Immune cells have the ability to recognize and destroy cancer cells. CD4+ T cells may target cancer cells by modulating the tumor microenvironment [20, 21]. CD4+ T lymphocytes are predominantly T regulatory cells in CRC tissues, and they express several IC molecules such as PD-1, CTLA-4, TIM-3, and LAG-3 [8]. High levels of Treg-related markers were observed in the TME in CRC patients, suggesting their potential effects in carcinogenesis [8, 22, 23]. In contrast to other solid tumors, high levels of tumor-infiltrating FoxP3+ Tregs were related with increased survival in CRC patients [9, 24]. Recent studies reported that tumor-infiltrating CD39+ Tregs in CRC patients expressed different markers such as OX-40, CTLA-4 and ICOS, implicating their high immunosuppressive abilities in inhibiting anti-tumor immune responses [25, 26]. Correale et al., showed that a higher level of FoxP3+ T-lymphocyte tumor infiltration in CRC patients receiving chemotherapy or chemo-immunotherapy was a favorable prognostic marker [27]. Moreover, a high frequency of FoxP3+ Tregs within tumor lead to a promising outcome in CRC, suggesting that FoxP3+ Tregs are one of the most useful indicators for predicting the prognosis of CRC [7, 28,29,30]. Another study found that CD8+:FoxP3+ cell ratios were significantly correlated with distant-recurrence-free survival (DRFS) in the CRC TME [31]. Also, they reported that high numbers of FoxP3+ cells were associated with longer overall survival (OS) and DFS, although non-significantly [31]. Interestingly, FoxP3+ T cells only have a positive effect on survival in colon tumors that have low levels of CD8+ T-cell infiltration [32]. Another study found that TGF-β, which is produced by tumors, has been linked to an increase in the number of intratumoral FoxP3+ Tregs [33]. Moreover, they found that intratumoral CD8+ T cell:FoxP3+ Treg ratio positively correlated with longer DFS and OS [33]. Nevertheless, functionally different subgroups of tumor-infiltrating FoxP3+ Tregs contribute in opposing ways to determining CRC disease prognosis [6].
Importantly, it has been demonstrated that circulating Tregs are effective in suppressing antitumor immunity, leading to an adverse outcome of CRC patients [10, 34]. We found that higher frequencies of CD4+FoxP3+ Tregs in circulation were associated with shorter DFS, implicating the harmful effect of these suppressive cells in inhibiting anti-tumor immune responses in circulation. However, high frequencies of this Treg subset in the TME were associated with longer DFS, indicating the beneficial anti-inflammatory role of CD4+FoxP3+ Tregs in the TME of CRC patients.
Expression of FoxP3 was highly and positively correlated with the expression of Helios on T cells within PBMCs and TILs in cancer patients [17]. In tumor tissues, the majority of Tregs co-expressed both FoxP3 and Helios, suggesting higher immunosuppressive potentials than cells with single expressions of FoxP3 or Helios [17]. Classification of FoxP3+ Tregs into subsets helps to investigate Treg cell differentiation in both normal and disease conditions, as well as to alter immune responses by modulating specific FoxP3+ Treg subpopulations [35]. Our group has also recently proposed that FoxP3+Helios+ Tregs constitute a functional subset of Tregs with higher suppressive characteristics [36]. Tumor tissues in CRC patients were characterized by high levels of Helios+ Tregs compared to PBMCs and normal colon tissues [24, 37], suggesting their potential roles in CRC progression [12]. In this study, we found that a high frequency of CD4+FoxP3+Helios+ Tregs in blood was associated with shorter DFS, suggesting the potential role of this highly immunosuppressive Treg subset in inhibiting anti-tumor immune responses, and consequently worsening clinical outcomes. Importantly, it is hypothesized that the TME could enhance the induction of the FoxP3+Helios+ Treg subset from the FoxP3−Helios+ T cell subset [17]. Interestingly, we found that high frequencies of CD4+FoxP3−Helios− TILs were significantly associated with worse DFS, suggesting that these cells could induce inflammation in CRC TME.
In TME, certain tumor ligands bind to inhibitory molecules on T cells, such as CTLA-4, PD-1, TIM-3, and LAG-3 and others, which in turn produce immune-suppressive mediators, leading to the failure of cancer elimination [18, 38]. CD4+PD-1+ T cells were predominantly detectable in tumor tissues of CRC patients [39], which may lead to T-cell exhaustion and cancer progression [40, 41]. Additionally, CRC patients with high expression of PD-1 had worse TNM staging and DFS, compared with those with low expression [42]. In agreement with these studies, we found that high frequencies of FoxP3+Helios−PD-1+ were associated with shorter DFS in circulation. More samples might be required to determine possible associations of these subsets with DFS in tumor tissues.
