TERT links telomere length to cancer risk by integrating genomic instability and immune modulation

Abstract

Background

Telomere homeostasis serves as a key regulatory mechanism linking aging and cancer. While telomere attrition imposes a proliferative barrier by inducing cellular senescence, abnormal telomere elongation circumvents this constraint, thereby granting malignant cells unlimited replicative capacity. This study systematically explores the causal relationship between telomere length and cancer risk, with the goal of elucidating the molecular pathways involved in telomere-driven tumorigenesis.

Method

Mendelian randomization (MR) was employed to establish a causal link between telomere length and pan-cancer susceptibility. Multiple MR models were employed to ensure the robustness of the results. Telomere-associated genes were identified through SNPense analysis, followed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment. Key gene regulatory networks were identified and visualized using the MCODE and cytoHubba algorithms. The expression profile of TERT across diverse cancer types was analyzed using TCGA datasets, and its diagnostic potential was evaluated via receiver operating characteristic (ROC) curve analysis. Correlations between TERT expression and immune cell infiltration were further explored.

Results

Longer telomere length was significantly associated with an increased risk of 33 cancer types (IVW OR = 1.27–1.41, all P < 0.001). A total of 143 telomere-related genes were identified, with functional enrichment highlighting their involvement in genome integrity and telomere maintenance. TERT emerged as the most influential hub gene (MCC score = 169). Transcriptomic analyses from TCGA demonstrated widespread TERT overexpression in 16 cancers (e.g., CHOL, LIHC, LUAD), a finding corroborated by RT-qPCR validation. TERT exhibited high diagnostic performance (AUC > 0.85 in 16 cancers, peaking at AUC = 0.97 in LUSC). Immune infiltration analysis revealed a positive correlation with Th2 cells (r = 0.42) but negative correlations with dendritic cells (r = − 0.38) and macrophages (r = − 0.31).

Conclusion

This study proposes a comprehensive framework linking telomere length regulation to cancer progression through the TERT axis. Telomere dysfunction contributes to tumorigenesis via two key mechanisms: promoting genomic instability and altering the immune microenvironment. These findings offer new insights into telomere-driven oncogenesis and lay a conceptual foundation for precision diagnostics and targeted therapies in oncology.

Introduction

Telomeres are specialized structures located at chromosome ends that preserve genomic stability by preventing DNA degradation and chromosomal fusion [1–3]. They progressively shorten with each cell division, serving as a mitotic clock that restricts cellular lifespan and promotes aging-related pathologies [4–6]. The relationship between telomere length and tumorigenesis in cancer is multifaceted. While telomere shortening in normal cells can inhibit cancer development by limiting cell replication, aberrant telomere elongation in precancerous lesions provides cancer cells with immortal proliferative potential, driving cancer progression [7, 8]. Recent studies have identified telomere length as a significant biomarker for cancer aggressiveness, prognosis, and treatment response. In most types of cancer, telomerase activity is significantly elevated. This phenomenon highlights the therapeutic promise of telomerase regulation, suggesting it as a potential target for curbing cancer cell proliferation [9, 10].

Genes involved in telomere maintenance and DNA repair are essential for preserving genomic integrity [11, 12]. Dysregulation of these genes has been implicated in various cancers, offering new insights into cancer development and treatment strategies. The shelterin complex, comprising proteins such as TRF1, TRF2, POT1, TPP1, TIN2, and RAP1, plays a pivotal role in protecting telomeres and preventing genomic instability [13]. The CST complex, consisting of CTC1, STN1, and TEN1, also contributes to telomere stability by promoting telomeric C-strand fill-in reactions [14, 15].

Mendelian randomization (MR) has emerged as a powerful tool in cancer research, offering a way to assess causal relationships by using genetic variants as instrumental variables [16–18]. This approach helps to minimize confounding factors and provides robust evidence for causal inference. Pan-cancer studies, which analyze multiple cancer types simultaneously, have become increasingly important in understanding the shared and distinct mechanisms across different cancers [19–21].

This study aims to systematically investigate the causal relationship between telomere length and the risk of various cancers using MR analyses. We also explore the underlying molecular mechanisms by identifying telomere-related genes and their functional pathways. Furthermore, we examine the expression of the key hub gene across diverse cancer types and evaluate its diagnostic utility and immunomodulatory functions within the tumor microenvironment.

