Ferroptosis-related lncRNA signature predicts prognosis and treatment response in colon cancer

Introduction

Colon cancer is the third most-diagnosed cancer and the second leading cause of cancer-associated mortality globally (1). Despite continuous progress in the treatment and management of colon cancer, the outcome of advanced patients is poor, with a 5-year overall survival (OS) rate of 12.5% (2,3). Currently, the tumor-node-metastasis (TNM) staging system is considered as the cornerstone for prognosis prediction and treatment decision-making for colon cancer (4). Nevertheless, the prognosis and response to treatment among colon cancer patients may vary significantly even for patients in the same stage due to the high level of heterogeneity (4,5). Hence, there is an urgent demand to find reliable models or biomarkers to predict prognosis accurately and optimize treatment selection for patients with colon cancer.

Ferroptosis, an iron-dependent lipid peroxidation-induced form of regulated cell death, is implicated in the progression of colorectal cancer (6,7). Inducing ferroptosis is regarded as an effective therapeutic strategy for cancer treatment (8). Histone lactylation enhances GCLC expression and thus promotes chemoresistance of colorectal cancer stem cells through inhibiting ferroptosis (9). Icariin promotes ferroptosis by activating mitochondrial dysfunction to inhibit colorectal cancer and synergistically enhances the efficacy of programmed cell death protein 1 (PD-1) inhibitors (10).

Long non-coding RNAs (lncRNAs) are a class of noncoding RNAs of more than 200 nucleotides in length (11). Emerging studies suggest that lncRNAs play crucial roles in ferroptosis regulation. Liao et al. reported that LINC00942 activates complex I and inhibits ferroptosis through interacting with GRSF1 in liver cancer (12). LINC01833 drives gemcitabine resistance in non-small cell lung cancer by shielding SLC7A11 from WWP1-mediated ubiquitination and inhibiting ferroptosis (13). Yang et al. found that lncRNA MIR4435-2HG regulates bladder cancer progression and ferroptosis through the IQGAP3/Ras/ERK/CREB axis (14). Nevertheless, research concentrating on ferroptosis-related lncRNAs (FeRLs) in colon cancer is limited, and the prognostic value of FeRLs in colon cancer remains unclear. Thus, it is important to explore key FeRLs with prognostic significance in colon cancer patients.

Here, we established and validated a prognostic model based on FeRLs. Furthermore, the functional enrichment, immune landscape, mutation status and response to immunotherapy and chemotherapy were further analyzed to assess the underlying mechanism of FeRLs in colon cancer. Finally, in vitro experiments were performed to validate the expression of FeRLs. Our results may help to improve prognosis prediction and provide references for clinical chemotherapy and immunotherapy in colon cancer. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2155/rc).

Methods

Patient data acquisition

Colon adenocarcinoma (COAD) RNA sequencing (RNA-seq) data and corresponding clinical information were retrieved from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). Clinical data, including sex, age, TNM stage and follow-up data, were collected. Patients with no follow-up information or incomplete clinicopathological characteristics were excluded. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Identification of FeRLs

A total of 259 ferroptosis-related genes (FeRGs) were sourced from FerrDb (15). mRNA and lncRNA annotations were derived from GENCODE Release 29 (GRCh38.p12). FeRLs were identified by Pearson correlation analysis (thresholds: |R| >0.5, P<0.001) between FeRG and lncRNA expression levels.

Prognostic signature construction

Candidate prognostic lncRNAs were initially screened via univariate Cox regression (P<0.01) and further analyzed by multivariate Cox regression. A risk model was optimized using the minimal Akaike information criterion (AIC). Individual risk scores were computed as: Risk score = Σ(Coef_i × Exp_i) where Coef_i denotes the coefficient of lncRNA_i, and Exp_i represents its expression. Patients were stratified into high-/low-risk groups based on median risk scores. OS differences were evaluated by Kaplan-Meier/log-rank tests, while predictive accuracy was assessed via time-dependent receiver operating characteristic (ROC) analysis.

Nomogram development and validation

A nomogram was established based on the independent prognostic factors, which were determined through univariate and multivariate Cox regression analyses. Predictive performance was quantified using the concordance index (C-index), calibration curves, and area under the receiver operating characteristic curve (AUROC) values.

LncRNA-mRNA co-expression network

The co-expression network between FeRGs and FeRLs was visualized using Cytoscape 3.7.2. Prognostic lncRNA-mRNA relationships (risk/protective) were illustrated via a Sankey diagram (R package ‘ggalluvial’).

Functional enrichment analysis

Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) and gene set enrichment analysis (GSEA) analyses were performed using the ‘clusterProfiler’ package.

