Proteome-wide analysis links circulating proteins to kidney disease biology
Chronic kidney disease is not one disorder. Its clinical face shifts across diabetic nephropathy, glomerular disease, vascular injury, tubulointerstitial damage, and mixed systemic syndromes, and that heterogeneity blunts the value of broad biomarker lists that treat every patient as if the same pathways are active. This integrative analysis used a proteome-wide Mendelian randomization framework to move past correlation and ask a sharper question: which circulating proteins appear to drive kidney disease risk, and which are simply reflecting downstream injury?
Investigators instrumented cis-acting protein quantitative trait loci for 2,807 circulating proteins and tested those genetically proxied protein signals against detailed kidney function outcomes in 464,631 participants from the Million Veteran Program. The scale matters because kidney disease phenotypes are noisy in ordinary practice, especially when electronic health record data must separate chronic decline, acute injury patterns, and mixed clinical labels. By anchoring proteins to inherited variation rather than to a single blood draw, the analysis sought to identify causal protein drivers rather than passive biomarkers.
A clear hierarchy emerged. Ninety-three proteins reached proteome-wide significance for kidney outcomes, creating a prioritized set that can be followed into functional validation, therapeutic repurposing, and target discovery. The study did not stop at a generic CKD endpoint. It mapped proteins onto nine kidney disease subtypes defined through genome-wide association meta-analyses spanning the Million Veteran Program, UK Biobank, and FinnGen, then compared renal and extra-renal signals through phenome-wide association studies across 1,020 traits. A signal tied to one kidney subtype but not another can point to disease-specific biology instead of a universal renal stress response.
Subtype-specific signals matter more than a single CKD label
The subtype framework is the most clinically useful part of this work. CKD often gets handled as one umbrella diagnosis in practice, yet the mechanisms behind reduced estimated glomerular filtration rate are not interchangeable. A protein that tracks with one subtype may sit upstream of immune activation, endothelial injury, or fibrotic remodeling, while another may reflect compensatory responses that rise only after nephron loss has already occurred.
By aligning protein signals with nine genetically informed kidney disease subtypes, the analysis exposed where biology appears concentrated and where it fans outward into broader systemic effects. Some proteins showed associations confined to renal outcomes, while others also connected to extra-renal traits across the phenome. This separation is clinically important because it helps distinguish a protein with therapeutic promise for kidney disease from one that may carry pleiotropic effects across cardiovascular, metabolic, or inflammatory pathways. A treatment strategy built on the wrong protein target can miss the kidney entirely or produce off-target effects outside the renal system.
The study’s design also strengthens causal interpretation. Cis-acting protein quantitative trait loci reduce the risk of some forms of reverse causation and confounding that complicate conventional observational proteomics. In kidney disease, where inflammation, medication exposure, and organ dysfunction can all shift circulating proteins at once, that genetic tether is essential. A protein that remains associated with kidney outcomes through inherited regulation deserves a different level of attention than one that merely rises in parallel with disease severity.
From association to drug target nomination
The therapeutic implication is straightforward: this framework prioritizes proteins that may be druggable or repurposable, rather than leaving the field with a long list of associative biomarkers. The analysis explicitly maps candidate proteins to treatment hypotheses for drug development and repurposing, which is where causal proteomics can move from academic cataloging to intervention design. If a circulating protein is genetically linked to kidney injury, and if that same protein also sits within a biologically coherent subtype, it becomes a plausible entry point for mechanism-based therapy.
That translational logic is especially relevant in kidney medicine, where many patients are diagnosed late and where existing treatments do not fully address the diversity of pathogenic pathways. A protein target tied to a specific kidney disease subtype could help stratify patients for future trials, refine enrichment strategies, and reduce the dilution that often weakens CKD studies. A broad CKD label may obscure the very biology a precision therapy needs to hit.
The phenome-wide component adds another layer. By testing 1,020 traits, the study separated kidney-relevant associations from broader systemic biology. In practical terms, that means a target with a clean renal profile may be more attractive than one whose effects spill into many organs and disease domains. Conversely, a protein with both kidney and extra-renal relevance may represent a shared pathway of vascular or inflammatory injury, which could matter for patients whose CKD coexists with cardiometabolic disease.
What the 464,631-participant scale changes
Large biobank-scale data do more than inflate statistical power. They allow kidney subtypes, rare protein effects, and cross-phenotype patterns to be seen at a resolution that smaller studies cannot achieve. The Million Veteran Program cohort gave this analysis enough breadth to interrogate both common and clinically heterogeneous kidney outcomes while preserving the genetic structure needed for Mendelian randomization.
Still, the output is a prioritization map, not proof of clinical benefit. Proteome-wide significance does not automatically mean that inhibiting or augmenting a protein will improve renal outcomes in patients. Genetic evidence can nominate a target, but drug development still has to deal with pharmacology, tissue delivery, safety, and the possibility that a protein has different roles across disease stages. A causal signal is a strong starting point; it is not the finish line.
The most immediate value lies in target triage. Ninety-three proteins now sit at the top of a shortlist for mechanistic follow-up, subtype testing, and repurposing assessment. For a field burdened by heterogeneous disease definitions and overlapping systemic pathology, that is a meaningful narrowing of the search space. The next step is not simply to confirm that these proteins matter, but to determine which ones can be modulated safely enough to change the course of chronic kidney disease in specific biologic subtypes.
- • PubMed PMID 41282674: https://pubmed.ncbi.nlm.nih.gov/41282674/