Methods

Mendelian randomization analysis

We used summary-level data from large-scale Genome-Wide Association Study (GWAS) to perform MR analyses, employing the inverse-variance weighted (IVW) method, MR Egger regression, and weighted mode as primary and sensitivity analyses. The analysis method is the same as that reported in our previous articles. The selection criterion of IVs was P < 5e-06. The liftOver tool was used for coordinate conversion and have detailed the SNP quality control criteria, including P-value thresholds, minor allele frequency (MAF) thresholds, and LD pruning parameters, in the methods section [22]. GWAS information of this study were showed in Table 1. The analysis method is the same as that reported in our previous articles [16]. These analyses were conducted to assess the causal relationship between telomere length and pan-cancer risk. The scatterplots of MR analyses were generated to visualize the consistent directional effects across all three statistical approaches. Leave-one-out sensitivity analysis was performed to confirm the robustness of MR results, and funnel plot symmetry was evaluated to validate the absence of significant pleiotropic bias among instrumental variables. SNPs linked to telomere length (exposure) were obtained from the MRCIEU GWAS database (https://gwas.mrcieu.ac.uk/). This database contains 472,174 samples. SNPs associated with pan-cancer obtained from the Finnish database (https://www.finngen.fi/en). The pan-cancer dataset focuses on malignant neoplasms, with controls defined as individuals without any cancer diagnosis (ICD 0-3), and includes 121,495 cases (Table 2). All statistical analysis was performed with the “TwoSampleMR” package for MR analyses in R version 4.2.2 (The R Foundation for Statistical Computing, Vienna, Austria) [23, 24].

Table 1.

GWAS information of this study

Trait	Data sources	Population	Sample Size	Sex	Year	Genome reference version
Telomere length	MRCIEU (ieu-b-4879)	European	472,174	Males and Females	2021	GRCh37
Pan-cancer	FinnGen (R11)	European	500,244	Males and Females	2024	GRCh38

Table 2.

Endpoints included in the pan-cancer GWAS dataset

Endpoints
HEAD_AND_NECK_EXALLC
RESPIRATORY_INTRATHORACIC_EXALLC
DIGESTIVE_ORGANS_EXALLC
BONE_CARTILAGE_EXALLC
SKIN_EXALLC
MESTOTHEL_SOFTTISSUE_EXALLC
BREAST_EXALLC
FEMALE_GENITAL_EXALLC
MALE_GENITAL_EXALLC
URINARY_TRACT_EXALLC
BLADDER_EXALLC
EYE_BRAIN_NEURO_EXALLC
ENDOCRINE_EXALLC
UNCERTAIN_SECONDARY_EXALLC
PRIMARY_LYMPHOID_HEMATOPOIETIC
INDEPENDENT_MULTIPLE_SITES

Linkage disequilibrium (LD)

Compared to random chance, LD results in a higher likelihood of alleles from two or more loci co-occurring on a single chromosome. Our screening criteria were consistent with previous work [22]. The criteria include: (1) kb > 10,000; (2) r² < 0.001. Here, “kb” represented the extent of regional LD, while the evaluation of “r²” spanned from 0 to 1.

Elimination of weak IVs

Weak IVs were evaluated in the same way as in our previous article [22]. Briefly, the F-value was used for screening, when it was less than 10, IV was classified as a weak IV. The F-value was calculated using the following formula:

N represents the sample size of the exposure variable, K represents the number of IVs, MAF indicates the minor allele frequency, β indicates the allele effect size, and SD represents the standard deviation.

Protein–protein interaction network construction and analysis

To elucidate the mechanisms linking telomere length to pan-cancer pathogenesis, we annotated 121 SNPs using SNPense, identifying 143 protein-coding genes [25]. A protein–protein interaction (PPI) network was constructed via the STRING database, followed by module screening using MCODE and cytoHubba. Using the plugin MCODE module, the filtering conditions were set with Degree Cutoff = 2, Node Score Cutoff = 0.2, K-Core = 2, MAX. Depth = 100 in the whole PPI network. The top seven molecules in PPI were screened for analysis using cytoHubba, Betweenness, Degree, Closeness, and BottleNeck in cytoHubba module. Maximal Clique Centrality (MCC) is an algorithm for maximal clique centrality, which evaluates the centrality of a node by calculating the maximum number of cliques to which the node belongs. The maximum group refers to a subgraph in which each node is directly connected to all the other nodes.

