Shared genetic architecture between inflammatory proteins and eGFR difference identifies chronic inflammation risk factors for kidney function impairment

Shared genetic architecture between inflammatory proteins and eGFR difference identifies chronic inflammation risk factors for kidney function impairment | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Shared genetic architecture between inflammatory proteins and eGFR difference identifies chronic inflammation risk factors for kidney function impairment Yifei LIN, Shu Su, Tingting Fu, Nanyan Xiang, Le Wang, Buu Truong, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8627406/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Chronic kidney disease(CKD) is linked to systemic inflammation, yet the shared genetic mechanisms remain unclear. Emerging evidence indicates that the difference between cystatin C- and creatinine-based eGFR serves as an early biological indicator capturing inflammatory, cardiometabolic, and mortality-related risk signals that precede subclinical kidney injury. Using large-scale cohorts (~390,000 UK Biobank; ~36,000 deCODE), we integrated genome-wide association studies across 13 kidney function traits and 839 inflammation-related proteins. Cross-trait genetic analyses and Mendelian randomization highlighted that the eGFR difference is highly heritable and shares genetic components with cardiometabolic signals. Multiple upstream inflammatory proteins, including ALPI, FABP9, INSR, and ABO, showed consistent protein-to-trait effects, corroborated by protein-level associations in 50,000 UK Biobank participants. Proteomic polygenic risk scores improved prediction of eGFR difference by 61% beyond demographic factors, achieved performance comparable to later-life clinical factors. Together, these findings provide the first genome-wide evidence delineating the genetic and inflammatory architecture of the eGFR difference, offering new biological insights for Cardiovascular-Kidney-Metabolic(CKM) risk stratification and transplant monitoring. Health sciences/Risk factors Health sciences/Diseases/Kidney diseases/Chronic kidney disease Health sciences/Biomarkers/Predictive markers Health sciences/Medical research/Epidemiology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Chronic kidney disease (CKD) affects over 13.4% of the global population 1 and is projected to become the fifth leading cause of death by 2040 2 . Early-stage kidney impairment is frequently asymptomatic; diagnosis is often delayed, thereby limiting opportunities for intervention. 3 . Although newly developed combined eGFR equations 4 have improved risk assessment, they remain grounded in physiological markers that reflect established damage. Increasing attention has turned to the difference between cystatin C- and creatinine-based (eGFRdiff). This metric has been shown across multiple studies to capture systemic inflammation, metabolic stress, and elevated cardiometabolic and mortality risks beyond what conventional kidney function markers can detect 5,6 . However, the genetic architecture of these markers and their association with chronic low-grade inflammation remains unknown 7,8 . Systemic chronic inflammation (SCI) is likely a key driver of this discordance and a central contributor to CKD pathogenesis 9 . Elevated levels of inflammatory cytokines, such as tumor necrosis factor and interleukin-6, are strongly associated with kidney function decline 10,11 and also closely linked to comorbid conditions, including cardiovascular-kidney-metabolic (CKM) syndrome, obesity, and malnutrition 12,13 . As a long-standing systemic process, SCI may shape kidney function trajectories long before clinical decline becomes evident 9 . However, most studies to date remain observational and cross-sectional, lacking insight into the genetics of these associations and limiting their translation to early-stage risk assessment 14 . Because genetic variants are fixed from birth, their effects on inflammatory proteins can proxy long-term inflammation and enable causal inference 15 . Recent advances in genome-wide association studies (GWAS) for kidney-related traits and large-scale proteomic profiling now enable population-scale investigations into how genetic variation shapes both protein expression and kidney function impairment 16,17 . Building on this property, their independence from environmental confounding and reverse causation make them powerful instrumental variables for causal inference through Mendelian randomization. Leveraging large-scale GWAS, polygenic risk scores (PRS) for circulating inflammatory proteins could be used to identify inflammation markers predictive of kidney functions and could be verify by the association between observed inflammatory protein levels and kidney functions 18 . Here, we use large-scale genomic and proteomic data from the UK Biobank (~390,000 participants) and deCODE Iceland (~36,000 participants) to address this gap ( Fig. 1 ). We conducted GWAS for 13 kidney function-related traits—including three updated eGFR equations 4,19 , four derived indicators, two metabolic waste markers, and four urinary ratios( Supplementary Table 1 )—and assessed their genetic correlations and causal relationships with 839 circulating inflammatory proteins, primarily from the Olink inflammation panel, the SomaScan platforms and manually curated markers across multiple international cohorts 20-24 . We verified the findings using observed protein levels and kidney functions in the UK Biobank. We then evaluated whether inflammatory protein-based PRS could improve the prediction of kidney function variation. Collectively, this work provides a comprehensive evaluation of the biological basis and translational relevance of the eGFR difference. Together, these analyses may bridge molecular inflammation and kidney physiology, providing an integrative framework to disentangle inflammation’s genetic architecture and its translational implications for subclinical kidney dysfunction. Results 1. Kidney function gene mapping We first analyzed 13 kidney function–related traits through GWAS in ~390,000 UK Biobank participants with European ancestry( Fig. 2A–D, Supplementary Table 1 ). These traits encompassed glomerular filtration indicators (eGFRcr, eGFRcys, eGFRcc), metabolic waste markers (urea, urate), urinary ratios (UACR, UK/UCR, UNA/UCR, UNA/UK), and eGFR-derived indices (eGFRdiff, eGFRdiff.mean, eGFRdiff.pct.cr and eGFRmean). The eGFR-derived indices capture the divergence between cystatin C- and creatinine-based estimates: eGFRdiff is their difference; eGFRdiff.mean is the normalized form obtained by dividing eGFRdiff by their eGFRmean(average); and eGFRdiff.pct.cr is the percentage form calculated by standardizing eGFRdiff to the creatinine-based estimate. ( Supplementary Table 1 ). Functional mapping and annotation (FUMA) 25 linked the identified loci to 110–3,160 genes and prioritized 30–612 candidate variants for each trait. The top loci for conventional eGFR traits (eGFRcys, eGFRcr, eGFRcc) corresponded to well-established genes, including CST2, SPATA5L1 , and SHROOM3 , all previously reported in the GWAS Catalog(https://www.ebi.ac.uk/gwas/) 26 , confirming the validity of our results. Metabolic and urinary traits, such as urate and UK/UCR, mapped to fewer loci but showed associations with known metabolic genes such as CPS1 and SLC2A9 . Among the derived eGFR difference traits, eGFRdiff, eGFRdiff.mean, and eGFRdiff.pct.cr consistently mapped to the CSTL1 /CST1 loci, whereas eGFRmean was associated with CST4 , highlighting the contribution of the cystatin gene family to kidney function variation. Further cross-referencing with GWAS Catalog revealed additional overlaps between kidney function–related loci and a range of complex traits, including cardiometabolic and inflammatory phenotypes. Supplementary Table 2 summarizes overlapped lead SNPs separately by different traits. Specifically, even after excluding SNPs previously reported for kidney function traits, the remaining top loci for eGFRdiff, eGFRdiff.mean, and eGFRdiff.pct.cr predominantly mapped to cardiovascular and metabolic traits (9 out of 10) and inflammatory traits, such as platelet count. The derived eGFRmean trait also showed similar annotation, with 8 variants related to cardiometabolic traits and 2 linked to inflammatory pathways. Collectively, these findings indicated both shared and trait-specific genetic architectures connecting kidney function with metabolic, cardiovascular, and inflammatory pathways. 2. Heritability and Genetic Correlations We estimated SNP-based heritability (h²SNP) using linkage disequilibrium score regression (LDSC). Heritability estimates were obtained for all 13 studied traits( Fig. 2 E, Supplementary Table 1 ). Overall, we observed moderate to high heritability across kidney-related traits, with the highest estimates for eGFRcc (h² = 0.342, SE = 0.039) and lowest estimates for UACR(h² = 0.052, SE = 0.003). We estimated genome-wide genetic correlation between 893 inflammation-related plasma proteins (Supplementary Table 3) and 13 kidney function traits using cross-trait LDSC. Through bi-clustering analysis using genetic correlation estimate(Rg), we identified six distinct inflammation clusters with coherent and trait-specific patterns of shared genetic architecture. (Fig.3A-B, Supplementary Table 4-5) . Cluster 1, consisting of three proteins, showed uniformly strong correlations with kidney filtration traits, led by RNASE1, which demonstrated the strongest negative genetic correlation with eGFRcys (Rg = –1, FDR = 0.001) and a strong positive correlation with urate (Rg = 0.526, FDR = 0.001), while the RNASE1–eGFRcr pair represented the most statistically significant association within this cluster (Rg = –0.684, FDR = 5.78× 10⁻ 4 ). Cluster 2, comprising 216 proteins, was characterized by broad negative correlations with creatinine-based kidney function, with the PRELP–eGFRcr association showing the highest significance ( Rg = –0.469, FDR = 2.07 × 10⁻³⁰). Cluster 3, containing 253 proteins, displayed preferential overlap with urate-related pathways, highlighted by ADGRD1–urate ( Rg = 0.293, FDR = 4.92 × 10⁻¹³). Cluster 4, composed of 142 proteins, showed enriched associations with creatinine-derived filtration measures, exemplified by CD58–eGFRcr ( Rg = –0.287, FDR = 4.27 × 10⁻¹⁶). Cluster 5, consisting of 62 proteins, was enriched for correlations with cystatin-C–based kidney function, with ENPP5–eGFRcys emerging as the strongest signal (Rg = 0.236, FDR = 2.16 × 10⁻¹⁵). Cluster 6, comprising 79 proteins, demonstrated broad negative correlations across multiple kidney traits, with IL18BP–eGFRmean showing the most pronounced overall association (Rg = –0.447, FDR = 8.34 × 10⁻⁴⁴). Together, these clusters reveal heterogeneous yet structured axes of genetic overlap between systemic inflammation and kidney function and nominate biologically coherent protein groups for downstream causal investigation. 3. Candidate causal relationships between inflammation-related proteins and kidney function Across 839 inflammatory proteins and 13 kidney function traits, we identified consistent significant protein-trait associations supported by multiple MR methods (≥3 of 5 methods, FDR < 0.05, Fig. 4 ， Supplementary Table 6-9 ). Specifically, one basic eGFR trait(eGFRcys), 3 derived eGFR traits(eGFRdiff, eGFRdiff.mean, and eGFRdiff.pct.cr), all metabolic waste traits(urea, urate), and one urinary ratio trait(UK/UCR) showed significant protein-to-trait effects driven by inflammatory proteins, without observable reverse effects from these kidney function traits to proteins. No significant protein-to-trait or trait-to-protein associations were found for UNA/UCR, UACR, and UNA/UK. The predominant direction of causality was therefore from inflammation to kidney function, underscoring inflammation as a likely upstream driver of renal variation. For basic eGFR traits, eGFRcys demonstrated exclusive protein-to-trait effects, implicating seven significant proteins (ACP1, CXCL10, ECM1, FGF5, MGMT, MST1, and RALB), involved in immune regulation, cell signaling, or tissue remodeling. However, no evidence supported the reverse direction from eGFRcys to proteins. In contrast, eGFRcr exhibited bidirectional relationships, identifying 11 protein-to- eGFRcr and 52 eGFRcr -to-protein effects. The combined index eGFRcc showed partial overlap, identifying 12 protein-to-eGFRcc and 15 eGFRcc-to-protein effects. Among them, IL12RB1 (immune signaling) was unique to eGFRcr, while CASP9 (apoptosis regulation) was unique to eGFRcc. Among the derived eGFR traits, eGFRdiff, eGFRdiff.mean, eGFRdiff.pct.cr, and eGFRmean exhibited consistent significant protein-to-trait effects, each associated with 7–12 significant proteins. (all FDR of at least 4 MR methods <0.05). These traits showed substantial overlap across eGFRdiff, eGFRdiff.mean, and eGFRdiff.pct.cr traits, sharing a core set of proteins—including ALPI, FABP9, PNLIPRP2, SERPINA1, APOC1, IL31, and PDGFD—implicated in lipid metabolism, inflammation, and tissue remodeling. Within the eGFRmean trait, MGMT and ALPI both showed significant associations across all five MR methods(all FDR <0.05), reinforcing their robustness as potential upstream contributors of kidney function variation. Metabolic waste markers revealed distinct signatures: Urea demonstrated protein-to-urea effects for 11 proteins (e.g., ABO, ACP1, CXADR, FGF19, HEG1, INSR), whereas urate was associated with four (e.g., FGF5, SCGB3A1), with no overlap between the two kidney function traits. Consistent results across all five MR methods identified HEG1 as unique to urea, while a subset of proteins, such as CXADR, LGALS4, INSR, TBCA, ACP1, THSD1, S100A13, and ABO, showed cross-trait relevance to both urea and both basic and derived eGFR indicators. FGF5, also shared by urate and eGFR indicators, demonstrated robust associations across all MR models. For urinary traits, eight proteins (e.g., ABO, CD34, CFHR4, CXADR, EPHA4, INSR, LGALS4, RELB) showed significant protein-to-trait associations with UK/UCR, reflecting biological roles in vascular regulation, immune signaling, and metabolic balance. Four of these proteins (RELB, CFHR4, CD34, EPHA4) were unique to UK/UCR. No significant causal proteins were identified for UACR or UNA/UK based on at least 4 MR methods. In the reverse direction, a limited set of trait-to-protein effects was observed, primarily involving eGFRcr, eGFRcc, and eGFRmean. eGFRcr was associated with 52 downstream proteins, while eGFRcc and eGFRmean were linked to 15 and 12, respectively. Among these, RNASE4 was the most significant shared downstream protein for both eGFRcc and eGFRmean (mode-based estimate (MBE)MR: β= from –0.59 to –0.62, FDR < 0.05) and RNASE1 (MBE: β = –0.615, 95% CI: –0.951 to –0.278, FDR < 0.05) showed a unique association to eGFRcc, suggesting RNA metabolism and immune regulation as possible pathways influenced by kidney dysfunction. Two additional proteins—BTN3A2 and CLEC7A— displayed consistent trait-to-protein associations within the eGFRmean, indicating potential roles in immune signaling and kidney-inflammatory feedback. 4. Phenotypic corroboration of MR-inferred relationships We corroborated the MR-inferred relationships of proteins and traits through cross-sectional phenotypic analyses stratified by CKD status ( Supplementary Table 10 ), using linear regression adjusted for age and sex ( Fig. 5 ). Among the significant protein–to-trait direction pairs( Supplementary Table 11 ), 16 pairs showed consistent associations in both CKD and non-CKD groups, while three pairs (TBCA-eGFRcr, FGF5-Urate, SCGB3A1-Urate) exhibited opposite correlation directions across the two strata(all |β| >5, all FDR < 0.05). One pair (ABO-Urea) was significant only in the CKD group (β = 0.219, FDR = 0.021), whereas 19 pairs were uniquely significant in the non-CKD group, including a strong association between NCF2 and urate (β = 4.052, FDR = 3.20 × 10⁻³⁴). For trait-to-protein directions( Supplementary Table 12) , all tested pairs were significant in both strata, with effect sizes consistently larger in the CKD group. Among them, the well-known eGFRcr-NPHS1(nephrosis protein 1) 27 pair demonstrated the strongest positive association(β= 19.243, FDR=2.74E-31), whereas the eGFRcc-RNASE4 pair showed the strongest negative association (β= -37.63, FDR=3.15E-134). 