Integrative Genomics Establishes GNL3 as a Pleiotropic Hub and Causal Gene for Osteoarthritis

Integrative Genomics Establishes GNL3 as a Pleiotropic Hub and Causal Gene for Osteoarthritis | 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 Research Article Integrative Genomics Establishes GNL3 as a Pleiotropic Hub and Causal Gene for Osteoarthritis Tianhao Qu, Yan Zhong, Yonghuan Zhou, Lin Liu, Zheng Ye This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7630649/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 16 Feb, 2026 Read the published version in Human Genomics → Version 1 posted 10 You are reading this latest preprint version Abstract Background: Osteoarthritis (OA) is a leading cause of disability, yet there are no approved disease-modifying therapies (DMOADs) capable of halting its progression. While genome-wide association studies (GWAS) have robustly linked the 3q21locus to OA, the causal effecator gene and its underlying mechanism have remained elusive, hindering translational progress. Methods: We implemented a multi-tiered, systematic integrative genomics strategy to proceed from genetic association to causal mechanism. By integrating summary statistics from large-scale OA GWAS with multi-tissue molecular quantitative trait loci (QTL) and single-cell expression QTL (sc-eQTL) data, we employed a combination of Bayesian colocalization and Mendelian randomization (MR) to establish a robust chain of causal evidence. Results: Our analysis definitively identifies GNL3 as the causal effector gene at the 3q21 locus. MR analysis formally demonstrated that genetically predicted higher GNL3 expression is causally protective against both knee and hip OA (e.g., Knee OA: Odds Ratio = 0.85, p = 4.4e-4). Further single-cell causal inference pinpointed this protective effect to specific cellular contexts, most prominently in activated T-cells, neural cells under cellular stress, and developmental progenitors . Moreover, we establish this locus as a pleiotropic hub, showing that its causal variants are shared with systemic OA risk factors, including Body Mass Index (BMI) and vitamin D levels. Conclusion: This study establishes, for the first time, a complete chain of evidence from a shared causal variant to the regulation of GNL3 expression in specific cell types and, ultimately, to a causal impact on OA risk. Our findings converge on a novel mechanistic model: GNL3 acts as a master regulator of systemic homeostasis, where its genetically determined expression level modulates immune and neural cell responses to stress, thereby dictating an individual's susceptibility to OA. This work validates GNL3 as a high-confidence therapeutic target and provides a new framework for developing DMOADs aimed at reinforcing systemic homeostatic pathways. Osteoarthritis GNL3 Mendelian Randomization Integrative Genomics Integrative Genomics Pleiotropy Figures Figure 1 Figure 2 1. Introduction Osteoarthritis (OA), the most common form of arthritis, is a leading cause of chronic pain and disability globally, imposing a significant socioeconomic burden[ 1 ]. Pathologically, OA involves a complex interplay of articular cartilage degradation, synovial inflammation, and subchondral bone remodeling, driven by genetic, metabolic, and mechanical factors[ 2 ]. Genome-wide association studies (GWAS) have successfully identified hundreds of genetic loci associated with OA susceptibility[ 3 , 4 ]. However, the majority of these associated variants are located in non-coding regions, making the identification of their target effector genes and downstream functional mechanisms a critical bottleneck in translating genetic discoveries into clinical applications. A prominent risk locus on chromosome 3q21 has been consistently implicated in OA across multiple large-scale GWAS[ 5 ]. This region is gene-dense, containing several plausible candidates such as GNL3 , ITIH1 , and PBRM1 . Differentiating the causal gene(s) from bystander genes within the same linkage disequilibrium (LD) block is a non-trivial task that requires integrative functional genomic approaches[ 6 ]. Pioneering work by Gee et al. provided the first functional evidence at this locus, using allelic expression imbalance (AEI) analysis in joint tissues from OA patients. Their study revealed that the OA-associated allele at this locus correlated with significantly lower expression of both GNL3 and SPCS1[ 5 ]. While these initial findings were crucial, GNL3's definitive causal role, the full extent of its systemic influence, and the precise molecular mechanisms through which it impacts OA pathogenesis remained to be elucidated. GNL3 is a highly conserved nucleolar GTP-binding protein known to play fundamental roles in ribosome biogenesis, cell cycle regulation, and stem cell maintenance[ 7 ], but a comprehensive understanding of its function in a complex degenerative disease like OA requires a broader, more systematic approach. However, these foundational findings remained correlational, and two critical questions were left unanswered: is the effect of GNL3 on OA truly causal, and in which specific cellular contexts and states does this genetic regulation exert its pathogenic influence? GNL3 , a highly conserved nucleolar GTP-binding protein, is known to play fundamental roles in ribosome biogenesis and cell cycle regulation, but its function within the systemic pathology of a complex degenerative disease like OA has remained poorly understood. To bridge the gap from genetic association to causal mechanism, we designed a systematic, multi-omics strategy(Fig. 1) to test the central hypothesis that shared genetic variants driving OA risk do so by causally modulating GNL3 expression in specific, disease-relevant cellular contexts. We leveraged large-scale GWAS and molecular QTL data to first build a robust, multi-layered evidence chain for GNL3 's causal role through comprehensive colocalization and Mendelian Randomization (MR). Crucially, we extended this framework to the single-cell level to dissect its function in distinct cell types and activation states. This integrative approach allowed us not only to formally validate GNL3 as the causal effector gene but also to uncover its role as a pleiotropic hub that integrates systemic metabolic and immune signals. Our findings converge to propose a unifying model where genetically determined GNL3 expression dictates OA susceptibility by modulating cellular responses to stress, thereby providing a new mechanistic framework for the disease. 2. Materials and Methods 2.1 Data Sources This study is based entirely on the analysis of publicly available, summary-level data from large-scale genome-wide association studies (GWAS) and molecular quantitative trait locus (QTL) consortia. All data utilized were from studies that had previously obtained the necessary patient consent and ethical approvals( Supplement Materials 1 ). 2.1.1 GWAS Summary Statistics for Osteoarthritis Summary statistics for osteoarthritis (OA) and its joint-specific subtypes were obtained from several large-scale GWAS. All included cohorts consisted of individuals of European ancestry. Our primary analysis leveraged summary data from the most extensive OA genetics meta-analysis to date, published by Boer et al. (2021)[8]. This study provided robust genetic association data for hip osteoarthritis (N = 353,388) and for a combined knee and/or hip osteoarthritis phenotype (N = 490,345). To ensure comprehensive coverage, we also incorporated data from the UK Biobank, as analyzed by Tachmazidou et al. (2019)[9], which furnished summary statistics for hip osteoarthritis (N = 393,873) and for osteoarthritis of the hip or knee (N = 417,596). For complementary purposes, our analyses also included data from the foundational OA GWAS conducted by Zeggini et al. (2012)[10], which comprised 18,419 individuals. 2.1.2 Molecular Quantitative Trait Locus (QTL) Data A comprehensive collection of molecular QTL data was assembled to interrogate the functional consequences of OA-associated variants. Bulk-Tissue eQTL and sQTL Data: Expression QTL (eQTL) and splicing QTL (sQTL) data for 21 tissues—including adipose tissue, cultured fibroblasts, muscle, nerve, and whole blood—were sourced from the Genotype-Tissue Expression (GTEx) project v8[11]. For a more powered analysis in blood, we also utilized cis-eQTL data from the eQTLGen consortium[12] (n = 31,684). Brain-Specific QTL Data: To enrich our analysis of neurological tissues, we integrated several brain-specific datasets. These included cis-eQTL data from the BrainMeta study (cortex, n = 2,865)[13], the PsychENCODE project (prefrontal cortex, n = 1,387)[14], and the GTEx-brain-std study (n=233). Furthermore, methylation QTL (mQTL) data for the brain cortex were obtained from the Religious Orders Study and Memory and Aging Project (ROSMAP)[15], specifically the ROSMAP_CMC dataset accessed via the Brain xQTL Serve. Single-Cell eQTL Data: To explore gene regulation at cellular resolution, we utilized the scQTLbase portal[16], which aggregates sc-eQTLs from numerous datasets covering a wide range of cell types and states. This resource enabled the investigation of cell-type-specific regulatory effects. 2.1.3 GWAS Summary Statistics for Immune Cell Phenotypes For our Mendelian Randomization analysis of the immune system, we leveraged summary statistics from a large-scale GWAS of 731 immune phenotypes (Orrù et al., 2020)[17]. The source study was based on a cohort of 3,757 individuals of European ancestry from the SardiNIA Project. In that study, a comprehensive panel of immune traits was quantified using flow cytometry, encompassing absolute and relative cell counts, median fluorescence intensities (MFIs), and morphological parameters. The publicly available summary data, which we accessed via the IEU OpenGWAS database, provided the genetic instruments for our investigation. 2.1.4 TCGA Data for Pan-Cancer Analysis To investigate the expression and methylation patterns of GNL3 across various cancer types, we utilized publicly available data from The Cancer Genome Atlas (TCGA) project[18]. All processed data were downloaded from the UCSC Xena functional genomics browser, which provides uniformly processed TCGA datasets. Specifically, we retrieved pan-cancer gene expression data (RNA-Seq, quantified as log2(norm_count+1)) and DNA methylation data (Beta-values from the Illumina Human Methylation 450k platform) for the TCGA Pan-Cancer (PANCAN) cohort. Data for the GNL3 gene were extracted across all available primary tumor (sample type code: 01) and adjacent solid tissue normal (sample type code: 11) samples for differential analysis. 