Enrichment analysis

Local network cluster enrichment analysis of the 143 genes was performed to prioritize key biological processes and molecular functions. This analysis highlighted the involvement of these genes in telomere extension by telomerase and the CST complex, as well as telomere cap complex formation. Gene Ontology (GO) enrichment analysis was conducted to further characterize the cellular component (CC), biological process (BP), and molecular function (MF) of telomere-related genes. Human Phenotype enrichment analysis and Reactome pathway analysis were used to assess the association of these genes with telomere length and hematological phenotypes.

Data acquisition and processing of pan-cancer

Pan-cancer RNA-seq data were retrieved from the TCGA database (https://portal.gdc.cancer.gov) and processed following the STAR alignment workflow. The RNA-seq data were extracted in TPM format. No data filtering was applied, and the data were processed using log2(value + 1) transformation.

Expression analysis of TERT in pan-cancer

The expression of TERT was examined using the TCGA pan-cancer database. Radar charts were generated to show the expression trend of TERT in pan-cancer. The statistical analysis was performed using R (version 4.2.1) with the following packages: ggplot2 (3.4.4) [26], stats (4.2.1) [27], and car (3.1-0) [27]. The Wilcoxon rank sum test was used for statistical comparisons, as implemented in the stats and car packages. Statistical tests were conducted only when the underlying assumptions were met; otherwise, analyses were omitted. The expression levels of TERT were visualized across different cancer types by the ggplot2 package.

Cell lines and TERT mRNA expression analysis

To assess the relative expression of TERT mRNA across a range of human cancers, we employed 16 tumor cell lines representing malignancies from the urinary, breast, hepatobiliary, gastrointestinal, respiratory, nervous, cutaneous, and reproductive systems. Each tumor cell line was paired with a corresponding non-malignant cell line of the same tissue origin to serve as a normal control. The tumor-control cell line pairs were as follows: UM-UC-3 (bladder cancer, Pricella, CL-0463) and SV-HUC-1 (normal uroepithelial cells, Pricella, CL-0222); MDA-MB-468 (triple-negative breast cancer, Pricella, CL-0290) and MCF 10A (non-tumorigenic breast epithelial cells, Pricella, CL-0525); HuCCT1 (intrahepatic cholangiocarcinoma, Pricella, CL-0725) and HIBEpiC (normal intrahepatic biliary epithelial cells, Pricella, CP-H042); HCT116 (colorectal carcinoma, Pricella, CL-0096) and NCM460 (normal colon mucosal epithelial cells, FENGHUISHENGWU, CL-0393); KYSE-30 (esophageal squamous cell carcinoma, Pricella, CL-0577) and Het-1A (normal esophageal epithelial cells, FENGHUISHENGWU, CL-0139); U251 (glioblastoma, Pricella, CL-0237) and Human Astrocyte cells (normal human astrocytes, Pricella, CP-H122); FaDu (hypopharyngeal squamous cell carcinoma, Pricella, CL-0083) and HOK (human oral keratinocytes, WANWUSHENGWU, Delf-16831); Caki-2 (clear cell renal cell carcinoma, Pricella, CL-0326) and HK-2 (human renal proximal tubular epithelial cells, Pricella, CL-0109); ACHN (metastatic renal adenocarcinoma, Pricella, CL-0021) and HK-2; HepG2 (hepatocellular carcinoma, Pricella, CL-0103) and THLE-2 (normal human liver cells, Pricella, CL-0833); A549 (lung adenocarcinoma, Pricella, CL-0016) and BEAS-2B (normal bronchial epithelial cells, SUNNCELL, SNL-203); H226 (lung squamous cell carcinoma,Pricella, CL-0396) and BEAS-2B; LNCaP (androgen-sensitive prostate cancer, Pricella, CL-0143) and RWPE-1 (normal prostate epithelial cells, Pricella, CL-0200); SW1463 (rectal squamous cell carcinoma, Pricella, CL-0742) and Human Rectal Mucosal Epithelial Cells (human intestinal epithelial cells, CP-H041); AGS (gastric adenocarcinoma, Pricella, CL-0022) and GES-1 (non-tumorigenic gastric epithelial cells, IMMOCELL, IM-H084); and A375 (cutaneous melanoma, Pricella, CL-0014) and NHEM (normal human epidermal melanocytes, Pricella, CP-H108).

