Screening of shared molecular markers between cervical cancer and major depressive disorder and the mechanism of CRAT/CLIC4 regulating EMT in cervical cancer cells.

BACKGROUND: Cervical cancer (CC) ranks among the most prevalent malignant neoplasms affecting women worldwide. Tumor recurrence, distant metastases, and chemotherapy resistance significantly hinder long-term clinical survival and therapeutic outcomes. Clinical studies indicate a heightened prevalence of depressive symptoms and major depressive disorder (MDD) among CC patients, suggesting the possibility of bidirectional adverse biological interactions between cervical tumor progression and depressive states. However, the shared molecular signatures underlying both cervical carcinoma and depressive disorders have yet to be fully elucidated, highlighting the need for identifying reliable molecular markers for supplementary diagnosis and prognostic stratification of CC. METHODS: Two datasets (GSE98793 for MDD and GSE63514 for CC) were retrieved from the Gene Expression Omnibus (GEO) database to identify shared differentially expressed genes (DEGs). Gene clustering analysis of the overlapping DEGs was conducted utilizing Metascape. Venn analysis facilitated the identification of overlaps between shared DEGs and extracellular vesicle (EV)-related genes. Six machine learning algorithms were employed to pinpoint core comorbid genes and develop a diagnostic model. Functional enrichment, drug sensitivity, and survival analyses were performed to assess gene functions, therapeutic potential, and prognostic relevance. In vitro experiments provided additional validation of the effects of core genes on CC malignant phenotypes and epithelial-mesenchymal transition (EMT). Transcriptomic and clinical data of cervical cancer were retrieved from TCGA. Corrected Pearson correlation analysis and PROGENy pathway activity analysis were performed via R packages to dissect the pathway regulatory correlations of CRAT and CLIC4. RESULTS: A total of 85 intersecting DEGs were identified from the transcriptomes of CC and female MDD. Gene cluster analysis indicated that these genes are primarily linked to biological processes such as immune homeostasis modulation, cell cycle progression, PD-1/PD-L1 immune checkpoint signaling, and dopaminergic synaptic transmission. Machine learning techniques identified four candidate hub genes-EZR, MAOA, CLIC4, and CRAT. This four-gene signature exhibited a high discriminatory capacity for CC and moderate diagnostic performance for MDD samples. Bioinformatic predictions implied that these genes are involved in tumor metabolic reprogramming regulation, cell cycle control, EMT activation, and anti-tumor immune responses. Cobimetinib was recognized as a potential small-molecule agent targeting CLIC4, EZR, and MAOA. Survival analysis indicated that CRAT may function as a favorable prognostic marker, whereas increased expression of CLIC4 correlates with adverse clinical outcomes. Risk scores derived from the four hub genes could serve as independent prognostic indicators for CC. In vitro cellular experiments partially confirmed that CRAT knockdown enhances malignant transformation and EMT in CC cells, while CLIC4 interference inhibits cell proliferation, migration, and EMT progression. Multi-pathway correlation networks demonstrated a positive association between CRAT and the JAK-STAT signaling cascade. Linear correlation analysis further substantiated that CRAT expression was significantly positively correlated with the levels of JAK1 and STAT3. Concurrently, CLIC4 showed a positive correlation with the TGF-β signaling pathway, and single-gene correlation analysis confirmed a pronounced linear positive association between CLIC4 and its downstream molecules, TGFBR1 and SMAD2. CONCLUSION: This study examines the shared molecular signatures between CC and MDD. The four-gene model, comprising EZR, MAOA, CLIC4, and CRAT, shows promise for differential diagnosis and independent prognostic assessment of cervical cancer. In vitro experiments indicate that elevated CRAT expression is associated with extended survival, whereas increased CLIC4 expression is linked to a poor prognosis. These two genes influence malignant phenotypes and EMT through distinct pathways. This research provides preliminary evidence of molecular crosstalk between the two disorders and offers a theoretical foundation for the identification of prognostic biomarkers and therapeutic targets for CC.