Abstract: FR-OR20
Genome-Wide Association Study Across Biobanks Identifies New Susceptibility Loci for Urinary Tract Infections
Session Information
- Development, Organoids, and Genetic Models of Kidney Diseases
November 03, 2023 | Location: Room 105, Pennsylvania Convention Center
Abstract Time: 05:51 PM - 06:00 PM
Category: Genetic Diseases of the Kidneys
- 1202 Genetic Diseases of the Kidneys: Non-Cystic
Authors
- Xu, Katherine, Columbia University Irving Medical Center, New York, New York, United States
- Khan, Atlas, Columbia University Irving Medical Center, New York, New York, United States
- Shang, Ning, Columbia University Irving Medical Center, New York, New York, United States
- Berrouet, Cecilia C., Columbia University Irving Medical Center, New York, New York, United States
- Shen, Tian, Columbia University Irving Medical Center, New York, New York, United States
- Barasch, Jonathan M., Columbia University Irving Medical Center, New York, New York, United States
- Sims, Peter A., Columbia University Irving Medical Center, New York, New York, United States
- Kiryluk, Krzysztof, Columbia University Irving Medical Center, New York, New York, United States
Background
Urinary tract infections (UTIs) represent the most common bacterial infections worldwide and if untreated, can lead to pyelonephritis, urosepsis, and severe complications. Urologic abnormalities, immunosuppression, sexual activity, and other known risk factors do not completely explain why some individuals are more susceptible to severe or recurrent UTIs. Genetic predisposition to UTIs has not been well studied.
Methods
We designed an electronic phenotype for UTIs and defined 75,485 cases and 634,538 controls across several biobanks, including eMERGE, UKBB, BioVU, and All of Us. Cases were defined by ICD9/10 codes by ≥2 events ±1 week apart in the absence of known clinical risk factors. We performed a GWAS for UTI across diverse ancestries using REGENIE, controlling for age, sex, and genetic ancestry, followed by a fixed effects meta-analysis across individual biobanks. We also performed single cell RNA-seq (scRNAseq, 10x Genomics) of kidneys from mice ±UTI (n=10). We used UPEC isolate UTI89 in bacterial growth rate and recombinant protein binding assays.
Results
Our UTI GWAS meta-analysis defined several genome-wide significant loci, including the PSCA locus (P=2.9E-11). This gene is specifically expressed in urothelium and kidney papilla. Based on the latest kidney eQTL datasets, the risk allele at this locus was associated with lower PSCA expression in the tubulointerstitial compartment (P=8.8E-39). Our scRNAseq data demonstrated its expression in the kidney pelvic epithelial cells in both infected and noninfected mice, and RNA in situ localized kidney PSCA exclusively to the ascending thin limbs. In binding assays, we found E. coli bound to human PSCA, a heavily N-glycosylated protein (initial calculations show >1000 PSCA molecules/bacterium). Co-incubation with PSCA in vitro limited bacterial growth in a dose-dependent manner (0.35μM PSCA + 1E7 CFUs: 21% reduction at 8 hours (OD600 0.71±0.01 vs 0.56±0.02, p=0.004)).
Conclusion
Our large-scale multi-biobank-based GWAS combined with kidney scRNAseq data and bacteria binding assays identified PSCA as a candidate causal gene in human UTIs. PSCA is expressed in various urinary tract epithelia and inhibits bacterial growth in vitro. Our findings suggest that PSCA plays an important role in human innate immunity against bacterial UTIs.
Funding
- NIDDK Support