5. Prediction of inflammation proteomics for kidney function using PRS To evaluate how long-term inflammation status contributes to kidney function variation, we applied LDpred2-auto to compute polygenic risk scores (PRS) 28 for each of 839 inflammatory proteins as genetic proxies for chronic inflammation, using genome-wide SNP weights derived from meta-analytic summary statistics of proteomic GWAS 28 . Each protein PRS represents the inherited tendency for elevated circulating levels of that protein, serving as a proxy for lifelong systemic inflammation. These protein PRSs were jointly modeled using elastic net regression with 10-fold cross-validation to identify key inflammatory markers and quantify their combined predictive value ( Fig. 6, Supplementary Table 13-14 ). Two models were constructed: Model 1 adjusted for age, sex, and genetic principal components, while Model 2 additionally included clinical covariates (hypertension, diabetes, high cholesterol, and BMI). The inclusion of PRS improved prediction across most kidney function traits. For derived eGFR traits, predictive correlations increased by 0.036 to 0.118, while conventional eGFR traits showed smaller gains (0.027 to 0.061). Interestingly, for the derived eGFR traits, the gain in predictive performance from incorporating proteomic PRS in Model 1 was of similar magnitude to the improvement achieved by adding clinical covariates in Model 2, indicating that proteomic genetic profiles can provide predictive information of a similar scale to traditional risk factors. More modest improvements( Supplementary Figure 1 ) were observed for metabolic waste markers (0.010 to 0.017) and urinary biomarkers (0 to 0.006). Notably, for eGFRdiff, proteomic PRS improved predictive performance by 61% beyond demographic factors alone(0.193 to 0.311, Δ=0.118, +61%), achieving an effect size comparable to that of established later-life clinical risk factors. Moreover, when added on top of these clinical variables, proteomic PRS provided a further 24% improvement in prediction(0.317 to 0.391, Δ=0.075，+24%), indicating substantial incremental value beyond traditional risk stratification. Discussion 1. Summary of Main Findings Leveraging data from large prospective cohorts, we systematically integrated genome-wide, proteomic, and genetic risk modeling approaches to dissect the causal link between systemic inflammation and kidney function. Using LDSC and MR, we established shared genetic architecture and directional effects between 13 kidney function traits and 839 inflammation-related proteins, indicating a predominant causal influence from inflammation to kidney function variation. Among derived eGFR traits, particularly the eGFR difference (eGFRdiff) and its normalized forms( eGFRdiff.mean and eGFRdiff.pct.cr), we identified consistent protein-to-trait associations involving a core set of inflammatory and metabolic proteins (e.g., ALPI, FABP9, SERPINA1, APOC1, and PDGFD). This suggests these indices are sensitive captures of subclinical, inflammation-related kidney function impairment. Additional associations with metabolic (urea, urate) and urinary (UK/UCR) markers implicated pathways of vascular regulation, lipid metabolism, and RNA processing (e.g., RNASE family). Finally, by modeling PRS of inflammatory proteins as proxies for lifelong exposure to the inflammation, we demonstrated that genetically inferred inflammation profiles significantly explain variation in kidney function. This was particularly evident for eGFRdiff-related traits, highlighting their translational value for risk stratification and monitoring of Cardiovascular-Kidney-Metabolic(CKM) health. 2. From Clinical Observation to Genetic Mechanism: The eGFR Difference Previous studies have demonstrated that the directionality of eGFRdiff is clinically meaningful. For instance, a negative eGFRdiff (eGFRcys eGFRcr) correlates with more favorable health status and reduced risks of adverse events, such as falls, hospitalization, and mortality 30 . Therefore, our study provides the first genome-wide evidence elucidating the genetic architecture underlying this discrepancy, building upon these clinical observations. We found that eGFRdiff and its normalized derivatives possess significant SNP-based heritability ( h 2 = 0.24), exceeding that of conventional biomarkers, i.e., eGFRcr. This indicates that eGFRdiff captures a broader range of genetically driven variation in renal function, reflecting non-GFR determinants such as muscle mass, nutritional status, and systemic inflammation, thereby offering complementary value for risk stratification. By integrating inflammatory proteomic PRS, we further improved the predictive performance of eGFRdiff, achieving model correlations comparable to clinical risk factors. These findings indicate that eGFRdiff is not merely a marker of glomerular filtration discordance but a genetically informed indicator of systemic inflammatory states. Compared with single-marker approaches, integrating eGFRdiff with inflammation-related proteomics enables earlier recognition of subclinical dysfunction, even among individuals with preserved global renal function. Across eGFRdiff, eGFRdiff.mean, eGFRdiff.pct.cr traits, a consistent inflammation-related protein signature—comprising ALPI, FABP9, PNLIPRP2, SERPINA1, APOC1, IL31, and PDGFD—was identified. These proteins demonstrated robust protein-to-trait associations across multiple MR models, suggesting upstream involvement in inflammation-mediated renal variation. Among them, SERPINA1, PDGFD, IL31, and APOC1 have been previously implicated in various renal pathologies, including diabetic kidney disease, glomerulosclerosis, interstitial fibrosis, CKD-associated pruritus, and IgA nephropathy. Their consistent associations across multiple derived eGFR traits highlight their biological and clinical relevance, indicating these proteins as mechanistic indicators of inflammation-related kidney dysfunction and potential exploratory therapeutic targets 31,32 . In contrast, ALPI, FABP9, and PNLIPRP2 have not yet been directly reported to link to kidney dysfunction previously. Specifically, ALPI is known to as a central player in microbial homeostasis 33 , suggesting a potential role in the gut-kidney axis. FABP9, a fatty acid-binding protein, and PNLIPRP2, a pancreatic lipase-related protein related to coronary atherosclerosis 34,35 , might reflect the potential role of inflammatory metabolic states or metainflammation 36 in kidney function decline. Collectively, these novel associations highlight the promise of these proteins as predictive markers for early-stage kidney impairment. 3. Novel Classification of Kidney Function Based on Inflammatory Proteins by Causal Direction To refine the biological understanding of kidney traits, we propose a novel classification framework grounded in the causal directionality between inflammatory proteins and kidney function. This causal framework is conceptually aligned with the LDSC-based bi-clustering results, which reveal coherent groups of inflammatory proteins sharing consistent genetic relationships with specific kidney traits. Specifically, the 13 kidney-related traits can be categorized into two groups: Inflammation-driven traits , including eGFRdiff, eGFRcys, urea, urate, and UK/UCR, exhibited protein-to-trait causal associations with inflammatory proteins without corresponding reverse effects. Inflammation-interactive traits , including eGFRcr, eGFRcc, and eGFRmean, showed both protein-to-trait and trait-to-protein associations, indicating complex, feedback-driven interactions between systemic inflammation and kidney function. 3.1 Inflammation-Driven Traits Provide Insights into the Cardio-Kidney-Metabolic (CKM) Axis Inflammation-driven traits offer enhanced predictive value by capturing subclinical renal changes not detected through conventional biomarkers, highlighting their utility as complementary tools for clinical risk assessment. Consistent with prior evidence, eGFRcys demonstrated a strong protein-to-trait association pattern with upstream inflammatory drivers. Extending this framework, our analysis identified additional inflammation-driven kidney traits, indicating broader involvement of immune and metabolic pathways in renal regulation. For blood-based traits , GWAS Catalog cross-referencing revealed substantial genetic overlaps between derived eGFR metrics (e.g., eGFRdiff, eGFRdiff.eGFRmean and eGFRdiff.pct.cr) and known cardiovascular, metabolic, and inflammatory traits, as discussed above. This shared genetic architecture underscores the interconnected biological basis of these conditions, reinforcing the clinical relevance of derived kidney metrics for early CKM risk prediction and stratification. For urinary traits , particularly UK/UCR, although not directly associated with conventional kidney function measures, we identified four inflammatory proteins(ABO, CXADR, LGALS4, INSR ) showing consistent protein-to-trait associations shared with established kidney function indicators such as urea, eGFRcc and eGFRcr. Among these, INSR, a critical regulator of insulin signaling and metabolic homeostasis, highlights UK/UCR’s potential as an intermediary biomarker within cardio-kidney-metabolic(CKM) syndrome. Additionally, CD34, a marker of endothelial activation, exhibited a unique protein-to-trait association with UK/UCR and previously linked to acute kidney injury 37 , diabetes and diabetic nephropathy 38 , further supports its role in capturing early endothelial and immune perturbations along the CKM axis. Given its non-invasive nature and ease of measurement, UK/UCR represents a promising alternative to current CKM diagnostic frameworks, which often rely on heterogeneous clinical criteria. 3.2 Inflammation-Interactive traits reflect both kidney injury and potential kidney-brain axis involvement Among inflammation-interactive traits, the RNASE family stood out for its strong and consistent trait-to-protein associations. The eGFRcc-RNASE4 association showed the most robust negative correlation at the protein-level, while LDSC genetic correlations also supported RNASE1’s link to eGFRcc and RNASE4’s association with eGFRcys. Beyond their established involvements in kidney impairment 39,40 , RNASE family proteins have been implicated in protecting against neuronal degeneration and maintaining blood-brain barrier integrity, suggesting a molecular basis for the kidney-brain axis 41,42 . Moreover, the eGFRcr-NPHS1 pair demonstrated the strongest positive trait-to-protein association. NPHS1 (Nephrosis Protein 1), a well-established protein in nephrotic syndromes, has also been implicated in neurodevelopmental processes 27 , supporting the biological interplay between kidney impairment and neurological outcomes such as uremic encephalopathy. Furthermore, the eGFRmean trait revealed 2 uniquely associated trait-to-protein associations , CLEC7A and BTN3A2, that were absent in basic eGFR traits. Both have been previously linked to kidney injury 43,44 , and emerging evidence also associates them with neurodegenerative disease 45,46 . In general, these findings highlight that inflammation-interactive traits may capture interconnected molecular signatures bridging renal and neural inflammation, potentially reflecting early kidney–brain crosstalk. 4. Clinical Implications for Kidney Transplantation Beyond genetic susceptibility, trait-to-protein associations reveal how kidney function itself modulates systemic immune signaling. Among these, the ABO protein emerged as the most broadly connected upstream factor in our Mendelian randomization analysis, showing consistent protein-to-trait associations with multiple kidney function indicators and involvement in up to eight mechanistic associations. Beyond its established role in blood group compatibility, ABO variation was found to influence post-transplant outcomes through glycosylation-dependent and immune-mediated mechanisms 47-49 . Our findings extend the significance of ABO beyond immunological compatibility, suggesting that ABO polymorphisms may also influence post-transplant kidney function through glycosylation-related and inflammatory mechanisms. Similarly, several trait-to-protein associations, including HLA-E, LGALS9, CD200, and CD7, suggested that impaired filtration might trigger compensatory immunoregulatory responses. Clinically, these proteins could serve as dynamic biomarkers for transplant monitoring and individualized immunosuppression, bridging molecular mechanisms with patient-level management. 5. Opposite Associations by CKD Status Reveal State-Dependent Inflammatory Regulation In the protein-level validation analysis, the genetic causal proteins TBCA, FGF5, and SCGB3A1 exhibited opposite-level correlation trends with eGFRcr and urate between the CKD and non-CKD populations, suggesting that they may play state-dependent regulatory roles under different kidney function conditions. Specifically, TBCA was negatively associated with eGFRcr in CKD but positively in non-CKD individuals. Although current studies mainly focus on the role of TBCA in kidney cancer 50 , its function in non-tumor kidney diseases remains underexplored and warrants further investigation. Similarly, the opposite associations of FGF5 and SCGB3A1 with urate levels in different populations also suggest that they may be involved in the dynamic network regulating kidney function. Notably, FGF5 emerged as a shared causal factor across eGFR-based, derived eGFR, and metabolic waste traits, while FGF19 was uniquely associated with urea. Both have been previously implicated in the progression of kidney dysfunction 43,51 , and belong to the fibroblast growth factor (FGF) family, which also includes FGF23, a known mediator of cardiac remodeling in CKD 52 . These findings also highlight a broader role of the FGF signaling in kidney function decline and its potential relevance to the CKM syndrome. 7. Strength and Limitation The main strength of this study is its systematic integration of genomic, proteomic, and causal inference methods to investigate early kidney function decline, enabling both biological discovery and translational insight. First, we present the first genome-wide association analysis of the eGFRcys–eGFRcr differential marker (eGFRdiff), uncovering novel inflammation-related genetic signals that may reflect early kidney dysfunction. Second, by integrating protein-level associations with genetic risk, we identified several inflammation-related proteins (e.g., ABO, RNASE family) with potential causal roles in kidney decline, offering promising targets for biomarker development and therapeutic intervention. Third, we propose UK/UCR as a sensitive and non-invasive marker to evaluate subclinical kidney stress, particularly relevant for early-stage detection in high-risk populations. Fourth, we pioneer the application of proteomics-informed PRS to improve prediction of kidney outcomes, enhancing risk stratification for early nephroprotective intervention. Together, these contributions provide a framework for future translational research in personalized nephrology, bridging genomic discovery, protein validation, and clinical applicability. The limitations of this study should be acknowledged. First, although the genetic prediction model was developed using individuals of European ancestry to enhance precision under a homogeneous genetic background, this may limit the generalizability of our findings to other populations. Future research should extend this work to multi-ancestry populations to improve equity and applicability. Second, external validation and experimental confirmation were not available. Nonetheless, stratified analyses by CKD status and consistent associations across complementary kidney biomarkers (e.g., blood- and urine-based indicators) helped support the robustness of our findings. Third, due to the asymptomatic nature of early CKD and the limited number of clinically confirmed cases, we relied on continuous kidney function measures rather than binary disease outcomes. This approach improved statistical power and enabled cross-marker validation. Fourth, our proteomic coverage did not include certain kidney injury markers such as KIM-1, which are part of the Olink Oncology panel rather than the inflammation panel. However, multiple inflammation-related proteins identified here converge with recently reported kidney injury pathways 53 , highlighting potential biological convergence. Fifth, we did not account for the use of SGLT2 inhibitors, which may influence biomarker levels and disease progression. Even so, the inclusion of genetic information substantially improved the prediction of renal traits, suggesting that the observed associations capture underlying disease susceptibility beyond clinical treatment effects. In conclusion, our study uncovered shared genetic mechanisms and bidirectional causality between kidney function and systemic inflammation, offering insights for targeted interventions and personalized prevention strategies. Findings underscore the potential of proteomic signatures in mitigating CKD progression, though validation in diverse populations remains critical. Methods Study Population Our study leveraged two large-scale European population-based datasets to investigate the genetic relationships between kidney function and chronic inflammation. The UK Biobank (UKB) cohort comprised approximately 390,000 participants of European ancestry recruited between 2006-2010, while the deCODE Genetics study included 35,892 Icelandic individuals with biomarker measurements collected from 2000-2019. Both studies obtained appropriate ethical approvals, and participants provided informed consent. Genetic data underwent stringent quality control procedures, ensuring robust genetic analyses. This study was carried out under UKBiobank project 45052. Exposures assessment of inflammation traits The primary exposure of interest is the comprehensive assessment of