Development of a liquid-liquid phase separation-related genes signature to predict clinical outcomes, tumor immune microenvironment and chemotherapeutic response in head and neck squamous cell carcinoma

Objective

Liquid-liquid phase separation (LLPS) has been linked to oncogenesis, prompting research into LLPS-related genes as prognostic markers. Here, our objective was to develop a prognostic signature specific to head and neck squamous cell carcinoma (HNSCC) utilizing genes associated with LLPS.

Design

The LRG signature was developed using transcriptional sequence data obtained from the TCGA and GEO databases, employing univariate Cox regression, Kaplan-Meier analysis, and LASSO algorithms. Additionally, we developed an LRG nomogram that integrates the signature with specific clinicopathological features, identified using multivariate Cox regression. Clinical value of LRG signature in predicting chemotherapeutic drug sensitivity and immune infiltration was evaluated by bioinformatics approaches. One of the genes in the signature was knocked down to confirm its role in vitro.

Results

The LRG signature demonstrated prognostic performance across multiple independent cohorts. Moreover, LRG signature scores exhibited a negative correlation with the presence of tumor-infiltrating immune cells and were associated with the sensitivities of various chemotherapeutic agents. LRP12 silencing markedly inhibited the proliferation, migration, and invasion of HNSCC cells, while also inducing apoptosis in these cells.

Conclusion

The findings of our study underscore the efficacy of the LRG signature and nomogram in forecasting prognosis and response to chemotherapy, suggesting a favorable outlook for their application in clinical settings.