Multiomic Profiling of Putative Urinary Biomarkers for the Diagnosis of Pyelonephritis

Background: Pyelonephritis, a severe upper urinary tract infection, is the leading cause of acute kidney injury and renal scarring in childhood. Despite its clinical significance, non-invasive biomarkers for early diagnosis, progression, and target intervention have not been well characterized. To address this gap, we constructed a multi-omics atlas of acute pyelonephritis incorporating spatial transcriptomic, single-cell RNAseq, bulk microarray, secretomic and urinary proteomics, to identify candidate urinary biomarkers. Methods: We performed spatial transcriptomics (10X Visium) on kidneys from adult female C3H/HeOuJ infected with Uropathogenic Escherichia coli (UPEC) at 0,1, 3, 5, 7, 28 days post-infection. Infection-associate niches (IANs) were identified across the course of UPEC infection. Differentially expressed genes (DEGs) from IANs vs. adjacent uninfected regions were compared with DEGs from spatial and bulk datasets, the mouse secretome, and human urinary proteomes (AKI, CKD, controls) to identify putative PN urinary biomarkers. Those spatial, single cell and bulk transcriptomic temporal data was used to distinguish early from late biomarkers. Selected candidates were validated in urine from children with febrile UTI (14 patients vs. 30 controls). Results: Canonical AKI-CKD inflammatory (e.g., Lcn2, S100a8, S100a9) and fibrosis markers (e.g., Tgfb1, Col1a1, Fn1) were enriched in IANs, while tubular injury markers (Havcr1, Igfbp7) localized elsewhere, indicating spatial injury heterogeneity. We identified 1,369 DEGs in IANs (994 upregulated, 375 downregulated). Cross-platform integration from micro-array, single cells, secreted proteins, and human urinary proteomes yielded 64 conserved and 163 secreted-specific biomarkers, including known (Vcam1, Lcn2, S100a9, Ctss) and novel (Arg1, Ccl7, Serpine1) candidates. Spatial mapping showed region-specific expression, and urinary validation confirmed elevated VCAM1 (P<0.0001), CCL7(P <0.001), and ARG1(P<0.001) in febrile pediatric UTI patients. Conclusion: Our mult-omic atlas provided a valuable resource for identifying PN biomarkers, potentially improve the non-invasive diagnosis of patients with PN. Further correlation with imaging (DMSA, CT) is needed to assess predictive value. Funding: NIDDK Support