作者
Xiaomin Qi,Caijuan Bai,Dong Li,Anqi Wang,Chaojun Wei,Yonghong Li,Mengyuan Zhao,Chongge You
摘要
Abstract Colorectal cancer (CRC) is a highly lethal gastrointestinal tract malignancy whose pathogenesis and molecular drivers are not fully understood. This study focused on searching for genes that are differentially expressed in cancer versus normal mucosa, with the goal of identifying molecular patterns of expression that are mechanistically linked to CRC pathogenesis. We analyzed 585 CRC samples and 329 normal samples from the Gene Expression Omnibus (GEO) database, creating a weighted gene co-expression network (WGCNA) across 24,069 genes. Through this approach, five modules associated with CRC were identified, which were enriched in MAPK signaling and cholesterol metabolism pathways. Using LASSO regression, we selected 13 hub genes (ABCB5, AOC1, ARHGAP44, CACNG3, DBNDD1, GAS7, GTF2IRD1, PRSS22, SCN4A, TTC22, DLX6, PDK4, and SLC13A2) from these modules. Survival analysis indicated that higher expression of DBNDD1 correlated with worse overall survival in CRC patients. Machine learning validation confirmed the stability of these genetic markers. Experimental validation demonstrated increased levels of DBNDD1 and GDF15 in CRC, promoting constant NF-κB activation via DBNDD1-dependent GDF15 induction. Knocking down DBNDD1 inhibited cell proliferation, migration, and invasion in vitro (DLD1/HCT116 cells), alongside decreased GDF15 expression and reduced p-NF-κB p65/p-IκB signaling. Additionally, DBNDD1 knockdown resulted in reduced tumor growth in vivo, highlighting that DBNDD1-GDF15-NF-κB signaling pathway drives CRC pathogenesis. Implications: This study highlights the crucial role of the DBNDD1/GDF15/NF-κB signaling pathway in CRC development, positioning DBNDD1 as a promising target for precision medicine strategies aimed at enhancing patient outcomes.