All cell lines were cultured in appropriate growth media recommended by the supplier, supplemented with 10% fetal bovine serum (FBS, Gibco, 26140079), and maintained at 37 °C in a humidified incubator with 5% CO₂. Total RNA was extracted from 1 × 10⁶ cells using TRIzol™ reagent (Invitrogen) and treated with DNase I (TaKaRa) to remove genomic DNA contamination. RNA purity and concentration were assessed using a NanoDrop spectrophotometer. First-strand cDNA was synthesized from 1 μg of total RNA using the HiScript III All-in-One RT SuperMix (Vazyme), with a no-RT control included to confirm the absence of genomic DNA. Quantitative real-time PCR (qRT-PCR) was performed using the ChamQ SYBR Master Mix (Vazyme) on an Applied Biosystems real-time PCR system. Each reaction contained 20 ng of cDNA and 200 nM of gene-specific primers. Amplification specificity and primer efficiency were validated through melt curve analysis and standard curve evaluation, respectively. GAPDH was used as an internal reference gene, and relative mRNA expression levels were determined using the 2^(-ΔΔCt) method. All experiments were conducted in triplicate to ensure reproducibility.

Pan-cancer immune infiltration analysis

The immune infiltration analysis of TERT across pan-cancer data was performed using R (version 4.2.1) with the ggplot2 package (3.4.4). Spearman correlation analysis was performed between EGFR expression, immune cell infiltration, and other matrix data. For visualization, heatmaps were generated. Using the GSVA R package (1.46.0) and based on the single-sample gene set enrichment analysis (ssGSEA) algorithm, the infiltration of 24 kinds of immune cells was calculated. The immune cell markers used were obtained from the Immunity article (Bindea, Gabriela, et al., 2013), including aDC, B cells, CD8 T cells, Cytotoxic cells, DC, Eosinophils, iDC, Macrophages, Mast cells, Neutrophils, NK CD56bright cells, NK CD56dim cells, NK cells, pDC, T cells, T helper cells, Tcm, Tem, TFH, Tgd, Th1 cells, Th17 cells, Th2 cells, TReg in the tumor immune microenvironment.

Diagnostic potential of TERT in pan-cancer

The diagnostic value of TERT was evaluated using ROC curve analysis, with AUC values calculated to assess its ability to distinguish between cancer types.

Results

Positive correlation between telomere length and pan-cancer risk

We first compiled GWAS datasets related to telomere length and pan-cancer risk (Fig. 1A). The characteristics of the included datasets are summarized in Table 1. MR analyses using multiple methods demonstrated that increased telomere length significantly elevates pan-cancer risk (Fig. 1B). Specifically, the IVW OR was 1.339 (95% CI: 1.270–1.417), MR Egger OR was 1.413 (95% CI: 1.276–1.566), and Weight Mode OR was 1.234 (95% CI: 1.105–1.378). Scatter plots from MR analyses showed consistent effect directions across all three statistical models (Fig. 1C). The heterogeneity and MR-PRESSO test were shown in Table 3. Leave-one-out sensitivity analysis confirmed the robustness of MR results: sequential exclusion of individual SNPs and reapplication of the IVW method to remaining SNPs showed no substantial deviation from the overall effect size, supporting the reliability of our findings (Fig. 1D and E). Funnel plot symmetry further validated the absence of significant pleiotropic bias among instrumental variables (Fig. 1F).

Fig. 1.

Positive Association Between Telomere Length and Pan-Cancer Risk. A Conceptual framework illustrating the MR approach. Created with BioRender.com. B Summary statistics of MR analyses assessing the association between telomere length and pan-cancer risk. C Scatter plot displaying SNP effects on telomere length (x-axis) and cancer risk (y-axis). D Forest plot of MR leave-one-out sensitivity analysis showing the effect of telomere length on pan-cancer risk across different SNPs. E Forest plot of individual MR estimates for the effect of telomere length on pan-cancer risk, using the inverse-variance weighted method. F Funnel plot for MR analysis, assessing heterogeneity and potential directional pleiotropy.

Table 3.

