Abstract: |
Recently, the mRNA presence of pregnancy-specific glycoproteins (PSGs) in cancer biopsies has been shown to be associated with poor survival. Given the pregnancy-related function of PSGs, we hypothesized that PSGs might act in a sex-dependent behavior in cancer patients. A differential sex effect of PSG genes with respect to tumor immune landscape and cancer outcomes was investigated using statistical, bioinformatic, and machine learning analyses in The Cancer Genome Atlas (TCGA) data. The resulting findings were then validated in the Clinical Proteomic Tumor Analysis Consortium (CPTAC) data. In a pan-cancer TCGA data analysis, the strongest PSG-related sex difference for the prognostic association was found in lung adenocarcinoma (LUAD). Kaplan-Meier analysis revealed that expression of PSG genes is strongly associated with overall survival rate in the female group on the TCGA, but not in the male group. This sex-specific association was validated in an independent dataset from the CPTAC study. A combination of PSG3, PSG7, and PSG8 expression was most significantly linked to poor prognosis in females (P = 8.67E-06 in TCGA and P =. 0382 in CPTAC). Pathway analysis revealed enrichment of the 'KRAS Signaling Down' pathway in the high-risk female group. A predictive model showed good predictive performance for the female group (validated C-index = 0.78 in CPTAC), but poor predictive performance for the male group. These findings suggest that PSGs may have a sex-specific negative impact on survival in female LUAD patients, and the mechanism may be related to KRAS signaling pathway modulation. © 2025 The Author(s). |