but immune checkpoint receptors [ICR; such as programmed death 1 (PD-1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4)], immunosuppressive molecules such as CD39 and LAP have been shown to play critical roles in their functions (13, 14). CTLA-4 (CD152) is constitutively expressed by Tregs where it enhances Treg suppressive functions. Studies in human and murine models showed that blocking CTLA-4 signaling by monoclonal antibodies increased proliferation of T effector cells and impaired Treg suppressive activity (15, 16). PD-1 (CD279), like CTLA-4, is highly expressed on Tregs and helps in Treg immunosuppression (17), and blocking PD-1 pathway with anti-PD-1 monoclonal antibody prevented Treg-mediated suppression of effector T cells in vitro (18). Several other studies have also reported the elevated levels of CD39, an ectonucleotidase, on the cell surface of Tregs (13, 19, 20). CD39 is involved in the conversion of ATP to adenosine which is a potent immunosuppressive molecule and blocks proliferation and function of effector T cells (21). Schuler et al. have characterized Treg subsets based on the presence or absence of CD39, FoxP3, and CD25, and showed that CD4⁺CD39⁺FoxP3⁺CD25⁺ Tregs suppressed proliferation of effector T cells (22).

Regulatory T cells have been implicated in progression of cancers due to their ability to suppress other T lymphocytes and therefore are considered to be a major barrier for antitumor immunotherapies. Several studies have shown Treg infiltration at tumor sites in various cancers including colorectal cancer (CRC) and their enrichment within tumor-infiltrating lymphocytes (TILs) (1, 20, 23, 24). CRC is one of the most common malignancies worldwide (25). CRC is mainly an inflammation-associated cancer, and Tregs are expanded in tumor microenvironment (TME) and play important roles in the pathogenesis of CRC (26). Though accumulation of FoxP3⁺ Tregs in majority of human carcinomas is associated with tumor growth and poor outcome, some studies in CRC have shown the presence of Tregs in tumors is associated with better outcome (27), suggesting that in CRC Treg functions might be antitumorigenic by suppressing inflammation at the early stages of the disease. However, some other studies have shown that Tregs in colorectal tumor tissue (TT) might contribute to disease progression by suppressing antitumor immune responses (28, 29). These seemingly contradicting results suggest the presence of different Treg subpopulations with varying degrees of suppression (30). Saito et al. (31) showed that these functionally heterogeneous Treg subpopulations affected prognosis of CRC in different ways, as CD45RA⁻FoxP3^lo non-Treg subset secreted pro-inflammatory cytokines and patients with significant accumulation of this Treg subset had better prognosis compared to those with highly suppressive CD45RA⁻FoxP3^hi Treg subset. They also reported that pro-inflammatory FoxP3^lo Tregs have lower expression of CTLA-4 and TIGIT (T cell immunoreceptor with Ig and ITIM domains), indicating that immunosuppressive molecules on T cells and Tregs affect the outcome of CRC disease. Therefore, studies focusing on characterization of Tregs in the colorectal TME with potential implications for CRC treatment are warranted.

In this preliminary study, we investigated the levels and phenotypes of immune cells infiltrating colorectal TT and compared them with adjacent non-tumor normal tissue (NT) and peripheral blood from the same patients. We also characterized different subsets of TILs and assessed the expression of PD-1, CTLA-4, and CD39 on different FoxP3 and Helios Treg subsets. We found that CD4⁺FoxP3⁺Helios⁺ Tregs are expanded in the TME of CRC patients compared with normal tissue and peripheral blood. Additionally, we observed increased frequency of PD-1/CTLA-4 and PD-1/CD39 co-expressing cells within FoxP3⁺Helios⁺ and FoxP3⁺Helios⁻ Treg subsets. CD4⁺FoxP3⁺Helios⁺ and CD4⁺FoxP3⁺Helios⁻ T cells expressed similar relative percentages of ICRs/CD39, but the absolute percentages were significantly higher in CD4⁺FoxP3⁺Helios⁺ Treg subset, indicating the potent immunosuppressive potential of CD4⁺FoxP3⁺Helios⁺ Treg subset.