Mutation profiling

Somatic mutation data from TCGA-COAD were processed via ‘TCGAbiolinks’, with variant annotation and visualization performed using ‘maftools’.

Immune microenvironment assessment

The Estimation of STromal and Immune cells in MAlignant Tumours using Expression data (ESTIMATE) method was performed to calculate the immune scores and stromal scores of COAD patients (16). Immune cell infiltration disparities between risk groups were quantified via single sample gene set enrichment analysis (ssGSEA).

Therapy response prediction

Chemotherapy drug sensitivity was estimated using ‘pRRophetic’, based on the Genomics of Drug Sensitivity in Cancer (GDSC) database. Tumor Immune Dysfunction and Exclusion (TIDE) is a calculation method that integrates data from 12 immune checkpoint inhibitor (ICI) clinical studies and uses gene expression profiles to predict immune checkpoint blockade (ICB) response, and higher TIDE scores represent a higher likelihood of immune evasion, indicating that patients are less likely to benefit from ICI therapy (17). Immunotherapy efficacy was predicted using the TIDE algorithm.

Quantitative real-time polymerase chain reaction (qRT-PCR) validation

This study utilized normal human intestinal epithelial cells (NCM460) and colon cancer cells (HCT116), both obtained from the American Type Culture Collection (ATCC, Manassas, USA). Cells were maintained under standard conditions (37 °C, 5% CO₂) in appropriate culture media. For gene expression analysis, total RNA was isolated using a commercial extraction reagent (Yeasen, Shanghai, China; #10606ES60), followed by reverse transcription into cDNA (Yeasen, #11139ES10). qRT-PCR was performed on a Roche LightCycler 480 system with SYBR Green Master Mix (Yeasen, #11201ES03). Relative gene expression was determined using the 2^−ΔΔCt method, normalized to β-actin as an internal control. Primer sequences are provided in Table S1.

Statistical analysis

All analyses were conducted using R software (version 4.1.1). Variables were compared using Student’s t test or Wilcoxon test. Kruskal-Wallis or one-way ANOVA tests were performed to compare differences among three or more groups. Kaplan-Meier method and log-rank test were used to evaluate survival differences. Statistical significance was defined as P<0.05, and all P values were two-tailed.

Results

Identification of prognostic FeRLs

A total of 259 FeRGs were downloaded from FerrDb (as detailed in https://cdn.amegroups.cn/static/public/TCR-2025-2155-1.xlsx); the expression data for 238 of these genes were available in the TCGA COAD dataset. To ensure high-quality analysis, mRNAs and lncRNAs with low expression were filtered using the average expression in all samples >0.5 as the threshold. A total of 849 FeRLs were identified through Pearson correlation analysis with |R²| >0.5 and P<0.001 as thresholds. Next, through univariate Cox regression analysis, 20 lncRNAs associated with OS were screened (P<0.01, Figure 1). Among these 20 lncRNAs, 2 were poor prognostic factors, and 18 were favorable prognostic factors.

Figure 1 Forest plot indicating the HR (95% CI) and P values of selected lncRNAs determined using univariate Cox proportional hazards analysis. CI, confidence interval; HR, hazard ratio; lncRNA, long non-coding RNA.

Construction and validation of a prognostic FeRL signature

Subsequently, these 20 lncRNAs were analyzed using multivariate stepwise Cox regression, and 9 lncRNAs were determined according to the lowest AIC value. Therefore, we constructed a 9-FeRL prognostic signature. The risk score of each COAD patient was calculated as follows:

0.5160 × AC068580.6 + 0.2338 × AC113189.5 + 0.3052 × AP006621.5 + 0.4521 × CTD-2547G23.4 + (−0.3049) × CTD-2547L24.3 + 0.2676 × PCAT6 + 0.4278 × RP11-797A18.6 + (−0.2122) × RP11-79H23.3 + 0.1883 × RP11-815M8.1.

Patients were assigned to the high/low-risk group based on the median risk score (Figure 2). Kaplan-Meier curves demonstrated that patients in the high-risk group had shorter survival (Figure 2A, P<0.001), indicating that this novel signature effectively predicts prognosis. The distribution plot of the risk score, survival status and expression of the 9 lncRNAs is shown in Figure 2D. The heatmap displayed the expression patterns of lncRNAs between the two risk groups (Figure 2G).

Figure 2 Construction and validation of the ferroptosis-related lncRNA signature. (A-C) Kaplan-Meier curves for the overall survival of patients in the high- and low-risk groups in the training cohort, test cohort, and overall cohort. (D-F) The distribution of risk scores and overall survival status for each patient in the training cohort, test cohort, and overall cohort. (G-I) Heatmap showing the expression of the ferroptosis-related lncRNAs in the training in the training cohort, test cohort, and overall cohort. lncRNA, long non-coding RNA.