inflammatory biomarkers and genetic data for participants. We have collected GWAS summary data with inflammatory protein biomarkers and genetics from multiple sources, which include 839 chronic inflammatory biomarkers curated from both Olink and SomaScan platforms, selected based on their established roles in inflammatory pathways in multiple international cohorts 20-24 and availability across multiple cohorts to ensure sufficient statistical power according to the study of deCODE and UK Biobank 54 . Outcome assessment of kidney function traits To capture a multidimensional profile of renal function, we evaluated 13 quantitative kidney-related traits from the UK Biobank, spanning both blood- and urine-based biomarkers. These traits were categorized into four functional groups: estimated glomerular filtration rate (eGFR) indicators, metabolic waste markers, urinary ratio measures, and eGFR-derived indices. Together, they provide a comprehensive view of renal filtration, metabolic clearance, electrolyte handling, and biomarker discordance. First, we included three core eGFR indicators using the latest equations. eGFR based on serum creatinine (eGFRcr) was calculated using the 2021 CKD-EPI creatinine-based equation, while eGFR based on cystatin C (eGFRcys) followed the 2012 CKD-EPI cystatin C equation. A combined eGFR measure (eGFRcc) that incorporates both creatinine and cystatin C was also derived using the 2021 CKD-EPI combined formula. These markers reflect overall glomerular filtration capacity from complementary biochemical sources. Second, to assess renal clearance of metabolic waste, we evaluated serum concentrations of urea and urate. As end-products of protein and purine metabolism, respectively, both biomarkers are filtered by the kidneys and serve as sensitive indicators of renal excretory function. Third, we included four urinary ratio markers reflecting electrolyte and protein excretion. These comprised the urinary potassium-to-creatinine ratio (UK/UCR), sodium-to-creatinine ratio (UNA/UCR), sodium-to-potassium ratio (UNA/UK), and the urinary albumin-to-creatinine ratio (UACR). All ratios were standardized to urinary creatinine to account for variation in urine concentration. UACR, in particular, is widely used to detect early glomerular injury and predict CKD progression. Lastly, we derived four composite traits to quantify the difference between cystatin C– and creatinine–based eGFR estimates. The difference (eGFRdiff) was calculated as eGFRcys minus eGFRcr. The percent difference (eGFRdiff.mean.cr) was defined as eGFRdiff divided by eGFRcr and multiplied by 100. A normalized difference (eGFRdiff.mean) was calculated as eGFRdiff divided by the mean of the two values, with the eGFRmean itself defined as (eGFRcr + eGFRcys)/2. These derived traits have been proposed as sensitive markers of early functional deviation and discordant filtration behavior. All kidney function traits were measured under standardized protocols and subjected to rigorous quality control as part of the UK Biobank infrastructure. Covariates We considered a structured set of covariates tailored to each analysis stage to control for potential confounding and improve model precision. For genome-wide association analyses (GWAS), we adjusted for age, sex, age squared (AGE²), assessment center, genotyping array, and the first 20 genetic principal components (PCs) to account for population stratification. For polygenic risk score (PRS) analyses, we applied two covariate models: a baseline model including 24 variables—age, sex, assessment center, genotyping array, and the first 20 PCs—and an extended model with 28 variables that further incorporated body mass index (BMI), hypertension status, diabetes status, and high cholesterol as additional clinical risk factors. For cross-sectional protein-level validation analyses, only age and sex were included as covariates. All covariates were obtained through standardized procedures at the time of recruitment and were harmonized across participants to ensure data consistency and comparability. Statistical Analysis We employed a multi-analytical framework integrating cutting-edge genetic epidemiological approaches to examine the genetic relationships between inflammatory biomarkers and kidney function. We performed GWASs for all the kidney function traits using a linear mixed model (LMM) implemented in BOLT-LMM v2.3.4 55 to account for cryptic relatedness and potential population stratification. Based on European ancestry, the analysis adjusted for age, age squared, sex, genotyping array, assessment center and 20 ancestry principal components to assess the association between the inverse normally transformed phenotype residuals and imputed genotype dosages. Using the public released proteomics GWAS summary data from UK Biobank and deCODE 54 , we conducted meta-analyses for the same circulating protein biomarkers using METAL with the inverse-variance-weighted method. We performed linkage disequilibrium score regression (LDSC) to estimate genetic correlations between traits using GWAS summary statistics. This method leverages patterns of linkage disequilibrium across the genome to distinguish true polygenic signals from confounding factors such as population stratification. Genetic correlation coefficients (Rg) were considered significant at Rg > 0.1 with p < 0.05 after false discovery rate (FDR) correction. Following the initial LDSC analysis, we conducted hierarchical bi-clustering performed in the R environment using Euclidean distance with average linkage for both inflammatory proteomics and kidney function biomarkers. Also, we generated heatmaps to visualize patterns of shared genetic architecture among the biomarkers, identifying potential biological modules and pathways underlying the observed correlations. For causal inference, we implemented bidirectional two-sample Mendelian randomization (MR) analyses using genetic variants as instrumental variables. This approach mimics randomized controlled trials through Mendel's second law of inheritance, effectively addressing confounding present in observational studies. To ensure robust causal estimates, we employed five complementary MR methods: mode-based estimate (MBE) as primary analysis, supplemented by weighted median, MR-Egger, robust adjusted profile score (RAPS), and inverse-variance weighted (IVW) approaches. To visualize the complex web of causal relationships, we constructed causal networks using Cytoscape (v3.9.1), where nodes represented traits and edges represented significant causal associations (False Discovery Rate, FDR < 0.05 in at least three MR methods). This network analysis helped identify hub biomarkers with particularly influential roles in the kidney-inflammation axis. To corroborate the MR-inferred protein–trait relationships at the observational level, we performed cross-sectional association analyses in the UK Biobank. Protein levels were regressed on corresponding kidney function traits using linear regression models stratified by CKD status and adjusted for age and sex. Standardized β coefficients were estimated to quantify the direction and magnitude of associations. This analysis aimed to evaluate concordance between genetically inferred and phenotypic relationships rather than to validate causal effects. Finally, to assess how long-term inflammation contributes to kidney function variation, we developed protein-based polygenic risk scores (PRS) as genetic proxies for chronic inflammation. We used LDpred2-auto 28 to estimate posterior SNP effect sizes based on genome-wide association summary statistics from a meta-analysis of 839 inflammatory proteins, incorporating linkage disequilibrium (LD) information from the 1000 Genomes Project European reference panel. LDpred2 models LD structure to infer posterior SNP effect sizes under a point–normal mixture prior, enabling genome-wide PRS construction without requiring external tuning parameters. Individual-level PRSs were subsequently calculated by aggregating weighted SNP effects across the genome. To account for correlations among protein PRSs and select informative markers, we applied elastic net regression with 10-fold cross-validation, which jointly modeled their contribution to kidney function outcomes while optimizing regularization parameters. Model 1 adjusted for age, sex, and genetic principal components(demographic factors), whereas Model 2 additionally included hypertension, diabetes, high cholesterol, and BMI(traditional clinical factors). Predictive performance was evaluated using cross-validated correlations between observed and predicted trait values (ΔR reported where applicable). All statistical analyses were conducted using R (v4.0.3). All analyses were adjusted for multiple comparisons using the False Discovery Rate (FDR) method to control for type I errors and to enhance the reliability of the findings. Declarations Data availability The individual-level genotype and phenotype data used in this study are available through controlled-access application procedures. UK Biobank data are available to bona fide researchers upon approval of a data access application via the UK Biobank Access Management System (https://www.ukbiobank.ac.uk), under project number 45052. Summary-level genome-wide association statistics for inflammatory protein biomarkers were obtained from publicly released proteomics GWAS datasets generated using the Olink and SomaScan platforms, including data from UK Biobank and deCODE Genetics. Access to deCODE summary statistics is subject to the data use policies of deCODE Genetics. All analyses were conducted in accordance with the relevant ethical approvals and data access agreements. Code availability Analyses were performed using a combination of publicly available software and custom scripts. Genome-wide association analyses were conducted using PLINK (v1.9) and BOLT-LMM (v2.3.4). ( https://alkesgroup.broadinstitute.org/BOLT-LMM/BOLT-LMM_manual.html). SNP-based heritability and genetic correlations were estimated using LD Score Regression (LDSC) (https://github.com/bulik/ldsc). Mendelian randomization analyses were performed using the MendelianRandomization and MR-RAPS R packages. Proteomic polygenic risk scores were constructed using LDpred2-auto, which estimates posterior SNP effect sizes under a point–normal mixture prior while accounting for linkage disequilibrium, using European ancestry reference panels from the 1000 Genomes Project. Elastic net regression models were fitted using the glmnet R package. Custom scripts used for data preprocessing, quality control, and downstream analyses are available from the corresponding authors upon reasonable request, subject to institutional data governance and participant privacy regulations. Acknowledgements This work was conducted during the author’s academic practicum training at the Harvard T.H. Chan School of Public Health. This work was supported by the National Natural Science Foundation of China (32471519 and 32571690) and 1.3.5 project for disciplines of excellence from West China Hospital of Sichuan University (ZYGD23039). References Lv, J.C. & Zhang, L.X. Prevalence and Disease Burden of Chronic Kidney Disease. Advances in experimental medicine and biology 1165 , 3-15 (2019). Global, regional, and national burden of chronic kidney disease, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet (London, England) 395 , 709-733 (2020). Webster, A.C., Nagler, E.V., Morton, R.L. & Masson, P. Chronic Kidney Disease. Lancet (London, England) 389 , 1238-1252 (2017). Inker, L.A. , et al. New creatinine-and cystatin C–based equations to estimate GFR without race. New England Journal of Medicine 385 , 1737-1749 (2021). Chen, D.C. , et al. Association of intraindividual difference in estimated glomerular filtration rate by creatinine vs cystatin C and end-stage kidney disease and mortality. JAMA Network Open 5 , e2148940-e2148940 (2022). Fu, E.L. , et al. Accuracy of GFR Estimating Equations in Patients with Discordances between Creatinine and Cystatin C-Based Estimations. 34 , 1241-1251 (2023). Tangri, N. , et al. Multinational Assessment of Accuracy of Equations for Predicting Risk of Kidney Failure: A Meta-analysis. Jama 315 , 164-174 (2016). Wuttke, M. , et al. A catalog of genetic loci associated with kidney function from analyses of a million individuals. Nat Genet 51 , 957-972 (2019). Furman, D. , et al. Chronic inflammation in the etiology of disease across the life span. Nat Med 25 , 1822-1832 (2019). Amdur, R.L. , et al. Inflammation and Progression of CKD: The CRIC Study. Clinical journal of the American Society of Nephrology : CJASN 11 , 1546-1556 (2016). Kadatane, S.P., Satariano, M., Massey, M., Mongan, K. & Raina, R. The Role of Inflammation in CKD. Cells 12 (2023). Stenvinkel, P. , et al. Chronic Inflammation in Chronic Kidney Disease Progression: Role of Nrf2. Kidney international reports 6 , 1775-1787 (2021). Romagnani, P. , et al. Chronic kidney disease. Nature Reviews Disease Primers 11 , 8 (2025). Carrasco-Zanini, J. , et al. Proteomic signatures improve risk prediction for common and rare diseases. Nature medicine 30 , 2489-2498 (2024). Davey Smith, G. & Hemani, G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet 23 , R89-98 (2014). Yu, Z. , et al. Polygenic risk scores for kidney function and their associations with circulating proteome, and incident kidney diseases. Journal of the American Society of Nephrology 32 , 3161-3173 (2021). Carrasco-Zanini, J. , et al. Proteomic signatures improve risk prediction for common and rare diseases. Nat Med 30 , 2489-2498 (2024). Chen, Y.J. , et al. Polygenic Risk Score Predicts Earlier-Onset Adult Systemic Lupus Erythematosus and First-Year Renal Diseases in a Taiwanese Cohort. RMD open 10 (2024). Inker, L.A. , et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 367 , 20-29 (2012). Ferkingstad, E. , et al. Large-scale integration of the plasma proteome with genetics and disease. Nature Genetics 53 , 1712-1721 (2021). Sun, B.B. , et al. Genomic atlas of the human plasma proteome. Nature 558 , 73-79 (2018). Suhre, K. , et al. Connecting genetic risk to disease end points through the human blood plasma proteome. Nature Communications 8 , 14357 (2017). Ahola-Olli, A.V. , et al. Genome-wide Association Study Identifies 27 Loci Influencing Concentrations of Circulating Cytokines and Growth Factors. The American Journal of Human Genetics 100 , 40-50 (2017). Ferreira, M.A.R. , et al. Quantitative Trait Loci for CD4:CD8 Lymphocyte Ratio Are Associated with Risk of Type 1 Diabetes and HIV-1 Immune Control. The American Journal of Human Genetics 86 , 88-92 (2010). Watanabe, K., Taskesen, E., Van Bochoven, A. & Posthuma, D. Functional mapping and annotation of genetic associations with FUMA. Nature communications 8 , 1826 (2017). Sollis, E. , et al. The NHGRI-EBI GWAS Catalog: knowledgebase and deposition resource. Nucleic Acids Res 51 , D977-D985 (2023). Moeller, M.J., Sanden, S.K., Soofi, A., Wiggins, R.C. & Holzman, L.B. Two gene fragments that direct podocyte-specific expression in transgenic mice. J Am Soc Nephrol 13 , 1561-1567 (2002). Prive, F., Arbel, J. & Vilhjalmsson, B.J. LDpred2: better, faster, stronger. Bioinformatics 36 , 5424-5431 (2021). Zhang, L.W. , et al. The Association of Intraindividual Difference Between Cystatin- and Creatinine-Based Estimated GFR and Contrast-Associated Acute Kidney Injury. Clinical interventions in aging 19 , 411-420 (2024). Potok, O.A. , et al. The Difference Between Cystatin C- and Creatinine-Based Estimated GFR and Associations With Frailty and Adverse Outcomes: A Cohort Analysis of the Systolic Blood Pressure Intervention Trial (SPRINT). American journal of kidney diseases : the official journal of the National Kidney Foundation 76 , 765-774 (2020). Brondani, L.A. , et al. Urinary peptidomics and bioinformatics for the detection of diabetic kidney disease. Sci Rep 10 , 1242 (2020). Fan, Z. , et al. Immune-related SERPINA3 as a biomarker involved in diabetic nephropathy renal tubular injury. Front Immunol 13 , 979995 (2022). Nunez-Sanchez, M.A. , et al. Microbial bile salt hydrolase activity influences gene expression profiles and gastrointestinal maturation in infant mice. Gut Microbes 14 , 2149023 (2022). Scholz, M. , et al. Genome-wide meta-analysis of phytosterols reveals five novel loci and a detrimental effect on coronary atherosclerosis. Nature communications 13 , 143 (2022). Rodriguez-Calvo, R. , et al. Role of the fatty acid-binding protein 4 in heart failure and cardiovascular disease. J Endocrinol 233 , R173-R184 (2017). Li, C. , et al. Macrophage polarization and meta-inflammation. Transl Res 191 , 29-44 (2018). Zhang, M. , et al. A rapamycin-loading platelet membrane hybrid liposome with anti-inflammation effect and long-lasting repair capability for acute kidney injury. J Control Release 380 , 927-942 (2025). Awal, H.B. , et al. Linagliptin, when compared to placebo, improves CD34+ ve endothelial progenitor cells in type 2 diabetes subjects with chronic kidney disease taking metformin and/or insulin: a randomized controlled trial. Cardiovascular diabetology 19 , 1-15 (2020). Schwartz, L. , et al. Repurposing HDAC inhibitors to enhance ribonuclease 4 and 7 expression and reduce urinary tract infection. Proc Natl Acad Sci U S A 120 , e2213363120 (2023). Ozbek, D.A. , et al. A comparative urinary proteomic and metabolomic analysis between renal aa amyloidosis and membranous nephropathy with clinicopathologic correlations. Journal of Proteomics 293 , 105064 (2024). Kramer, T.J. , et al. Ribonuclease-1 treatment after traumatic brain injury preserves blood-brain barrier integrity and delays secondary brain damage in mice. Sci Rep 12 , 5731 (2022). Li, S. , et al. Ribonuclease 4 protects neuron degeneration by promoting angiogenesis, neurogenesis, and neuronal survival under stress. Angiogenesis 16 , 387-404 (2013). Si, S., Liu, H., Xu, L. & Zhan, S. Identification of novel therapeutic targets for chronic kidney disease and kidney function by integrating multi-omics proteome with transcriptome. Genome Med 16 , 84 (2024). Wang, Y. , et al. Clec7a expression in inflammatory macrophages orchestrates progression of acute kidney injury. Front Immunol 13 , 1008727 (2022). Rasmussen, J. , et al. An early proinflammatory transcriptional response to tau pathology is age-specific and foreshadows reduced tau burden. Brain Pathol 32 , e13018 (2022). Kang, J. , et al. Plasma proteomics identifies proteins and pathways associated with incident depression in 46,165 adults. Sci Bull (Beijing) 70 , 573-586 (2025). Su, Z. & Wan, Q. Potential therapeutic targets for membranous nephropathy: proteome-wide Mendelian randomization and colocalization analysis. Front Immunol 15 , 1342912 (2024). Yang, M. , et al. ABO blood type is associated with renal outcomes in patients with IgA nephropathy. Oncotarget 8 , 73603-73612 (2017). Lu, Y. , et al. Metabolic Syndrome and Risk of Upper Tract Urothelial Carcinoma: A Case-Control Study From Surveillance, Epidemiology and End Results-Medicare-Linked Database. 