Systematic analysis of markers for T-cell differentiation revealing CD95 as a novel biomarker for prognosis and immunotherapeutic efficacy in colon cancer

Abstract

Colon cancer is the third most frequently diagnosed cancer worldwide. Considerable progress has been made in the therapeutic strategies and the accuracy of predicting the patients' prognosis. In the past decade, immunotherapy, represented by programmed cell death 1 (PD-1) and programmed cell death 1 ligand 1 (PD-L1) signaling inhibitors have made remarkable advances in many solid tumors, but limited effect in colon cancer. In particular, colon cancer demonstrates a remarkably poor response to immunotherapy compared with other cancers, with a notable exception of tumors harboring high microsatellite instability (MSI-H) or deficient mismatch repair (dMMR).¹ MSI-H or dMMR tumors are characterized by high mutational/neo-antigen burden, and an inflammatory tumor microenvironment with abundant tumor-infiltrating lymphocytes (TILs).² Although better efficacy could be achieved in a small group of patients with MSI-H or dMMR tumors, a proportion of other patients could benefit from immunotherapy. Thus, more appropriate biomarkers are needed to be discovered to predict the patients' prognosis and immunotherapeutic efficacy.

Different T-cell subsets in the tumor immune microenvironment (TIME) are closely related to the antitumor response and the prognosis of tumor patients. Tumor-specific CD8⁺ T cells are the core cellular components that exert antitumor effects in TIME, by recognizing specific receptors on the tumor surface and secreting cytokines or via the Fas/FsaL pathway. Multiple T-cell subsets are characterized by respective differentiation markers. Currently, with the deepening research of T-cell subsets, the detection of relevant immune indicators and the formulation of therapeutic schedules have attracted increasing clinical attention. By monitoring the expression of T-cell differentiation markers in patients with colon cancer, the benefit of patients receiving immunotherapy can be further evaluated, thus providing a theoretical basis for judging the clinical prognosis and designing more feasible treatment plans.

In this research, we systematically analyzed the prognostic values of multiple markers for different T-cell subsets³ using the Cancer Genome Atlas (TCGA) data set. Among all markers for different T cell subsets, we found that CD95 expression was correlated with better overall survival (OS) and progression-free survival (PFS), while BCL2 and CD122 were only correlated with PFS (Figure 1A,B). Therefore, CD95 was speculated to be associated with prognosis in colon cancer and was selected for further investigation and validation. The analysis of single-cell RNA-sequencing data set confirmed that CD95 was highly expressed in immune cells, especially in various T-cell subtypes (Figure S1A–C). CD95 is a member of the tumor necrosis factor receptor (TNF-R) superfamily and treated as characteristic marker of memeroy cells. It could also be used as the most classical death receptor to induce apoptosis. The CD95-induced apoptosis pathway is a powerful way for the immune system to clear virus-infected cells or tumor cells, therefore CD95 was initially identified as a tumor suppressor gene.

Next, to understand the potential biological role of CD95 in colon cancer, we applied Gene Set Enrichment Analysis (GSEA) in the TCGA data set. The results showed that CD95 was significantly related to multiple vital immune-related processes, such as allograft rejection, inflammatory response, and interferon-gamma response (Figure 1C and Table S1). Given the tight correlation between CD95 and immune-related processes in colon cancer, we next assessed the correlation between CD95 and comprehensive immunomodulators consisting of chemokines, receptors, MHC molecules, immunoinhibitors, and immunostimulators, which were associated with features of TIME and the responses to immunotherapy. The results showed that CD95 was positively correlated with most immunomodulators (Figure 1D and Table S2). In addition, we also analyzed the correlation between CD95 and T-cell subsets calculated by the ImmuCellAI tool, and the results showed that the strongest positive correlation was observed between CD95 and cytotoxic T cells (Table S3). All findings suggested CD95 was highly correlated with the inflamed TIME. In particular, substantial evidence suggests that patients with MSI-H/dMMR colon cancer benefit more from immunotherapy and have an optimistic immunotherapeutic efficacy compared with patients with proficient MMR (pMMR). Subsequently, we used three validated cohorts, namely the TCGA cohort, the GSE39582 cohort, and the CPTAC2 cohort, and the results indicated that CD95 was upregulated in tumors with MSI-H/dMMR status (Figure S2A–C). Moreover, the correlations between CD95 and several common mutations associated with immunotherapeutic response, including NF1, POLE, ARID1A, and BRAF, were also assessed. The results exhibited that tumor with gene mutations expressed higher CD95 (Figure S3A–D). Furthermore, a small-scale immunotherapy cohort consisting of 29 colon cancer specimens receiving immunotherapy was recruited to further validate the association between CD95 expression and immunotherapeutic response. Immunofluorescence exhibited that CD95 was expressed in immune cells and tumor cells (Figure 1E). Given the function of CD95 in T-cell differentiation, only CD95-positive immune cells were calculated. The results suggested that patients with well immunotherapeutic responses expressed high CD95 compared with those with poor response (Figure 1E). Finally, survival analysis showed patients with high CD95⁺ immune cell infiltration exhibited well prognosis than that with low CD95⁺ cell infiltration (Figure 1F). To sum up, these results supported that CD95 could serve as a novel and promising biomarker of benefit to immunotherapy in colon cancer.