2.2 Statistical Analysis All statistical analyses were conducted using R (version 4.3.0). The "TwoSampleMR"[19] and "coloc"[20] packages were central to the Mendelian randomization and colocalization analyses, respectively. 2.2.1 Colocalization Analysis To test for shared causal variants between OA risk loci and molecular QTLs, we employed two complementary approaches. Bulk-tissue colocalization: The Bayesian colocalization method, COLOC, was used to calculate the posterior probability (PP) for five mutually exclusive hypotheses: H0 (no association), H1 (association with OA only), H2 (association with QTL only), H3 (distinct causal variants), and H4 (a shared causal variant). We used the default conservative prior probabilities (p1 = 1x10⁻⁴, p2 = 1x10⁻⁴, p12 = 1x10⁻⁵). Strong evidence for colocalization was defined by a high posterior probability for a shared variant (PP4 ≥ 0.8). To avoid false positives from low-power studies, SNPs with p > 1x10⁻⁴ in either the GWAS or QTL dataset were excluded from the analysis. Summary-based Mendelian Randomization (SMR) & HEIDI Test: As a complementary approach, Summary-based Mendelian Randomization (SMR)[21] was used to identify pleiotropic associations between gene expression (and other molecular traits) and OA risk, using the top QTL variant as an instrument. A Benjamini-Hochberg false discovery rate (FDR < 0.05) was applied to correct for multiple testing. To distinguish true pleiotropy from confounding by linkage disequilibrium (LD), we performed the Heterogeneity in Dependent Instruments (HEIDI) test. Loci with a HEIDI p-value > 0.05 were considered to have a low probability of heterogeneity, supporting a shared causal variant. 2.2.2 Mendelian Randomization (MR) Analysis We conducted three distinct sets of MR analyses to investigate causal relationships from different perspectives. Causal Effect of GNL3 Expression on Osteoarthritis: To assess the direct causal effect of GNL3 expression on OA risk, we used summary statistics from the eQTLGen consortium for cis-eQTLs of GNL3 in whole blood. Since the lead eQTL for GNL3 represented a single, strong instrumental variable after LD clumping, the causal effect was estimated using the Wald Ratio method[22]. As this analysis utilized a single instrument, sensitivity analyses to assess horizontal pleiotropy (e.g., MR-Egger[22]) were not applicable. Causal Effects of Immune Cells on Osteoarthritis: To systematically investigate the causal roles of immune cells in OA, we performed a broader MR analysis using 729 immune cell phenotypes as exposures and four distinct OA datasets as outcomes. Instrument Selection: For each immune phenotype, we selected independent, genome-wide significant SNPs (p < 5 × 10⁻⁸) to serve as instrumental variables. To ensure independence, we performed LD clumping using a strict r² threshold of < 0.001 within a 10,000 kb window, based on the 1000 Genomes European reference panel. Primary and Sensitivity Analyses: Our choice of MR method depended on the number of available instrumental variables for each immune phenotype. For phenotypes with multiple independent SNPs, the Inverse-Variance Weighted (IVW) [23] method was used as the primary method to estimate the causal effect. To assess the robustness of these findings, we employed a suite of sensitivity analyses, including MR-Egger regression [24] , the Weighted Median method[25], the Weighted Mode method[26], and Bayesian Weighted Mendelian Randomization [27] . In cases where only a single SNP remained as a valid instrument after clumping, the causal effect was estimated using the Wald Ratio method. Heterogeneity and Pleiotropy Assessment: For instruments comprising two or more SNPs, we conducted several diagnostic tests. Heterogeneity was assessed using Cochran's Q statistic . Directional horizontal pleiotropy was evaluated using the MR-Egger intercept test , where a p-value < 0.05 indicates significant pleiotropy. We also generated scatter plots and leave-one-out plots to visually inspect the influence of individual SNPs. Single-cell MR and colocalization analysis to identify cellular mechanisms of GNL3 : To dissect the cell-type-specific causal effects of GNL3 expression, we extended the MR framework to single-cell eQTL (sc-eQTL) data. Instrumental variables were selected and harmonized following the same procedure described for the immune cell screen. Causal effects were estimated using the IVW method for multi-SNP instruments and the Wald Ratio for single-SNP instruments.The robustness of our single-cell findings was confirmed using the sensitivity analyses described previously (e.g., MR-Egger intercept). Additionally, we applied the Steiger test to verify the causal direction from gene expression to OA risk and rule out reverse causation. To account for multiple testing across the numerous cellular contexts, we applied the False Discovery Rate (FDR) correction, with a corrected p-value (FDR < 0.05) considered statistically significant. Finally, to test for a shared causal variant between the sc-eQTL and OA GWAS signals, we performed a Bayesian colocalization analysis for each cellular context, defining strong evidence as PP.H4 ≥ 0.8. 2.2.3 Functional Annotation of Genetic Variants To functionally characterize genetic variants of interest, we used the ANNOVAR software. Variants were annotated using a comprehensive set of databases, including gene-based annotations (RefSeq, Ensembl), population frequency data (gnomAD, 1000 Genomes), and data on clinical and functional significance (ClinVar, dbNSFP, GWAS Catalog). 3. Results 3.1. Widespread eQTL and mQTL Colocalization Provides Robust Evidence for GNL3 To identify the most likely effector gene at the 3q21 locus, we performed a comprehensive colocalization analysis across multiple osteoarthritis (OA) GWAS datasets and molecular QTLs. Among all genes in the region, GNL3 was uniquely distinguished by widespread, consistent, and robust evidence of colocalization, firmly establishing it as the lead candidate gene. A detailed summary of all colocalization results is provided in Supplementary Materials 2 (Figure S1-Figure S5) . Our investigation began with the foundational OA GWAS by Zeggini et al. (2012; P01779), which revealed exceptionally strong colocalization signals between the OA risk locus and GNL3 expression QTLs (eQTLs). This colocalization was observed in a broad array of tissues, with 17 diverse tissues showing strong evidence of a shared causal variant (PP.H4 > 0.8), most notably in Cultured Fibroblasts (PP.H4 = 0.98), Testis (PP.H4 = 0.97), and Skeletal Muscle (PP.H4 = 0.97). The critical question was whether this signal was robust. We confirmed this by replicating the analysis using larger, more recent GWAS datasets. In the UK Biobank analysis by Tachmazidou et al. (2019), strong evidence of colocalization (PP.H4 > 0.8) for GNL3 eQTLs was consistently observed for both hip OA (P02060) and combined hip/knee OA (P02059) in many of the same key tissues, such as Breast Mammary Tissue and Testis (Table 1) . This pattern of shared signals across independent cohorts was further validated using the latest large-scale meta-analysis by Boer et al. (2021; P02070, P02071). While the posterior probabilities in this latter dataset did not always strictly exceed the 0.8 threshold in every tissue, the signals remained consistently high (e.g., PP.H4 ≈ 0.7), reinforcing the overall conclusion. This slight variation, likely due to differences in cohort composition and statistical power, demonstrates the stability of the signal rather than contradicting it. The remarkable consistency of this eQTL colocalization across different OA study populations and a diverse panel of tissues provides powerful evidence that this association is not a result of chance or confounding by linkage disequilibrium. To explore upstream regulatory mechanisms, we found that the OA risk signal also strongly and consistently colocalized with methylation QTLs (mQTLs) that control DNA methylation at CpG sites within the GNL3 locus. This finding was also robust across the different OA datasets. For instance, the mQTLs for probes cg18595196, cg00845626, and cg11041457 all robustly colocalized with hip and knee OA risk from both the Tachmazidou et al. (2019) and Boer et al. (2021) studies, with posterior probabilities consistently exceeding 0.85 (Table 2) . This suggests that the genetic effect on OA is likely mediated, at least in part, through the epigenetic modification of the GNL3 gene, which in turn regulates its transcription. Collectively, the highly consistent colocalization results across multiple independent OA datasets, diverse tissue eQTLs, and distinct regulatory layers (mQTLs) build a powerful and solid evidentiary foundation, prioritizing GNL3 as the primary functional gene at this locus. 3.2 GNL3 Expression is Negatively Correlated with Promoter Methylation To further elucidate the regulatory mechanism linking DNA methylation to GNL3 expression, we performed a pan-cancer correlation analysis using data from The Cancer Genome Atlas (TCGA). We specifically examined the correlation between GNL3 mRNA expression and methylation levels (Beta-values). Our analysis focused on two key individual CpG sites, cg18595196 and cg11041457, which were identified as significant in our mQTL colocalization analysis, as well as the average methylation value across 17 CpG sites within the GNL3 locus. As shown in the gene schematic ( Figure S1, Figure S6 ), these 17 probes are strategically clustered within key regulatory regions of the GNL3 gene, including its promoter, the 5' UTR, and the first exon. Many of these probes are also located within or adjacent to a CpG island, positioning them to directly influence its transcriptional activity. The analysis revealed a consistent and statistically significant negative correlation between GNL3 expression and the methylation levels of both the individual CpG sites and their aggregate average across a wide spectrum of cancer types ( Figure 2, Figure S7 ). This inverse relationship, where higher methylation is associated with lower gene expression, was observed for all three methylation metrics and was particularly pronounced in numerous cancers, including Liver Hepatocellular Carcinoma (LIHC), Stomach Adenocarcinoma (STAD), Lung Adenocarcinoma (LUAD), and Lung Squamous Cell Carcinoma (LUSC). This pan-cancer evidence provides strong, independent support for the hypothesis derived from our mQTL analysis: that DNA methylation within the GNL3 promoter acts as a repressive regulatory mechanism, contributing to the silencing of its expression. This finding mechanistically links the genetic variants (mQTLs) to a functional outcome (GNL3 expression levels), strengthening the overall causal argument. 