To comprehensively assess the predictive performance of the 9-lncRNA signature, we validated the risk stratification through distribution patterns, expression heatmaps, and survival analysis across both validation and combined cohorts. The risk groups demonstrated consistent stratification in the validation set (Figure 2E,2H) and overall population (Figure 2F,2I), with high-risk patients exhibiting significantly poorer survival outcomes compared to low-risk patients (Figure 2B,2C).

Correlation between the signature and clinicopathological features

To assess the potential involvement of the 9-lncRNA signature in colon cancer progression, we analyzed its correlation with key clinicopathological parameters. The results demonstrated statistically significant associations between risk scores and T stage, N stage, M stage and TNM stage (Figure 3A-3D). Notably, elevated risk scores consistently correlated with advanced disease stages, implying a potential role of this molecular signature in tumor advancement.

Figure 3 The ferroptosis-related lncRNA signature is associated with the clinicopathological characteristics of patients with colon cancer. The boxplots show that T stage (A), N stage (B), M stage (C), and TNM stage (D) are significantly associated with the risk score. lncRNA, long non-coding RNA; M, metastasis; N, node; T, tumor; TNM, tumor-node-metastasis.

Furthermore, we explored the prognostic predictive efficacy of the 9-lncRNA signature in different clinical subgroups. Patients were assigned to two groups based on age, sex, TNM stage, T stage, N stage, and M stage. We found that the high-risk group had significantly worse survival than the low-risk group for each subgroup (Figure 4). The results showed that the signature remained a significant prognostic indicator after stratification by clinicopathological variables, indicating the independent association of the signature with the prognosis.

Figure 4 Kaplan-Meier survival plots according to the expression of the 9 lncRNAs in the signature in patients with different clinical characteristics. lncRNA, long non-coding RNA; M, metastasis; N, node; T, tumor.

Establishment and evaluation of the nomogram

Univariate and multivariate Cox regression analyses were conducted to determine independent prognostic factors for OS. The results showed that age, M stage, N stage and risk score were independent prognostic factors for OS (Figure 5A,5B). Then, ROC analysis revealed the risk score’s superior predictive capability, with 3-year AUROC reaching 0.796, significantly outperforming clinicopathological parameters (Figure 5C). In addition, the AUROCs for 1-, 3- and 5-year of the risk score were 0.773, 0.796 and 0.787, respectively (Figure 5D).

Figure 5 Prognostic value of the ferroptosis-related lncRNA signature. Univariate (A) and multivariate (B) Cox regression analyses of relationships between clinical parameters (including the 9 lncRNA signature) and OS. (C) ROC curve analyses for predicting OS at 3-year for the signature and clinicopathological features. (D) ROC curve analyses of the signature for predicting OS at 1-, 3- and 5-year. AUC, area under the curve; lncRNA, long non-coding RNA; M, metastasis; N, node; OS, overall survival; ROC, receiver operating characteristic; TNM, tumor-node-metastasis.

To enhance clinical utility, we constructed a predictive nomogram by integrating these prognostic factors (Figure 6A). The concordance index of this nomogram was 0.816 [95% confidence interval (CI): 0.761–0.871]. The AUCs for 1-, 3- and 5-year of the nomogram were 0.823, 0.826 and 0.837, respectively (Figure 6B). Compared with the traditional model based on clinicopathological variables, the nomogram displayed a significantly improved AUROC (Figure 6C, 0.826 vs. 0.748, P<0.001). The calibration plots of the nomogram showed that there was a good agreement between the nomogram-predicted and the actual survival (Figure 6D). The decision curve analysis indicated that the nomogram provided greater net benefits across most likelihood thresholds (Figure 6E).

Figure 6 Construction and evaluation of a predictive nomogram. (A) Nomogram for predicting the OS at 1, 3, and 5 years. (B) ROC curve analyses of the nomogram for predicting OS at 1-, 3- and 5-year. (C) Comparison of the ROC curve between the nomogram and traditional model in the entire cohort. (D) The calibration curve of the nomogram. (E) Decision curve analysis of the nomogram. AUC, area under the curve; CI, confidence interval; M, metastasis; N, node; OS, overall survival; ROC, receiver operating characteristic.