10 (2021). Zhang, P. , et al. Tubulin cofactor A functions as a novel positive regulator of ccRCC progression, invasion and metastasis. Int J Cancer 133 , 2801-2811 (2013). Kuro, O.M. The Klotho proteins in health and disease. Nature reviews. Nephrology 15 , 27-44 (2019). Grabner, A. , et al. Activation of Cardiac Fibroblast Growth Factor Receptor 4 Causes Left Ventricular Hypertrophy. Cell metabolism 22 , 1020-1032 (2015). Schmidt, I.M. , et al. Plasma proteomics of acute tubular injury. Nature communications 15 , 7368 (2024). Eldjarn, G.H. , et al. Large-scale plasma proteomics comparisons through genetics and disease associations. Nature 622 , 348-358 (2023). Loh, P.R. , et al. Efficient Bayesian mixed-model analysis increases association power in large cohorts. Nat Genet 47 , 284-290 (2015). Additional Declarations There is NO Competing Interest. Supplementary Files Fig.S1.tif Supplementary figure 1. Pearson’s correlation coefficients between predicted and observed urinary ratios and metabolism waste traits using polygenic risk scores (PRS) of inflammatory proteins. Prediction models were adjusted for genetic background using genotype array and the first 20 genetic principal components. Two covariate models were evaluated: a baseline model (age + sex) and an extended model (age + sex + clinical factors including BMI, hypertension, diabetes, and high cholesterol). Bars indicate Pearson’s correlation coefficients under each model, grouped and color-coded by trait category: (A) Urinary ratios (UACR, UK/UCR, UNA/UCR, UNA/UK); (B) Metabolic waste markers (Urea, Urate). supplementaryTables20251229.xlsx Supplementary Tables Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8627406","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":577481583,"identity":"2ddf2788-ce23-4a08-9105-261e9a75ee76","order_by":0,"name":"Yifei LIN","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6klEQVRIiWNgGAWjYDACdsYGho8NDAwGDDxQkQMMDBJ4tTAzNjDOJFELEPGSpMXgMHPbY9sddXnmDLwHP1e22eXzHWA+eJsHrxbGduPcM4eLLRv4kiXPtiVbzjzAlmxNQEubdG7bgcQNB3gMJBvOMBsYHOAxkyaoxbKtDqTF+GfDmXqgFv5vhLUwtjGDtJhJNlQcBtnChleLJFCLZG/b4WKDwzxmlg0Vxw0kD7MZW87Bo4XvePsziZ9tdXkGx3uMbzYYVBvwHW9+eOMNHi0KByB0AiiCIIAZh1IYkG+AaRkFo2AUjIJRgAsAAGwdTLD9IOVeAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-3184-9213","institution":"Harvard School of Public Health","correspondingAuthor":true,"prefix":"","firstName":"Yifei","middleName":"","lastName":"LIN","suffix":""},{"id":577481584,"identity":"f2272fea-c816-495c-bf63-ad22f0586543","order_by":1,"name":"Shu Su","email":"","orcid":"","institution":"West China Hostpital, Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Shu","middleName":"","lastName":"Su","suffix":""},{"id":577481585,"identity":"93d9f263-d5f4-4931-a233-3b6e3546fed5","order_by":2,"name":"Tingting Fu","email":"","orcid":"","institution":"West China Hostpital, Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Tingting","middleName":"","lastName":"Fu","suffix":""},{"id":577481586,"identity":"c0bcb75e-384d-499f-ba3e-0643f842225c","order_by":3,"name":"Nanyan Xiang","email":"","orcid":"","institution":"West China Hostpital, Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Nanyan","middleName":"","lastName":"Xiang","suffix":""},{"id":577481587,"identity":"2a044610-1df9-47f4-ad0d-80ea5354d0da","order_by":4,"name":"Le Wang","email":"","orcid":"","institution":"West China Hostpital, Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Le","middleName":"","lastName":"Wang","suffix":""},{"id":577481588,"identity":"2af3f868-d176-49d9-9dcd-c1f410621c5d","order_by":5,"name":"Buu Truong","email":"","orcid":"","institution":"Harvard T.H. Chan School of Public Health","correspondingAuthor":false,"prefix":"","firstName":"Buu","middleName":"","lastName":"Truong","suffix":""},{"id":577481589,"identity":"a31e7b5b-ba0c-4bbf-bf72-ab31a903802b","order_by":6,"name":"Huan Yun","email":"","orcid":"","institution":"Harvard T.H. Chan School of Public Health","correspondingAuthor":false,"prefix":"","firstName":"Huan","middleName":"","lastName":"Yun","suffix":""},{"id":577481590,"identity":"671049a1-32d0-4dcb-a70d-18052f26a84d","order_by":7,"name":"Xingyan Wang","email":"","orcid":"","institution":"Harvard T. H. Chan School of Public Health","correspondingAuthor":false,"prefix":"","firstName":"Xingyan","middleName":"","lastName":"Wang","suffix":""},{"id":577481591,"identity":"f4d30164-21d8-4057-84fd-949fb5b9ce1f","order_by":8,"name":"Anand Srivastava","email":"","orcid":"","institution":"Division of Nephrology, Department of Medicine, University of Illinois Chicago","correspondingAuthor":false,"prefix":"","firstName":"Anand","middleName":"","lastName":"Srivastava","suffix":""},{"id":577481592,"identity":"6937f060-efbb-40d1-954f-829c1186c4fb","order_by":9,"name":"Ashish Verma","email":"","orcid":"","institution":"Section of Nephrology Department of Medicine Boston University Chobanian \u0026 Avedisian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Ashish","middleName":"","lastName":"Verma","suffix":""},{"id":577481593,"identity":"198500af-ee74-405a-be7b-38a60c2eb6c6","order_by":10,"name":"Sushrut Waikar","email":"","orcid":"https://orcid.org/0000-0003-4004-326X","institution":"Boston University","correspondingAuthor":false,"prefix":"","firstName":"Sushrut","middleName":"","lastName":"Waikar","suffix":""},{"id":577481594,"identity":"95ed5a59-0a0c-4ea7-ab82-a1aad20fb87b","order_by":11,"name":"Jin Huang","email":"","orcid":"","institution":"West China Hostpital, Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Jin","middleName":"","lastName":"Huang","suffix":""},{"id":577481595,"identity":"b4f0d8e1-5888-4228-ae96-aeb58eae4974","order_by":12,"name":"Liming Liang","email":"","orcid":"https://orcid.org/0000-0001-8261-3174","institution":"Harvard T.H. Chan School of Public Health","correspondingAuthor":false,"prefix":"","firstName":"Liming","middleName":"","lastName":"Liang","suffix":""}],"badges":[],"createdAt":"2026-01-17 16:50:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8627406/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8627406/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":101214507,"identity":"8796f8d5-ec62-438e-b3b8-6f16942f246b","added_by":"auto","created_at":"2026-01-27 10:35:36","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3707296,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFlow diagram of study design\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/de394e885ad135f276d6e371.png"},{"id":101214501,"identity":"24867502-21f4-40d3-a0fe-2e3eff1f09d0","added_by":"auto","created_at":"2026-01-27 10:35:35","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":4923208,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGWAS of 4 derived eGFRs(with the conventional eGFR inside) and Heritability of 13 kidney function traits.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGenome-wide association results and SNP-based heritability of kidney function traits.\u003c/p\u003e\n\u003cp\u003e(A–D) Circular Manhattan plots showing genome-wide association study (GWAS) results for four derived kidney traits:\u003c/p\u003e\n\u003cp\u003e(A) eGFRdiff = eGFRcys − eGFRcr,\u003c/p\u003e\n\u003cp\u003e(B) eGFRdiff.mean = (eGFRcys − eGFRcr)/ [(eGFRcys + eGFRcr)/2],\u003c/p\u003e\n\u003cp\u003e(C) eGFRdiff.pct.cr = (eGFRcys − eGFRcr)/eGFRcr (%),\u003c/p\u003e\n\u003cp\u003e(D) eGFRmean = (eGFRcys + eGFRcr)/2.\u003c/p\u003e\n\u003cp\u003eEach circular plot contains three inner 3 concentric rings representing the association p-values of the trait calculated using eGFRcys (inner ring), eGFRcr (middle ring), and eGFRcc (outer ring). Each point represents a single nucleotide polymorphism (SNP), mapped to its genomic position across chromosomes (Chr1–Chr22). The y-axis is represented by radial height as −log₁₀(p-value), and genome-wide significance threshold is indicated.\u003c/p\u003e\n\u003cp\u003e(E) SNP-based heritability (h²) estimates for all 13 kidney function traits calculated using LD score regression. Each point represents the h² estimate with 95% confidence intervals. Trait categories are color-coded: eGFR-based (orange), derived eGFR metrics (green), urinary biomarkers (purple), and metabolic waste (pink). The dashed red line indicates the null expectation (h² = 0).\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/5b4af14f69568845508ec9e2.png"},{"id":101297161,"identity":"7e24e544-fa13-4a32-b488-61a20efed0a2","added_by":"auto","created_at":"2026-01-28 09:25:45","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":3733723,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGenetic correlations between inflammation-related proteins and kidney function traits across biologically defined clusters.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A-B)\u003c/strong\u003e Hierarchy bi-clustering analysis and heatmap analysis of genetic correlation coefficient(Rg). The left heatmap displays the full spectrum of genetic correlations between 893 inflammation-associated proteins and 13 kidney-related traits, with hierarchical bi-clustering applied to both proteins (rows) and traits (columns). We identified six distinct inflammation clusters with coherent and trait-specific patterns of shared genetic architecture. Color represents the direction and strength of genetic correlation (red: positive; blue: negative), with darker shades indicating stronger correlations. The right heatmap presents only the statistically significant correlations based on the applied significance threshold (e.g., FDR\u0026lt; 0.05), with non-significant associations shown in grey. \u003cstrong\u003e(C–F) \u003c/strong\u003eDensity plots of Rg distributions for proteins associated with each of the four kidney trait categories:(C) Basic eGFR traits (eGFRcr, eGFRcys, eGFRcc),(D) Derived difference traits (eGFRdiff, eGFRdiff.mean, eGFRdiff.pct.cr and eGFRmean), (E) Urinary ratios (UACR, UK/UCR, UNA/UCR, UNA/UK),(F) Metabolic waste markers (Urea, Urate). Each curve shows the distribution of Rg values for all protein–trait pairs within the respective group, color-coded by trait.\u003c/p\u003e","description":"","filename":"Fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/bbd2a9908ce5195b9534c747.png"},{"id":101214506,"identity":"e854df0c-89c2-4aa1-8067-71d53630ca48","added_by":"auto","created_at":"2026-01-27 10:35:36","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":10906009,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNetwork of genetic causal proteins and outcome proteins with different kidney function traits.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the circular layout, proteins significant in 3 MR methods are shown in the outermost ring; those significant in 4 and 5 MR methods are displayed in black and red font, respectively. The nodes in the network represent either kidney function traits or inflammatory proteins, differentiated by background colors and label colors.\u003c/p\u003e\n\u003cp\u003e· Kidney function groups are highlighted with green, orange, purple, or pink backgrounds.\u003c/p\u003e\n\u003cp\u003e· Inflammatory proteins are categorized based on the number of Mendelian randomization (MR) validation methods: those confirmed by 5 MR methods appear with a blue background and red font, those validated by 4 MR methods have a blue background with black font, and proteins supported by 3 MR methods are shown with a grey background and black font.\u003c/p\u003e\n\u003cp\u003e· The rightmost light blue area represents the trait-to-protein group; the middle-colored area denotes group-specific, protein-to-trait relationships, and the leftmost dark blue area indicates group-shared, protein-to-trait relationships.\u003c/p\u003e\n\u003cp\u003e· The edges connecting the nodes illustrate causal associations, with blue edges representing pathways validated by 4 or 5 MR methods and grey edges indicating those confirmed by 3 MR methods.\u003c/p\u003e","description":"","filename":"Fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/60a6817b624b140e63581d67.png"},{"id":101214503,"identity":"b6f316b4-3948-47ef-ae91-17863696043d","added_by":"auto","created_at":"2026-01-27 10:35:35","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1441534,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScatterplot of pair-wise association effect of protein level and specific kidney function traits stratified by CKD status.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) The effect sizes of protein-to-trait effects stratified by their statistical significance in CKD and non-CKD groups.\u003c/p\u003e\n\u003cp\u003e(B) The effect sizes of trait-to-protein effects stratified by their statistical significance in CKD and non-CKD groups.\u003c/p\u003e\n\u003cp\u003eThe pathways are represented by distinct markers: red circles indicate pathways significant in both populations, blue solid triangles denote pathways significant only in CKD, green hollow triangles represent pathways significant only in non-CKD, and grey circles mark pathways that were not significant in either group.\u003c/p\u003e","description":"","filename":"Fig6.png","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/b692a41b91fe7b0fe3df2caf.png"},{"id":101214500,"identity":"f081a09a-9a28-4d8d-afb9-ff2d948ada24","added_by":"auto","created_at":"2026-01-27 10:35:35","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":750320,"visible":true,"origin":"","legend":"\u003cp\u003ePearson’s correlation coefficients between predicted and observed basic and derived eGFRs using polygenic risk scores (PRS) of inflammatory proteins.\u003c/p\u003e\n\u003cp\u003ePrediction models were adjusted for genetic background using genotype array and the first 20 genetic principal components. Two covariate models were evaluated: a baseline model (age + sex) and an extended model (age + sex + clinical factors including BMI, hypertension, diabetes, and high cholesterol). Bars indicate Pearson’s correlation coefficients under each model, grouped and color-coded by trait category:\u003c/p\u003e\n\u003cp\u003e(A) Basic eGFR traits (eGFRcr, eGFRcys, eGFRcc);\u003c/p\u003e\n\u003cp\u003e(B) Derived difference traits (eGFRdiff, eGFRdiff.mean, eGFRdiff.pct.cr and eGFRmean);\u003c/p\u003e","description":"","filename":"FIg5.png","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/5475700f914265463ef246fa.png"},{"id":101751096,"identity":"18d78dfd-d041-4f5d-a2ca-dcf648565e40","added_by":"auto","created_at":"2026-02-03 10:13:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":26006882,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/3df29666-5db0-480f-b52d-07f3bc138011.pdf"},{"id":101397681,"identity":"8c0f1581-64d6-49bb-bdef-1878c8b9fccf","added_by":"auto","created_at":"2026-01-29 09:35:06","extension":"tif","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":6343842,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary figure 1\u003c/strong\u003e. Pearson’s correlation coefficients between predicted and observed urinary ratios and metabolism waste traits using polygenic risk scores (PRS) of inflammatory proteins.