Renal Biomarker Evaluation using Data Integration in the Age of Precision Medicine

Abstract

Decades of work have given nephrology a handful of strong biomarkers of acute kidney injury (AKI), including clinical serum creatinine, blood urea nitrogen levels, and urine proteinuria. To refine the location of AKI to different regions of the nephron, proteins such as Albumin, NGAL, and KIM‐1 have been critical. Yet, many of our classical nephrology biomarkers are rarely evaluated considering broad patient pathologies and environmental factors such as viral infection. Millions of data points can inform on AKI in the data age. Thus, we have performed a systematic data analysis of ~10,000 blood sample precision transcriptomes for broad pathologies, ~200 RNA‐Seq samples for human AKI, the Comparative Toxicogenomics Database, Human Protein Atlas, GTEx, and our custom cell culture models of kidney response followed by advanced sequencing including direct RNA‐Seq. This data integration work elucidated biomarkers such as KIM‐1 can be elevated in blood due to viral infections and environmental contributors. In some patients, the elevation of KIM‐1 in the urine occurs due to blood activation and urine proteinuria instead of tubule secretion. However, refining KIM‐1 data with blood RNA‐Seq allows for identifying isoform switches that correlate to AKI and tubule damage.Further work in our data integration shows that RNA specific to tubule cells or podocytes can be observed in patient blood during AKI, suggesting a promising observable measure of detailed cell types damaged that may contribute to AKI. Finally, integrating our human GWAS and expression quantitative trait loci (eQTLs) for kidney traits elucidates the ability to map response eQTLs and mechanisms from easier to obtain blood samples that can be critical in defining AKI and chronic kidney diseases. Thus, continued revaluation of kidney biomarkers in the big data age can bring nephrology into the forefront of precision medicine.