Summary of heterogeneity and MR-PRESSO test results

Exposure	Outcome	Heterogeneity tests (Q pval)		Horizontal pleiotropy (pval)
		MR Egger	IVW	Egger_intercept
Telomere length	Pan-cancer	0.193	0.341	0.228
Pan-cancer	Telomere length	0.730	0.852	0.447

Protein–protein interaction network identifies core telomere length related genes

To elucidate mechanisms linking telomere length to pan-cancer pathogenesis, 121 SNPs were annotated using SNPense, identifying 143 protein-coding genes. A PPI network was constructed via the STRING database, followed by module screening using MCODE and cytoHubba (Fig. 2A). These genes are critically involved in signaling, cell adhesion, and apoptosis, with dysregulation implicated in carcinogenesis. MCODE model extracted a core network comprising TERT, POT1, TRF1, TRF2, RTEL1, STN1, and PARN, forming a highly interactive telomere regulatory module (Fig. 2B). Subsequent centrality scoring by cytoHubba identified TERT was identified as the central hub, showing extensive connectivity with key proteins involved in telomere maintenance, chromosomal stability, and DNA repair (Fig. 2C). This centrality underscores TERT's pivotal role in genomic stability and cellular longevity, positioning it as a potential driver of telomere-mediated oncogenesis.

Fig. 2.

Protein–protein interaction network identifies core telomere length related genes. A Interaction network of Telomere-Length-Related Genes. B Six centrality algorithms were used to extract the core gene set

Enrichment analysis of telomere related genes

Local network cluster enrichment analysis of the 143 genes (Fig. 2A) prioritized “Telomere Extension by Telomerase and CST Complex” and “Telomere Cap Complex,” highlighting their involvement in telomeric structure stabilization (Fig. 3A). Concurrent enrichment in “Pyrimidine Nucleoside Monophosphate Metabolic Process” implicated these genes in DNA synthesis/repair. GO enrichment analysis encompasses three aspects: cellular component (CC), biological process (BP), and molecular function (MF). In CC, telomere-related genes are significantly involved in key functions such as Chromosome, telomeric region, Chromosomal region, and Nuclear telomere cap complex (Fig. 3B). Within MF, they are enriched in telomeric DNA binding, telomerase RNA binding, and DNA polymerase activity, indicating these proteins play an important role in maintaining telomere stability and genomic integrity. In BP, the top three are Regulation of telomere maintenance via telomere lengthening, Regulation of telomere maintenance, and Negative regulation of telomere maintenance (Fig. 3C). Within MF, the main enrichment features are Single-stranded telomeric DNA binding, Telomeric DNA binding, and Single-stranded DNA binding (Fig. 3D). Human Phenotype enrichment analysis shows that these genes are significantly associated with Telomere length and impact hematological phenotypes like Erythrocyte indices, Mean reticulocyte volume, and Platelet count. Reactome pathway analysis further associated these genes with telomere extension, DNA repair, cell cycle (e.g., “G1/S Transition”), and nucleotide metabolism, implicating them in proliferative and DNA damage response pathways central to carcinogenesis (Fig. 3F). Considering telomere-related enrichment, these genes may affect hematopoietic system function by influencing telomere length and genomic stability, and play a role in some cancers and aging-related diseases.

Fig. 3.

Enrichment analysis of telomere-related genes. A Local Network Cluster enrichment. B Cellular Component enrichment. C Biological Process enrichment. D Molecular Function enrichment. E Human Phenotype enrichment. F Reactome Pathways enrichment

Pan-cancer expression analysis and experimental verification of TERT

In summary, the telomere-related gene TERT drives the core protein interaction network, participating in telomere maintenance, DNA repair, and cellular immortalization, thus potentially playing a key role in the development and progression of pan-cancer. Therefore, we examined the expression of this gene through the TCGA pan-cancer database (Fig. 4A). In the TCGA pan-cancer database, TERT is significantly highly expressed in tumor tissues (P < 0.001), including bladder urothelial carcinoma (BLCA), breast invasive carcinoma (BRCA), cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), glioblastoma multiforme (GBM), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ), stomach adenocarcinoma (STAD), and skin cutaneous melanoma (SKCM). The radar chart shows the expression trend of TERT in pan-cancer. In most cancer types, the expression level of TERT in tumor tissues is higher than that in normal tissues (Fig. 4B). To further verify the expression of TERT, mRNA was extracted from 16 tumor cell lines for further verification. TERT was confirmed to be significantly overexpressed at the mRNA level in all 16 tumor cell lines (Fig. 4C). Correlation analysis revealed that TERT expression was significantly associated with diverse immune cell populations across cancer types, suggesting its immunomodulatory role in distinct tumor microenvironments (Fig. 5). Notably, TERT demonstrated prominent correlations with T helper 2 cells (Th2 cells), DCs (dendritic cells), and macrophages in tumor immune microenvironments. Moreover, the negative association between TERT and macrophages (r =  − 0.31) indicates that elevated TERT expression correlates with reduced macrophage abundance or activity, likely through its regulation of immunosuppressive signaling pathways that globally inhibit macrophage-mediated antitumor functions. Collectively, these findings underscore the pivotal role of TERT in tumor immune evasion through three interrelated mechanisms: enhancing Th2-mediated immunosuppression, impairing DC–driven antigen presentation, and inhibiting macrophage function.