\u003c/p\u003e\n\u003cp\u003ePrediction models were adjusted for genetic background using genotype array and the first 20 genetic principal components. Two covariate models were evaluated: a baseline model (age + sex) and an extended model (age + sex + clinical factors including BMI, hypertension, diabetes, and high cholesterol). Bars indicate Pearson’s correlation coefficients under each model, grouped and color-coded by trait category:\u003c/p\u003e\n\u003cp\u003e(A) Urinary ratios (UACR, UK/UCR, UNA/UCR, UNA/UK);\u003c/p\u003e\n\u003cp\u003e(B) Metabolic waste markers (Urea, Urate).\u003c/p\u003e","description":"","filename":"Fig.S1.tif","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/15ff1d354dfc1ead78f79eb2.tif"},{"id":101214504,"identity":"151f9a4e-998d-4ec7-9996-dd91cb2e7d48","added_by":"auto","created_at":"2026-01-27 10:35:36","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":13086526,"visible":true,"origin":"","legend":"Supplementary Tables","description":"","filename":"supplementaryTables20251229.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8627406/v1/fa158dd2f64e217535e8c3fa.xlsx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Shared genetic architecture between inflammatory proteins and eGFR difference identifies chronic inflammation risk factors for kidney function impairment","fulltext":[{"header":"Introduction","content":"\u003cp\u003e\u0026nbsp;Chronic kidney disease (CKD) affects over 13.4% of the global population\u003csup\u003e1\u003c/sup\u003e and is projected to become the fifth leading cause of death by 2040\u003csup\u003e2\u003c/sup\u003e. Early-stage kidney impairment is frequently asymptomatic; diagnosis is often delayed, thereby limiting opportunities for intervention.\u003csup\u003e3\u003c/sup\u003e. Although newly developed combined eGFR equations\u003csup\u003e4\u003c/sup\u003e have improved risk assessment, they remain grounded in physiological markers that reflect established damage. Increasing attention has turned to the difference between cystatin C- and creatinine-based (eGFRdiff). This metric has been shown across multiple studies to capture systemic inflammation, metabolic stress, and elevated cardiometabolic and mortality risks beyond what conventional kidney function markers can detect\u003csup\u003e5,6\u003c/sup\u003e. However, the genetic architecture of these markers and their association with chronic low-grade inflammation remains unknown\u003csup\u003e7,8\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eSystemic chronic inflammation (SCI) is likely a key driver of this discordance and a central contributor to CKD pathogenesis\u003csup\u003e9\u003c/sup\u003e. Elevated levels of inflammatory cytokines, such as tumor necrosis factor and interleukin-6, are strongly associated with kidney function decline\u003csup\u003e10,11\u003c/sup\u003e and also closely linked to comorbid conditions, including cardiovascular-kidney-metabolic (CKM) syndrome, obesity, and malnutrition\u003csup\u003e12,13\u003c/sup\u003e. As a long-standing systemic process, SCI may shape kidney function trajectories long before clinical decline becomes evident\u003csup\u003e9\u003c/sup\u003e. However, most studies to date remain observational and cross-sectional, lacking insight into the genetics of these associations and limiting their translation to early-stage risk assessment\u003csup\u003e14\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eBecause genetic variants are fixed from birth, their effects on inflammatory proteins can proxy long-term inflammation and enable causal inference\u003csup\u003e15\u003c/sup\u003e.\u0026nbsp;Recent advances in genome-wide association studies (GWAS) for kidney-related traits and large-scale proteomic profiling now enable population-scale investigations into how genetic variation shapes both protein expression and kidney function impairment\u003csup\u003e16,17\u003c/sup\u003e. Building on this property, their independence from environmental confounding and reverse causation make them powerful instrumental variables for causal inference through Mendelian randomization. Leveraging large-scale GWAS, polygenic risk scores (PRS) for circulating inflammatory proteins could be used to identify inflammation markers predictive of kidney functions and could be verify by the association between observed inflammatory protein levels and kidney functions\u003csup\u003e18\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eHere, we use large-scale genomic and proteomic data from the UK Biobank (~390,000 participants) and deCODE Iceland (~36,000 participants) to address this gap (\u003cstrong\u003eFig. 1\u003c/strong\u003e). We conducted GWAS for 13 kidney function-related traits\u0026mdash;including three updated eGFR equations\u003csup\u003e4,19\u003c/sup\u003e, four derived indicators, two metabolic waste markers, and four urinary ratios(\u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e)\u0026mdash;and assessed their genetic correlations and causal relationships with 839 circulating inflammatory proteins, primarily from the Olink inflammation panel, the SomaScan platforms and manually curated markers across multiple international cohorts\u003csup\u003e20-24\u003c/sup\u003e. We verified the findings using observed protein levels and kidney functions in the UK Biobank. We then evaluated whether inflammatory protein-based PRS could improve the prediction of kidney function variation. Collectively, this work provides a comprehensive evaluation of the biological basis and translational relevance of the eGFR difference. Together, these analyses may bridge molecular inflammation and kidney physiology, providing an integrative framework to disentangle inflammation\u0026rsquo;s genetic architecture and its translational implications for subclinical kidney dysfunction.\u0026nbsp;\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003e1. Kidney function gene mapping\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe first analyzed 13 kidney function\u0026ndash;related traits\u0026nbsp;through GWAS in ~390,000 UK Biobank participants with European ancestry(\u003cstrong\u003eFig. 2A\u0026ndash;D,\u003c/strong\u003e \u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e).\u0026nbsp;These traits encompassed glomerular filtration indicators (eGFRcr, eGFRcys, eGFRcc), metabolic waste markers (urea, urate), urinary ratios (UACR, UK/UCR, UNA/UCR, UNA/UK), and eGFR-derived indices (eGFRdiff, eGFRdiff.mean, eGFRdiff.pct.cr and eGFRmean). The eGFR-derived indices capture the divergence between cystatin C- and creatinine-based estimates:\u0026nbsp;eGFRdiff is their difference; eGFRdiff.mean is the normalized form obtained by dividing eGFRdiff by their eGFRmean(average); and eGFRdiff.pct.cr is the percentage form calculated by standardizing eGFRdiff to the creatinine-based estimate. (\u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Functional mapping and annotation (FUMA)\u003csup\u003e25\u003c/sup\u003e linked the identified loci to 110\u0026ndash;3,160 genes and prioritized 30\u0026ndash;612 candidate variants for each trait. The top loci for conventional eGFR traits (eGFRcys, eGFRcr, eGFRcc) corresponded to well-established genes, including \u003cem\u003eCST2, SPATA5L1\u003c/em\u003e, and \u003cem\u003eSHROOM3\u003c/em\u003e, all previously reported in the GWAS Catalog(https://www.ebi.ac.uk/gwas/)\u003csup\u003e26\u003c/sup\u003e, confirming the validity of our results. Metabolic and urinary traits, such as urate and UK/UCR, mapped to fewer loci but showed associations with known metabolic genes such as \u003cem\u003eCPS1\u003c/em\u003e and \u003cem\u003eSLC2A9\u003c/em\u003e. Among the derived eGFR difference traits, eGFRdiff, eGFRdiff.mean, and eGFRdiff.pct.cr consistently mapped to the \u003cem\u003eCSTL1\u003c/em\u003e/CST1 loci, whereas eGFRmean was associated with \u003cem\u003eCST4\u003c/em\u003e, highlighting the contribution of the cystatin gene family to kidney function variation.\u003c/p\u003e\n\u003cp\u003eFurther cross-referencing with GWAS Catalog revealed additional overlaps between kidney function\u0026ndash;related loci and a range of complex traits, including cardiometabolic and inflammatory phenotypes.\u003cstrong\u003e\u0026nbsp;Supplementary Table 2\u003c/strong\u003e summarizes overlapped lead SNPs separately by different traits. Specifically, even after excluding SNPs previously reported for kidney function traits, the remaining top loci for eGFRdiff, eGFRdiff.mean, and eGFRdiff.pct.cr predominantly mapped to cardiovascular and metabolic traits (9 out of 10) and inflammatory traits, such as platelet count. The derived eGFRmean trait also showed similar annotation, with 8 variants related to cardiometabolic traits and 2 linked to inflammatory pathways. Collectively, these findings indicated both shared and trait-specific genetic architectures connecting kidney function with metabolic, cardiovascular, and inflammatory pathways.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. Heritability and Genetic Correlations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; We estimated SNP-based heritability (h\u0026sup2;SNP) using\u0026nbsp;linkage disequilibrium score regression (LDSC). Heritability estimates were obtained for all 13 studied traits(\u003cstrong\u003eFig. 2\u003c/strong\u003eE, \u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e). Overall, we observed moderate to high heritability across kidney-related traits, with the highest estimates for eGFRcc (h\u0026sup2; = 0.342, SE = 0.039) and lowest estimates for UACR(h\u0026sup2; = 0.052, SE = 0.003).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; We estimated genome-wide genetic correlation between 893 inflammation-related plasma proteins\u003cstrong\u003e\u0026nbsp;(Supplementary Table 3)\u003c/strong\u003e and 13 kidney function traits using cross-trait LDSC. Through bi-clustering analysis using genetic correlation estimate(Rg), we identified six distinct inflammation clusters with coherent and trait-specific patterns of shared genetic architecture. \u003cstrong\u003e(Fig.3A-B, Supplementary Table 4-5)\u003c/strong\u003e. Cluster 1, consisting of three proteins, showed uniformly strong correlations with kidney filtration traits, led by RNASE1, which demonstrated the strongest negative genetic correlation with eGFRcys (Rg = \u0026ndash;1, FDR = 0.001) and a strong positive correlation with urate (Rg = 0.526, FDR = 0.001), while the RNASE1\u0026ndash;eGFRcr pair represented the most statistically significant association within this cluster (Rg = \u0026ndash;0.684, FDR = 5.78\u0026times; 10⁻\u003csup\u003e4\u003c/sup\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCluster 2, comprising 216 proteins, was characterized by broad negative correlations with creatinine-based kidney function, with the PRELP\u0026ndash;eGFRcr association showing the highest significance ( Rg = \u0026ndash;0.469, FDR = 2.07 \u0026times; 10⁻\u0026sup3;⁰). Cluster 3, containing 253 proteins, displayed preferential overlap with urate-related pathways, highlighted by ADGRD1\u0026ndash;urate ( Rg = 0.293, FDR = 4.92 \u0026times; 10⁻\u0026sup1;\u0026sup3;). Cluster 4, composed of 142 proteins, showed enriched associations with creatinine-derived filtration measures, exemplified by CD58\u0026ndash;eGFRcr ( Rg = \u0026ndash;0.287, FDR = 4.27 \u0026times; 10⁻\u0026sup1;⁶). Cluster 5, consisting of 62 proteins, was enriched for correlations with cystatin-C\u0026ndash;based kidney function, with ENPP5\u0026ndash;eGFRcys emerging as the strongest signal (Rg = 0.236, FDR = 2.16 \u0026times; 10⁻\u0026sup1;⁵). Cluster 6, comprising 79 proteins, demonstrated broad negative correlations across multiple kidney traits, with IL18BP\u0026ndash;eGFRmean showing the most pronounced overall association (Rg = \u0026ndash;0.447, FDR = 8.34 \u0026times; 10⁻⁴⁴). Together, these clusters reveal heterogeneous yet structured axes of genetic overlap between systemic inflammation and kidney function and nominate biologically coherent protein groups for downstream causal investigation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3. Candidate causal relationships between inflammation-related proteins and kidney function\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAcross 839 inflammatory proteins and 13 kidney function traits, we identified consistent significant protein-trait associations supported by multiple MR methods (\u0026ge;3 of 5 methods, FDR \u0026lt; 0.05, \u003cstrong\u003eFig. 4\u003c/strong\u003e\u003cstrong\u003e，\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eSupplementary Table 6-9\u003c/strong\u003e). Specifically, one basic eGFR trait(eGFRcys), 3 derived eGFR traits(eGFRdiff, eGFRdiff.mean, and eGFRdiff.pct.cr), all metabolic waste traits(urea, urate), and one urinary ratio trait(UK/UCR) showed significant protein-to-trait effects driven by inflammatory proteins, without observable reverse effects from these kidney function traits to proteins. No significant protein-to-trait or trait-to-protein associations were found for UNA/UCR, UACR, and UNA/UK. The predominant direction of causality was therefore from inflammation to kidney function, underscoring inflammation as a likely upstream driver of renal variation.\u003c/p\u003e\n\u003cp\u003eFor basic eGFR traits, eGFRcys demonstrated exclusive protein-to-trait\u0026nbsp;effects, implicating seven significant proteins (ACP1, CXCL10, ECM1, FGF5, MGMT, MST1, and RALB), involved in immune regulation, cell signaling, or tissue remodeling. However, no evidence supported the reverse direction from eGFRcys to proteins. In contrast, eGFRcr exhibited bidirectional relationships, identifying 11 protein-to- eGFRcr and 52 eGFRcr -to-protein effects. The combined index eGFRcc showed partial overlap, identifying 12 protein-to-eGFRcc and 15 eGFRcc-to-protein effects. Among them, IL12RB1 (immune signaling) was unique to eGFRcr, while CASP9 (apoptosis regulation) was unique to eGFRcc.\u003c/p\u003e\n\u003cp\u003eAmong the derived eGFR traits, eGFRdiff, eGFRdiff.mean, eGFRdiff.pct.cr, and eGFRmean exhibited consistent significant protein-to-trait effects, each associated with 7\u0026ndash;12 significant proteins. (all FDR of at least 4 MR methods \u0026lt;0.05). These traits showed substantial overlap across eGFRdiff, eGFRdiff.mean, and eGFRdiff.pct.cr traits, sharing a core set of proteins\u0026mdash;including ALPI, FABP9, PNLIPRP2, SERPINA1, APOC1, IL31, and PDGFD\u0026mdash;implicated in lipid metabolism, inflammation, and tissue remodeling. Within the eGFRmean trait, MGMT and ALPI both showed significant associations across all five MR methods(all FDR \u0026lt;0.05), reinforcing their robustness as potential upstream contributors of kidney function variation.\u003c/p\u003e\n\u003cp\u003eMetabolic waste markers revealed distinct signatures: Urea demonstrated protein-to-urea effects for 11 proteins (e.g., ABO, ACP1, CXADR, FGF19, HEG1, INSR), whereas urate was associated with four (e.g., FGF5, SCGB3A1), with no overlap between the two kidney function traits. Consistent results across all five MR methods identified HEG1 as unique to urea, while a subset of proteins, such as CXADR, LGALS4, INSR, TBCA, ACP1, THSD1, S100A13, and ABO, showed cross-trait relevance to both urea and both basic and derived eGFR indicators.\u0026nbsp;FGF5, also shared by urate and eGFR indicators, demonstrated robust associations across all MR models.\u003c/p\u003e\n\u003cp\u003eFor urinary traits, eight proteins (e.g., ABO, CD34, CFHR4, CXADR, EPHA4, INSR, LGALS4, RELB) showed significant protein-to-trait associations with UK/UCR, reflecting biological roles in vascular regulation, immune signaling, and metabolic balance. Four of these proteins (RELB, CFHR4, CD34, EPHA4) were unique to UK/UCR. No significant causal proteins were identified for UACR or UNA/UK based on at least 4 MR methods.\u003c/p\u003e\n\u003cp\u003eIn the reverse direction, a limited set of trait-to-protein effects was observed, primarily involving eGFRcr, eGFRcc, and eGFRmean. eGFRcr was associated with 52 downstream proteins, while eGFRcc and eGFRmean were linked to 15 and 12, respectively. Among these, RNASE4 was the most significant shared downstream protein for both eGFRcc and eGFRmean (mode-based estimate (MBE)MR: \u0026beta;= from \u0026ndash;0.59 to \u0026ndash;0.62, FDR \u0026lt; 0.05) and RNASE1 (MBE: \u0026beta; = \u0026ndash;0.615, 95% CI: \u0026ndash;0.951 to \u0026ndash;0.278, FDR \u0026lt; 0.05) showed a unique association to eGFRcc, suggesting RNA metabolism and immune regulation as possible pathways influenced by kidney dysfunction. Two additional proteins\u0026mdash;BTN3A2 and CLEC7A\u0026mdash;\u0026nbsp;displayed consistent trait-to-protein associations within the eGFRmean, indicating potential roles in immune signaling and kidney-inflammatory feedback.