Clonal hematopoiesis of indeterminate potential is associated with acute kidney injury

AKI affects more than one in five hospitalized adults worldwide and is associated with high costs and mortality. It involves inflammatory and fibrotic responses to injury, with heterogeneous outcomes; only about half of AKI cases return to baseline function within 3 months. Known risk factors are largely nonmodifiable, such as age, and no genetic predispositions to AKI or AKI outcomes had been identified. Clonal hematopoiesis of indeterminate potential (CHIP), an age-related expansion of hematopoietic clones with somatic mutations (commonly in DNMT3A, TET2, ASXL1, JAK2), has been linked to higher mortality and extra-hematopoietic diseases (cardiovascular, pulmonary, liver, and inflammatory conditions) and may affect kidney health, being associated with lower kidney function, faster decline, and CKD progression. Experimental models show Tet2-mutant hematopoietic cells can replace resident macrophages in kidney and liver, and myeloid cells are key in kidney injury and repair. The study hypothesized that CHIP increases AKI risk and impairs recovery through aberrant inflammatory responses, particularly for non-DNMT3A driver mutations.

Prior work established CHIP as common in aging and associated with adverse outcomes including atherosclerotic cardiovascular disease, stroke, pulmonary disease (COPD), severe COVID-19, chronic liver disease, gout, and metabolic disease. Within nephrology, CHIP has been associated with worse cross-sectional kidney function, accelerated kidney function decline in the general population, and adverse CKD outcomes. Mouse studies have demonstrated that Tet2-deficient hematopoietic cells can engraft and replace resident macrophage populations in tissues, contributing to inflammatory phenotypes relevant to organ injury. These findings motivated investigating CHIP as a putative genetic risk factor for AKI incidence and recovery, focusing on non-DNMT3A drivers such as TET2 and JAK2.

Human epidemiology: CHIP was ascertained using whole-exome sequencing in UK Biobank (UKB; n≈428,793) and whole-genome sequencing in TOPMed cohorts ARIC (n=10,570) and CHS (n=2,790) with somatic variant calling via Mutect2 and established filtering criteria. CHIP subtypes were defined by the gene with largest variant allele fraction (VAF). In ASSESS-AKI and BioVU, targeted panel sequencing (high sensitivity for small clones) was used, and 'large CHIP' was defined as VAF ≥10%. Incident AKI outcomes were identified using ICD-9/10 codes (with additional CHS codes and chart review); severe AKI with dialysis (AKI-D) was defined in UKB by AKI plus a dialysis procedure code within 30 days. Individuals with baseline eGFR <15 ml/min/1.73 m², ESKD, or without baseline kidney function were excluded. Cox proportional hazards models adjusted for age, age², sex, baseline eGFR, smoking, diabetes, hypertension, and ancestry (PCs in UKB; self-reported ethnicity in TOPMed) estimated AKI risk; fixed- or random-effects meta-analyses combined results. Recovery analyses in ASSESS-AKI (n=321 with DNA) and BioVU (n=454) classified AKI as resolving vs nonresolving within 72 hours; logistic regression (adjusted for demographics, baseline creatinine, AKI stage, smoking, ethnicity, diabetes, hypertension, cardiovascular disease) estimated odds of nonresolving AKI. Time-to-event models assessed the composite of ESKD or ≥50% eGFR decline over 5 years in ASSESS-AKI. Mendelian randomization: Two-sample MR used genome-wide significant independent instruments for CHIP from large GWAS (excluding TERT due to pleiotropy) and AKI outcome summary statistics from ASSESS-AKI and BioVU. Primary analysis used multiplicative random-effects IVW with outlier checks; sensitivity analyses included weighted median, weighted mode, and MR-Egger. Mouse models: Non-DNMT3A CHIP mechanisms were probed using (1) Tet2-CHIP bone marrow chimeras (20% Tet2−/− CD45.2 with 80% WT CD45.1 into lethally irradiated recipients; controls received 20% WT CD45.2 + 80% WT CD45.1), and (2) hematopoietic-specific Jak2 V617F inducible mice. AKI was induced via unilateral ischemia–reperfusion injury (IRI) with contralateral nephrectomy (acute) or delayed nephrectomy (chronic), and via unilateral ureteral obstruction (UUO). Kidney injury and inflammation were assessed by BUN, serum creatinine, mRNA (qPCR), protein (immunoblotting, immunofluorescence/IHC), flow cytometry, histology (PAS, Masson trichrome, Picrosirius red), tubular injury scoring, and macrophage markers (CD68, F4/80) including NLRP3/IL-1β. scRNA-seq of CD45+ kidney cells (10x Genomics) with standard preprocessing (Cell Ranger, Seurat, SoupX, DoubletFinder, Harmony), cell type annotation via kidney atlas, DGE via DESeq2 on metacells, pathway analysis (clusterProfiler GSEA), and cell–cell communication (CellChat). Statistical analyses for animal studies used two-tailed t-tests or two-way ANOVA with post hoc tests; n≥4–7 mice per group as specified.