In this study, we focused on phenotypic characterization of intratumoral Tregs in CRC patients and observed the presence of different immunosuppressive molecules including CTLA-4, PD-1, and CD39. CRC being an inflammatory disease is characterized by the accumulation of different types of immune cells within the TME. We observed significant accumulation of CD45⁺CD3⁺ cells within colorectal TT, which is in line with previous studies (35). Majority of these infiltrating CD3⁺ cells were CD4⁺, and this selective infiltration of CD4⁺ T cells within TT might be partly due to the requirement of CD4⁺ T cells for efficient priming and activation of CD8⁺ T cells (36). This also suggests a pro-tumorigenic function of infiltrating CD4⁺ T cells in CRC patients, as accumulation of CD8⁺ T cells would have indicated a tumor-specific immune response as a protective function against cancer. Moreover, the increase in CD45⁺CD3⁻ cells within TT indicated the presence of cells of myeloid origin and might be of importance in colorectal tumorigenesis. We have recently shown that granulocytic cells of myeloid origin were elevated in peripheral blood and the TME of this CRC patients’ cohort and their levels correlated with tumor stage and histological grades of the disease (32). These tumor-infiltrating CD4⁺ T cells expressed high CD25, which indicated the presence of Tregs, and this can be a result of conversion of CD4⁺CD25⁻ cells into CD4⁺CD25⁺ cells in which TGF-β is involved (37). These cells also expressed relatively higher levels of CD39 and LAP compared with NILs, indicating their suppressive characteristics.

Immune checkpoint receptors such as PD-1 and CTLA-4 not only play important roles in immune homeostasis, but these inhibitory pathways are also utilized by tumor cells for immune response evasion. In the present study, we observed relatively higher frequency of CD8⁺ T cells expressing PD-1 in TILs, indicating their “exhaustive state,” which might have rendered them unresponsive in mounting any cell-mediated immune responses against cancer and blocking PD-1/PD-L1 inhibitory pathway restored their functions (38). Zhang et al. (39) reported high PD-1 expression on dysfunctional CD8⁺ T cells from PBMC and tissue of human non-small cell lung cancer, and blocking PD-1/PD-L1 pathway using PD-L1-specific antibody partially restored proliferation and cytokine production in these cells. Dunne et al. (23) also observed a significant increase in PD-1 expressing CD8⁺ T cells within colorectal TT compared with adjacent NT. They also reported a similar increase in CD4⁺PD-1⁺ T cells in TT relative to NT, which we did not observe. However, a significant difference in PD-1-expressing CD4⁺ T cells between NILs and TILs was observed if absolute percentages were compared. Overall, these contrasting results indicate that PD-1 might downmodulate the activating signals in CD8⁺ T cells in TT, but its role in CD4⁺ T cells requires further investigation.

Majority of the studies have reported the expression of either CTLA-4 or PD-1 on immune cells in cancer; however, limited studies described the possible associations between ICR pathways in cancer. Duraiswamy et al. (40) showed the co-expression of PD-1 and CTLA-4 was associated with antigen-specific T-cell dysfunction in xenograft mouse model and reported that dual blockade of these ICR pathways restored T-cell functions. Here, we showed a high frequency of CD4⁺ T cells co-expressed PD-1 and CTLA-4 in TILs, which might synergistically dampen T-cell activation and function in CRC patients. The over-expression of CD39 but not PD-1 on CD4⁺ T cells in NILs suggested that CD39 or PD-1 alone might not contribute to tumor-specific immune suppression, and that co-expression of CD39 with PD-1 in TILs indicated a T-cell dysfunction mechanism used by tumor cells to evade host immune response.

The marked increase in CD4⁺FoxP3⁺ Treg levels in TT compared to NT and peripheral blood of CRC patients in our study was in accordance with previously published reports from other groups, where an expansion in Treg levels in TILs in various cancers including CRC was demonstrated (23, 26, 33). Other studies, however, also reported an increase in circulating Treg levels in CRC patients compared to HDs (24, 29), which was not observed in our patient’s cohort. This disparity could again be due to the limited sample size. A role of FoxP3-expressing Tregs has remained ambiguous in colorectal tumor progression, as some studies showed a protective role of FoxP3-expressing Tregs in CRC (27), while other studies have shown it to be associated with an adverse outcome of the patients (28, 29). The reasons for these discrepancies, though not clear, might be due to the methods employed for identifying Tregs in tissue, or due to the presence of functionally distinct subpopulations of Tregs with different levels of immunosuppression as reported by Miyara et al. (30). Another reason might be due to the presence of different immunosuppressive molecules such as Helios, PD-1, CTLA-4, and CD39 in Tregs. Helios is upregulated in Tregs (41) and is essential for the stable Treg inhibitory functions (10). Moreover, our group has recently suggested that FoxP3⁺Helios⁺ Tregs represent a functional subset with stronger suppressive characteristics (11). Consistent with these data, we observed a strong positive correlation between FoxP3 and Helios expression in TILs. Majority of FoxP3⁺ Tregs co-expressed Helios in these TILs, indicating these with greater immunosuppressive potential than either FoxP3 or Helios expressing Treg subset.