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4. Phenotypic corroboration of MR-inferred relationships\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe corroborated the MR-inferred relationships of proteins and traits through cross-sectional phenotypic analyses stratified by CKD status (\u003cstrong\u003eSupplementary Table 10\u003c/strong\u003e), using linear regression adjusted for age and sex (\u003cstrong\u003eFig. 5\u003c/strong\u003e). Among the significant protein\u0026ndash;to-trait direction pairs(\u003cstrong\u003eSupplementary Table 11\u003c/strong\u003e), 16 pairs showed consistent associations in both CKD and non-CKD groups, while three pairs (TBCA-eGFRcr, FGF5-Urate, SCGB3A1-Urate) exhibited opposite correlation directions across the two strata(all |\u0026beta;| \u0026gt;5, all FDR \u0026lt; 0.05). One pair (ABO-Urea) was significant only in the CKD group (\u0026beta; = 0.219, FDR = 0.021), whereas 19 pairs were uniquely significant in the non-CKD group, including a strong association between NCF2 and urate (\u0026beta; = 4.052, FDR = 3.20 \u0026times; 10⁻\u0026sup3;⁴). For trait-to-protein directions(\u003cstrong\u003eSupplementary Table 12)\u003c/strong\u003e, all tested pairs were significant in both strata, with effect sizes consistently larger in the CKD group. Among them, the well-known eGFRcr-NPHS1(nephrosis protein 1)\u003csup\u003e27\u003c/sup\u003e pair demonstrated the strongest positive association(\u0026beta;= 19.243, FDR=2.74E-31), whereas the eGFRcc-RNASE4 pair showed the strongest negative association (\u0026beta;= -37.63, FDR=3.15E-134).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5. Prediction of inflammation proteomics for kidney function using PRS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate how long-term inflammation status contributes to kidney function variation, we applied LDpred2-auto to compute polygenic risk scores (PRS)\u003csup\u003e28\u003c/sup\u003e for each of 839 inflammatory proteins as genetic proxies for chronic inflammation,\u0026nbsp;using genome-wide SNP weights derived from meta-analytic summary statistics of proteomic GWAS\u003csup\u003e28\u003c/sup\u003e.\u0026nbsp;Each protein PRS represents the inherited tendency for elevated circulating levels of that protein, serving as a proxy for lifelong systemic inflammation. These protein PRSs were jointly modeled using elastic net regression with 10-fold cross-validation to identify key inflammatory markers and quantify their combined predictive value (\u003cstrong\u003eFig. 6, Supplementary Table 13-14\u003c/strong\u003e). Two models were constructed: Model 1 adjusted for age, sex, and genetic principal components, while Model 2 additionally included clinical covariates (hypertension, diabetes, high cholesterol, and BMI). The inclusion of PRS improved prediction across most kidney function traits. For derived eGFR traits, predictive correlations increased by 0.036 to 0.118, while conventional eGFR traits showed smaller gains (0.027 to 0.061).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eInterestingly, for the derived eGFR traits, the gain in predictive performance from incorporating proteomic PRS in Model 1 was of similar magnitude to the improvement achieved by adding clinical covariates in Model 2, indicating that proteomic genetic profiles can provide predictive information of a similar scale to traditional risk factors. More modest improvements(\u003cstrong\u003eSupplementary Figure 1\u003c/strong\u003e) were observed for metabolic waste markers (0.010 to 0.017) and urinary biomarkers (0 to 0.006). Notably, for eGFRdiff, proteomic PRS improved predictive performance by 61% beyond demographic factors alone(0.193 to 0.311, \u0026Delta;=0.118, +61%), achieving an effect size comparable to that of established later-life clinical risk factors. Moreover, when added on top of these clinical variables, proteomic PRS provided a further 24% improvement in prediction(0.317 to 0.391, \u0026Delta;=0.075，+24%), indicating substantial incremental value beyond traditional risk stratification.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e\u003cstrong\u003e1. Summary of Main Findings\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLeveraging data from large prospective cohorts, we systematically integrated genome-wide, proteomic, and genetic risk modeling approaches to dissect the causal link between systemic inflammation and kidney function. Using LDSC and MR, we established shared genetic architecture and directional effects between 13 kidney function traits and 839 inflammation-related proteins, indicating a predominant causal influence from inflammation to kidney function variation. Among derived eGFR traits, particularly the eGFR difference (eGFRdiff) and its normalized forms( eGFRdiff.mean and eGFRdiff.pct.cr), we identified consistent protein-to-trait associations involving a core set of inflammatory and metabolic proteins (e.g., ALPI, FABP9, SERPINA1, APOC1, and PDGFD). This suggests these indices are sensitive captures of subclinical, inflammation-related kidney function impairment.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; Additional associations with metabolic (urea, urate) and urinary (UK/UCR) markers implicated pathways of vascular regulation, lipid metabolism, and RNA processing (e.g., RNASE family). Finally, by modeling PRS of inflammatory proteins as proxies for lifelong exposure to the inflammation, we demonstrated that genetically inferred inflammation profiles significantly explain variation in kidney function. This was particularly evident for eGFRdiff-related traits, highlighting their translational value for risk stratification and monitoring of Cardiovascular-Kidney-Metabolic(CKM) health.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. From Clinical Observation to Genetic Mechanism: The eGFR Difference\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePrevious studies have demonstrated that the directionality of eGFRdiff is clinically meaningful. For instance, a negative eGFRdiff (eGFRcys \u0026lt; eGFRcr) is associated with poorer health outcomes and a higher risk of acute kidney injury\u003csup\u003e5,29\u003c/sup\u003e, whereas a positive eGFRdiff (eGFRcys \u0026gt; eGFRcr) correlates with more favorable health status and reduced risks of adverse events, such as falls, hospitalization, and mortality\u003csup\u003e30\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTherefore, our study provides the first genome-wide evidence elucidating the genetic architecture underlying this discrepancy, building upon these clinical observations. We found that eGFRdiff and its normalized derivatives possess significant SNP-based heritability (\u003cem\u003eh\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 0.24), exceeding that of conventional biomarkers, i.e., eGFRcr. This indicates that eGFRdiff captures a broader range of genetically driven variation in renal function, reflecting non-GFR determinants such as muscle mass, nutritional status, and systemic inflammation, thereby offering complementary value for risk stratification.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; By integrating inflammatory proteomic PRS, we further improved the predictive performance of eGFRdiff, achieving model correlations comparable to clinical risk factors. These findings indicate that eGFRdiff is not merely a marker of glomerular filtration discordance but a genetically informed indicator of systemic inflammatory states. Compared with single-marker approaches, integrating eGFRdiff with inflammation-related proteomics enables earlier recognition of subclinical dysfunction, even among individuals with preserved global renal function.\u003c/p\u003e\n\u003cp\u003eAcross eGFRdiff, eGFRdiff.mean, eGFRdiff.pct.cr traits, a consistent inflammation-related protein signature\u0026mdash;comprising ALPI, FABP9, PNLIPRP2, SERPINA1, APOC1, IL31, and PDGFD\u0026mdash;was identified. These proteins demonstrated robust protein-to-trait associations across multiple MR models, suggesting upstream involvement in inflammation-mediated renal variation. Among them, SERPINA1, PDGFD, IL31, and APOC1 have been previously implicated in various renal pathologies, including diabetic kidney disease, glomerulosclerosis, interstitial fibrosis, CKD-associated pruritus, and IgA nephropathy. Their consistent associations across multiple derived eGFR traits highlight their biological and clinical relevance, indicating these proteins as mechanistic indicators of inflammation-related kidney dysfunction\u0026nbsp;and potential exploratory therapeutic targets\u003csup\u003e31,32\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn contrast, ALPI, FABP9, and PNLIPRP2 have not yet been directly reported to link to kidney dysfunction previously. Specifically, ALPI is known to as a central player in microbial homeostasis\u003csup\u003e33\u003c/sup\u003e, suggesting a potential role in the gut-kidney axis. FABP9, a fatty acid-binding protein, and PNLIPRP2, a pancreatic lipase-related protein related to coronary atherosclerosis\u003csup\u003e34,35\u003c/sup\u003e, might reflect the potential role of inflammatory metabolic states\u0026nbsp;or metainflammation\u003csup\u003e36\u003c/sup\u003e in kidney function decline. Collectively, these novel associations highlight the promise of these proteins as predictive markers for early-stage kidney impairment.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3. Novel Classification of Kidney Function Based on Inflammatory Proteins by Causal Direction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo refine the biological understanding of kidney traits,\u0026nbsp;we propose a novel classification framework grounded in the causal directionality between inflammatory proteins and kidney function.\u0026nbsp;This causal framework is conceptually aligned with the LDSC-based bi-clustering results, which reveal coherent groups of inflammatory proteins sharing consistent genetic relationships with specific kidney traits. Specifically, the 13 kidney-related traits can be categorized into two groups:\u0026nbsp;\u003cstrong\u003eInflammation-driven traits\u003c/strong\u003e, including eGFRdiff, eGFRcys, urea, urate, and UK/UCR, exhibited protein-to-trait causal associations with inflammatory proteins without corresponding reverse effects. \u003cstrong\u003eInflammation-interactive traits\u003c/strong\u003e, including eGFRcr, eGFRcc, and eGFRmean, showed both protein-to-trait and trait-to-protein associations, indicating complex, feedback-driven interactions between systemic inflammation and kidney function.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.1 Inflammation-Driven Traits Provide Insights into the Cardio-Kidney-Metabolic (CKM) Axis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInflammation-driven traits offer enhanced predictive value by capturing subclinical renal changes not detected through conventional biomarkers, highlighting their utility as complementary tools for clinical risk assessment. Consistent with prior evidence, eGFRcys demonstrated a strong protein-to-trait association pattern with upstream inflammatory drivers. Extending this framework, our analysis identified additional inflammation-driven kidney traits, indicating broader involvement of immune and metabolic pathways in renal regulation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFor blood-based traits\u003c/strong\u003e, GWAS Catalog cross-referencing revealed substantial genetic overlaps between derived eGFR metrics (e.g., eGFRdiff, eGFRdiff.eGFRmean and eGFRdiff.pct.cr) and known cardiovascular, metabolic, and inflammatory traits, as discussed above. This shared genetic architecture underscores the interconnected biological basis of these conditions, reinforcing the clinical relevance of derived kidney metrics for early CKM risk prediction and stratification.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFor urinary traits\u003c/strong\u003e, particularly UK/UCR,\u0026nbsp;although not directly associated with conventional kidney function measures, we identified four inflammatory proteins(ABO, CXADR, LGALS4, INSR\u003cem\u003e)\u003c/em\u003e showing consistent \u003cstrong\u003eprotein-to-trait associations\u003c/strong\u003e shared with established kidney function indicators such as\u0026nbsp;urea, eGFRcc and eGFRcr. Among these, INSR, a critical regulator of insulin signaling and metabolic homeostasis, highlights UK/UCR\u0026rsquo;s potential as an intermediary biomarker within cardio-kidney-metabolic(CKM) syndrome. Additionally, CD34, a marker of endothelial activation, exhibited a unique \u003cstrong\u003eprotein-to-trait association\u003c/strong\u003e with UK/UCR and previously linked to acute kidney injury\u003csup\u003e37\u003c/sup\u003e, diabetes and diabetic nephropathy\u003csup\u003e38\u003c/sup\u003e, further supports its role in capturing early endothelial and immune perturbations along the CKM axis. Given its non-invasive nature and ease of measurement, UK/UCR represents a promising alternative to current CKM diagnostic frameworks, which often rely on heterogeneous clinical criteria.\u003c/p\u003e\n\u003cp\u003e3.2 \u003cstrong\u003eInflammation-Interactive traits reflect both kidney injury and potential kidney-brain axis involvement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAmong inflammation-interactive traits, the RNASE family stood out for its strong and consistent trait-to-protein associations. The eGFRcc-RNASE4 association showed the most robust negative correlation at the protein-level, while LDSC genetic correlations also supported RNASE1\u0026rsquo;s link to eGFRcc and RNASE4\u0026rsquo;s association with eGFRcys. Beyond their established involvements in kidney impairment\u003csup\u003e39,40\u003c/sup\u003e, RNASE family proteins have been implicated in protecting against neuronal degeneration and maintaining blood-brain barrier integrity, suggesting a molecular basis for the kidney-brain axis\u003csup\u003e41,42\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMoreover, the eGFRcr-NPHS1 pair demonstrated the strongest positive trait-to-protein association. NPHS1 (Nephrosis Protein 1), a well-established protein in nephrotic syndromes, has also been implicated in neurodevelopmental processes\u003csup\u003e27\u003c/sup\u003e, supporting the biological interplay between kidney impairment and neurological outcomes such as uremic encephalopathy. Furthermore, the eGFRmean trait revealed 2 uniquely associated \u003cstrong\u003etrait-to-protein associations\u003c/strong\u003e, CLEC7A and BTN3A2, that were absent in basic eGFR traits. Both have been previously linked to kidney injury\u003csup\u003e43,44\u003c/sup\u003e, and emerging evidence also associates them with neurodegenerative disease\u003csup\u003e45,46\u003c/sup\u003e. In general, these findings highlight that inflammation-interactive traits may capture interconnected molecular signatures bridging renal and neural inflammation, potentially reflecting early kidney\u0026ndash;brain\u0026nbsp;crosstalk.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4. Clinical Implications for Kidney Transplantation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBeyond genetic susceptibility, trait-to-protein associations reveal how kidney function itself modulates systemic immune signaling. Among these, the ABO protein emerged as the most broadly connected upstream factor in our Mendelian randomization analysis, showing consistent protein-to-trait associations with multiple kidney function indicators and involvement in up to eight mechanistic associations. Beyond its established role in blood group compatibility, ABO variation was found to influence post-transplant outcomes through glycosylation-dependent and immune-mediated mechanisms\u003csup\u003e47-49\u003c/sup\u003e. Our findings extend the significance of ABO beyond immunological compatibility, suggesting that ABO polymorphisms may also influence post-transplant kidney function through glycosylation-related and inflammatory mechanisms. Similarly, several trait-to-protein associations, including HLA-E, LGALS9, CD200, and CD7, suggested that impaired filtration might trigger compensatory immunoregulatory responses. Clinically, these proteins could serve as dynamic biomarkers for transplant monitoring and individualized immunosuppression, bridging molecular mechanisms with patient-level management.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5. Opposite Associations by CKD Status Reveal State-Dependent Inflammatory Regulation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the protein-level validation analysis, the genetic causal proteins TBCA, FGF5, and SCGB3A1 exhibited opposite-level correlation trends with eGFRcr and urate between the CKD and non-CKD populations, suggesting that they may play state-dependent regulatory roles under different kidney function conditions. Specifically,\u0026nbsp;TBCA was negatively associated with eGFRcr in CKD but positively in non-CKD individuals. Although current studies mainly focus on the role of TBCA in kidney cancer\u003csup\u003e50\u003c/sup\u003e, its function in non-tumor kidney diseases remains underexplored and warrants further investigation.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Similarly, the opposite associations of FGF5 and SCGB3A1 with urate levels in different populations also suggest that they may be involved in the dynamic network regulating kidney function. Notably, FGF5 emerged as a shared causal factor across eGFR-based, derived eGFR, and metabolic waste traits, while FGF19 was uniquely associated with urea. Both\u0026nbsp;have been previously implicated in the progression of kidney dysfunction\u003csup\u003e43,51\u003c/sup\u003e, and belong to the fibroblast growth factor (FGF) family, which also includes FGF23, a known mediator of cardiac remodeling in CKD\u003csup\u003e52\u003c/sup\u003e. These findings also highlight a broader role of the FGF signaling in kidney function decline and its potential relevance to the CKM syndrome.