Urinary cell transcriptomics: a non-invasive expression readout of kidney genes of relevance to hypertension

Objective: Hypertension affects 35% of the global adult population and is the leading preventable risk factor for premature death globally. The kidney is the key organ of blood pressure regulation in the body. Here, we compare the gene expression profile from renal cells shed naturally into the urine and kidney tissue to determine if urine can provide a non-invasive readout of kidney gene expression and whether hypertension-associated genes can be quantified in urine. Design and Method: Urinary cell and kidney tissue samples were collected from 33 human participants and were both profiled by poly-A RNA-sequencing, generating an average of 30 million paired reads per sample. These were quantified using the standard Genotype Tissue Expression (GTEx) project pipeline and were compared against 43 different human tissues and other bodily fluids using transcriptomic correlation. RNA-sequencing quality metrics were calculated by RNA-SeQC software. Gene set overrepresentation analysis for Gene Ontology (GO) terms and Kyoto Encyclopaedia of Genes and Genomes (KEGG) employed Fishers exact test. Enriched and enhanced kidney genes were collected from the Human Protein Atlas (HPA). Results: Our RNA-sequencing metric analysis revealed that urinary cells can generate robust data of comparable or superior quality to that of saliva at similar read coverage. The top one hundred most highly expressed urinary cell genes show an enrichment for biological themes shared with the renal transcriptome, including immunity (P = 1.2x10 –7 ), glucose metabolism (P = 1.3x10 –5 ) and renal mineral reabsorption (P = 9.8x10 –4 ). In an analysis all protein-coding genes across 43 human tissues/cell-types, urinary cells showed the highest level of transcriptomic correlation with kidney cortex (r 2 = 0.65) and kidney medulla (r 2 = 0.64). This correlation between urinary cells and kidney tissue was particularly strong (r 2 = 0.72) in an analysis restricted to highly specific kidney genes (including uromodulin and the Na-K-Cl cotransporter NKCC2 [loop diuretic target]). 98% (176 out of 179) of kidney genes with a known causal association to blood pressure were expressed in urinary cells. Their urinary expression demonstrated strong correlation with their abundance in kidney cortex (r 2 = 0.68) and medulla (r 2 = 0.64). Conclusions: Standard poly-A RNA-sequencing of cells harvested from urinary sediments produces robust gene expression profiles. These profiles provide a non-invasive insight into transcriptome of the kidney and permit measuring expression of kidney genes of relevance to BP regulation and hypertension.