GPC3 as a potential diagnostic and prognostic marker for lung adenocarcinoma

Abstract

We designed this study to identify potential diagnostic and prognostic markers for lung adenocarcinoma (LUAD). Four gene expression profiles were downloaded from GEO and merged into a training cohort, and those genes that were differentially expressed between LUAD and normal samples were selected. We performed LASSO regression, SVM-RFE, and ROC curve analyses, and external validations were conducted using the GSE115002 dataset, TCGA + GTEx datasets, and tissue microarrays (TMAs) of 56 patients with LUAD from our hospital. Subsequent analyses included survival analyses, Cox regression, CIBERSORT, CCK-8, colony formation, and flow cytometry analyses. We observed that the expression of SH3GL2, CAV1, and GPC3 was reduced in LUAD samples in comparison to normal samples. Therefore, these genes were identified as diagnostic markers for LUAD, and validated using the GSE115002 and TCGA + GTEx datasets (AUC > 0.86 for all three). Notably, lower GPC3 expression levels were associated with worse prognoses. And GPC3 is an independent prognostic factor for LUAD; TMA analysis confirmed low GPC3 expression in tumors. Moreover, GPC3 overexpression inhibited cell viability and clonality but not cell apoptosis. In conclusion, we infer that GPC3 could serve as both a diagnostic and prognostic marker for LUAD.