3.3 Mendelian Randomization Confirms a Causal, Protective Role for GNL3 Expression Having established GNL3 as the likely causal gene through colocalization, we performed a two-sample Mendelian Randomization (MR) analysis to formally test for a causal relationship and determine the direction of its effect on OA risk. Using genetically predicted GNL3 expression in whole blood as the exposure, our analysis revealed a significant and consistent protective effect against major forms of osteoarthritis. Specifically, a genetically predicted one-standard-deviation increase in GNL3 expression was associated with a significantly lower risk of knee osteoarthritis (OR = 0.85, 95% CI [0.78, 0.93], p = 4.4e-4) and hospital-diagnosed hip osteoarthritis (OR = 0.63, 95% CI [0.46, 0.86], p = 3.8e-3). Furthermore, a significant protective effect was also observed for a broader, localized OA phenotype (PheCode 740.1; OR = 0.81, 95% CI [0.70, 0.93], p = 3.5e-3). While not all tested OA-related traits reached statistical significance, these consistent findings across distinct, large-scale OA datasets provide strong statistical support for the hypothesis that maintaining higher levels of GNL3 expression is causally protective against the development of osteoarthritis( Table 3 ). 3.4 Exploratory Analysis of Splicing QTLs Suggests Tissue-Specific Effects We next conducted an exploratory analysis to investigate if alternative splicing could constitute an additional regulatory mechanism. A broad colocalization analysis between OA risk and GNL3 splicing QTLs (sQTLs) did not reveal a widespread pattern of shared causal variants comparable to the eQTL results ( Supplementary Materials 3 ). However, a more sensitive, secondary analysis using SMR did identify significant, albeit highly tissue-specific, associations ( Table 4 ). For example, a distinct splicing event in the Thyroid was causally associated with a protective effect on OA risk, while a different splicing event in the same tissue was associated with increased risk. These preliminary, tissue-specific findings suggest that while the regulation of total gene expression appears to be the primary mechanism at this locus, alternative splicing of GNL3 may represent a secondary, context-dependent regulatory layer, representing a potential avenue for future investigation into its tissue-specific functions. 3.5 Systematic single-cell analysis pinpoints causal mechanisms to specific immune, neural, and progenitor cells To dissect the precise cellular context of GNL3 's causal effect, we performed a systematic colocalization and Mendelian randomization analysis using single-cell eQTL data. Our systematic MR analysis of 47 available single-cell contexts provided robust and consistent evidence of causality. After correcting for multiple testing, all significant associations revealed a protective effect of higher GNL3 expression on knee OA risk, with the effects concentrated in specific, state-dependent cellular contexts ( Table 5 , Supplementary Materials 4 ). This causal effect was particularly strong in the immune system. For instance, genetically predicted GNL3 expression in activated CD4+ Memory T-cells showed a robust protective signal (OR = 0.81, 95% CI [0.73, 0.90], FDR = 0.003) . We also uncovered a novel and significant link to neural lineages under cellular stress. GNL3 expression in stressed dopaminergic neurons was strongly protective (OR = 0.73, 95% CI [0.60, 0.88], FDR = 0.007) and showed highly suggestive evidence of a shared genetic architecture with OA risk (PP.H4 = 0.71) . Crucially, the most powerful evidence for a specific causal pathway emerged from a clear convergence of our analytical approaches in developmental progenitors. The signal in floor plate progenitors not only demonstrated a strong and significant causal protective effect (OR = 0.89, 95% CI [0.84, 0.95], FDR = 0.003) but was also the only context to surpass the high-confidence threshold for a shared genetic variant (PP.H4 = 0.81) . This dual, quantitative evidence provides a definitive link between a shared causal variant, the regulation of GNL3 in this specific cell type, and protection against osteoarthritis. 3.6 GNL3 Locus is a Pleiotropic Hub for Systemic Metabolic, Hematological, and Neurological Traits To understand the broader systemic impact of the GNL3 locus, we performed a GWAS-GWAS colocalization analysis using a body composition trait, impedance of the leg (a proxy for fat-free mass), as the anchor. This revealed that the same causal variant influencing GNL3 and OA risk is highly pleiotropic, significantly affecting a range of systemic traits and established OA risk factors ( Table 6 ). The most compelling evidence of pleiotropy was observed with key metabolic and endocrine traits. We found exceptionally strong colocalization with Body Mass Index (BMI; PP.H4 > 0.97) and waist-to-hip ratio (WHR; PP.H4 > 0.99), two critical indicators of metabolic health. A similarly robust signal was identified for 25-hydroxyvitamin D levels (PP.H4 > 0.98), a crucial factor in bone and joint health, with the evidence remaining strong even after conditioning on BMI. Intriguingly, the locus also showed significant, albeit more moderate, evidence of colocalization with other systemic phenotypes. These included hematological traits such as hematocrit (PP.H4 ≈ 0.80) and several neuro-behavioral traits, most notably neuroticism (PP.H4 > 0.96) and its related "worry" subcluster (PP.H4 ≈ 0.79). This pattern of shared genetic architecture across metabolic, endocrine, hematological, and neurological domains positions GNL3 as a central genetic hub. These findings strongly reinforce the view of OA as a disease with significant systemic components, driven by genes that operate far beyond the local joint environment. Key pleiotropic colocalization results are summarized in Table 6, with a comprehensive list provided in Supplementary Materials 5 . 3.7 Mendelian Randomization Reveals Causal Roles of Specific Immune Cell Subsets in OA Pathogenesis To dissect the systemic inflammatory component of OA pathogenesis, we conducted a comprehensive two-sample MR analysis to systematically evaluate the causal effects of 729 genetically predicted immune cell phenotypes on four distinct OA outcomes. This analysis identified a complex landscape of both protective and risk-increasing roles for specific innate and adaptive immune cell populations, providing strong causal evidence for their involvement in the disease ( Table 7; full results in Supplementary Materials 6 ). Within the adaptive immune system, we found compelling evidence for the involvement of both T and B cell lineages. Notably, higher genetically predicted levels of naive CD4+ T-cells were causally associated with a reduced risk of localized OA (OR = 0.78, 95% CI [0.69, 0.89], p = 1.5e-4). A strong protective signal was also observed for IgD- CD27- B-cells, which significantly lowered the risk of hip OA (OR = 0.66, 95% CI [0.54, 0.80], p = 3.0e-5), an association that remained robust after multiple testing correction. In contrast, higher levels of effector memory CD8+ T-cells were associated with an increased risk of knee and hip OA (OR = 1.33, 95% CI [1.05, 1.68], p = 1.7e-02). In the innate immune compartment, we identified striking and opposing effects. A genetically predicted increase in classical monocytes (CD14+ CD16-) demonstrated a strong protective effect against knee OA (OR = 0.44, 95% CI [0.27, 0.70], p = 6.7e-4). Conversely, a robust risk-increasing signal was observed for several markers of granulocyte activation. The most significant of these was the expression of CD86 on granulocytes, which was causally associated with a more than two-fold increase in the risk of knee OA (OR = 2.07, 95% CI [1.36, 3.14], p = 6.7e-04). Collectively, these MR results provide strong causal evidence that the balance between specific immune cell subsets is a critical determinant of OA risk, highlighting distinct cell-mediated pathways for joint-specific pathology. 4. Discussion Our systematic, multi-modal genomic analysis has definitively established GNL3 as the causal effector gene at the 3q21 osteoarthritis (OA) risk locus . Moving beyond this primary identification, our findings converge to propose a unifying mechanistic model: GNL3 functions as a master regulator of systemic homeostasis . We posit that genetically determined GNL3 expression levels act as a systemic set point, modulating the response of immune, neural, and progenitor cells to metabolic and inflammatory stressors, thereby dictating an individual's susceptibility to developing OA. 4.1 A multi-layered evidence chain establishes GNL3 as the causal gene The foundation of our argument is the exceptionally consistent evidence of shared causal variants between OA risk and the regulation of GNL3 . The genetic signal for OA strongly and repeatedly colocalized with GNL3 expression QTLs (eQTLs) across a diverse panel of 17 tissues and was robustly replicated across independent, large-scale GWAS cohorts (Zeggini et al. 2012, Tachmazidou et al. 2019, and Boer et al. 2021), providing powerful evidence against confounding by chance or linkage disequilibrium ( Section 3.1 ; Supplementary Materials S2 ). We further strengthened this link by demonstrating that the OA risk locus also consistently colocalized with methylation QTLs (mQTLs) controlling CpG sites within the GNL3 promoter, and functionally validated this by showing a significant negative correlation between promoter methylation and GNL3 expression ( Section 3.2 ). While our exploratory analysis of splicing QTLs did not suggest a widespread role ( Section 3.4 ), our Mendelian Randomization analysis provided formal proof of causality, demonstrating that genetically predicted higher expression of GNL3 is significantly protective against both knee and hip OA ( Section 3.3 ). This multi-layered evidence chain forms an unshakable foundation for GNL3 's causal role. 4.2 A unified mechanistic hypothesis: GNL3 as a master regulator of cellular stress response in OA A critical insight from our study is that the function of GNL3 in OA is not confined to the joint but is deeply embedded in systemic biology. Our pleiotropy analysis firmly establishes the GNL3 locus as a hub sharing causal variants with key metabolic and endocrine mediators, including BMI and 25-hydroxyvitamin D ( Section 3.6 ). This provides a direct genetic underpinning for the modern paradigm of OA as a systemic, metabolic disease characterized by "metaflammation"[28-30]. Similarly, Vitamin D is essential for bone health and has known immunomodulatory functions[31, 32]. Our systematic single-cell analyses provide the crucial next step, dissecting precisely how these systemic risks are executed at a cellular level through the action of GNL3 . 