Plasma miRNAs in predicting radiosensitivity in non-small cell lung cancer

Abstract

Background: Radioresistance of thoracic radiotherapy is a major bottleneck in the treatment of non-small cell lung cancer (NSCLC). Until now, there have been no effective biomarkers to predict the radiosensitivity.

Purposes: Based on miRNA profile screened from NSCLC cell lines with different radiosensitivity, this study was conducted to explore the correlation between plasma miRNAs and radiotherapy response in NSCLC patients, and to identify biomarkers of the radiosensitivity in NSCLC.

Methods: Differentially expressed genes were acquired from time-series gene expression profiles of radioresistant H1299 and radiosensitive H460 lung cancer cells (GSE20549). Potential miRNAs were screened from these differentially expressed genes by combining bioinformatics with GO analysis, pathway analysis, and miRNA prediction. A clinical observational study was performed to explore the correlation between candidate miRNAs and radiotherapy response. Stage IIIa-IV NSCLC patients who received two to four cycles of previous chemotherapy and underwent thoracic radiotherapy alone were included. Total RNA was purified from peripheral blood before radiotherapy, and plasma miRNAs were detected by real-time PCR (qRT-PCR). Then, tumor response, progression-free survival (PFS), and overall survival (OS) were acquired. Four miRNAs significantly different between effective and ineffective groups were further analyzed to obtain cutpoints from receiver operating characteristic (ROC) curves and the predictive value of radiosensitivity.

Results: Candidate miRNAs included 14 miRNAs screened from radioresistant genes and five from radiosensitive genes. From Jan., 2013 to Dec., 2014, 54 eligible patients were enrolled with a median follow-up of 15.3 months (range 4.6 to 31.4) by the deadline of Aug. 31, 2015. Totally, there were no case of complete response (CR), 15 of partial response (PR), 35 of stable disease (SD), and 4 of progressive disease (PD). Eight patients had no progression and 19 patients were still alive. The median PFS and OS were 6.6 months (range 2.3 to 29.3) and 15.3 months (range 4.6 to 31.4), respectively. Four miRNAs (hsa-miR-98-5p, hsa-miR-302e, hsa-miR-495-3p, and hsa-miR-613) demonstrated a higher expression in effective group (CR + PR, 15 cases) than in ineffective group (SD + PD, 39 cases). Based on each cutpoint, objective response rate (ORR) was higher in miR-high group than in miR-low group. No miRNA showed correlation with median PFS or OS.

Conclusion: Bioinformatical analysis and clinical verification reveal the correlation between plasma miRNAs and radiosensitivity in NSCLC patients. Plasma miRNAs represent novel biomarkers to predict radiotherapy response clinically.

Keywords: Bioinformatics; Biomarker; NSCLC; Plasma; Radiosensitivity; miRNAs.

Fig. 1

Bioinformatics of miRNA associated with radiosensitivity in NSCLC. a miRNA-gene-network (14 miRNAs associated with radioresistant genes). b miRNA-gene-network (5 miRNAs associated with radiosensitive genes). Red rectangles represent miRNAs, blue circles represent gene ontology, and straight lines represent the regulation between miRNAs and gene ontology

Fig. 2

Study design and workflow. a Study design. b Flow chart of patient selection

Fig. 3

Tumor response and radiosensitivity prediction by plasma miRNAs. a Waterfall map of maximum tumor response in 54 patients. b Differentially expressed plasma miRNAs based on tumor response. Four miRNAs have a significant difference (*P < 0.05). c Cutpoints in receiver operating characteristic (ROC) curves of plasma miRNAs (*P < 0.05). d Tumor response based on cutpoints of plasma miRNAs (*P < 0.05, **P < 0.01)

Fig. 4

PFS and OS comparison based on miRNA level. Kaplan-Meier analysis of PFS and OS based on plasma miRNA level. No miRNA is associated with PFS or OS