MicroRNAs (18-22 nucleotide long non-coding RNAs) are emerging as promising biomarkers in multiple diseases. A single microRNA can target multiple messenger(m)RNAs and regulate gene expression post-transcriptionally. MicroRNAs are becoming increasingly important in understanding gene expression regulation. Our analysis of developing and adult rodent and human tissues identified that the gallbladder and the brain also produce insulin. Since microRNAs are necessary for normal pancreas development/function, it is important to generate a database of microRNA expression profile in human insulin-producing tissues. The aim of this research is to generate a miRNome from human brain, gallbladder, pancreatic islets and other tissues.
We investigated a biobank of 284 different human tissues for eight mRNAs (ins, gcg, sst, pdx1, mafA, ngn3, 18s and gapdh) and 754 known and validated microRNAs using TaqMan™ real-time quantitative (q) PCR. Data were normalized to housekeeping genes (for mRNA) or internal controls followed by global normalization (for microRNAs) and analysed using penalized logistic regression analyses. Islet hormones were visualized by confocal microscopy.
Confocal microscopy and qPCR analysis for gene transcripts confirmed human endothelium, muscle, liver and skin did not contain insulin transcripts, whilst brain, gallbladder and islets transcribed insulin gene with increasing efficiency. Bidirectional hierarchical clustering identified microRNAs enriched in each tissue and those associated with high level of insulin transcript. Logistic regression analysis of 150 islet samples (compared to insulin-negative tissues) led to discovery set of 12 microRNAs associated with high levels of insulin transcript. This miRNA signature predicted insulin expression in a validation set of 88 blinded tissue profiles with 97% efficiency.
We present the first large resource of microRNA expression analyses from human insulin-producing tissues. MicroRNA signature provides key evidence to understand their role in influencing insulin gene expression. Generating such signatures will help in identifying beta cell differentiation / regeneration or death in individuals progressing to diabetes.