CD95 and CD95L are widely expressed in colon cancer, previous studies in colon cancer showed that switching CD95 signaling from apoptotic to nonapoptotic ERK/NF-κB signaling facilitates tumor incidence, local growth, and the ability to metastasize to the liver, which depends on expression of CD95L. Furthermore, existing evidence suggests that CD95 mainly implements nonapoptotic signaling pathways such as NF-κB, MAPK, and PI3K that are involved in cell migration, differentiation, survival, and cytokine secretion. Not only in colon cancer, but in lung cancer, glioblastoma, ovarian cancer, and liver cancer, CD95 expression in some conditions conversely activates its nonapoptotic pathway, which plays an important role in tumorigenesis and progression.⁴ In addition, the latest research showed that targeting CD95 in melanoma increased the levels of cytotoxic T cells and dendritic cells.⁵ We hypothesized that CD95 may mediate the immunogenic death of tumor cells. However, the concrete effects of CD95 on tumor biological behavior and their relevance to immunotherapy efficacy are unclear. Based on the above finding that CD95 was positively associated with multiple immune-related processes and immunomodulators in the TIME, we speculated that in colon cancer CD95 was associated with inflamed TIME and identified immuno-hot tumors which favored antitumor immunity. Moreover, we also found that CD95 was upregulated in tumors with MSI-H/dMMR status. Colon tumors with dMMR status have association with a higher density of CD8⁺ TILs and higher tumor mutation burden (TMB), which are qualities that seem to be biomarkers of immunotherapeutic response and demonstrated a positive prognostic value. Considering the high incidence and mortality of colon cancer and the dilemma in obtaining satisfactory treatment results, CD95 as an appropriate biomarker is needed to help identify candidates that could benefit from immunotherapy and can appraise clinical outcomes and therapeutic responses, which may make up for the deficiency of existing immunotherapy strategy.

In conclusion, the current study identified potential biomarkers to predict well clinical outcomes and response to immunotherapy in colon cancer. After systematic analysis, CD95, a memory marker of T cells, might be an appropriate and accurate biomarker in clinical practice to predict the prognosis and immunotherapeutic responses. Scientifically, the identification of novel predictive biomarkers will greatly expand our understanding of the mechanisms of colon cancer immunity, providing information on the decision of treatment strategies in individuals.

Yuxin Shi: Data curation (equal); formal analysis (equal); methodology (equal); resources (equal); visualization (equal); writing—original draft (equal). Jie Mei: Data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); validation (equal); writing—original draft (equal). Rui Hou: Formal analysis (supporting); validation (supporting). Hao Wang: Formal analysis (supporting); resources (supporting). Junli Ding: Conceptualization (equal); funding acquisition (supporting); project administration (equal); writing—review and editing (equal). Junying Xu: Conceptualization (equal); funding acquisition (lead); project administration (equal); writing—review and editing (equal). All authors have read and approved the final manuscript.

The authors declare no conflict of interest.

Ethical approval was granted by the Clinical Research Ethics Committee of Nanjing Medical University (2022-383). This is a retrospective study, and trial registration number is not needed. Written informed consent was obtained from all participants.