4.2.1 A Novel Neuro-Skeletal Axis: Linking Central Nervous System Integrity to Joint Pathogenesis One of the most striking and novel findings of our study is the strong causal protective effect of GNL3 in neural cells, particularly under conditions of cellular stress. This discovery provides compelling genetic evidence for the emerging concept of a "neuro-skeletal axis" in OA, moving beyond the traditional view of pain as a mere consequence of joint damage, and positioning the nervous system as an active participant in the disease process[33]. Our analysis revealed a significant protective signal for GNL3 expression in rotenone-treated dopaminergic neurons and astrocytes, suggesting that individuals with genetically lower GNL3 expression may have a reduced capacity to buffer their neural systems against metabolic or oxidative stress. We propose this genetic vulnerability could manifest in two distinct, yet interconnected, pathways relevant to OA: Promotion of Neurogenic Inflammation and Structural Joint Damage: The nervous system is not a passive bystander to joint pathology. Peripheral sensory neurons that innervate the joint can, when stressed or damaged, release pro-inflammatory neuropeptides such as Calcitonin Gene-Related Peptide (CGRP) and Substance P directly into the synovial space[34]. This process, known as neurogenic inflammation, can directly drive OA pathogenesis by promoting synovitis, increasing vascular permeability, and stimulating the production of matrix-degrading enzymes by chondrocytes and synoviocytes. Our findings suggest a novel genetic basis for this mechanism: genetically determined lower GNL3 expression may render these sensory neurons more susceptible to stress-induced dysfunction, leading to a heightened state of neurogenic inflammation that actively contributes to the structural degradation of the joint. This provides a direct, mechanistic link from a genetic variant to the cellular processes that cause cartilage breakdown. Modulation of Central Sensitization and the Pain Experience: A well-known paradox in OA is the often-poor correlation between the degree of joint damage visible on X-rays and the level of pain experienced by the patient[35]. This is largely explained by central sensitization, a process where the central nervous system becomes hyperexcitable, amplifying pain signals[36]. Our findings provide a potential genetic explanation for an individual's predisposition to central sensitization. The causal effect of GNL3 in central nervous system cells (astrocytes, dopaminergic neurons) suggests that its expression level influences the overall "tone" and resilience of these circuits. Lower GNL3 expression could lead to a state of heightened neural excitability or impaired glial cell function, creating a permissive environment for the development of chronic, amplified pain. This aligns intriguingly with our pleiotropy finding linking the GNL3 locus to neuroticism, suggesting that this single genetic variant may simultaneously influence the structural integrity of the joint (via neurogenic inflammation) and the central nervous system's processing of pain and stress. In summary, our data reframe the role of the nervous system in OA from a simple pain reporter to a critical, genetically-influenced modulator of the disease. The GNL3 locus appears to be a key factor governing the resilience of the neuro-skeletal axis, where its protective, high-expression alleles may help maintain both peripheral neural homeostasis and central pain circuit stability, thereby protecting against both the structural and symptomatic progression of osteoarthritis. 4.2.2 Modulating Immune Homeostasis and Cellular Senescence Our systematic analysis confirmed a robust causal role for GNL3 in activated T-cells ( Section 3.5 ), providing a mechanistic anchor for the well-documented immune component of OA, which our broader immune cell MR screen showed was defined by a delicate balance between protective (e.g., naive T-cells) and pathogenic (e.g., activated effector cells) populations ( Section 3.7 ). We hypothesize that GNL3 , a known regulator of the p53-mediated stress response[37, 38], acts as a critical checkpoint for maintaining immune cell homeostasis and preventing premature exhaustion. In individuals with risk-associated, low-expression alleles, T-cells, upon activation by systemic inflammatory triggers (such as those from adipose tissue in obesity), may have a lower threshold for entering a state of activation-induced senescence. These senescent T-cells are known to adopt a pro-inflammatory senescence-associated secretory phenotype (SASP), releasing a cocktail of cytokines and chemokines that can directly contribute to synovitis and cartilage degradation in the joint[39]. 4.2.3 Impaired Regenerative Capacity in Progenitor Cells The strongest convergence of our causal and colocalization evidence was found in developmental progenitor cells ( Section 3.5 ). This finding is profound because it connects the OA risk locus directly to GNL3 's fundamental biological role in ribosome biogenesis, cell cycle control, and stem cell maintenance. While the specific progenitor lineage remains to be identified, this result supports a compelling hypothesis: genetically determined lower GNL3 expression could impair the regenerative potential of certain cell populations crucial for joint homeostasis. This could include mesenchymal stem cells responsible for cartilage repair, or even hematopoietic progenitors that give rise to balanced immune cell populations. Under conditions of chronic micro-trauma or inflammatory stress, this reduced regenerative capacity could lead to a net loss of tissue integrity over time, providing a plausible, cell-intrinsic mechanism for the initiation and progression of OA. 4.3 Strengths, Limitations, and Future Directions The primary strength of this study is its rigorous, multi-modal design, which constructs a comprehensive evidence chain from a GWAS locus to a causal gene and its specific, systemic, mechanistic context. By systematically integrating multiple layers of evidence—colocalization, MR, pan-cancer analysis, and, crucially, systematic single-cell causal inference—we have built a robust and multifaceted case for the role of GNL3 in OA. The convergence of our MR and colocalization results in a specific cell type is a major strength that provides a high-confidence starting point for experimental validation. However, certain limitations must be acknowledged. Our study relies on publicly available summary statistics, and while we have used the largest and most well-powered datasets available, the exposure QTL data are not always derived from the most disease-relevant tissue (i.e., synovium or cartilage). While the strong signals in immune cells and fibroblasts provide a compelling systemic link, direct validation in joint tissues is a critical next step. Future research should be directed by the specific, testable hypotheses generated by our study. First, in vitro studies using iPSC-derived neurons, T-cells, and mesenchymal stem cells with CRISPR-mediated modulation of GNL3 expression are needed to functionally validate its role in buffering against metabolic and inflammatory stressors. Second, the development of cell-type-specific conditional GNL3 knockout mouse models (e.g., in sensory neurons or T-lymphocytes) will be essential to dissect its contribution to OA pathogenesis in vivo, assessing not only joint degradation but also pain behaviors. Finally, from a translational perspective, exploring whether pharmacological agents can safely and effectively upregulate GNL3 expression could represent a novel therapeutic avenue, targeting the fundamental cellular stress pathways that appear to underpin the genetic risk of OA. 5. Conclusion In summary, our rigorous, multi-modal integrative genomic analysis establishes GNL3 as a causal gene for osteoarthritis. The weight of the evidence moves beyond a simple genetic association to propose a coherent mechanistic model where GNL3 acts as a systemic regulatory hub. Genetically determined lower expression of GNL3 appears to sensitize the host to metabolic and inflammatory insults, creating a permissive environment for immune, neural, and progenitor cell dysregulation that culminates in joint degradation. This work validates GNL3 as a high-confidence therapeutic target and reframes our understanding of OA, highlighting the potential for novel therapeutic strategies aimed at modulating systemic homeostatic and cellular stress-response pathways to preserve joint health. Declarations Author Disclosure Statement Acknowledgements Not applicable. Ethics and Consent to Participate Not applicable. Consent to Publish Not applicable. Funding Not Applicable Declaration of Competing Interest All authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Author Contributions Tianhao Qu: Conceptualization, Methodology, Formal Analysis, Data Curation, Writing – Original Draft, Visualization. Yan Zhong: Methodology, Software, Validation. Yonghuan Zhou: Investigation, Resources, Supervision. Lin Liu: Software, Visualization. Zheng Ye: Conceptualization, Project Administration, Funding Acquisition, Writing – Review & Editing, Supervision. Data Availability Statement All data analyzed in this study are publicly available, summary-level data obtained from the sources as described in the Materials and Methods section of the manuscript. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. References J.D. Steinmetz, G.T. Culbreth, L.M. Haile, Q. Rafferty, J. Lo, K.G. Fukutaki, J.A. Cruz, A.E. Smith, S.E. Vollset, P.M. Brooks, Global, regional, and national burden of osteoarthritis, 1990–2020 and projections to 2050: a systematic analysis for the Global Burden of Disease Study 2021, The Lancet Rheumatology, 5 (2023) e508-e522. S.a. Tang, C. Zhang, W.M. Oo, K. Fu, M.A. Risberg, S.M. Bierma-Zeinstra, T. Neogi, I. Atukorala, A.