Construction of the mRNA-lncRNA co-expression network

Using Cytoscape, we constructed a co-expression network integrating ferroptosis-related mRNAs and lncRNAs (Table S2). The network comprised 43 significant lncRNA-mRNA interactions, with AC113189.5 demonstrating the highest connectivity (14 gene partners), followed by PCAT6 (12 gene partners) (Figure 7A). Sankey diagram visualization (Figure 7B) revealed distinct correlation patterns among lncRNAs, mRNAs, and risk stratification. Notably, multivariate analysis identified CTD-2547L24.3 and RP11-79H23.3 as independent prognostic factors associated with improved OS.

Figure 7 Construction of the ferroptosis-related lncRNA-mRNA co-expression network. (A) A co-expression network of FeRLs visualized using Cytoscape. Orange rectangles indicate lncRNAs, and purple rectangles represent mRNAs. (B) Sankey diagram showing the associations between ferroptosis-related lncRNAs, mRNAs and the risk type. FeRL, ferroptosis-related lncRNA; lncRNA, long non-coding RNA.

Functional enrichment analyses

Functional enrichment analysis was conducted on the differentially expressed genes (DEGs) identified between comparison groups. GO results revealed significant enrichment in stromal-related biological processes, including extracellular matrix organization, external encapsulating structure organization, and collagen fibril organization (Figure 8A). The DEGs were also enriched in stromal-associated cellular components, including collagen-containing extracellular matrix and focal adhesion (Figure 8B). In addition, the DEGs were enriched in stromal-associated molecular functions, such as extracellular matrix structural constituent and extracellular matrix binding (Figure 8C). Moreover, the KEGG results showed that the DEGs were enriched in stroma-related pathways, including focal adhesion and ECM-receptor interaction (Figure 8D). GSEA analyses demonstrated that some oncogenic pathways were enriched in high-risk group, including EMT and apical junction (Figure 8E,8F).

Figure 8 Results of the functional enrichment analyses. GO analyses showed that DEGs between the two groups were enriched in stroma-related biological processes (A), cellular components (B) and molecular functions (C). KEGG analyses showed that the DEGs were enriched in stroma-related pathways (D). GSEA analyses showed that EMT (E) and apical junction (F) were enriched in high-risk group. DEG, differentially expressed gene; EMT, epithelial-mesenchymal transition; GO, Gene Ontology; GSEA, gene set enrichment analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes; NES, normalized enrichment score.

Immune landscape analyses

To explore the association between immune cell infiltration and the ferroptosis-related signature, ssGSEA was performed to compare the proportions of 24 immune cell types between groups (available online: https://cdn.amegroups.cn/static/public/TCR-2025-2155-2.xlsx). The results showed that a high signature score was associated with decreased infiltration of activated mast cells, activated NK cells, activated memory CD4 T cells, CD8 T cells and gamma-delta T cells, and increased infiltration of fibroblasts and naive CD4 T cells (Figure 9A). In addition to immune cells, stromal cells also regulate the tumor immune phenotype (18). For example, cancer-associated fibroblasts (CAFs) both inhibit immune cell access to the TME and inhibit their functions within the tumor, therefore exerting a direct immunosuppressive effect (19). Further analysis of stroma-related pathways demonstrated significantly higher scores of epithelial-mesenchymal transition (EMT), pan-fibroblast transforming growth factor beta response (Pan-F-TBRS) and TMEscoreB (stromal-relevant) in high-risk patients (20,21), which was consistent with the increased enrichment of fibroblasts in patients with high risk scores (Figure 9B). In addition, the ESTIMATE results validated that the stromal activity in tumor microenvironment (TME) was significantly higher in the high-risk group (Figure 9C, P=0.03).

Figure 9 Comparison of the TME between the two groups. (A) ssGSEA score of 24 immune cells grouped by risk score. (B) Difference in the enrichment of stroma-related signatures between the high- and low-risk groups. (C) Boxplots of stromal scores from ESTIMATE analysis of the two groups. *, P<0.05; **, P<0.01; ****, P<0.0001. ESTIMATE, Estimation of STromal and Immune cells in MAlignant Tumours using Expression data; ssGSEA, single sample gene set enrichment analysis; TME, tumor microenvironment.