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e7. Strength and Limitation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe main strength of this study is its systematic integration of genomic, proteomic, and causal inference methods to investigate early kidney function decline, enabling both biological discovery and translational insight. First, we present the first genome-wide association analysis of the eGFRcys\u0026ndash;eGFRcr differential marker (eGFRdiff), uncovering novel inflammation-related genetic signals that may reflect early kidney dysfunction. Second, by integrating protein-level associations with genetic risk, we identified several inflammation-related proteins (e.g., ABO, RNASE family) with potential causal roles in kidney decline, offering promising targets for biomarker development and therapeutic intervention. Third, we propose UK/UCR as a sensitive and non-invasive marker to evaluate subclinical kidney stress, particularly relevant for early-stage detection in high-risk populations. Fourth, we pioneer the application of proteomics-informed PRS to improve prediction of kidney outcomes, enhancing risk stratification for early nephroprotective intervention. Together, these contributions provide a framework for future translational research in personalized nephrology, bridging genomic discovery, protein validation, and clinical applicability.\u003c/p\u003e\n\u003cp\u003eThe limitations of this study should be acknowledged. First, although the genetic prediction model was developed using individuals of European ancestry to enhance precision under a homogeneous genetic background, this may limit the generalizability of our findings to other populations. Future research should extend this work to multi-ancestry populations to improve equity and applicability. Second, external validation and experimental confirmation were not available. Nonetheless, stratified analyses by CKD status and consistent associations across complementary kidney biomarkers (e.g., blood- and urine-based indicators) helped support the robustness of our findings. Third, due to the asymptomatic nature of early CKD and the limited number of clinically confirmed cases, we relied on continuous kidney function measures rather than binary disease outcomes. This approach improved statistical power and enabled cross-marker validation. Fourth, our proteomic coverage did not include certain kidney injury markers such as KIM-1, which are part of the Olink Oncology panel rather than the inflammation panel. However, multiple inflammation-related proteins identified here converge with recently reported kidney injury pathways\u003csup\u003e53\u003c/sup\u003e, highlighting potential biological convergence. Fifth, we did not account for the use of SGLT2 inhibitors, which may influence biomarker levels and disease progression. Even so, the inclusion of genetic information substantially improved the prediction of renal traits, suggesting that the observed associations capture underlying disease susceptibility beyond clinical treatment effects.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; In conclusion, our study uncovered shared genetic mechanisms and bidirectional causality between kidney function and systemic inflammation, offering insights for targeted interventions and personalized prevention strategies. Findings underscore the potential of proteomic signatures in mitigating CKD progression, though validation in diverse populations remains critical. \u0026nbsp;\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eStudy Population\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur study leveraged two large-scale European population-based datasets to investigate the genetic relationships between kidney function and chronic inflammation. The UK Biobank (UKB) cohort comprised approximately 390,000 participants of European ancestry recruited between 2006-2010, while the deCODE Genetics study included 35,892 Icelandic individuals with biomarker measurements collected from 2000-2019. Both studies obtained appropriate ethical approvals, and participants provided informed consent.\u0026nbsp;Genetic data underwent stringent quality control procedures, ensuring robust genetic analyses.\u0026nbsp;This study was carried out under UKBiobank project 45052.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eExposures assessment of inflammation traits\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe primary exposure of interest is the comprehensive assessment of inflammatory biomarkers and genetic data for participants. We have collected GWAS summary data with inflammatory protein biomarkers and genetics from multiple sources, which include 839 chronic inflammatory biomarkers curated from both Olink and SomaScan platforms, selected based on their established roles in inflammatory pathways in multiple international cohorts\u003csup\u003e20-24\u003c/sup\u003e and availability across multiple cohorts to ensure sufficient statistical power according to the study of deCODE and UK Biobank\u003csup\u003e54\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOutcome assessment of kidney function traits\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;To capture a multidimensional profile of renal function, we evaluated 13 quantitative kidney-related traits from the UK Biobank, spanning both blood- and urine-based biomarkers. These traits were categorized into four functional groups: estimated glomerular filtration rate (eGFR) indicators, metabolic waste markers, urinary ratio measures, and eGFR-derived indices. Together, they provide a comprehensive view of renal filtration, metabolic clearance, electrolyte handling, and biomarker discordance. First, we included three core eGFR indicators using the latest equations. eGFR based on serum creatinine (eGFRcr) was calculated using the 2021 CKD-EPI creatinine-based equation, while eGFR based on cystatin C (eGFRcys) followed the 2012 CKD-EPI cystatin C equation. A combined eGFR measure (eGFRcc) that incorporates both creatinine and cystatin C was also derived using the 2021 CKD-EPI combined formula. These markers reflect overall glomerular filtration capacity from complementary biochemical sources.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; Second, to assess renal clearance of metabolic waste, we evaluated serum concentrations of urea and urate. As end-products of protein and purine metabolism, respectively, both biomarkers are filtered by the kidneys and serve as sensitive indicators of renal excretory function. Third, we included four urinary ratio markers reflecting electrolyte and protein excretion. These comprised the urinary potassium-to-creatinine ratio (UK/UCR), sodium-to-creatinine ratio (UNA/UCR), sodium-to-potassium ratio (UNA/UK), and the urinary albumin-to-creatinine ratio (UACR). All ratios were standardized to urinary creatinine to account for variation in urine concentration. UACR, in particular, is widely used to detect early glomerular injury and predict CKD progression.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; Lastly, we derived four composite traits to quantify the difference between cystatin C\u0026ndash; and creatinine\u0026ndash;based eGFR estimates. The difference (eGFRdiff) was calculated as eGFRcys minus eGFRcr. The percent difference (eGFRdiff.mean.cr) was defined as eGFRdiff divided by eGFRcr and multiplied by 100. A normalized difference (eGFRdiff.mean) was calculated as eGFRdiff divided by the mean of the two values, with the eGFRmean itself defined as (eGFRcr + eGFRcys)/2. These derived traits have been proposed as sensitive markers of early functional deviation and discordant filtration behavior. All kidney function traits were measured under standardized protocols and subjected to rigorous quality control as part of the UK Biobank infrastructure.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCovariates\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; We considered a structured set of covariates tailored to each analysis stage to control for potential confounding and improve model precision. For genome-wide association analyses (GWAS), we adjusted for age, sex, age squared (AGE\u0026sup2;), assessment center, genotyping array, and the first 20 genetic principal components (PCs) to account for population stratification. For polygenic risk score (PRS) analyses, we applied two covariate models: a baseline model including 24 variables\u0026mdash;age, sex, assessment center, genotyping array, and the first 20 PCs\u0026mdash;and an extended model with 28 variables that further incorporated body mass index (BMI), hypertension status, diabetes status, and high cholesterol as additional clinical risk factors. For cross-sectional protein-level validation analyses, only age and sex were included as covariates. All covariates were obtained through standardized procedures at the time of recruitment and were harmonized across participants to ensure data consistency and comparability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe employed a multi-analytical framework integrating cutting-edge genetic epidemiological approaches to examine the genetic relationships between inflammatory biomarkers and kidney function.\u003c/p\u003e\n\u003cp\u003eWe performed GWASs for all the kidney function traits using a linear mixed model (LMM) implemented in BOLT-LMM v2.3.4\u003csup\u003e55\u003c/sup\u003e to account for cryptic relatedness and potential population stratification. Based on European ancestry, the analysis adjusted for age, age squared, sex, genotyping array, assessment center and 20 ancestry principal components to assess the association between the inverse normally transformed phenotype residuals and imputed genotype dosages. Using the public released proteomics GWAS summary data from UK Biobank and deCODE\u003csup\u003e54\u003c/sup\u003e,\u0026nbsp;we conducted meta-analyses for the same circulating protein biomarkers using METAL with the inverse-variance-weighted method.\u003c/p\u003e\n\u003cp\u003eWe performed linkage disequilibrium score regression (LDSC) to estimate genetic correlations between traits using GWAS summary statistics. This method leverages patterns of linkage disequilibrium across the genome to distinguish true polygenic signals from confounding factors such as population stratification. Genetic correlation coefficients (Rg) were considered significant at Rg \u0026gt; 0.1 with p \u0026lt; 0.05 after false discovery rate (FDR) correction. Following the initial LDSC analysis, we conducted hierarchical bi-clustering\u0026nbsp;performed in the R environment using Euclidean distance with average linkage for both inflammatory proteomics and kidney function biomarkers. Also, we generated heatmaps to visualize patterns of shared genetic architecture among the biomarkers, identifying potential biological modules and pathways underlying the observed correlations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor causal inference, we implemented bidirectional two-sample Mendelian randomization (MR) analyses using genetic variants as instrumental variables. This approach mimics randomized controlled trials through Mendel\u0026apos;s second law of inheritance, effectively addressing confounding present in observational studies. To ensure robust causal estimates, we employed five complementary MR methods: mode-based estimate (MBE) as primary analysis, supplemented by weighted median, MR-Egger, robust adjusted profile score (RAPS), and inverse-variance weighted (IVW) approaches. To visualize the complex web of causal relationships, we constructed causal networks using Cytoscape (v3.9.1), where nodes represented traits and edges represented significant causal associations (False Discovery Rate, FDR \u0026lt; 0.05 in at least three MR methods). This network analysis helped identify hub biomarkers with particularly influential roles in the kidney-inflammation axis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo corroborate the MR-inferred protein\u0026ndash;trait relationships at the observational level, we performed cross-sectional association analyses in the UK Biobank. Protein levels were regressed on corresponding kidney function traits using linear regression models stratified by CKD status and adjusted for age and sex. Standardized \u0026beta; coefficients were estimated to quantify the direction and magnitude of associations. This analysis aimed to evaluate concordance between genetically inferred and phenotypic relationships rather than to validate causal effects.\u003c/p\u003e\n\u003cp\u003eFinally, to assess how long-term inflammation contributes to kidney function variation, we developed protein-based polygenic risk scores (PRS) as genetic proxies for chronic inflammation. We used LDpred2-auto\u003csup\u003e28\u003c/sup\u003e to estimate posterior SNP effect sizes based on genome-wide association summary statistics from a meta-analysis of 839 inflammatory proteins, incorporating linkage disequilibrium (LD) information from the 1000 Genomes Project European reference panel. LDpred2 models LD structure to infer posterior SNP effect sizes under a point\u0026ndash;normal mixture prior, enabling genome-wide PRS construction without requiring external tuning parameters. Individual-level PRSs were subsequently calculated by aggregating weighted SNP effects across the genome.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo account for correlations among protein PRSs and select informative markers, we applied elastic net regression with 10-fold cross-validation, which jointly modeled their contribution to kidney function outcomes while optimizing regularization parameters. Model 1 adjusted for age, sex, and genetic principal components(demographic factors), whereas Model 2 additionally included hypertension, diabetes, high cholesterol, and BMI(traditional clinical factors). Predictive performance was evaluated using cross-validated correlations between observed and predicted trait values (\u0026Delta;R reported where applicable).\u003c/p\u003e\n\u003cp\u003eAll statistical analyses were conducted using R (v4.0.3). All analyses were adjusted for multiple comparisons using the False Discovery Rate (FDR) method to control for type I errors and to enhance the reliability of the findings.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe individual-level genotype and phenotype data used in this study are available through controlled-access application procedures. UK Biobank data are available to bona fide researchers upon approval of a data access application via the UK Biobank Access Management System (https://www.ukbiobank.ac.uk), under project number 45052. Summary-level genome-wide association statistics for inflammatory protein biomarkers were obtained from publicly released proteomics GWAS datasets generated using the Olink and SomaScan platforms, including data from UK Biobank and deCODE Genetics. Access to deCODE summary statistics is subject to the data use policies of deCODE Genetics. All analyses were conducted in accordance with the relevant ethical approvals and data access agreements.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAnalyses were performed using a combination of publicly available software and custom scripts. Genome-wide association analyses were conducted using PLINK (v1.9) and BOLT-LMM (v2.3.4). ( https://alkesgroup.broadinstitute.org/BOLT-LMM/BOLT-LMM_manual.html). SNP-based heritability and genetic correlations were estimated using LD Score Regression (LDSC) (https://github.com/bulik/ldsc). Mendelian randomization analyses were performed using the MendelianRandomization and MR-RAPS R packages. Proteomic polygenic risk scores were constructed using LDpred2-auto, which estimates posterior SNP effect sizes under a point\u0026ndash;normal mixture prior while accounting for linkage disequilibrium, using European ancestry reference panels from the 1000 Genomes Project. Elastic net regression models were fitted using the glmnet R package. Custom scripts used for data preprocessing, quality control, and downstream analyses are available from the corresponding authors upon reasonable request, subject to institutional data governance and participant privacy regulations.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was conducted during the author\u0026rsquo;s academic practicum training at the Harvard T.H. Chan School of Public Health. This work was supported by the National Natural Science Foundation of China (32471519 and 32571690) and 1.3.5 project for disciplines of excellence from West China Hospital of Sichuan University (ZYGD23039).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eLv, J.C. \u0026amp; Zhang, L.X. Prevalence and Disease Burden of Chronic Kidney Disease. \u003cem\u003eAdvances in experimental medicine and biology\u003c/em\u003e \u003cstrong\u003e1165\u003c/strong\u003e, 3-15 (2019).