The landscape of high-confidence microrna - gene interactions operating in the human kidney and their role in hypertension

Objectives: MicroRNAs (miRNAs) are short non-coding RNAs acting as post-transcriptional regulators of mRNAs. It is becoming increasingly clear that they play an important role as molecular drivers/mediators of complex human diseases including hypertension. Given the key role of the kidney in the pathogenesis and treatment of hypertension we sought (i) to comprehensively characterise the network of renal miRNA-mRNA interactions and (ii) evaluate their relevance to human hypertension. Design and method: Using poly-A RNA-sequencing and small RNA-sequencing we have identified 22,558 genes and 1,604 miRNAs in the discovery analysis of 328 human kidney samples (Human Kidney Tissue Resource). The outputs from these experiments have been normalised and adjusted by linear regression. We conducted a replication analysis of the identified miRNA-gene pairs in normal control kidney samples from two National Institutes of Health cohorts - CPTAC (n = 66) and TCGA (n = 83). We then examined the extent to which the expression of kidney genes known as targets of blood pressure lowering medications can be modulated by microRNAs. Finally, we investigated whether the effects of these interactions are observed at the protein level (CPTAC dataset). Results: We identified 186,192 unique statistically significant inverse miRNA-gene correlations in our discovery resource. Of these, 4,319 were replicated in both NIH kidney resources. 32.5% of kidney protein-coding genes and 12.2% of kidney long non-coding RNAs had at least one such high-confidence miRNA associated with their expression. SLC12A3 (Solute Carrier Family 12 Member 3) - kidney gene target for thiazides, SLC12A1 (Solute Carrier Family 12 Member 1) - gene target for loop diuretics and SLC5A2 (Solute Carrier Family 12 Member 2) - gene target for SGLT2 inhibitors showed consistent, negative correlations with four, five and five kidney miRNAs in each of three independent datasets, respectively. Some of the identified miRNAs showed negative correlations with more than one gene-drug targets. For example, miR-31-5p showed a strong, significant inverse association with SLC5A2 (R = -0.67, P = 9.41E-10) and SLC12A3 (R = -0.70, P = 5.80E-11) in the CPTAC resource. We confirmed that miRNA-31-5p also negatively correlates with the renal expressions of SLC12A3 and SLC5A2 at the protein level (R = -0.58, P = 3.5E-7 and R = -0.45, P = 1.51E-4, respectively). Conclusion: We characterised highly credible miRNA-gene interactions in the human kidney and mapped them onto genes and proteins of established relevance to blood pressure lowering treatment. These kidney miRNAs represent exciting targets for the development of novel antihypertensives.