-M. Malfait, C. Ding, Osteoarthritis, Nature Reviews Disease Primers, 11 (2025) 1-22. J. Zhu, W. Chen, Y. Hu, Y. Qu, H. Yang, Y. Zeng, C. Hou, F. Ge, Z. Zhou, H. Song, Physical activity patterns, genetic susceptibility, and risk of hip/knee osteoarthritis: a prospective cohort study based on the UK Biobank, Osteoarthritis and cartilage, 30 (2022) 1079-1090. B.G. Faber, M. Frysz, C.G. Boer, D.S. Evans, R. Ebsim, K.A. Flynn, M. Lundberg, L. Southam, A. Hartley, F.R. Saunders, The identification of distinct protective and susceptibility mechanisms for hip osteoarthritis: findings from a genome-wide association study meta-analysis of minimum joint space width and Mendelian randomisation cluster analyses, EBioMedicine, 95 (2023). F. Gee, C.F. Clubbs, E.V. Raine, L.N. Reynard, J. Loughlin, Allelic expression analysis of the osteoarthritis susceptibility locus that maps to chromosome 3p21 reveals cis-acting eQTLs at GNL3 and SPCS1, BMC medical genetics, 15 (2014) 1-7. M.D. Gallagher, A.S. Chen-Plotkin, The post-GWAS era: from association to function, The American Journal of Human Genetics, 102 (2018) 717-730. G. Quiroga-Artigas, D. de Jong, C.E. Schnitzler, GNL3 is an evolutionarily conserved stem cell gene influencing cell proliferation, animal growth and regeneration in the hydrozoan Hydractinia, Open biology, 12 (2022) 220120. C.G. Boer, K. Hatzikotoulas, L. Southam, L. Stefánsdóttir, Y. Zhang, R.C. de Almeida, T.T. Wu, J. Zheng, A. Hartley, M. Teder-Laving, Deciphering osteoarthritis genetics across 826,690 individuals from 9 populations, Cell, 184 (2021) 4784-4818. e4717. I. Tachmazidou, K. Hatzikotoulas, L. Southam, J. Esparza-Gordillo, V. Haberland, J. Zheng, T. Johnson, M. Koprulu, E. Zengini, J. Steinberg, Identification of new therapeutic targets for osteoarthritis through genome-wide analyses of UK Biobank data, Nature genetics, 51 (2019) 230-236. E. Zeggini, K. Panoutsopoulou, L. Southam, N. Rayner, A. Day-Williams, M. Lopes, V. Boraska, T. Esko, E. Evangelou, A. Hoffman, Identification of new susceptibility loci for osteoarthritis (arcOGEN): a genome-wide association study, Lancet (London, England), 380 (2012). N.R. Gay, M. Gloudemans, M.L. Antonio, N.S. Abell, B. Balliu, Y. Park, A.R. Martin, S. Musharoff, A.S. Rao, F. Aguet, Impact of admixture and ancestry on eQTL analysis and GWAS colocalization in GTEx, Genome biology, 21 (2020) 1-20. U. Võsa, A. Claringbould, H.-J. Westra, M.J. Bonder, P. Deelen, B. Zeng, H. Kirsten, A. Saha, R. Kreuzhuber, S. Yazar, Large-scale cis-and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression, Nature genetics, 53 (2021) 1300-1310. T. Qi, Y. Wu, H. Fang, F. Zhang, S. Liu, J. Zeng, J. Yang, Genetic control of RNA splicing and its distinct role in complex trait variation, Nature genetics, 54 (2022) 1355-1363. A. Jourdon, S. Scuderi, D. Capauto, A. Abyzov, F.M. Vaccarino, PsychENCODE and beyond: transcriptomics and epigenomics of brain development and organoids, Neuropsychopharmacology, 46 (2021) 70-85. D.A. Bennett, A.S. Buchman, P.A. Boyle, L.L. Barnes, R.S. Wilson, J.A. Schneider, Religious orders study and rush memory and aging project, Journal of Alzheimer’s disease, 64 (2018) S161-S189. R. Ding, Q. Wang, L. Gong, T. Zhang, X. Zou, K. Xiong, Q. Liao, M. Plass, L. Li, scQTLbase: an integrated human single-cell eQTL database, Nucleic Acids Research, 52 (2024) D1010-D1017. V. Orrù, M. Steri, C. Sidore, M. Marongiu, V. Serra, S. Olla, G. Sole, S. Lai, M. Dei, A. Mulas, Complex genetic signatures in immune cells underlie autoimmunity and inform therapy, Nature genetics, 52 (2020) 1036-1045. J.N. Weinstein, E.A. Collisson, G.B. Mills, K.R. Shaw, B.A. Ozenberger, K. Ellrott, I. Shmulevich, C. Sander, J.M. Stuart, The cancer genome atlas pan-cancer analysis project, Nature genetics, 45 (2013) 1113-1120. Z. Ye, Y. Gao, J. Yuan, F. Chen, P. Xu, W. Liu, The Role of Gut Microbiota in Modulating Brain Structure and Psychiatric Disorders: A Mendelian Randomization Study, NeuroImage, (2025) 121292. C. Giambartolomei, D. Vukcevic, E.E. Schadt, L. Franke, A.D. Hingorani, C. Wallace, V. Plagnol, Bayesian test for colocalisation between pairs of genetic association studies using summary statistics, PLoS genetics, 10 (2014) e1004383. Y. Wu, J. Zeng, F. Zhang, Z. Zhu, T. Qi, Z. Zheng, L.R. Lloyd-Jones, R.E. Marioni, N.G. Martin, G.W. Montgomery, Integrative analysis of omics summary data reveals putative mechanisms underlying complex traits, Nature communications, 9 (2018) 918. D.A. Lawlor, R.M. Harbord, J.A. Sterne, N. Timpson, G. Davey Smith, Mendelian randomization: using genes as instruments for making causal inferences in epidemiology, Statistics in medicine, 27 (2008) 1133-1163. S. Burgess, A. Butterworth, S.G. Thompson, Mendelian randomization analysis with multiple genetic variants using summarized data, Genetic epidemiology, 37 (2013) 658-665. J. Bowden, G. Davey Smith, S. Burgess, Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression, International journal of epidemiology, 44 (2015) 512-525. J. Bowden, G. Davey Smith, P.C. Haycock, S. Burgess, Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator, Genetic epidemiology, 40 (2016) 304-314. F.P. Hartwig, G. Davey Smith, J. Bowden, Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption, International journal of epidemiology, 46 (2017) 1985-1998. J. Zhao, J. Ming, X. Hu, G. Chen, J. Liu, C. Yang, Bayesian weighted Mendelian randomization for causal inference based on summary statistics, Bioinformatics, 36 (2020) 1501-1508. Y. Jiang, Osteoarthritis year in review 2021: biology, Osteoarthritis and cartilage, 30 (2022) 207-215. C.R. Scanzello, Role of low-grade inflammation in osteoarthritis, Current opinion in rheumatology, 29 (2017) 79-85. D. Azamar-Llamas, G. Hernandez-Molina, B. Ramos-Avalos, J. Furuzawa-Carballeda, Adipokine contribution to the pathogenesis of osteoarthritis, Mediators of inflammation, 2017 (2017) 5468023. X.-X. Zhang, S.-H. He, X. Liang, W. Li, T.-F. Li, D.-F. Li, Aging, cell senescence, the pathogenesis and targeted therapies of osteoarthritis, Frontiers in Pharmacology, 12 (2021) 728100. M. Aribi, F.J. Mennechet, C. Touil-Boukoffa, The role of vitamin D as an immunomodulator, in, Frontiers Media SA, 2023, pp. 1186635. J. Yu, L. Ji, Y. Liu, X. Wang, J. Wang, C. Liu, Bone-brain interaction: mechanisms and potential intervention strategies of biomaterials, Bone Research, 13 (2025) 38. J.D. Richardson, M.R. Vasko, Cellular mechanisms of neurogenic inflammation, The Journal of pharmacology and experimental therapeutics, 302 (2002) 839-845. J. Bedson, P.R. Croft, The discordance between clinical and radiographic knee osteoarthritis: a systematic search and summary of the literature, BMC musculoskeletal disorders, 9 (2008) 116. C.J. Woolf, Central sensitization: implications for the diagnosis and treatment of pain, pain, 152 (2011) S2-S15. R.Y. Tsai, R.D. McKay, A nucleolar mechanism controlling cell proliferation in stem cells and cancer cells, Genes & development, 16 (2002) 2991-3003. H. Ma, T. Pederson, Nucleostemin: a multiplex regulator of cell-cycle progression, Trends in cell biology, 18 (2008) 575-579. O.H. Jeon, C. Kim, R.-M. Laberge, M. Demaria, S. Rathod, A.P. Vasserot, J.W. Chung, D.H. Kim, Y. Poon, N. David, Local clearance of senescent cells attenuates the development of post-traumatic osteoarthritis and creates a pro-regenerative environment, Nature medicine, 23 (2017) 775-781. Tables Table 1: Widespread Colocalization of GNL3 Expression QTLs (eQTLs) with Osteoarthritis Risk across Diverse Human Tissues. Legend: Colocalization analysis of GWAS signals for Osteoarthritis (P01779) and eQTLs for GNL3 gene expression across 44 tissues from the GTEx v8 project. The table lists tissues with a posterior probability of a shared causal variant (PP.H4) greater than 0.8, sorted in descending order of PP.H4. This widespread colocalization across multiple organ systems suggests a systemic role for GNL3 in OA pathogenesis. Tissue QTL Type Colocalization Probability (PP.H4) GWAS Top SNP eQTL Top SNP Cells Cultured Fibroblasts eQTL 0.981 rs12488461 rs6976 Testis eQTL 0.975 rs12488461 rs2028216 Muscle Skeletal eQTL 0.970 rs12488461 rs11714419 Breast Mammary Tissue eQTL 0.967 rs12488461 rs8906 Adipose Subcutaneous eQTL 0.950 rs12488461 rs2028216 Heart Left Ventricle eQTL 0.944 rs12488461 rs2028216 Esophagus Gastroesophageal J. eQTL 0.937 rs12488461 rs11714419 Nerve Tibial eQTL 0.932 rs12488461 rs6976 Colon Transverse eQTL 0.919 rs12488461 rs2028216 Pancreas eQTL 0.918 rs12488461 rs2028216 Artery Aorta eQTL 0.888 rs12488461 rs2028216 Artery Coronary eQTL 0.873 rs12488461 rs8906 Heart Atrial Appendage eQTL 0.864 rs12488461 rs1866268 Brain Cerebellum eQTL 0.839 rs12488461 rs13063160 Artery Tibial eQTL 0.834 rs12488461 rs3796353 PsychENCODE (Prefrontal Cortex) eQTL 0.898 P00954_rs2336147 rs2336147 ROSMAP (Brain Cortex) eQTL 0.823 P00954_rs7646741 rs7646741 Table 2: Strong Evidence of Colocalization between GNL3 Methylation QTLs (mQTLs) and Hip Osteoarthritis. Legend: Colocalization analysis of GWAS signals for hip osteoarthritis and total joint replacement with methylation QTLs (mQTLs) for CpG sites at the GNL3 locus. The results from different brain-related datasets (ROSMAP, FB_Brain) and blood provide evidence for an epigenetic regulatory mechanism linking genetic risk to OA. Trait Description PMID Tissue/Source Methylation Probe Colocalization Probability (PP.H4) Hip osteoarthritis 30664745 ROSMAP (Brain Cortex) cg18595196 0.948 Total joint replacement 34822786 ROSMAP (Brain Cortex) cg18595196 0.948 Hip osteoarthritis 30664745 ROSMAP (Brain Cortex) cg00845626 0.926 Total joint replacement 34822786 ROSMAP (Brain Cortex) cg00845626 0.916 Hip osteoarthritis 30664745 FB_Brain (Brain) cg11041457 0.896 Osteoarthritis 22763110 US_Blood (Whole Blood) cg08332332 0.943 Table 3: Mendelian Randomization Analysis of GNL3 Expression on Osteoarthritis and Related Traits. Legend: Results from two-sample Mendelian randomization (MR) analysis assessing the causal effect of genetically predicted GNL3 expression on various OA-related outcomes. A negative beta and an Odds Ratio (OR) less than 1 indicate a protective effect. Statistically significant results (p < 0.05) are highlighted in bold. Outcome GWAS ID Beta Std. Error P-value Odds Ratio (95% CI) Knee osteoarthritis ebi-a-GCST007090 -0.159 0.045 4.38E-04 0.85 [0.78, 0.93] Hip osteoarthritis (hospital diagnosed) ebi-a-GCST005810 -0.457 0.158 0.0038 0.63 [0.46, 0.86] Osteoarthritis; localized (PheCode 740.1) GCST90044591 -0.216 0.074 0.0035 0.81 [0.70, 0.93] Osteoarthritis GCST90134288 -0.567 0.273 0.0380 0.57 [0.33, 0.98] Osteoarthritis GCST90134279 -0.055 0.034 0.1038 0.95 [0.88, 1.01] Osteoarthritis of the hip or knee ebi-a-GCST007092 -0.092 0.095 0.3357 0.91 [0.76, 1.10] Osteoarthritis (self-reported) ebi-a-GCST005811 +0.056 0.069 0.4125 1.06 [0.92, 1.21] Note: Data was sourced from the IEU OpenGWAS database for studies with an 'ebi-a-' prefix in the GWAS ID, and from the EBI GWAS Catalog FTP server (http://ftp.ebi.ac.uk/pub/databases/gwas/summary_statistics/) for all others. Table 4: Bidirectional Causal Effects of GNL3 Splicing Events on Osteoarthritis Risk from SMR Analysis. Legend: Summary-data-based Mendelian Randomization (SMR) analysis using splicing QTLs (sQTLs) as instruments. The results reveal distinct splicing events with opposing causal effects on OA risk. Results are considered robust when the HEIDI test for pleiotropy is non-significant (p_heidi > 0.05). Trait Description Tissue Top sQTL SNP Effect Size (b_smr) Causal P-value (p_smr) Pleiotropy