Kidney Week

Abstract: FR-OR042

Comprehensive Identification of Therapeutic Targets in CKD: Integrating Population Genetics with High-Throughput Clustered Regularly Interspaced Short Palindromic Repeats Interference (CRISPRi) Perturb-seq

Session Information

• Exciting Developments in CKD Research: From Diagnostics to Management and More
  October 23, 2026 | Location: Mile High Ballroom 1D, Convention Center
  Abstract Time: 05:20 PM - 05:30 PM

Category: CKD (Non-Dialysis)

• 2203 CKD (Non-Dialysis): Mechanisms

Authors

• Shungin, Dmitry, Novo Nordisk A/S Global Research, Måløv, Capital Region of Denmark, Denmark

Dmitry Shungin, PhD

• Solovyev, Andrey, Novo Nordisk A/S Global Research, Måløv, Capital Region of Denmark, Denmark

Andrey Solovyev, MSc

• Fadista, João Paulo Santos, Novo Nordisk A/S Global Research, Måløv, Capital Region of Denmark, Denmark

João Paulo Santos Fadista, PhD

• Han, Seungmin, Novo Nordisk A/S Global Research, Måløv, Capital Region of Denmark, Denmark

Seungmin Han, PhD

• Liang, Xiujie, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, United States

Xiujie Liang, MD, PhD

• Susztak, Katalin, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, United States

Katalin Susztak, MD, PhD

• Karihaloo, Anil K., Novo Nordisk A/S Global Research, Måløv, Capital Region of Denmark, Denmark

Anil K. Karihaloo, PhD

Background

We present a comprehensive CKD target discovery effort, uniquely combining large-scale population genetics, multi-omics and high-throughput functional screening to identify and prioritize causal therapeutic candidates.

Methods

We performed systematic Mendelian randomization (MR; gIVW) on eGFR across two molecular layers: blood proteins measured in N≈137K individuals using cis-pQTLs (+/-500kb & P<5x10^-8) from Olink (UK Biobank N≈49K; SCALLOP N≈38K; EPIC-Norfolk N≈1K) & SomaScan (deCODE N≈35K; Fenland N≈10K; ARIC N≈7K); expression QTLs from tubular (N≈667) and glomerular (N≈543) tissues. Bayesian colocalization (HyprColoc) was applied to prioritize genes.
Candidates were functionally validated using CRISPRi Perturb-seq with single-cell transcriptomic readout in immortalized human proximal tubule epithelial cells (RPTEC/TERT1), under basal and injury-inducing conditions (TNFα + IFNγ). Knockdown efficiency was assessed using mixed linear models (glmmTMB); perturbation signatures and continuous knockdown dosages were quantified via Mixscape/Mixscale, and phenotypic effects were quantified using latent-space recovery axis projections. Transcriptional programs were mapped using GSFA, with differentially expressed genes (LFSR<0.05) annotated against KEGG, Hallmark, and canonical pathway databases. Experimental readouts were triangulated with MR results to identify candidates with convergent genetic and functional evidence.

Results

MR and colocalization identified 225 genes with putative causal associations with eGFR (MR P<2.5e-6; PP.H4≥0.7), with consistent effect directions across multiple studies and QTL sources. Of these, 193 gene perturbations passed stringent quality control in CRISPRi Perturb-seq assays. Triangulation of dose-dependent transcriptomic shifts with knockdown efficiency and MR causal directionality yielded high-confidence therapeutic candidates converging on shared biological programs, including NF-κB stress-resolution coupled with EMT suppression, and cell-cycle checkpoint activation coupled with inflammatory dampening. Known CKD-relevant targets (MANBA, CALR, DPEP1) were recovered, alongside multiple novel putative therapeutic candidates.

Conclusion

We propose potential novel targets for CKD and outline approach to combine MR and colocalization with high throughput CRISPRi Perturb-seq assays for efficient data-driven drug target prioritization.