Identification of reference genes for circulating microRNA analysis in colorectal cancer.

Yanqin Niu,Deming Gou,Jinyong Huang,Yike Wu,Kang Kang,Qing Li,Furong Li,Junle Qu

doi:10.1038/srep35611

Yanqin Niu, Deming Gou + Show 6 more

Open Access

https://doi.org/10.1038/srep35611

Copy DOI

Abstract

Quantitative real-time PCR (qPCR) is the most frequently used method for measuring expression levels of microRNAs (miRNAs), which is based on normalization to endogenous references. Although circulating miRNAs have been regarded as potential non-invasive biomarker of disease, no study has been performed so far on reference miRNAs for normalization in colorectal cancer. In this study we tried to identify optimal reference miRNAs for qPCR analysis across colorectal cancer patients and healthy individuals. 485 blood-derived miRNAs were profiled in serum sample pools of both colorectal cancer and healthy control. Seven candidate miRNAs chosen from profiling results as well as three previous reported reference miRNAs were validated using qPCR in 30 colorectal cancer patients and 30 healthy individuals, and thereafter analyzed by statistical algorithms BestKeeper, geNorm and NormFinder. Taken together, hsa-miR-93-5p, hsa-miR-25-3p and hsa-miR-106b-5p were recommended as a set of suitable reference genes. More interestingly, the three miRNAs validated from 485 miRNAs are derived from a single primary transcript, indicting the cluster may be highly conserved in colorectal cancer. However, all three miRNAs differed significantly between healthy individuals and non-small cell lung cancer or breast cancer patients and could not be used as reference genes in the two types of cancer.

Highlights

In this article, we aim to develop miRNA profiling data normalization approach based on the selection from large scale of candidate miRNAs in colorectal cancer
We found that there was no significant difference for hsa-miR-27a-3p in colorectal cancer patient serum compared to the healthy controls when data were normalized to cel-miR-54-5p (P = 0.1253), hsa-miR-93-5p (P = 0.3755), hsa-miR-25-3p (P = 0.3711) and hsa-miR-106b-5p (P = 0.3990), respectively; hsa-miR-144-3p was significantly up-regulated in colorectal cancer patient group when data were normalized to the four normalizers but the tendency was more significant when hsa-miR-106b-5p was used as reference (P < 0.01, P < 0.01, P < 0.01 and P < 0.001, respectively); for hsa-miR-223-3p, the expression levels differed significantly when cel-miR-54-5p (P < 0.05), hsa-miR-93-5p (P < 0.05) and hsa-miR-106b-5p (P < 0.05) were served as normalizers respectively, but did not differ significantly when data were normalized to hsa-miR-25-3p (P = 0.5226) (Fig. 4)
The accuracy of qPCR depends on suitable normalization of data inasmuch as inappropriate use of reference genes can significantly alter the results of target miRNA quantification

Summary

Introduction

We aim to develop miRNA profiling data normalization approach based on the selection from large scale of candidate miRNAs in colorectal cancer. 485 human miRNAs were selected and subsequently quantified with a sensitive qPCR-based assay, the S-Poly(T) Plus assay[21], of which, 372 miRNAs that could be detected in human serum or plasma were obtained from QIAGEN website (http://www.sabiosciences.com/genetable.php?pcatn=MIHS-3106Z); 113 other blood-derived miRNAs were extracted from literatures with the key words “microRNA/miRNA”, “serum/plasma/blood” in combination with “human/hsa-”. BestKeeper, geNorm and NormFinder algorithms were applied to identify the most stable reference genes. The recommended normalizer(s) was (were) validated in colorectal cancer as well as non-small cell lung cancer (NSCLC) and breast cancer

Objectives

Methods

Results

Conclusion