Test (p_heidi) Conclusion Hip or Knee OA (P02059) Thyroid rs2710323 -0.125 6.65e-05 0.134 Significant Protective Effect Osteoarthritis (P01779) Thyroid rs2710323 -0.303 2.85e-04 0.300 Significant Protective Effect Osteoarthritis (P01779) Thyroid rs9853056 +0.229 6.93e-04 0.061 Significant Risk Effect Osteoarthritis (P01779) Lung rs1010554 +0.240 7.42e-04 0.031 Association likely due to pleiotropy Table 5: Systematic single-cell analysis identifies causal cellular contexts for GNL3 in osteoarthritis. Legend: To identify the precise cellular drivers of GNL3's effect on osteoarthritis, we performed a systematic Mendelian randomization (MR) and colocalization analysis across all available single-cell eQTL datasets. The table shows all cellular contexts with a significant causal protective effect (Odds Ratio < 1) of GNL3 expression on knee OA after False Discovery Rate (FDR) correction. The corresponding Posterior Probability of a shared causal variant (PP.H4) is also shown. The convergence of a significant MR result and a high-confidence colocalization signal (PP.H4 > 0.8) in floor plate progenitors provides the strongest evidence for a specific causal pathway. Study Source Biological Domain Cellular Context Causal Effect on Knee OA (MR) FDR-corr. P Coloc. (PP.H4) Jerber et al. 2021 Progenitor Floor Plate Progenitors 0.89 (0.84 - 0.95) 0.003 0.81 Soskic et al. 2022 Immune CD4+ Memory T-Cell (Activated) 0.81 (0.73 - 0.90) 0.003 0.41 Soskic et al. 2022 Immune CD4+ Naive T-Cell (Activated) 0.57 (0.41 - 0.79) 0.007 0.22 Jerber et al. 2021 Neural Dopaminergic Neurons (Stress) 0.73 (0.60 - 0.88) 0.007 0.71 Jerber et al. 2021 Neural Astrocyte-like Cells (Stress) 0.80 (0.68 - 0.93) 0.024 0.34 Table 6: Key Pleiotropic Effects of the GNL3 Locus Identified Through GWAS-GWAS Colocalization. Legend: Representative results from GWAS-GWAS colocalization analysis showing shared genetic signals between body composition traits (a proxy for the OA-associated locus) and key systemic traits. Only the most significant findings for major metabolic, endocrine, and neurological traits are presented. PP.H4 indicates the posterior probability of a shared causal variant. Anchor Trait (PMID) Colocalizing Trait (PMID) Colocalization Probability (PP.H4) Impedance of leg (right) (31768069) Body mass index (BMI) (30124842) 0.998 Impedance of leg (left) (31768069) Waist-to-hip ratio adjusted for BMI (25673412) 0.990 Impedance of leg (right) (31768069) 25-hydroxyvitamin D (32242144) 0.998 Impedance of leg (right) (31768069) Neuroticism (30867560) 0.979 Impedance of leg (left) (31768069) Hematocrit (32888493) 0.796 Table 7: Causal Effects of Genetically Predicted Immune Cell Traits on Osteoarthritis Risk from Mendelian Randomization. Legend: Representative results from two-sample MR analysis showing causal associations between immune phenotypes and OA risk. Results with p < 0.01 are shown. OR (Odds Ratio) 1 indicates a risk-increasing effect. Exposure (Immune Cell Trait) Outcome Method OR (95% CI) P-value Protective Associations CD14+ CD16- monocyte Absolute Count Knee OA Wald ratio 0.44 (0.27-0.70) 6.7e-04 Naive CD4+ T cell Absolute Count Localized OA Wald ratio 0.78 (0.69-0.89) 1.5e-04 IgD- CD27- B cell %lymphocyte Hip OA Wald ratio 0.66 (0.54-0.80) 3.0e-05 CD25++ CD4+ T cell Absolute Count Localized OA Wald ratio 0.89 (0.82-0.96) 3.8e-03 Central Memory CD4+ T cell %CD4+ T cell Localized OA Wald ratio 0.83 (0.73-0.94) 3.6e-03 Risk-Increasing Associations CD86 on granulocyte Knee OA Wald ratio 2.07 (1.36-3.14) 6.7e-04 CD39 on granulocyte Knee OA Wald ratio 2.05 (1.36-3.11) 6.7e-04 CCR2 on granulocyte Knee OA Wald ratio 1.71 (1.25-2.32) 6.7e-04 CD19 on switched memory B cell Knee OA Wald ratio 1.54 (1.12-2.13) 8.2e-03 Effector Memory CD8+ T cell Abs. Count Knee & Hip OA Wald ratio 1.33 (1.05-1.68) 1.7e-02 Additional Declarations No competing interests reported. Supplementary Files SupplementMaterials1DataSource.xlsx SupplementMaterials2eQTLmQTL.xlsx SupplementMaterials3sQTLcoloc.xlsx SupplementMaterials4sceQTL.xlsx SupplementMaterials5GWASGWAScoloc.xlsx SupplementMaterials6ImmunophenotypeMR.xlsx SupplementFigures.docx Cite Share Download PDF Status: Published Journal Publication published 16 Feb, 2026 Read the published version in Human Genomics → Version 1 posted Editorial decision: Revision requested 19 Dec, 2025 Reviews received at journal 06 Dec, 2025 Reviews received at journal 18 Nov, 2025 Reviewers agreed at journal 17 Nov, 2025 Reviewers agreed at journal 17 Nov, 2025 Reviewers agreed at journal 08 Oct, 2025 Reviewers invited by journal 18 Sep, 2025 Editor assigned by journal 17 Sep, 2025 Submission checks completed at journal 17 Sep, 2025 First submitted to journal 16 Sep, 2025 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-7630649","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":521738708,"identity":"1bd757b5-8663-4a4d-bdbe-8e36bd68e227","order_by":0,"name":"Tianhao Qu","email":"","orcid":"","institution":"People's Hospital of Dongxihu District","correspondingAuthor":false,"prefix":"","firstName":"Tianhao","middleName":"","lastName":"Qu","suffix":""},{"id":521738709,"identity":"e5ffadf2-0659-404d-920f-adf8b490f67c","order_by":1,"name":"Yan Zhong","email":"","orcid":"","institution":"People's Hospital of Dongxihu District","correspondingAuthor":false,"prefix":"","firstName":"Yan","middleName":"","lastName":"Zhong","suffix":""},{"id":521738710,"identity":"d71be511-4641-407c-8a64-9e9daa70c017","order_by":2,"name":"Yonghuan Zhou","email":"","orcid":"","institution":"People's Hospital of Dongxihu District","correspondingAuthor":false,"prefix":"","firstName":"Yonghuan","middleName":"","lastName":"Zhou","suffix":""},{"id":521738711,"identity":"7a54894c-0f04-49d8-bc71-bb66799acb4d","order_by":3,"name":"Lin Liu","email":"","orcid":"","institution":"People's Hospital of Dongxihu District","correspondingAuthor":false,"prefix":"","firstName":"Lin","middleName":"","lastName":"Liu","suffix":""},{"id":521738712,"identity":"562f9c12-86dc-45a0-8b03-d6b401cd0678","order_by":4,"name":"Zheng Ye","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7UlEQVRIiWNgGAWjYDACCSjNz8DABqIZG4jWItlAshaDA8RqkZ/dfOzhF4bDcsbnlz97zMNgI7vhAPOzB/i0MM45lm4sw3DY2OzGg3RjHoY04w0H2MwN8Glhlsgxk5ZgOJy47caBY9I8QMaGAzxsEvi0sEnkfwNpqd8842AbUMt/wlp4JHLYJD8wHE4w4G9mA2o5QFiLhESamTQDQ7rhjBtsbJJzDJKNZx5mM8OrRX5G8jPJHwzW8vz9x59JvKmwk+073vwMrxYQYOb91wy0LwHIBAUVMyH1QMD4g6EOmGIOEKF0FIyCUTAKRiQAANe7Q4dxMxHeAAAAAElFTkSuQmCC","orcid":"","institution":"Guangzhou University","correspondingAuthor":true,"prefix":"","firstName":"Zheng","middleName":"","lastName":"Ye","suffix":""}],"badges":[],"createdAt":"2025-09-16 12:53:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7630649/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7630649/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s40246-026-00920-3","type":"published","date":"2026-02-16T15:57:32+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":92405004,"identity":"62d8dae8-88f5-48a7-821f-9973c8dd7d05","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"jpg","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":547314,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1IntegrativeGenomicsWorkflow01.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/ee9b888a02a3d04a61d468f1.jpg"},{"id":92405006,"identity":"865bb660-e124-41d2-ae8a-1c0828e8d3f6","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"jpg","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":838197,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.SignificantCorrelationHeatmap01.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/c781bf62c039e2c024dede9e.jpg"},{"id":92405008,"identity":"c1c3a265-5b66-4c11-9153-b939fc96c776","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":80291,"visible":true,"origin":"","legend":"","description":"","filename":"version5.docx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/9ece9df7864c0ff346bf6add.docx"},{"id":92405007,"identity":"2fbccad7-6270-4358-bfce-924a117556b9","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"json","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7538,"visible":true,"origin":"","legend":"","description":"","filename":"bca0fd8dab2a4adc94eeea14174f66b7.json","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/92047099e24aa0ff38fa4d33.json"},{"id":92406262,"identity":"f282bc2c-a12c-484a-ad14-06cf93d48aa6","added_by":"auto","created_at":"2025-09-29 11:14:17","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2820411,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/94068b712f09f0f60cf3cded.docx"},{"id":92405012,"identity":"6a31027c-3981-412f-a12c-99b203077916","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17269,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials1DataSource.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/066ad1e6b13903ea4d9ce8c2.xlsx"},{"id":92406643,"identity":"ae5d4577-ada9-4ea4-887f-ad19259e5efb","added_by":"auto","created_at":"2025-09-29 11:22:17","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":162761,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials2eQTLmQTL.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/ec262a76104da9cabbc17209.xlsx"},{"id":92407584,"identity":"165e7949-a843-4c0e-968b-854b3e2405cb","added_by":"auto","created_at":"2025-09-29 11:30:17","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":83560,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials3sQTLcoloc.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/8e20e0e16331375685f7b32e.xlsx"},{"id":92406644,"identity":"d9565979-e231-46a0-ae7e-1c29939aee16","added_by":"auto","created_at":"2025-09-29 11:22:17","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":284352,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials4sceQTL.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/deef2f24832253b812165ab8.xlsx"},{"id":92406258,"identity":"3b914f66-4568-4024-b6e6-a033a047f9ee","added_by":"auto","created_at":"2025-09-29 11:14:17","extension":"xlsx","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17203,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials5GWASGWAScoloc.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/ddaf7479f6bdf5d823e7992a.xlsx"},{"id":92405029,"identity":"9cc8c091-8a55-45ad-b930-b6fad7f38a2f","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"xlsx","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1038169,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials6ImmunophenotypeMR.