Signature predicts the response to immunotherapy and chemotherapy

Abnormal changes in the TME were associated with the treatment response (22). Therefore, the associations between risk score and the responses to immunotherapy and chemotherapy were analyzed. First, we selected 10 immune checkpoint genes (CD274, CTLA4, HAVCR2, LAG3, TGFB1, TGFB2, TGFB3, TIGIT, TNFRSF4 and TNFRSF9) to explore the expression differences between groups. The results demonstrated that the high-risk patients exhibited elevated expression of TGFB1, TGFB3 and TNFRSF4 (Figure 10A). The tumor mutation burden (TMB) was a known predictive biomarker of immunotherapy response (23), but no significant difference was observed in the two groups (Figure S1). Subsequently, we used the TIDE algorithm to investigate whether the risk score could predict the immunotherapeutic response. Detailed information on TIDE is shown in https://cdn.amegroups.cn/static/public/TCR-2025-2155-3.xlsx. We found that the number and percentage of responders to immunotherapy were significantly higher in the low-risk group (112/210) than in the high-risk group (73/210) (chi-square test, P<0.001, Figure 10B). In addition, patients in low-risk groups were more likely to be responders than non-responders (Student’s t test, P<0.001, Figure 10C). Subsequently, we explored the association between the signature and the sensitivity of patients to chemotherapeutic drugs. We found that the low-risk group had lower estimated half-maximal inhibitor concentration (IC₅₀) values for cisplatin and mitomycin C (Figure 10D,10E). Collectively, these findings suggested that the signature may serve as a potential predictor for immunotherapy and chemotherapy outcomes.

Figure 10 The risk score could predict the response to immunotherapeutics and chemotherapeutics. (A) Boxplots of the expression of immune checkpoint genes between the high- and low-risk groups. (B) The proportion of patients who responded to immunotherapy between the two groups. Red represents nonresponders; blue indicates responders. (C) The distribution of immunotherapeutic response status between two groups stratified by the risk score based on the TIDE algorithm (t-test). (D,E) The boxplots show the differences in the estimated IC₅₀ values for cisplatin (D) and mitomycin C (E) between two groups. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. IC₅₀, half-maximal inhibitor concentration; TIDE, Tumor Immune Dysfunction and Exclusion.

Gene mutation analysis

We then analyzed the difference in somatic mutations between the two groups. The results showed that APC (71%), TP53 (61%), TNN (53%), KRAS (42%), and MUC16 (31%) were the top five mutated genes in the high-risk group (Figure 11A), while APC (76%), TNN (50%), TP53 (47%), KRAS (41%), and PIK3CA (32%) were the top five mutated genes in the low-risk group (Figure 11B).

Figure 11 Association between the risk score and tumor somatic mutations. The waterfall plot of tumor somatic mutations showing the distribution of the top 10 mutated genes between the high- (A) and low-risk (B) groups. TMB, tumor mutation burden.

Validation of FeRLs expression

Among the nine FeRLs, five lncRNAs were differentially expressed between tumor and normal tissues. Therefore, we conducted qRT-PCR to evaluate the levels of five prognostic FeRLs in HCT116 cells and compared them with those in healthy NCM460 colon cells. The qRT-PCR analysis demonstrated that the expressions of AP006621.5, CTD-2547G24.3, PCAT6, and RP11-815M8.1 were remarkably elevated in HCT116 cells compared to normal cells (Figure 12A-12D), whereas the expression of CTD-2547L24.3 showed an opposite trend (Figure 12E).

Figure 12 Experimental validation of the expression level of the five ferroptosis-related lncRNAs in cell lines. qRT-PCR was used to detect the expression of (A) AP006621.5, (B) CTD-2547G24.3, (C) PCAT6, (D) RP11-815M8.1, and (E) CTD-2547L24.3 expression in normal human colon cells (NCM460) and colon cancer cells (HCT116). *, P<0.05; **, P<0.01; ***, P<0.001. lncRNA, long non-coding RNA; qRT-PCR, quantitative real-time polymerase chain reaction.

Discussion

Although the survival of colon cancer patients has improved in the past few decades, the outcome of patients with advanced disease is still poor (24). Moreover, the prognosis varies between patients with similar clinical features (4). Therefore, more biomarkers that could accurately predict the prognosis and treatment response need to be explored to improve survival benefits for more patients. LncRNAs have been proven to exert important influences on the occurrence and progression of diverse tumors (11,14), including colon cancer (25). Ferroptosis is characterized by iron-dependent lipid peroxidation (6). Accumulating evidence has shown that ferroptosis influences tumor progression and is associated with the chemotherapeutic and immunotherapeutic response of cancers (6,7). However, further investigation is required to identify FeRLs and explore their potential role in colon cancer. Thus, we aimed to construct an FeRLs-based prognostic signature that can improve prognosis and clinical treatment efficacy in patients with colon cancer.

