A Five-microrna Signature Identified From Genome-wide Serum Microrna Expression Profiling Serves As A Fingerprint For Gastric Cancer Diagnosis

Background:MicroRNA(miRNA), the endogenous non-coding RNAs consisting of19to23nucleotides in length, regulates target gene expression through binding to complementary sequences in the3’UTR of target mRNA specifically. A variety of evidences show that miRNA in various tissues have been associated with cancers. Recently, the studies on miRNA s focus on the relations between the serum miRNA which is stable in the serum and diseases. After testing the serum samples from the patients and the controls, some serum miRNA s have been found as the new biomarkers of the clinical diagnosis. As a new biomarker, serum miRNA could be the biomarker for early-stage cancer. Prognosis of patients with gastric cancer (GC) is generally poor due to the lack of non-invasive tools for GC detection. The purpose of present study was to identify a serum miRNA expression profile that can serve as a novel diagnostic biomarker for GC detection and to assess its clinical applications in monitoring disease progression.Methods: Serum samples were taken from164GC patients and127age-and gendermatched tumour-free controls. An initial screening of miRNA expression by Solexa sequencing was performed using serum samples pooled from20patients and20controls, respectively. In the biomarker selection phase, the expression miRNAs of the serum samples of22GC and22controls formed a training set was tested by using hydrolysis probe-based stem-loop quantitative reverse transcription polymerase chain reaction (qRT-PCR). In the validation selection phase, an additional142GC serum samples and105normal subjects formed a validation set was tested by using hydrolysis probe-based stem-loop quantitative reverse transcription polymerase chain reaction (qRT-PCR). Differential expression was validated using hydrolysis probe-based stem-loop quantitative reverse transcription polymerase chain reaction (qRT-PCR) in individuals samples, the samples were arranged in two phases.Results:The Solexa sequencing results demonstrated that19serum miRNAs were markedly upregulated in the GC patients compared to the controls. The qRT-PCR analysis further identified a profile of five serum miRNAs (miR-1, miR-20a, miR-27a, miR-34and miR-423-5p) as a biomarker for GC detection. The analysis results showed that the expression level of five serum miRNAs was correlated to tumour stage. The areas under the receiver operating characteristic (ROC) curve of this five-serum miRNA signature were0.879(95%confidence interval (CI)0.822-0.936) and0.831(95%CI0.767-0.898) for the two sets of serum samples, respectively, markedly higher than those of the biomarkers carcinoembryonic antigen (CEA)(0.503) and carbohydrate antigen19-9(CA19-9)(0.600).Conclusions:We identified five-miRNA signature for GC diagnosis by genome-wide serum miRNA expression profiling. Expression levels of this serum miRNA-based biomarker also indicate tumour progression stages.