\u003c/li\u003e\n\u003cli\u003eGlobal, regional, and national burden of chronic kidney disease, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. \u003cem\u003eLancet (London, England)\u003c/em\u003e \u003cstrong\u003e395\u003c/strong\u003e, 709-733 (2020).\u003c/li\u003e\n\u003cli\u003eWebster, A.C., Nagler, E.V., Morton, R.L. \u0026amp; Masson, P. Chronic Kidney Disease. \u003cem\u003eLancet (London, England)\u003c/em\u003e \u003cstrong\u003e389\u003c/strong\u003e, 1238-1252 (2017).\u003c/li\u003e\n\u003cli\u003eInker, L.A.\u003cem\u003e, et al.\u003c/em\u003e New creatinine-and cystatin C\u0026ndash;based equations to estimate GFR without race. \u003cem\u003eNew England Journal of Medicine\u003c/em\u003e \u003cstrong\u003e385\u003c/strong\u003e, 1737-1749 (2021).\u003c/li\u003e\n\u003cli\u003eChen, D.C.\u003cem\u003e, et al.\u003c/em\u003e Association of intraindividual difference in estimated glomerular filtration rate by creatinine vs cystatin C and end-stage kidney disease and mortality. \u003cem\u003eJAMA Network Open\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, e2148940-e2148940 (2022).\u003c/li\u003e\n\u003cli\u003eFu, E.L.\u003cem\u003e, et al.\u003c/em\u003e Accuracy of GFR Estimating Equations in Patients with Discordances between Creatinine and Cystatin C-Based Estimations. \u003cstrong\u003e34\u003c/strong\u003e, 1241-1251 (2023).\u003c/li\u003e\n\u003cli\u003eTangri, N.\u003cem\u003e, et al.\u003c/em\u003e Multinational Assessment of Accuracy of Equations for Predicting Risk of Kidney Failure: A Meta-analysis. \u003cem\u003eJama\u003c/em\u003e \u003cstrong\u003e315\u003c/strong\u003e, 164-174 (2016).\u003c/li\u003e\n\u003cli\u003eWuttke, M.\u003cem\u003e, et al.\u003c/em\u003e A catalog of genetic loci associated with kidney function from analyses of a million individuals. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, 957-972 (2019).\u003c/li\u003e\n\u003cli\u003eFurman, D.\u003cem\u003e, et al.\u003c/em\u003e Chronic inflammation in the etiology of disease across the life span. \u003cem\u003eNat Med\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 1822-1832 (2019).\u003c/li\u003e\n\u003cli\u003eAmdur, R.L.\u003cem\u003e, et al.\u003c/em\u003e Inflammation and Progression of CKD: The CRIC Study. \u003cem\u003eClinical journal of the American Society of Nephrology : CJASN\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 1546-1556 (2016).\u003c/li\u003e\n\u003cli\u003eKadatane, S.P., Satariano, M., Massey, M., Mongan, K. \u0026amp; Raina, R. The Role of Inflammation in CKD. \u003cem\u003eCells\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e(2023).\u003c/li\u003e\n\u003cli\u003eStenvinkel, P.\u003cem\u003e, et al.\u003c/em\u003e Chronic Inflammation in Chronic Kidney Disease Progression: Role of Nrf2. \u003cem\u003eKidney international reports\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 1775-1787 (2021).\u003c/li\u003e\n\u003cli\u003eRomagnani, P.\u003cem\u003e, et al.\u003c/em\u003e Chronic kidney disease. \u003cem\u003eNature Reviews Disease Primers\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 8 (2025).\u003c/li\u003e\n\u003cli\u003eCarrasco-Zanini, J.\u003cem\u003e, et al.\u003c/em\u003e Proteomic signatures improve risk prediction for common and rare diseases. \u003cem\u003eNature medicine\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 2489-2498 (2024).\u003c/li\u003e\n\u003cli\u003eDavey Smith, G. \u0026amp; Hemani, G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. \u003cem\u003eHum Mol Genet\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, R89-98 (2014).\u003c/li\u003e\n\u003cli\u003eYu, Z.\u003cem\u003e, et al.\u003c/em\u003e Polygenic risk scores for kidney function and their associations with circulating proteome, and incident kidney diseases. \u003cem\u003eJournal of the American Society of Nephrology\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 3161-3173 (2021).\u003c/li\u003e\n\u003cli\u003eCarrasco-Zanini, J.\u003cem\u003e, et al.\u003c/em\u003e Proteomic signatures improve risk prediction for common and rare diseases. \u003cem\u003eNat Med\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 2489-2498 (2024).\u003c/li\u003e\n\u003cli\u003eChen, Y.J.\u003cem\u003e, et al.\u003c/em\u003e Polygenic Risk Score Predicts Earlier-Onset Adult Systemic Lupus Erythematosus and First-Year Renal Diseases in a Taiwanese Cohort. \u003cem\u003eRMD open\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e(2024).\u003c/li\u003e\n\u003cli\u003eInker, L.A.\u003cem\u003e, et al.\u003c/em\u003e Estimating glomerular filtration rate from serum creatinine and cystatin C. \u003cem\u003eN Engl J Med\u003c/em\u003e \u003cstrong\u003e367\u003c/strong\u003e, 20-29 (2012).\u003c/li\u003e\n\u003cli\u003eFerkingstad, E.\u003cem\u003e, et al.\u003c/em\u003e Large-scale integration of the plasma proteome with genetics and disease. \u003cem\u003eNature Genetics\u003c/em\u003e \u003cstrong\u003e53\u003c/strong\u003e, 1712-1721 (2021).\u003c/li\u003e\n\u003cli\u003eSun, B.B.\u003cem\u003e, et al.\u003c/em\u003e Genomic atlas of the human plasma proteome. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e558\u003c/strong\u003e, 73-79 (2018).\u003c/li\u003e\n\u003cli\u003eSuhre, K.\u003cem\u003e, et al.\u003c/em\u003e Connecting genetic risk to disease end points through the human blood plasma proteome. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 14357 (2017).\u003c/li\u003e\n\u003cli\u003eAhola-Olli, A.V.\u003cem\u003e, et al.\u003c/em\u003e Genome-wide Association Study Identifies 27 Loci Influencing Concentrations of Circulating Cytokines and Growth Factors. \u003cem\u003eThe American Journal of Human Genetics\u003c/em\u003e \u003cstrong\u003e100\u003c/strong\u003e, 40-50 (2017).\u003c/li\u003e\n\u003cli\u003eFerreira, M.A.R.\u003cem\u003e, et al.\u003c/em\u003e Quantitative Trait Loci for CD4:CD8 Lymphocyte Ratio Are Associated with Risk of Type 1 Diabetes and HIV-1 Immune Control. \u003cem\u003eThe American Journal of Human Genetics\u003c/em\u003e \u003cstrong\u003e86\u003c/strong\u003e, 88-92 (2010).\u003c/li\u003e\n\u003cli\u003eWatanabe, K., Taskesen, E., Van Bochoven, A. \u0026amp; Posthuma, D. Functional mapping and annotation of genetic associations with FUMA. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 1826 (2017).\u003c/li\u003e\n\u003cli\u003eSollis, E.\u003cem\u003e, et al.\u003c/em\u003e The NHGRI-EBI GWAS Catalog: knowledgebase and deposition resource. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, D977-D985 (2023).\u003c/li\u003e\n\u003cli\u003eMoeller, M.J., Sanden, S.K., Soofi, A., Wiggins, R.C. \u0026amp; Holzman, L.B. Two gene fragments that direct podocyte-specific expression in transgenic mice. \u003cem\u003eJ Am Soc Nephrol\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 1561-1567 (2002).\u003c/li\u003e\n\u003cli\u003ePrive, F., Arbel, J. \u0026amp; Vilhjalmsson, B.J. LDpred2: better, faster, stronger. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e36\u003c/strong\u003e, 5424-5431 (2021).\u003c/li\u003e\n\u003cli\u003eZhang, L.W.\u003cem\u003e, et al.\u003c/em\u003e The Association of Intraindividual Difference Between Cystatin- and Creatinine-Based Estimated GFR and Contrast-Associated Acute Kidney Injury. \u003cem\u003eClinical interventions in aging\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 411-420 (2024).\u003c/li\u003e\n\u003cli\u003ePotok, O.A.\u003cem\u003e, et al.\u003c/em\u003e The Difference Between Cystatin C- and Creatinine-Based Estimated GFR and Associations With Frailty and Adverse Outcomes: A Cohort Analysis of the Systolic Blood Pressure Intervention Trial (SPRINT). \u003cem\u003eAmerican journal of kidney diseases : the official journal of the National Kidney Foundation\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e, 765-774 (2020).\u003c/li\u003e\n\u003cli\u003eBrondani, L.A.\u003cem\u003e, et al.\u003c/em\u003e Urinary peptidomics and bioinformatics for the detection of diabetic kidney disease. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 1242 (2020).\u003c/li\u003e\n\u003cli\u003eFan, Z.\u003cem\u003e, et al.\u003c/em\u003e Immune-related SERPINA3 as a biomarker involved in diabetic nephropathy renal tubular injury. \u003cem\u003eFront Immunol\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 979995 (2022).\u003c/li\u003e\n\u003cli\u003eNunez-Sanchez, M.A.\u003cem\u003e, et al.\u003c/em\u003e Microbial bile salt hydrolase activity influences gene expression profiles and gastrointestinal maturation in infant mice. \u003cem\u003eGut Microbes\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 2149023 (2022).\u003c/li\u003e\n\u003cli\u003eScholz, M.\u003cem\u003e, et al.\u003c/em\u003e Genome-wide meta-analysis of phytosterols reveals five novel loci and a detrimental effect on coronary atherosclerosis. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 143 (2022).\u003c/li\u003e\n\u003cli\u003eRodriguez-Calvo, R.\u003cem\u003e, et al.\u003c/em\u003e Role of the fatty acid-binding protein 4 in heart failure and cardiovascular disease. \u003cem\u003eJ Endocrinol\u003c/em\u003e \u003cstrong\u003e233\u003c/strong\u003e, R173-R184 (2017).\u003c/li\u003e\n\u003cli\u003eLi, C.\u003cem\u003e, et al.\u003c/em\u003e Macrophage polarization and meta-inflammation. \u003cem\u003eTransl Res\u003c/em\u003e \u003cstrong\u003e191\u003c/strong\u003e, 29-44 (2018).\u003c/li\u003e\n\u003cli\u003eZhang, M.\u003cem\u003e, et al.\u003c/em\u003e A rapamycin-loading platelet membrane hybrid liposome with anti-inflammation effect and long-lasting repair capability for acute kidney injury. \u003cem\u003eJ Control Release\u003c/em\u003e \u003cstrong\u003e380\u003c/strong\u003e, 927-942 (2025).\u003c/li\u003e\n\u003cli\u003eAwal, H.B.\u003cem\u003e, et al.\u003c/em\u003e Linagliptin, when compared to placebo, improves CD34+ ve endothelial progenitor cells in type 2 diabetes subjects with chronic kidney disease taking metformin and/or insulin: a randomized controlled trial. \u003cem\u003eCardiovascular diabetology\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 1-15 (2020).\u003c/li\u003e\n\u003cli\u003eSchwartz, L.\u003cem\u003e, et al.\u003c/em\u003e Repurposing HDAC inhibitors to enhance ribonuclease 4 and 7 expression and reduce urinary tract infection. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cstrong\u003e120\u003c/strong\u003e, e2213363120 (2023).\u003c/li\u003e\n\u003cli\u003eOzbek, D.A.\u003cem\u003e, et al.\u003c/em\u003e A comparative urinary proteomic and metabolomic analysis between renal aa amyloidosis and membranous nephropathy with clinicopathologic correlations. \u003cem\u003eJournal of Proteomics\u003c/em\u003e \u003cstrong\u003e293\u003c/strong\u003e, 105064 (2024).\u003c/li\u003e\n\u003cli\u003eKramer, T.J.\u003cem\u003e, et al.\u003c/em\u003e Ribonuclease-1 treatment after traumatic brain injury preserves blood-brain barrier integrity and delays secondary brain damage in mice. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 5731 (2022).\u003c/li\u003e\n\u003cli\u003eLi, S.\u003cem\u003e, et al.\u003c/em\u003e Ribonuclease 4 protects neuron degeneration by promoting angiogenesis, neurogenesis, and neuronal survival under stress. \u003cem\u003eAngiogenesis\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 387-404 (2013).\u003c/li\u003e\n\u003cli\u003eSi, S., Liu, H., Xu, L. \u0026amp; Zhan, S. Identification of novel therapeutic targets for chronic kidney disease and kidney function by integrating multi-omics proteome with transcriptome. \u003cem\u003eGenome Med\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 84 (2024).\u003c/li\u003e\n\u003cli\u003eWang, Y.\u003cem\u003e, et al.\u003c/em\u003e Clec7a expression in inflammatory macrophages orchestrates progression of acute kidney injury. \u003cem\u003eFront Immunol\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 1008727 (2022).\u003c/li\u003e\n\u003cli\u003eRasmussen, J.\u003cem\u003e, et al.\u003c/em\u003e An early proinflammatory transcriptional response to tau pathology is age-specific and foreshadows reduced tau burden. \u003cem\u003eBrain Pathol\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, e13018 (2022).\u003c/li\u003e\n\u003cli\u003eKang, J.\u003cem\u003e, et al.\u003c/em\u003e Plasma proteomics identifies proteins and pathways associated with incident depression in 46,165 adults. \u003cem\u003eSci Bull (Beijing)\u003c/em\u003e \u003cstrong\u003e70\u003c/strong\u003e, 573-586 (2025).\u003c/li\u003e\n\u003cli\u003eSu, Z. \u0026amp; Wan, Q. Potential therapeutic targets for membranous nephropathy: proteome-wide Mendelian randomization and colocalization analysis. \u003cem\u003eFront Immunol\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 1342912 (2024).\u003c/li\u003e\n\u003cli\u003eYang, M.\u003cem\u003e, et al.\u003c/em\u003e ABO blood type is associated with renal outcomes in patients with IgA nephropathy. \u003cem\u003eOncotarget\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 73603-73612 (2017).\u003c/li\u003e\n\u003cli\u003eLu, Y.\u003cem\u003e, et al.\u003c/em\u003e Metabolic Syndrome and Risk of Upper Tract Urothelial Carcinoma: A Case-Control Study From Surveillance, Epidemiology and End Results-Medicare-Linked Database. \u003cstrong\u003e10\u003c/strong\u003e(2021).\u003c/li\u003e\n\u003cli\u003eZhang, P.\u003cem\u003e, et al.\u003c/em\u003e Tubulin cofactor A functions as a novel positive regulator of ccRCC progression, invasion and metastasis. \u003cem\u003eInt J Cancer\u003c/em\u003e \u003cstrong\u003e133\u003c/strong\u003e, 2801-2811 (2013).\u003c/li\u003e\n\u003cli\u003eKuro, O.M. The Klotho proteins in health and disease. \u003cem\u003eNature reviews. Nephrology\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 27-44 (2019).\u003c/li\u003e\n\u003cli\u003eGrabner, A.\u003cem\u003e, et al.\u003c/em\u003e Activation of Cardiac Fibroblast Growth Factor Receptor 4 Causes Left Ventricular Hypertrophy. \u003cem\u003eCell metabolism\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 1020-1032 (2015).\u003c/li\u003e\n\u003cli\u003eSchmidt, I.M.\u003cem\u003e, et al.\u003c/em\u003e Plasma proteomics of acute tubular injury. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 7368 (2024).\u003c/li\u003e\n\u003cli\u003eEldjarn, G.H.\u003cem\u003e, et al.\u003c/em\u003e Large-scale plasma proteomics comparisons through genetics and disease associations. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e622\u003c/strong\u003e, 348-358 (2023).\u003c/li\u003e\n\u003cli\u003eLoh, P.R.\u003cem\u003e, et al.\u003c/em\u003e Efficient Bayesian mixed-model analysis increases association power in large cohorts. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 284-290 (2015).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8627406/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8627406/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Chronic kidney disease(CKD) is linked to systemic inflammation, yet the shared genetic mechanisms remain unclear. Emerging evidence indicates that the difference between cystatin C- and creatinine-based eGFR serves as an early biological indicator capturing inflammatory, cardiometabolic, and mortality-related risk signals that precede subclinical kidney injury. Using large-scale cohorts (~390,000 UK Biobank; ~36,000 deCODE), we integrated genome-wide association studies across 13 kidney function traits and 839 inflammation-related proteins. Cross-trait genetic analyses and Mendelian randomization highlighted that the eGFR difference is highly heritable and shares genetic components with cardiometabolic signals. Multiple upstream inflammatory proteins, including ALPI, FABP9, INSR, and ABO, showed consistent protein-to-trait effects, corroborated by protein-level associations in 50,000 UK Biobank participants. Proteomic polygenic risk scores improved prediction of eGFR difference by 61% beyond demographic factors, achieved performance comparable to later-life clinical factors. Together, these findings provide the first genome-wide evidence delineating the genetic and inflammatory architecture of the eGFR difference, offering new biological insights for Cardiovascular-Kidney-Metabolic(CKM) risk stratification and transplant monitoring.","manuscriptTitle":"Shared genetic architecture between inflammatory proteins and eGFR difference identifies chronic inflammation risk factors for kidney function impairment","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-27 10:35:11","doi":"10.21203/rs.3.rs-8627406/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"1663c13f-baf6-4ed0-9aba-79fe2ec5fcba","owner":[],"postedDate":"January 27th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":61439132,"name":"Health sciences/Risk factors"},{"id":61439133,"name":"Health sciences/Diseases/Kidney diseases/Chronic kidney disease"},{"id":61439134,"name":"Health sciences/Biomarkers/Predictive markers"},{"id":61439135,"name":"Health sciences/Medical research/Epidemiology"}],"tags":[],"updatedAt":"2026-02-15T20:25:21+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-27 10:35:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8627406","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8627406","identity":"rs-8627406","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}