Identification of biomarkers for Laryngeal squamous cell carcinoma through Mendelian randomization and integrated bioinformatics analysis.

BACKGROUND

Laryngeal squamous cell carcinoma (LSCC) is a common head and neck tumor with an increasing incidence rate and limited therapeutic efficacy in advanced stages.

METHODS

This study integrated multi-omics data from multiple databases to comprehensively analyze the role of pan-apoptosis-related genes (PANRGs) in the prognosis of LSCC. RNA-seq data from 114 LSCC and 12 normal samples were collected from the TCGA-HNSC cohort as a training set, while 270 LSCC samples were obtained from the GSE65858 dataset as a validation set. Additionally, copy number variation (CNV) data were retrieved from the UCSC Xena database, and single-cell RNA sequencing (scRNA-seq) data from two LSCC samples were obtained from the GSE150321 dataset. Differential expression analysis was performed using the R package "limma" to identify 33 differentially expressed PANRGs (DEPANRGs) with a significance threshold of 0.05. Potential pathways were explored through Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Metascape enrichment analyses. A prognostic model was constructed based on nine key genes (DNAJC3, FUNDC1, GATA3, NLRP3, PMAIP1, TGFB2, TIMP1, TIMP2, and TNFRSF1B) using the Lasso-Cox method, and its predictive accuracy was evaluated by Kaplan-Meier (KM) analysis and receiver operating characteristic (ROC) curves. The differences in immune landscapes between high-risk and low-risk groups were assessed using the CIBERSORT algorithm, and drug sensitivity was analyzed using the "oncoPredict" package. Finally, the expression patterns of prognostic genes in scRNA-seq data were explored through pseudotime analysis and gene set scoring methods.

RESULTS

Differential expression analysis identified 33 DEPANRGs, which were significantly enriched in immune response, inflammation, and apoptosis-related pathways. The constructed prognostic model exhibited robust predictive power, with area under the curve (AUC) values of 0.769 and 0.857 in the training and validation sets, respectively. Risk scores were significantly correlated with clinical factors such as gender and N stage (extent of regional lymph node metastasis). Immune landscape analysis revealed distinct patterns of immune cell infiltration and functional scores between high-risk and low-risk groups, with significant differences in the infiltration of monocytes and neutrophils. Drugs such as Acetalax, Entinostat, and OTX015 exhibited significant differences in efficacy between the high and low-risk groups. Single-cell RNA sequencing analysis identified monocytes as the cell type with the highest prognostic gene scores, which interacted significantly with other cell populations through pathways such as MIF, annexin, and CXCL. Differences in pathway activity and transcription factor (TF) activity were also observed between high-expression and low-expression groups. qRT-PCR experiments further validated the significant expression differences of prognostic genes between LSCC and control groups.

CONCLUSION

This integrative approach comprehensively elucidated the role of pan-apoptosis-related genes in LSCC. The constructed risk model has significant clinical application value in prognostic prediction, immune landscape assessment, and drug sensitivity analysis, and has the potential to guide precision treatment strategies for LSCC patients. These findings may help advance personalized treatment plans and improve the prognosis of patients with LSCC.