Fig. 4.

Pan-cancer expression analysis and experimental verification of TERT. A Paired expression map of TERT in pan-cancer. B Radar map of TERT expression in pan-cancer. C Heat map of TERT in pan-cancer immune infiltration. *P ≤ 0.05, **P < 0.01, ***P < 0.0001

Fig. 5.

Correlation analysis of immune cell infiltration. Asterisks (*) indicate statistically significant correlations (p < 0.05)

Diagnostic potential of TERT in pan-cancer

As shown in the Fig. 6, among the 16 types of cancer with high TERT expression, the ROC analysis indicates that TERT has good diagnostic power in most cancers. With an AUC of 0.85 as the threshold, TERT stands out in lung squamous cell carcinoma (LUSC, AUC = 0.971), lung adenocarcinoma (LUAD, AUC = 0.962), liver hepatocellular carcinoma (LIHC, AUC = 0.957), cholangiocarcinoma (CHOL, AUC = 0.952), glioblastoma multiforme (GBM, AUC = 0.933), uterine corpus endometrial cancer (UCEC, AUC = 0.939), head and neck squamous cell carcinoma (HNSC, AUC = 0.882), stomach adenocarcinoma (STAD, AUC = 0.882), bladder urothelial carcinoma (BLCA, AUC = 0.902), and prostate adenocarcinoma (PRAD, AUC = 0.867). This suggests its good discriminatory ability in these cancers and its potential as a reliable molecular marker for clinical diagnosis. Moreover, TERT shows diagnostic potential in kidney renal clear cell carcinoma (KIRC, AUC = 0.837), esophageal carcinoma (ESCA, AUC = 0.809), rectum adenocarcinoma (READ, AUC = 0.800), colon adenocarcinoma (COAD, AUC = 0.791), breast invasive carcinoma (BRCA, AUC = 0.77), and kidney renal papillary cell carcinoma (KIRP, AUC = 0.685). Overall, these results indicate that TERT may be clinically significant in multiple cancer types, offering new research directions for precision medicine and early screening.

Fig. 6.

Diagnostic efficacy of TERT in 16 cancers. The dashed diagonal line represents the performance of a random classifier (AUC = 0.5)

Discussion

Recent studies have highlighted the role of telomere length as a biomarker for cancer aggressiveness, prognosis, and therapeutic response [28]. The relationship between telomeres and cancer is complex and multifaceted. Telomere length maintenance is a critical factor in both replicative senescence and carcinogenesis, functioning as a double-edged sword. On one hand, telomere shortening in normal cells can suppress cancer development by limiting the replicative potential of cells [6]. On the other hand, in precancerous lesions, telomeric dysfunction due to telomere shortening may promote genomic instability, fueling cancer development [2]. Telomere metabolism involves specialized functional complexes such as shelterin and CST [14]. The shelterin complex, comprising TRF1, TRF2, POT1, TPP1, TIN2, and RAP1 proteins, inhibits inappropriate activation of the DNA damage response at telomeres, preventing genomic instability. The CST complex, consisting of CTC1, STN1, and TEN1, associates with single-stranded telomeric overhangs and promotes telomeric C-strand fill-in reactions.

Our study demonstrates a significant positive correlation between telomere length and pan-cancer risk using Mendelian randomization analyses, consistent with previous findings that link telomere length to cancer risk. We identified a core protein–protein interaction network centered on TERT, which plays a crucial role in telomere maintenance and genomic stability. This aligns with existing research highlighting TERT's pivotal role in cancer development and progression.

The enrichment analyses reveal that telomere-related genes are involved in key biological processes and molecular functions associated with carcinogenesis, such as telomere extension, DNA synthesis/repair, and cellular responses to DNA damage. These findings are supported by prior studies that have implicated telomere-related genes in cancer development.