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/aa3e91c3a20cc4c0c573f287.xlsx"},{"id":92405028,"identity":"6786b0cd-528b-43e0-95b0-bf86eac0757e","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"xml","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":153683,"visible":true,"origin":"","legend":"","description":"","filename":"bca0fd8dab2a4adc94eeea14174f66b71enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/e4e8f00f63b950a394985716.xml"},{"id":92405022,"identity":"01f2e650-a76e-4c04-af5c-b191d3e8dead","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"jpg","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":547314,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1IntegrativeGenomicsWorkflow01.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/4ea52482dc4d0a67d267cacc.jpg"},{"id":92405018,"identity":"6bf30d83-65fe-487b-ba0e-001e70837848","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"jpg","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":838197,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.SignificantCorrelationHeatmap01.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/7f0a73c2d4087845449a0805.jpg"},{"id":92406260,"identity":"48cf8770-5ca0-4785-b109-a0ec3c6c0fd5","added_by":"auto","created_at":"2025-09-29 11:14:17","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":241278,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure1IntegrativeGenomicsWorkflow01.png","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/b9a5e8bab2a4f31614c56b18.png"},{"id":92406264,"identity":"ea4c4a9c-6d54-46fd-a9e8-66fc6b1efa37","added_by":"auto","created_at":"2025-09-29 11:14:17","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":222652,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure2.SignificantCorrelationHeatmap01.png","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/4f3932507c7dc73cc756af08.png"},{"id":92405031,"identity":"87c4afbd-c942-4174-ba88-548ddf8cd31a","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"xml","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":154530,"visible":true,"origin":"","legend":"","description":"","filename":"bca0fd8dab2a4adc94eeea14174f66b71structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/cd4b9c75c8a64d46d7933c5c.xml"},{"id":92405030,"identity":"8c3107a7-5825-43d4-a972-a8d8c1bc52bb","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"html","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":165595,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/cbea80aa4f557bb83299600b.html"},{"id":92405003,"identity":"59282903-1aef-426f-bbe7-369cf757f1fe","added_by":"auto","created_at":"2025-09-29 11:06:16","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":547314,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA Systematic Integrative Genomics Workflow to Identify and Validate \u003c/strong\u003e\u003ccode\u003eGNL3\u003c/code\u003e\u003cstrong\u003e as a Causal Gene for Osteoarthritis. \u003c/strong\u003eThe workflow comprises four main stages. \u003cstrong\u003e(1) Data Integration and Candidate Gene Prioritization:\u003c/strong\u003e Summary-level data from osteoarthritis (OA) genome-wide association studies (GWAS) and multi-tissue molecular quantitative trait loci (xQTL) are integrated. Systematic colocalization analysis is then employed to prioritize \u003ccode\u003eGNL3\u003c/code\u003e as the lead candidate causal gene at the locus. \u003cstrong\u003e(2) Causal Inference and Mechanistic Validation:\u003c/strong\u003e Two-sample Mendelian randomization (MR) establishes a causal, protective effect of \u003ccode\u003eGNL3\u003c/code\u003e expression on OA risk. Independent analysis of pan-cancer TCGA data validates a negative correlation between \u003ccode\u003eGNL3\u003c/code\u003e promoter methylation and its expression, revealing a plausible epigenetic regulatory mechanism. \u003cstrong\u003e(3) Cellular Context and Systemic Impact:\u003c/strong\u003e The cellular and systemic context of \u003ccode\u003eGNL3\u003c/code\u003eregulation is explored. Single-cell eQTL analysis pinpoints the genetic regulation of \u003ccode\u003eGNL3\u003c/code\u003e to specific cell types and activation states. Further analysis establishes the \u003ccode\u003eGNL3\u003c/code\u003e locus as a pleiotropic hub associated with multiple systemic traits relevant to OA. \u003cstrong\u003e(4) Unifying Mechanistic Hypothesis:\u003c/strong\u003e Finally, the evidence converges to support a unifying mechanistic hypothesis. This model posits that \u003ccode\u003eGNL3\u003c/code\u003e acts as a systemic regulatory hub where genetically determined expression modulates metabolic and immune homeostasis, thereby influencing OA susceptibility and validating \u003ccode\u003eGNL3\u003c/code\u003e as a high-confidence therapeutic target.\u003c/p\u003e","description":"","filename":"Figure1IntegrativeGenomicsWorkflow01.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/11de2c574369f15422e16315.jpg"},{"id":92406254,"identity":"4c9c9629-e88d-4d0a-90f2-e38b49ab306b","added_by":"auto","created_at":"2025-09-29 11:14:17","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":838197,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePan-cancer analysis reveals a consistent negative correlation between GNL3 promoter methylation and gene expression. \u003c/strong\u003eThe heatmap illustrates the Pearson correlation between GNL3 mRNA expression and DNA methylation levels across various cancer types from The Cancer Genome Atlas (TCGA), analyzed separately for primary tumor and adjacent normal tissues. The analysis was performed for two individual CpG sites (cg18595196 and cg11041457) identified in our mQTL colocalization analysis, and for the average methylation Beta-value of 17 probes located in the GNL3 promoter region (labeled \"Value\" or \"Avg.Beta\"). The color of each point represents the Pearson's R correlation coefficient, with blue indicating a negative correlation and red indicating a positive correlation. The size of each point is proportional to the statistical significance (-log10 P-value). Only correlations with a p-value \u0026lt; 0.05 are shown. The consistent and strong negative correlation observed in numerous tumor types provides functional evidence that promoter methylation acts as a repressive mechanism for GNL3 expression.\u003c/p\u003e","description":"","filename":"Figure2.SignificantCorrelationHeatmap01.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/a21dab2e10934c57d21120b0.jpg"},{"id":103251380,"identity":"331351a8-cfc1-4cd4-ada7-d9ed6456bb15","added_by":"auto","created_at":"2026-02-23 16:08:32","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4016727,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/adecbf76-c218-4f5a-afcf-eabdffbf2c20.pdf"},{"id":92405002,"identity":"8d69b8eb-b31a-4dc8-8e43-ff1c22dff728","added_by":"auto","created_at":"2025-09-29 11:06:16","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":17269,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials1DataSource.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/5df6b5f2b27b0f48232fcc27.xlsx"},{"id":92406255,"identity":"3314906b-3e30-4650-b7a8-fdde2c34b090","added_by":"auto","created_at":"2025-09-29 11:14:17","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":162761,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials2eQTLmQTL.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/0d218f23672a725537a4d34a.xlsx"},{"id":92405011,"identity":"b38db90f-9a43-4774-992d-b4e7acdb3b5d","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":83560,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials3sQTLcoloc.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/165f28ca598d181690dc023f.xlsx"},{"id":92405015,"identity":"a7822426-421e-4492-a7e0-1271f9b4f768","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":284352,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials4sceQTL.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/81a3d675ffb93fa12295b224.xlsx"},{"id":92405009,"identity":"1e114313-7e99-4aaa-a91e-67da45eea1a6","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":17203,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials5GWASGWAScoloc.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/214ef1a40dea0acbaee7b3ef.xlsx"},{"id":92405026,"identity":"7a1cb8f3-c32f-4a91-a79f-eb7284c8e2d2","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":1038169,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementMaterials6ImmunophenotypeMR.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/5f884f3002e9c8a41ab37588.xlsx"},{"id":92405023,"identity":"d3e47953-03f3-4c3b-97d2-0e5a24a5afb5","added_by":"auto","created_at":"2025-09-29 11:06:17","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":2820411,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7630649/v1/a977662d9f772f927ffb1702.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Integrative Genomics Establishes GNL3 as a Pleiotropic Hub and Causal Gene for Osteoarthritis","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eOsteoarthritis (OA), the most common form of arthritis, is a leading cause of chronic pain and disability globally, imposing a significant socioeconomic burden[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Pathologically, OA involves a complex interplay of articular cartilage degradation, synovial inflammation, and subchondral bone remodeling, driven by genetic, metabolic, and mechanical factors[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Genome-wide association studies (GWAS) have successfully identified hundreds of genetic loci associated with OA susceptibility[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. However, the majority of these associated variants are located in non-coding regions, making the identification of their target effector genes and downstream functional mechanisms a critical bottleneck in translating genetic discoveries into clinical applications.\u003c/p\u003e\u003cp\u003eA prominent risk locus on chromosome 3q21 has been consistently implicated in OA across multiple large-scale GWAS[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. This region is gene-dense, containing several plausible candidates such as \u003cem\u003eGNL3\u003c/em\u003e, \u003cem\u003eITIH1\u003c/em\u003e, and \u003cem\u003ePBRM1\u003c/em\u003e. Differentiating the causal gene(s) from bystander genes within the same linkage disequilibrium (LD) block is a non-trivial task that requires integrative functional genomic approaches[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Pioneering work by Gee et al. provided the first functional evidence at this locus, using allelic expression imbalance (AEI) analysis in joint tissues from OA patients. Their study revealed that the OA-associated allele at this locus correlated with significantly lower expression of both GNL3 and SPCS1[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. While these initial findings were crucial, GNL3's definitive causal role, the full extent of its systemic influence, and the precise molecular mechanisms through which it impacts OA pathogenesis remained to be elucidated. GNL3 is a highly conserved nucleolar GTP-binding protein known to play fundamental roles in ribosome biogenesis, cell cycle regulation, and stem cell maintenance[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], but a comprehensive understanding of its function in a complex degenerative disease like OA requires a broader, more systematic approach.