TIM-3 is frequently overexpressed on exhausted CD4+ T cells in CRC patients, suggesting this could be associated with worse prognoses [43,44,45]. Moreover, TIM-3 was correlated with CRC progression and might be a possible therapeutic target [46]. Arai et al., found that TIM-3 expression on CD4+ T cells was significantly increased after CRC operation [47]. Additionally, they found that the production of IFN-γ was linked to TIM-3 and PD-1 expression on CD4+ and CD8+ T cells, suggesting that TIM-3+PD-1+CD4+ and CD8+ T cells are highly dysfunctional [47]. On the other hand, Zhang et al., found that TIM-3 expression either in the primary or metastatic tumor was associated with better progression-free survival (PFS) in renal cell carcinoma [48]. We found that high frequencies of TIM-3+ and FoxP3+Helios+TIM-3+ CD4+ T cells were associated with longer DFS in circulation. Of note, most studies investigated TIM-3 expression in bulk tumor tissues but not on specific T cell subsets. This is the first study to indicate that TIM-3 expression on T cells is associated with better DFS in CRC.
We have previously reported that mRNA level of CTLA-4 in tumor tissues was increased in advanced stages of CRC, suggesting their possible effects in CRC progression [12]. Moreover, we have shown that there was a significant elevation in levels of CD4+CTLA-4+ T cells only in PBMCs of CRC patients with advanced stages, suggesting that there is a relationship between increased levels of CTLA-4+ Tregs and CRC progression [8]. In this study, high frequencies of CTLA-4 expressed in different Treg subsets were associated with worse DFS in circulation. Therefore, targeting CTLA-4 on these subsets might have beneficial roles in CRC.
A recent study showed that overexpression of LAG-3 on tumor tissues was associated with worse prognosis in patients with microsatellite instability high (MSI-H) colon cancer [49]. In addition, we have previously reported that mRNA level of LAG-3 was higher in PBMCs of CRC patients than those of healthy controls [12]. Moreover, the frequency of LAG-3 in tumor tissues was associated with differentiation, lymph node metastasis, and invasion in CRC patients [50]. In our study, we found that there were no associations between frequencies of CD4+LAG-3+ T cells with DFS. Due to weak overall LAG-3 expression in original study [8], more samples are needed to investigate the role of LAG-3 in DFS in CRC patients.
CRC patients with mismatch repair deficiency (dMMR) usually have worse prognosis than patients without dMMR [51]. Notably, over-expressions of FoxP3, IL1β, IL17, TGF-β and IL6 were associated with the microsatellite stability MSS phenotype [52, 53]. It has been shown that a high frequency of tumor-infiltrating FoxP3+ Tregs predicts improved survival in mismatch repair-proficient CRC patients [53]. Moreover, in stage II MSS CRC, it has been found that low frequencies of both FoxP3+ and CD3+ TILs were associated with the highest progression risk [54]. In this study, we were not able to investigate such association because only 4 CRC patients had dMMR (12.5% of this study cohort, as shown in Table 1); the percentage of dMMR patients was low as expected in CRC patients [55].
Our study highlights the potential of some CD4+ T cell subsets as predictive biomarkers associated with worse DFS in CRC patients. Overall, high frequency of tumor-infiltrating FoxP3−Helios− T cells, and high frequencies of circulating FoxP3+, Helios+, FoxP3+Helios+, FoxP3+Helios−PD-1+, FoxP3+Helios+CTLA-4+, FoxP3+Helios−CTLA-4+ Tregs, and FoxP3−Helios+CTLA-4+ T cells, are associated with shorter DFS. Targeting these immune cell subsets in CRC patients could improve clinical outcomes. On the other hand, high frequencies of CD4+TIM-3+, FoxP3+Helios+TIM-3+ Tregs and FoxP3−Helios+TIM-3+ T cells in circulation are associated with longer DFS in CRC patients, suggesting that T cells expressing TIM-3 could be activated cells with improved anti-tumor activities.
Most available studies investigated expression of ICs in bulk tumor tissues but not on specific CD4+ and CD8+ T cell subsets. To date, this is the first study to investigate the associations of different CD4+ Treg subsets and immune checkpoints-expressing CD4+ T cells with DFS in CRC patients. In addition to CD4+ T cell subsets in this study, we have also investigated the association of CD8+ T cell subsets with DFS in CRC patients (Alsalman et al., submitted for publication). However, multi-center investigations are required to confirm these findings in larger cohorts of patients. Moreover, additional investigations are required to determine the exact function of these cell subsets in the TME and circulation of CRC patients. Our data suggest that different CD4+ Treg/T cell subsets in circulation or in the TME play different roles in DFS of CRC patients. Clearly, identification of the exact subpopulations contributing to clinical outcomes is critical for prognosis purposes and therapeutic approaches.