Identification and validation of a novel signature based on M2 macrophage co-expressed genes in bladder cancer.

Background: Extensive evidence has demonstrated a robust association between high infiltration of M2 macrophages and the prognosis of bladder cancer (BLCA). Nevertheless, no comprehensive analysis of co-expressed M2 genes in BLCA has been reported. We would like to develop a prognostic model for BLCA using M2 co-expressed genes.

Methods: Raw data and clinical characteristics were retrieved from public databases. We first used the "cibersort" package to determine the M2 macrophage infiltration coefficients of each BLCA sample in The Cancer Genome Atlas (TCGA). Subsequently, Pearson correlation was employed to screen co-expressed genes based on the coefficients. Least absolute shrinkage and selection operator (LASSO-COX) analysis was then employed to construct the prognostic gene signature, which was externally validated in the GSE13507 cohort. Further exploration of the signature involved tumor mutational burden (TMB) and drug sensitivity analyses. Finally, quantitative real-time polymerase chain reaction (qRT-PCR) was employed to validate the expression levels of the co-expressed genes.

Results: We developed a signature using two co-expressed genes and utilized it to categorize samples into two groups. Patients in the low-risk group exhibited more satisfactory outcomes (P=0.008) and higher TMB (P=0.04). Additionally, the high-risk group exhibited a substantial discrepancy in immune subtypes compared to the low-risk group, as indicated by the significantly elevated levels of resting CD4 memory T cells, M0, and M2.

Conclusions: We constructed a prognostic signature for BLCA based on two co-expressed genes. The performance of this signature was validated in both TCGA and GSE13507, indicating its potential usefulness in predicting the prognosis of BLCA and developing new therapeutic methods.