In this study, 259 FeRGs were obtained, and 849 FeRLs were identified using co-expression analysis. Subsequently, 20 prognostic lncRNAs were determined through univariate Cox regression. Then, we established a signature consisting of 9 lncRNAs with the lowest AIC using multivariate Cox regression. The lncRNA signature predicted OS accurately according to the AUC curve. Patients were divided into high- and low-risk groups according to the median risk score. Subsequently, univariate and multivariate Cox regression results showed that the signature and several clinicopathological features were independent prognostic factors. Then, a predictive nomogram was constructed by incorporating the and other clinicopathological features to estimate the 1-, 3-, and 5-year OS probabilities of colon cancer patients. The concordance index of the nomogram was 0.816, and the calibration curve indicated good performance in predicting actual mortality.

Among the 9 identified lncRNAs, CTD-2547G23.4 promotes hepatocellular carcinoma tumorigenesis and progression (26). Overexpression of PCAT6 is related to poor clinical outcomes in many cancers, including colorectal cancer (27-30). However, studies are lacking for the other 7 lncRNAs (RP11-815M8.1, AC068580.6, RP11-797A18.6, CTD-2547L24.3, AC113189.5, AP006621.5, RP11-79H23.3). Our study is the first report about these 7 lncRNAs on ferroptosis and colon cancer.

A growing amount of evidence has shown that the TME is closely linked to prognosis and immunotherapy response in multiple cancers (31-33). The role of ferroptosis in the TME remains unclear. Our immune infiltration analyses showed that the level of CD8⁺ T cells was significantly higher than that in the low-risk group, whereas the high-risk group exhibited greater fibroblast infiltration. Furthermore, we demonstrated that the stromal activity of the TME, as reflected by EMT, Pan-F-TBRS expression and TMEscoreB (stroma-related), was significantly higher than that in the high-risk group, suggesting that the high-risk group may exhibit a T cell suppressive and an immune-excluded phenotype (34,35). The GO/KEGG analyses showed that the signature was mainly associated with stroma-related biological processes or pathways. GSEA results showed that the EMT and apical junction were enriched in high-risk group. EMT is linked to carcinogenesis, invasiveness, metastasis, and drug resistance in cancer, which can be regulated through many signaling pathways. Ferroptosis is an iron-dependent programmed cell death caused by massive iron accumulation and lipid peroxidation. The induction of ferroptosis has become a promising cancer treatment approach, which triggers the death of cancer cells, particularly in malignancies resistant to conventional therapies. Studies have shown that mesenchymal cancer cells predispose to a higher susceptibility to ferroptotic cell death (36). Drug-resistant cancer cells are more easily eliminated by inducers of ferroptosis when they undergo EMT (37,38). Accumulating evidence suggests that different TME statuses might represent different immunotherapy responses (31-33). Thus, we explored the association between the signature and response to immunotherapy. We found that patients in the low-risk group were more likely to benefit from immunotherapy. The limited efficacy of immunotherapy observed in the high-risk group may be driven by the enhanced stromal activity and the immunosuppressive microenvironment. Therefore, for high-risk patients, monotherapy with either ICIs or ferroptosis inducers alone tends to be ineffective. Combined treatment with immune checkpoint inhibition and ferroptosis induction may help to eradicate cancer cells and overcome drug resistance.

In addition, the high-risk group was highly resistant to cisplatin and mitomycin C. Given the limited drug data in the GDSC database, the sensitivity of colon cancer to first-line chemotherapeutic agents (fluorouracil and oxaliplatin) needs to be analyzed to verify the capacity of the risk model in predicting chemosensitivity.

These results offer new perspectives on immunotherapy for patients with colon cancer.

Previous studies have identified FeRL signatures in cancer research (39,40). However, our study offers distinct advantages: (I) the newly developed signature demonstrates superior predictive performance and reveals novel associations with TME features; (II) unlike prior work, we systematically assessed the signature’s ability to predict responses to both immunotherapy and chemotherapy in colon cancer.

However, there are some limitations in this study. First, we used a single data source, and clinical validation is warranted to validate the 9-lncRNA signature’s predictive performance. Second, while a few experiments were performed to validate the expression of detected FeRLs, more in vitro experiments are warranted to elucidate the functional roles of these FeRLs in colon cancer.

Conclusions

In the current study, we constructed a new FeRL signature that could accurately predict the outcome in colon cancer. Importantly, this signature might indicate TME status and act as a potential predictor for immunotherapy efficacy.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2155/rc

Data Sharing Statement: Available at https://dss.amegroups.com/article/view/10.21037/tcr-2025-2155/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2155/prf

