Abstract
BACKGROUND: Alterations in histone modifications have been reported to be related to tumorigenicity and tumor progression. However, whether histone modification can aid the classification of patients or influence clinical behavior in patients with colon cancer remains unclear. Therefore, this study aimed to evaluate histone modifier expression patterns using the unsupervised clustering of the transcriptomic expressions of 88 histone acetylation and methylation regulators. RESULTS: In this study, by consensus clustering analysis based on the transcriptome data of 88 histone modification regulators, we identified four distinct expression patterns of histone modifiers associated with different prognoses, intrinsic fluorouracil sensitivities, biological pathways, and tumor microenvironment characteristics among 1372 colon cancer samples. In these four clusters, the HMC4 cluster represented a stroma activation phenotype characterized by both the worst prognosis and lowest response rates to fluorouracil treatment. Then, we established a scoring scheme comprising 155 genes designated as "HM_score" by using the Boruta algorithm to distinguish colon cancer patients within the HMC4 cluster. Patients with a high HM_score were considered to have high stromal pathway activation, high stromal fraction, and an unfavorable prognosis. Further analyses indicated that a high HM_score also correlated with reduced therapeutic benefits from fluorouracil chemotherapy. Moreover, through CRISPR library screening, ZEB2 was found to be a critical driver gene that mediates fluorouracil resistance, which is associated with histone modifier expression patterns. CONCLUSIONS: This study highlights that characterizing histone modifier expression patterns may help better understand the epigenetic mechanisms underlying tumor heterogeneity in patients with colon cancer and provide more personalized therapeutic strategies.