A telomere-associated molecular landscape reveals immunological, microbial, and therapeutic heterogeneity in colorectal cancer

Abstract

Background: Colorectal cancer (CRC) ranks among the most prevalent malignancies of the gastrointestinal tract and remains a leading cause of cancer-related mortality worldwide. Although telomere biology has been increasingly implicated in immune modulation and tumor progression, its clinical significance in CRC remains poorly understood.

Methods: We developed a telomere score, termed TELscore, by integrating transcriptomic and intratumoral microbiome profiles from publicly available colorectal cancer (CRC) cohorts. To comprehensively characterize TELscore subgroups, we performed pathway enrichment analysis, tumor immune microenvironment (TIME) profiling, and microbiome niche assessment. Whole-slide histopathological images (WSIs) and immunohistochemical (IHC) staining were utilized to visualize immune features, including tertiary lymphoid structures (TLSs), across subgroups. Patients were stratified into high and low TELscore categories, and the predictive robustness was validated across multiple independent training and validation cohorts. Chemotherapeutic drug sensitivity was evaluated using pharmacogenomic data from the Genomics of Drug Sensitivity in Cancer (GDSC) database. Furthermore, the predictive capacity of TELscore for immunotherapy response was independently assessed in an external cohort. Finally, single-cell RNA sequencing (scRNA-seq) analysis was conducted to further dissect the cellular landscape and immunological heterogeneity within the TME.

Results: TELscore stratified patients into two biologically and clinically distinct subgroups. The high TELscore group, which exhibited significantly shorter DFS, showed marked enrichment of tumorigenic pathways such as EMT, along with a distinctly immunosuppressive TME. This was reflected by elevated ESTIMATE/TIDE scores and corroborated by CIBERSORT, which revealed increased infiltration of M0 macrophages and upregulation of immunosuppressive signatures. In contrast, the low TELscore group was enriched for cell cycle related pathways, including E2F targets and the G2/M checkpoint, and demonstrated higher infiltration of pro-inflammatory M1 macrophages. 16S rRNA sequencing further revealed a divergent intratumoral microbiome between subgroups, the high TELscore group harbored significantly greater relative abundance of Selenomonas and Lachnoclostridium, two pathogenic genera previously associated with colorectal tumorigenesis. Complementary histopathological assessment via WSI demonstrated a marked absence of intraTLSs in high TELscore tumors. From a therapeutic standpoint, high TELscore tumors exhibited reduced sensitivity to standard chemotherapeutic agents-including Fluorouracil, Irinotecan, Oxaliplatin, and Docetaxel-as reflected by elevated IC50 values. Conversely, these tumors demonstrated increased susceptibility to MAPK pathway inhibitors, such as Selumetinib and Trametinib. Notably, TELscore also served as a robust predictor of immunotherapy response, which was validated in the IMvigor210 cohort. Finally, scRNA analysis highlighted profound cellular and functional divergence between TELscore subgroups. We identified intensified intercellular communication between inflammatory macrophages and fibroblasts, reinforcing the presence of an immunosuppressive niche.

Conclusion: TELscore is a robust stratification tool that captures the interplay between tumor biology, immune characteristics, and microbial ecology in colorectal cancer. By identifying clinically relevant subtypes with distinct therapeutic vulnerabilities, TELscore offers a powerful framework to advance personalized treatment and precision oncology.

Keywords: colorectal cancer; microbiome; prognosis; telomere; tumor microenvironment.

FIGURE 1

The overall workflow of this study.

FIGURE 2

Construction and validation of the TELscore. (a), Features selected by the LASSO Cox regression model. (b), Selected genes and their corresponding coefficients. (c), Kaplan–Meier survival analysis in the training cohort (TCGA-CRC). (d–h), Survival validation in external cohorts: GSE39582, GSE14333, GSE28722, GSE38832, and GSE41258. (i), Heatmap of Hallmark pathway enrichment between high and low TELscore groups.

FIGURE 3

Molecular and immune characteristics associated with TELscore. (a), Heatmap of immune cell infiltration across TELscore groups. (b), IOBR-derived expression profiles of immune checkpoint genes, immunosuppressive markers, exhaustion signatures, and exclusion indicators. (c), CIBERSORT based deconvolution analysis comparing immune cell compositions between high and low TELscore groups.

FIGURE 4

Immune microenvironment and histopathological features of TELscore groups. (a), Boxplot comparing enrichment of oncogenic pathways between TELscore groups. (b), ESTIMATE-derived stromal and immune scores across TELscore groups. (c–g), TME-related scores and TIDE-based immunotherapy prediction scores across groups. (h), Representative whole-slide imaging (WSI) for high and low TELscore tumors. (i), Immunohistochemical (IHC) staining of selected TELscore features in normal versus tumor tissues.

FIGURE 5

Immunotherapy response, drug sensitivity, microbiome niche and single cell analysis. (a,b), Comparison of drug sensitivity (IC50 values) between different TELscore groups. (c), Differential intratumoral microbiome composition based on 16S rRNA sequencing between high and low TELscore groups. (d), Prognostic value and immunotherapy response prediction of TELscore in the IMvigor210 cohort. (e–h), Single-cell transcriptomic analysis based on Zhang et al., including TELscore distribution and predicted immunotherapy responsiveness. (i), Cell–cell communication networks inferred from single-cell data.