Funding: This work was supported by the National Natural Science Foundation of China (No. 82303446); The Shenzhen High-level Hospital Construction Fund, and Peking University Shenzhen Hospital Scientific Research Fund (No. KYQD2025484); Guangdong Basic and Applied Basic Research Foundation (No. 2023A1515220200); Medical Scientific Research Foundation of Guangdong Province of China (No. A2024351); and Shenzhen Science and Technology Program (No. JCYJ20240813115900001).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2155/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Nishihara R, Wu K, Lochhead P, et al. Long-term colorectal-cancer incidence and mortality after lower endoscopy. N Engl J Med 2013;369:1095-105. [Crossref] [PubMed]
3. Pita-Fernández S, González-Sáez L, López-Calviño B, et al. Effect of diagnostic delay on survival in patients with colorectal cancer: a retrospective cohort study. BMC Cancer 2016;16:664. [Crossref] [PubMed]
4. Weiser MR, Gönen M, Chou JF, et al. Predicting survival after curative colectomy for cancer: individualizing colon cancer staging. J Clin Oncol 2011;29:4796-802. [Crossref] [PubMed]
5. Punt CJ, Koopman M, Vermeulen L. From tumour heterogeneity to advances in precision treatment of colorectal cancer. Nat Rev Clin Oncol 2017;14:235-46. [Crossref] [PubMed]
6. Dixon SJ, Lemberg KM, Lamprecht MR, et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell 2012;149:1060-72. [Crossref] [PubMed]
7. Tang D, Kang R, Berghe TV, et al. The molecular machinery of regulated cell death. Cell Res 2019;29:347-64. [Crossref] [PubMed]
8. Chen X, Ren Q, Zheng X, et al. Long non-coding RNA HCP5 accelerated malignant progression of ovarian cancer by inhibiting ferroptosis through interaction with polypyrimidine tract binding protein 1. J Ovarian Res 2025;18:271. [Crossref] [PubMed]
9. Deng J, Li Y, Yin L, et al. Histone lactylation enhances GCLC expression and thus promotes chemoresistance of colorectal cancer stem cells through inhibiting ferroptosis. Cell Death Dis 2025;16:193. [Crossref] [PubMed]
10. Haoyue W, Kexiang S, Shan TW, et al. Icariin promoted ferroptosis by activating mitochondrial dysfunction to inhibit colorectal cancer and synergistically enhanced the efficacy of PD-1 inhibitors. Phytomedicine 2025;136:156224. [Crossref] [PubMed]
11. Wang W, Jiang Z, Zhao K, et al. Long non-coding RNAs in the tumor immune microenvironment of non-small cell lung cancer: mechanisms and clinical translational perspectives. J Transl Med 2025;24:100. [Crossref] [PubMed]
12. Liao Y, Su Y, Liu Y, et al. Mitochondrial LINC00942 activates complex I and inhibits ferroptosis through interacting with GRSF1 in liver cancer. Sci China Life Sci 2026;69:438-54. [Crossref] [PubMed]
13. Ni L, Li H, Chen P, et al. Linc01833 drives gemcitabine resistance in non-small cell lung cancer by shielding SLC7A11 from WWP1-mediated ubiquitination and inhibiting ferroptosis. Drug Resist Updat 2026;85:101351. [Crossref] [PubMed]
14. Yang T, Guo L, Kong S, et al. The lncRNA MIR4435-2HG Modulates Bladder Cancer Progression and Ferroptosis Through the IQGAP3/Ras/ERK/CREB Pathway. FASEB J 2025;39:e71119. [Crossref] [PubMed]
15. Zhou N, Bao J. FerrDb: a manually curated resource for regulators and markers of ferroptosis and ferroptosis-disease associations. Database (Oxford) 2020;2020:baaa021. [Crossref] [PubMed]
16. Yoshihara K, Shahmoradgoli M, Martínez E, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 2013;4:2612. [Crossref] [PubMed]
17. Jiang P, Gu S, Pan D, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med 2018;24:1550-8. [Crossref] [PubMed]
18. Wu J, Li L, Zhang H, et al. A risk model developed based on tumor microenvironment predicts overall survival and associates with tumor immunity of patients with lung adenocarcinoma. Oncogene 2021;40:4413-24. [Crossref] [PubMed]
19. Denton AE, Roberts EW, Fearon DT. Stromal Cells in the Tumor Microenvironment. Adv Exp Med Biol 2018;1060:99-114. [Crossref] [PubMed]
20. Mariathasan S, Turley SJ, Nickles D, et al. TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature 2018;554:544-8. [Crossref] [PubMed]
