Sedigheh Gharbi; Fatemh Mirzadeh; Shahriar Khatrei; Mohammad Reza Soroush; Mahmood Tavallaie; Mohammad Reza Nourani; mehdi Sahmsara; Seyyed Javad Mowla
Abstract
MicroRNAs constitute a group of small non-coding RNAs that negatively regulate gene expression. Aside from their contribution to biological and pathological pathways, altered expression of microRNAs is reported in bio-fluid samples, such as serum. To employ serum's microRNAs as potential biomarkers, ...
Read More
MicroRNAs constitute a group of small non-coding RNAs that negatively regulate gene expression. Aside from their contribution to biological and pathological pathways, altered expression of microRNAs is reported in bio-fluid samples, such as serum. To employ serum's microRNAs as potential biomarkers, it is crucial to develop an efficient method for microRNA quantification, avoiding pre-analytical and analytical variations which could affect the accuracy of data analysis. Here, we optimized a real-time PCR quantification procedure for microRNA detection in serum samples. Serum's total RNA was extracted using two different RNA isolation methods, one based on phenol-chloroform and the other based on silica column. To investigate a potential PCR inhibitory effect, different RNA amounts were subjected to reverse transcription. Moreover to assess the enzymatic efficiency, synthetic exogenous microRNAs was spiked into the mixture. Moreover, to find a reliable internal control gene for normalizing the microRNA quantification, the amounts of 8 candidate non-coding RNAs including SNORD38B, SNORD49A, U6, 5S rRNA, miR-423-5p, miR-191, miR-16 and miR-103 were assessed on serum samples. Altogether, our data demonstrated that the silica-based method was more efficient for microRNA recovery. Furthermore, increasing the input volume of the extracted RNA would dramatically increase inhibitors' amounts which could end up in a larger Cq values. Therefore, the best input volume of RNA turned out to be 1.5 microliter/reaction. Among the 8 aforementioned internal controls, U6, SNORD38B and SNORD49A showed low levels of expression, and were undetectable in some samples. Amongst the others, 5s rRNA, had the biggest standard deviation which could significantly affect data analysis. MiR-103 with the least variation appeared to be the best normalizer gene.