miRNA Expression Signatures of Therapy Response in Squamous Cell Carcinomas

SIMPLE SUMMARY: miRNAs play role in various diseases and can also modulate therapy response. Our aim was to identify predictive miRNAs in platinum treated squamous cell carcinomas (SCC). Using a set of 266 squamous cancer samples we uncovered 16, 103, and 9 miRNAs correlated to chemotherapy response in the cervical, head and neck, and lung squamous cell carcinomas, respectively. By employing a logistic regression model, a signature comprising a set of six miRNAs was established capable to predict chemotherapy response with an AUC of 0.897. Our results show common molecular features of SCC tumors and pinpoint the most important miRNAs related to treatment outcome. ABSTRACT: Introduction: Squamous cell carcinomas (SCC) are a major subgroup of malignant tumors with a platinum-based first-line systematic chemotherapy. miRNAs play a role in various diseases and modulate therapy response as well. The aim of this study was to identify predictive miRNAs in platinum-treated SCCs. Methods: miRNA expression data of platinum-treated head and neck (HNSC), cervical (CESC) and lung (LUSC) cancer were collected from the TCGA repositories. Treatment response was defined based on presence or absence of disease progression at 18 months. Responder and nonresponder cohorts were compared using Mann–Whitney and Receiver Operating Characteristic tests. Logistic regression was developed to establish a predictive miRNA signature. Significance was set at FDR < 5%. Results: The integrated database includes 266 SCC patient samples with platinum-based therapy and available follow-up. We uncovered 16, 103, and 9 miRNAs correlated to chemotherapy response in the CESC, HNSC, and LUSC cohorts, respectively. Eight miRNAs overlapped between the CESC and HNSC subgroups, and three miRNAs overlapped between the LUSC and HNSC subgroups. We established a logistic regression model in HNSC and CESC which included six miRNAs: hsa-miR-5586 (Exp (B): 2.94, p = 0.001), hsa-miR-632 (Exp (B): 10.75, p = 0.002), hsa-miR-2355 (Exp (B): 0.48, p = 0.004), hsa-miR-642a (Exp (B): 2.22, p = 0.01), hsa-miR-101-2 (Exp (B): 0.39, p = 0.013) and hsa-miR-6728 (Exp (B): 0.21, p = 0.016). The model using these miRNAs was able to predict chemotherapy resistance with an AUC of 0.897. Conclusions: We performed an analysis of RNA-seq data of squamous cell carcinomas samples and identified significant miRNAs correlated to the response against platinum-based therapy in cervical, head and neck, and lung tumors.