One of the reasons for Treg accumulation at tumor site can be due to the production of chemokines by tumor cells and stroma in the TME, which mediates Treg influx within the TT. Furthermore, the conversion of FoxP3⁻ cells into FoxP3⁺ cells in presence of TGF-β or increased proliferation of Tregs can lead to their expansion (42). Our data suggested that the TME might enhance the conversion of FoxP3⁻Helios⁺ Treg subset into FoxP3⁺Helios⁺ Treg subset as FoxP3⁻Helios⁺ subset constituted more than half of different FoxP3 and Helios expressing Treg subset in NT and peripheral blood of cancer patients, but within the TME, FoxP3⁻Helios⁺ comprised of around 30% of Treg population and FoxP3⁺Helios⁺ Treg subset increased significantly and constituted more than 54% of the population among the three Treg subsets.

Apart from CTLA-4 and PD-1 immunoregulatory pathways, FoxP3⁺ Tregs can also suppress effector T cell functions through CD39/CD73 pathway, which involves adenosine hydrolysis from ATP (19). Nikolova et al. (43) showed that blocking of CD39 using neutralizing antibodies attenuated immunosuppressive functions of Tregs. In our current study, CD39 but not PD-1 expression was observed in FoxP3⁻Helios⁺ and FoxP3⁺Helios⁺ Treg subsets in normal tissue, whereas within TT, co-expression of PD-1 and CD39 was observed in all the three Treg subsets, indicating that CD39 expression alone is not sufficient for Treg-mediated immune suppression, but its co-expression with PD-1 is required for these Treg subsets for tumor-specific immune suppression. Furthermore, these FoxP3⁺Helios⁺ and FoxP3⁺Helios⁻ Tregs also co-expressed PD-1/CTLA-4, suggesting that synergistic associations between different immune inhibitory molecules might regulate immunosuppressive functions of these Treg subsets.

Studies have shown that blocking CTLA-4 pathway resulted in reduction in Treg-mediated immunosuppression and hence activation and proliferation of T cells, whereas PD-1 pathway blockade removes the “exhaustive” state of T cells and restores their antitumor activity. Blocking only one of the ICRs might not be very effective in CRC patients (44). In a recent Phase I clinical trial with anti-PD-1 blocking antibody alone, no objective responses were seen in CRC patients (45). Additionally, CTLA-4 monotherapy was also not very successful in CRC models, but its potent activity was seen in combination with other therapies (46). Preclinical studies and initial clinical trials have shown that the dual blockade of these ICRs might result in more effective antitumor immune responses (47–49). Overall, we report the co-expression of different immunosuppressive molecules in FoxP3⁺Helios⁺ and FoxP3⁺Helios⁻ Treg subsets, which might synergistically dampen T-cell activation and function and suppress tumor-specific immune responses in CRC patients. The effect of these different Treg subsets on the clinical outcome could not be validated in our patients’ cohort due to limited sample size. Our data suggest that dual blockade of these immunosuppressive molecules might have the potential to be used for more effective strategy for inducing antitumor immune responses in CRC patients.

Ethical approval (protocol no. 13/23-CRD 244/13) was obtained from Al Ain Medical District Human Research Ethics Committee. All participating individuals signed written informed consent prior to collection of samples. All experiments were performed in accordance with relevant guidelines and regulations.

ASSK performed experimental work, data analysis, and wrote the manuscript. ST performed experimental work and data analysis. EE conceived the idea, designed the study, obtained fund, analyzed and interpreted data, and wrote and revised the manuscript. HE contributed to sample collection and acquisition of patients’ clinical data. BA obtained funds and revised the manuscript. All authors were involved in the final approval of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.