• In UKB, CHIP was associated with a 34% higher risk of incident AKI (HR 1.34, 95% CI 1.29–1.40, P<0.0001); stronger for AKI requiring dialysis (AKI-D: HR 1.65, 95% CI 1.24–2.20).
• Across UKB, ARIC, CHS meta-analysis: CHIP increased AKI risk (HR 1.26, 95% CI 1.19–1.34). Risk increased with clone size among CHIP carriers (per 10% VAF: HR 1.19, 95% CI 1.13–1.25 in UKB; overall meta-analysis 1.18, 95% CI 1.13–1.24).
• Non-DNMT3A-CHIP showed stronger association with AKI (meta-analysis HR 1.48, 95% CI 1.36–1.60); DNMT3A-CHIP was not associated (HR 1.03, 95% CI 0.94–1.12).
• Gene-specific risks (meta-analyses): TET2 HR 1.20 (1.03–1.40); ASXL1 HR 1.36 (1.17–1.59); PPM1D HR 1.67 (1.23–2.27); TP53 HR 2.13 (1.44–3.16); SRSF2 HR 2.60 (1.86–3.62); JAK2 HR 3.08 (2.08–4.56).
• Mendelian randomization: Genetically predicted CHIP risk associated with higher odds of AKI (meta-analysis OR 1.24, 95% CI 1.02–1.51), consistent across MR methods.
• Recovery after AKI: In ASSESS-AKI and BioVU, non-DNMT3A-CHIP and large clones (VAF ≥10%) were more common in nonresolving AKI and associated with higher odds of nonresolving AKI (meta-analysis OR 2.36, 95% CI 1.41–3.93 for non-DNMT3A-CHIP; OR 2.40, 95% CI 1.17–4.93 for large CHIP). In ASSESS-AKI, large CHIP was associated with increased risk of the composite kidney outcome (HR 2.9, 95% CI 1.1–8.0) over 5 years.
• Mouse models: Tet2-CHIP and Jak2 V617F mice exhibited more severe AKI (higher BUN/creatinine), increased tubular injury (KIM-1, NGAL), enhanced infiltration of proinflammatory macrophages (F4/80, CD68), elevated kidney inflammatory cytokines/chemokines (Tnf, Il1b, Il6, Ccl2, Ccl3), heightened NLRP3 inflammasome activation and IL-1β, and greater interstitial fibrosis (increased profibrotic gene expression and collagen deposition) following IRI and UUO. scRNA-seq indicated global inflammatory and profibrotic upregulation in Tet2−/− renal macrophages and altered macrophage–tubule signaling.

The study demonstrates that CHIP, particularly non-DNMT3A driver mutations (for example, TET2 and JAK2), increases the risk of incident AKI and impairs recovery from AKI. Associations persisted after adjustment for traditional risk factors and were consistent across diverse cohorts, with clone size exhibiting a dose–response relationship. MR findings support a causal role of CHIP in AKI. Mechanistic mouse models reveal that CHIP-mutant hematopoietic cells promote an exaggerated renal macrophage–mediated inflammatory response to kidney injury, leading to more severe acute injury and increased fibrosis during recovery. These results identify a somatic, age-related hematopoietic mechanism underlying AKI susceptibility and progression, distinct from inherited common variants previously studied, and highlight macrophage inflammasome pathways as potential therapeutic targets.

CHIP is a genetic risk factor for AKI and is associated with impaired recovery via enhanced renal macrophage inflammation, particularly for non-DNMT3A driver mutations such as TET2 and JAK2. Integrating epidemiologic analyses, MR, and mechanistic mouse models, the study suggests that targeting inflammatory pathways (for example, NLRP3 inflammasome or downstream mediators) could mitigate AKI risk and progression to ESKD in individuals with CHIP. Future work should evaluate CHIP’s impact across diverse AKI etiologies (e.g., glomerulonephritis, drug-induced AKI), assess interventions modulating macrophage inflammasome activity, and explore clinical screening strategies using CHIP status and clone size to stratify AKI risk and guide preventive care.

Prospective analyses could not address all AKI subtypes; intrarenal injury models (e.g., glomerulonephritis, drug-induced AKI) were not studied in mice. Residual confounding may persist due to collinearity between age, CHIP, and AKI despite adjustment and modeling of age effects. Differences in sequencing approaches (exome/genome versus targeted panels) may affect CHIP detection thresholds across cohorts.