Pan-cancer expression analysis shows TERT's widespread overexpression in tumor tissues and its associations with immune cell infiltration, particularly Th2 cells, DCs, and macrophages. Given the central role of DCs in T cell priming and immune response regulation, TERT-mediated modulation of DC functionality may further compromise antitumor immunity. As a subset of adaptive immunity, Th2 cells critically promote immunosuppression and tumor microenvironment remodeling. The robust positive correlation between TERT and Th2 cells implies that TERT may amplify Th2 cell activity to suppress antitumor immunity, thereby fostering a permissive niche for cancer progression. Furthermore, TERT displayed broad correlations with DCs and their subtypes (e.g., aDC, iDC), suggesting its potential disruption of antigen-presenting functions and subsequent alteration of immune activation states in the tumor microenvironment. This suggests TERT's immunomodulatory role in the tumor microenvironment, potentially promoting cancer progression by altering immune responses. Previous research has also noted TERT's role in immune evasion and its potential as a therapeutic target. Aberrant TERT expression can disrupt the normal function of immune cells. For example, in patients with allergic rhinitis, the expression of TERT in dendritic cells increases, causing dendritic cells to inhibit the expression of IL-10 through the TERT-CMIP-IL-10 axis, thereby disrupting the immune tolerance characteristics of dendritic cells and triggering allergic reactions [29]. However, the mechanisms regulating TERT expression by immune cells remain incompletely understood. Some studies have shown that immune cells may affect the expression of TERT by secreting cytokines or directly interacting with tumor cells. For example, in the tumor microenvironment, the interaction between immune cells and tumor cells may lead to the upregulation or downregulation of TERT expression, thereby affecting the growth of tumor cells and immune escape.

The diagnostic potential of TERT across various cancers, as indicated by ROC analysis, suggests its clinical utility as a molecular marker. This is consistent with studies that have explored TERT as a diagnostic and therapeutic target in specific cancers like gliomas. Telomere-associated TERT is highly expressed in a variety of tumors and is significantly associated with immune regulation. This finding provides a new rationale for cancer treatment strategies targeting TERT and its interacting proteins.

While our study provides valuable insights into the relationship between telomere length and pan-cancer risk, several limitations should be acknowledged. First, the Mendelian randomization analyses rely on the assumption of pleiotropy, and although our sensitivity analyses suggested minimal pleiotropic bias, we cannot entirely rule out the possibility of unmeasured pleiotropic effects influencing the results. Second, the protein–protein interaction network and enrichment analyses are based on computational predictions and require further experimental validation to confirm the functional relevance of the identified genes and pathways. Third, the diagnostic potential of TERT identified through ROC analysis needs to be validated in independent cohorts and clinical settings. The importance of TERT was mainly determined by centrality algorithms, But centrality algorithm relies primarily on network topology and highly connected nodes, which does not necessarily equate to biological causal importance. Therefore, our future studies need to focus more on the biological validation of TERT. Future studies should focus on validation of TERT molecular mechanisms and clinical cohorts. Addressing these limitations in future research will help to further refine our understanding of telomere biology in cancer development and its clinical applications.

Conclusion

In conclusion, our study demonstrates a significant positive correlation between telomere length and pan-cancer risk using Mendelian randomization analyses. We identify a core protein–protein interaction network centered on TERT, which is crucial for telomere maintenance and genomic stability. Enrichment analyses highlight the involvement of telomere-related genes in key biological processes and molecular functions related to carcinogenesis. Our pan-cancer expression analysis shows TERT's widespread overexpression in tumor tissues and its associations with immune cell infiltration. Finally, TERT exhibits diagnostic potential across various cancers. Overall, this study enhances our understanding of telomere biology in cancer development and offers promising directions for precision medicine and early cancer screening.

Author contributions

JWZ, JYZ, and KLZ designed the research.JWZ, JYZ, KLZ, and MYW collected and analyzed the data. JWZ, JYZ, and KLZ performed the literature search. JWZ, JYZ, and KLZ drafted the article. WG, JDW, and SLL supervised the study.

Funding

This work was supported by the Young Scientists Fund of the National Natural Science Foundation of China (No. 82303289), and the Shanghai Sailing Program (No. 23YF1426000).

Data availability

The datasets generated during the current study are available in the IEU GWAS database (https://gwas.mrcieu.ac.uk/) and Finnish database (https://www.finngen.fi/en).

Declarations

Ethics approval and consent to participate

Not applicable. This study is based on publicly available datasets and does not involve human participants, animals, or tissue samples requiring ethical approval. Not applicable.

Consent for publication

All authors have read and approved the final manuscript and consent to its submission for publication.

Competing interests

The authors declare no competing interests.