21. Zeng D, Wu J, Luo H, et al. Tumor microenvironment evaluation promotes precise checkpoint immunotherapy of advanced gastric cancer. J Immunother Cancer 2021;9:e002467. [Crossref] [PubMed]
22. Rahal Z, El Darzi R, Moghaddam SJ, et al. Tumour and microenvironment crosstalk in NSCLC progression and response to therapy. Nat Rev Clin Oncol 2025;22:463-82. [Crossref] [PubMed]
23. Parseghian CM, Eluri M. Beyond the Burden: Elevating Immunotherapy in Colorectal Cancer. Clin Cancer Res 2026;32:999-1000. [Crossref] [PubMed]
24. Saber R, Carneiro M, Montagnon E, et al. Multi-omics fusion network for prediction of early recurrence in colorectal liver metastases. NPJ Precis Oncol 2026;10:61. [Crossref] [PubMed]
25. Lei JP, Wang L, Wang TY, et al. LINC02878/ZNF282/PYCR2 axis promotes proline synthesis and tumor progression in colorectal cancer. Cell Mol Life Sci 2025;83:25. [Crossref] [PubMed]
26. Wen DY, Lin P, Liang HW, et al. Up-regulation of CTD-2547G23.4 in hepatocellular carcinoma tissues and its prospective molecular regulatory mechanism: a novel qRT-PCR and bioinformatics analysis study. Cancer Cell Int 2018;18:74. [Crossref] [PubMed]
27. Ghafouri-Fard S, Khoshbakht T, Taheri M, et al. A review on the role of PCAT6 lncRNA in tumorigenesis. Biomed Pharmacother 2021;142:112010. [Crossref] [PubMed]
28. Sun X, Yuan Y, Li S, et al. Prostate cancer-associated transcript 6 (PCAT6) promotes epithelial-mesenchymal transition and stemness and worsens prognosis in patients with colorectal cancer. Acta Biochim Biophys Sin (Shanghai) 2024;56:866-78. [Crossref] [PubMed]
29. Han L, Li Z, Zhang P, et al. LncRNA PCAT6 is a predictor of poor prognosis of colorectal cancer. J Gastrointest Oncol 2024;15:190-202. [Crossref] [PubMed]
30. Wang F, Mu HF, Wang C, et al. LncRNA PCAT6 promotes progression and metastasis of colonic neuroendocrine carcinoma via MAPK pathway. World J Gastrointest Oncol 2025;17:96230. [Crossref] [PubMed]
31. Wang Y, Yue B, Ni H, et al. POSTN⁺ cancer-associated fibroblast-CCL3⁺ macrophage crosstalk defines the immune-excluded tumor microenvironment in clear cell renal cell carcinoma. Transl Oncol 2026;65:102682. [Crossref] [PubMed]
32. Jin Q, Jiang N, Chen G, et al. A prognostic signature derived from ac4C-associated genes stratifies survival and tumor immune microenvironment in cutaneous melanoma. Front Immunol 2025;16:1727135. [Crossref] [PubMed]
33. Zhou M, Niu H, Cui D, et al. Cancer stem cell-driven drug resistance in colorectal carcinoma: molecular aspects and therapeutic potentials. Mol Cancer 2026;25:54. [Crossref] [PubMed]
34. Zheng S, Wang W, Shen L, et al. Tumor battlefield within inflamed, excluded or desert immune phenotypes: the mechanisms and strategies. Exp Hematol Oncol 2024;13:80. [Crossref] [PubMed]
35. Ueshima E, Sofue K, Kodama T, et al. Gadoxetic Acid-Enhanced Magnetic Resonance Imaging Features Can Predict Immune-Excluded Phenotype of Hepatocellular Carcinoma. Liver Cancer 2025;14:271-85. [Crossref] [PubMed]
36. Schwab A, Rao Z, Zhang J, et al. Zeb1 mediates EMT/plasticity-associated ferroptosis sensitivity in cancer cells by regulating lipogenic enzyme expression and phospholipid composition. Nat Cell Biol 2024;26:1470-81. [Crossref] [PubMed]
37. Ren Y, Mao X, Xu H, et al. Ferroptosis and EMT: key targets for combating cancer progression and therapy resistance. Cell Mol Life Sci 2023;80:263. [Crossref] [PubMed]
38. Zhang C, Wu Q, Yang H, et al. Ferroptosis-related gene signature for predicting prognosis and identifying potential therapeutic drug in EGFR wild-type lung adenocarcinoma. Commun Biol 2024;7:1416. [Crossref] [PubMed]
39. Chen W, Chen Y, Liu L, et al. Comprehensive analysis of immune infiltrates of ferroptosis-related long noncoding RNA and prediction of colon cancer patient prognoses. J Immunol Res 2022;2022:9480628. [Crossref] [PubMed]
40. Wu Z, Lu Z, Li L, et al. Identification and validation of ferroptosis-related lncRNA signatures as a novel prognostic model for colon cancer. Front Immunol 2022;12:783362. [Crossref] [PubMed]