Prognostic signature associated with radioresistance in head and neck cancer via transcriptomic and bioinformatic analyses

Background: Radiotherapy is an indispensable treatment modality in head and neck cancer (HNC), while radioresistance is the major cause of treatment failure. The aim of this study is to identify a prognostic molecular signature associated with radio-resistance in HNC for further clinical applications. Methods: Affymetrix cDNA microarrays were used to globally survey different transcriptomes between HNC cell lines and isogenic radioresistant sublines. The KEGG and Partek bioinformatic analytical methods were used to assess functional pathways associated with radioresistance. The SurvExpress web tool was applied to study the clinical association between gene expression profiles and patient survival using The Cancer Genome Atlas (TCGA)-head and neck squamous cell carcinoma (HNSCC) dataset (n = 283). The Kaplan-Meier survival analyses were further validated after retrieving clinical data from the TCGA-HNSCC dataset (n = 502) via the Genomic Data Commons (GDC)-Data-Portal of National Cancer Institute. A panel maker molecule was generated to assess the efficacy of prognostic prediction for radiotherapy in HNC patients. Results: In total, the expression of 255 molecules was found to be significantly altered in the radioresistant cell sublines, with 155 molecules up-regulated 100 down-regulated. Four core functional pathways were identified to enrich the up-regulated genes and were significantly associated with a worse prognosis in HNC patients, as the modulation of cellular focal adhesion, the PI3K-Akt signaling pathway, the HIF-1 signaling pathway, and the regulation of stem cell pluripotency. Total of 16 up-regulated genes in the 4 core pathways were defined, and 11 over-expressed molecules showed correlated with poor survival (TCGA-HNSCC dataset, n = 283). Among these, 4 molecules were independently validated as key molecules associated with poor survival in HNC patients receiving radiotherapy (TCGA-HNSCC dataset, n = 502), as IGF1R (p = 0.0454, HR = 1.43), LAMC2 (p = 0.0235, HR = 1.50), ITGB1 (p = 0.0336, HR = 1.46), and IL-6 (p = 0.0033, HR = 1.68). Furthermore, the combined use of these 4 markers product an excellent result to predict worse radiotherapeutic outcome in HNC (p < 0.0001, HR = 2.44). Conclusions: Four core functional pathways and 4 key molecular markers significantly contributed to radioresistance in HNC. These molecular signatures may be used as a predictive biomarker panel, which can be further applied in personalized radiotherapy or as radio-sensitizing targets to treat refractory HNC.