Classification of colorectal carcinoma subtypes based on ferroptosis-associated molecular markers

Background Ferroptosis is associated with the development of many cancers; the molecular features of colorectal carcinoma (CRC) based on ferroptosis-related genes (FRGs) remain unknown. Herein, we aimed to identify ferroptosis-associated molecular subtypes of CRC based on the expression profiles of FRGs. Methods To explore ferroptosis-associated subtypes of CRC, the gene expression data and clinical information of 682 patients were extracted from The Cancer Genome Atlas and Gene Expression Omnibus databases. We performed consensus clustering to identify robust clusters of patients. Then the distribution of the subtypes in terms of prognosis significance, transcriptome features, immune microenvironment, drug sensitivity, gene mutations, and copy number alternations (CNAs) were evaluated respectively. In addition, we analyzed the correlation of these ferroptosis-associated molecular subtypes with the distribution of conventional clinical indicators in CRC. Results Four subtypes of CRC (C1, C2, C3, and C4) were identified in which the prognosis, immune cell infiltration, immune score, stromal score, and tumor purity were significantly different between the four subtypes. The C3 subtype had a higher infiltration of B cells, M2 macrophages, resting mast cells, monocytes, natural killer cells, plasma cells, and CD8 T cells. The C3 subtype had the highest immune and stromal scores and the lowest tumor purity. In contrast, the C4 subtype demonstrated the lowest immune and stromal scores and the highest tumor purity. Programmed cell death ligand 1 (PD-L1), an immune checkpoint protein, was differentially expressed in the four subtypes (P < 2e-16) and was significantly correlated with the expression of several FRGs in all subtypes. Significant differences in stem cell indices (P < 0.01) and drug sensitivity (P < 0.01) were observed in the four subtypes. Additionally, gene mutations analysis showed that FRGs such as TP53 had a high frequency of mutation in the four subtypes (49%, 62%, 61%, and 71%, respectively), and the CNAs showed significant difference among all subtypes (P < 0.001). Conclusion In summary, the ferroptosis-associated subtypes could serve as an independent biomarker for estimating oncological outcomes in patients with CRC. Our results demonstrated that the high level of heterogeneity in the expression of FRGs might be useful for the stratification of patients with CRC and the implementation of individualized therapeutic strategies.