\u003c/p\u003e\u003cp\u003e\u003cb\u003eHowever, these foundational findings remained correlational, and two critical questions were left unanswered: is the effect of\u003c/b\u003e \u003cb\u003eGNL3\u003c/b\u003e \u003cb\u003eon OA truly causal, and in which specific cellular contexts and states does this genetic regulation exert its pathogenic influence?\u003c/b\u003e \u003cem\u003eGNL3\u003c/em\u003e, a highly conserved nucleolar GTP-binding protein, is known to play fundamental roles in ribosome biogenesis and cell cycle regulation, but its function within the systemic pathology of a complex degenerative disease like OA has remained poorly understood.\u003c/p\u003e\u003cp\u003e\u003cb\u003eTo bridge the gap from genetic association to causal mechanism, we designed a systematic, multi-omics strategy(Fig.\u0026nbsp;1) to test the central hypothesis that shared genetic variants driving OA risk do so by causally modulating\u003c/b\u003e \u003cb\u003eGNL3\u003c/b\u003e \u003cb\u003eexpression in specific, disease-relevant cellular contexts.\u003c/b\u003e We leveraged large-scale GWAS and molecular QTL data to first build a robust, multi-layered evidence chain for \u003cem\u003eGNL3\u003c/em\u003e's causal role through comprehensive colocalization and Mendelian Randomization (MR). \u003cb\u003eCrucially, we extended this framework to the single-cell level to dissect its function in distinct cell types and activation states.\u003c/b\u003e This integrative approach allowed us not only to formally validate \u003cem\u003eGNL3\u003c/em\u003e as the causal effector gene but also to uncover its role as a pleiotropic hub that integrates systemic metabolic and immune signals. Our findings converge to propose a unifying model where genetically determined \u003cem\u003eGNL3\u003c/em\u003e expression dictates OA susceptibility by modulating cellular responses to stress, thereby providing a new mechanistic framework for the disease.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003ch4\u003e2.1 Data Sources\u003c/h4\u003e\n\u003cp\u003eThis study is based entirely on the analysis of publicly available, summary-level data from large-scale genome-wide association studies (GWAS) and molecular quantitative trait locus (QTL) consortia. All data utilized were from studies that had previously obtained the necessary patient consent and ethical approvals(\u003cstrong\u003eSupplement Materials 1\u003c/strong\u003e).\u003c/p\u003e\n\u003ch5\u003e2.1.1 GWAS Summary Statistics for Osteoarthritis\u003c/h5\u003e\n\u003cp\u003eSummary statistics for osteoarthritis (OA) and its joint-specific subtypes were obtained from several large-scale GWAS. All included cohorts consisted of individuals of European ancestry. Our primary analysis leveraged summary data from the most extensive OA genetics meta-analysis to date, published by Boer et al. (2021)[8]. This study provided robust genetic association data for hip osteoarthritis (N = 353,388) and for a combined knee and/or hip osteoarthritis phenotype (N = 490,345). To ensure comprehensive coverage, we also incorporated data from the UK Biobank, as analyzed by Tachmazidou et al. (2019)[9], which furnished summary statistics for hip osteoarthritis (N = 393,873) and for osteoarthritis of the hip or knee (N = 417,596). For complementary purposes, our analyses also included data from the foundational OA GWAS conducted by Zeggini et al. (2012)[10], which comprised 18,419 individuals.\u003c/p\u003e\n\u003ch5\u003e2.1.2 Molecular Quantitative Trait Locus (QTL) Data\u003c/h5\u003e\n\u003cp\u003eA comprehensive collection of molecular QTL data was assembled to interrogate the functional consequences of OA-associated variants.\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cstrong\u003eBulk-Tissue eQTL and sQTL Data:\u003c/strong\u003e Expression QTL (eQTL) and splicing QTL (sQTL) data for 21 tissues—including adipose tissue, cultured fibroblasts, muscle, nerve, and whole blood—were sourced from the Genotype-Tissue Expression (GTEx) project v8[11]. For a more powered analysis in blood, we also utilized cis-eQTL data from the eQTLGen consortium[12] (n = 31,684).\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eBrain-Specific QTL Data:\u003c/strong\u003e To enrich our analysis of neurological tissues, we integrated several brain-specific datasets. These included cis-eQTL data from the BrainMeta study (cortex, n = 2,865)[13], the PsychENCODE project (prefrontal cortex, n = 1,387)[14], and the GTEx-brain-std study (n=233). Furthermore, methylation QTL (mQTL) data for the brain cortex were obtained from the Religious Orders Study and Memory and Aging Project (ROSMAP)[15], specifically the ROSMAP_CMC dataset accessed via the Brain xQTL Serve.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eSingle-Cell eQTL Data:\u003c/strong\u003e To explore gene regulation at cellular resolution, we utilized the scQTLbase portal[16], which aggregates sc-eQTLs from numerous datasets covering a wide range of cell types and states. This resource enabled the investigation of cell-type-specific regulatory effects.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch5\u003e2.1.3 GWAS Summary Statistics for Immune Cell Phenotypes\u003c/h5\u003e\n\u003cp\u003eFor our Mendelian Randomization analysis of the immune system, we leveraged summary statistics from a large-scale GWAS of 731 immune phenotypes (Orrù et al., 2020)[17]. The source study was based on a cohort of 3,757 individuals of European ancestry from the SardiNIA Project. In that study, a comprehensive panel of immune traits was quantified using flow cytometry, encompassing absolute and relative cell counts, median fluorescence intensities (MFIs), and morphological parameters. The publicly available summary data, which we accessed via the IEU OpenGWAS database, provided the genetic instruments for our investigation.\u003c/p\u003e\n\u003ch5\u003e2.1.4 TCGA Data for Pan-Cancer Analysis\u003c/h5\u003e\n\u003cp\u003eTo investigate the expression and methylation patterns of \u003cem\u003eGNL3\u003c/em\u003e across various cancer types, we utilized publicly available data from The Cancer Genome Atlas (TCGA)\u0026nbsp;project[18]. All processed data were downloaded from the UCSC Xena functional genomics browser, which provides uniformly processed TCGA datasets. Specifically, we retrieved pan-cancer gene expression data (RNA-Seq, quantified as log2(norm_count+1)) and DNA methylation data (Beta-values from the Illumina Human Methylation 450k platform) for the TCGA Pan-Cancer (PANCAN) cohort. Data for the \u003cem\u003eGNL3\u003c/em\u003e gene were extracted across all available primary tumor (sample type code: 01) and adjacent solid tissue normal (sample type code: 11) samples for differential analysis.\u003c/p\u003e\n\u003ch4\u003e2.2 Statistical Analysis\u003c/h4\u003e\n\u003cp\u003eAll statistical analyses were conducted using R (version 4.3.0). The \"TwoSampleMR\"[19] and \"coloc\"[20] packages were central to the Mendelian randomization and colocalization analyses, respectively.\u003c/p\u003e\n\u003ch5\u003e2.2.1 Colocalization Analysis\u003c/h5\u003e\n\u003cp\u003eTo test for shared causal variants between OA risk loci and molecular QTLs, we employed two complementary approaches.\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cstrong\u003eBulk-tissue colocalization:\u003c/strong\u003e The Bayesian colocalization method, COLOC, was used to calculate the posterior probability (PP) for five mutually exclusive hypotheses: H0 (no association), H1 (association with OA only), H2 (association with QTL only), H3 (distinct causal variants), and H4 (a shared causal variant). We used the default conservative prior probabilities (p1 = 1x10⁻⁴, p2 = 1x10⁻⁴, p12 = 1x10⁻⁵). Strong evidence for colocalization was defined by a high posterior probability for a shared variant (PP4 ≥ 0.8). To avoid false positives from low-power studies, SNPs with p \u0026gt; 1x10⁻⁴ in either the GWAS or QTL dataset were excluded from the analysis.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eSummary-based Mendelian Randomization (SMR) \u0026amp; HEIDI Test:\u003c/strong\u003e As a complementary approach, Summary-based Mendelian Randomization (SMR)[21] was used to identify pleiotropic associations between gene expression (and other molecular traits) and OA risk, using the top QTL variant as an instrument. A Benjamini-Hochberg false discovery rate (FDR \u0026lt; 0.05) was applied to correct for multiple testing. To distinguish true pleiotropy from confounding by linkage disequilibrium (LD), we performed the Heterogeneity in Dependent Instruments (HEIDI) test. Loci with a HEIDI p-value \u0026gt; 0.05 were considered to have a low probability of heterogeneity, supporting a shared causal variant.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch5\u003e2.2.2 Mendelian Randomization (MR) Analysis\u003c/h5\u003e\n\u003cp\u003eWe conducted three distinct sets of MR analyses to investigate causal relationships from different perspectives.\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cstrong\u003eCausal Effect of GNL3 Expression on Osteoarthritis:\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;To assess the direct causal effect of GNL3 expression on OA risk, we used summary statistics from the eQTLGen consortium for cis-eQTLs of GNL3 in whole blood. Since the lead eQTL for GNL3 represented a single, strong instrumental variable after LD clumping, the causal effect was estimated using the Wald Ratio method[22]. As this analysis utilized a single instrument, sensitivity analyses to assess horizontal pleiotropy (e.g., MR-Egger[22]) were not applicable.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eCausal Effects of Immune Cells on Osteoarthritis:\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;To systematically investigate the causal roles of immune cells in OA, we performed a broader MR analysis using 729 immune cell phenotypes as exposures and four distinct OA datasets as outcomes.\u003cul\u003e\n \u003cli\u003e\u003cstrong\u003eInstrument Selection:\u003c/strong\u003e For each immune phenotype, we selected independent, genome-wide significant SNPs (p \u0026lt; 5 × 10⁻⁸) to serve as instrumental variables. To ensure independence, we performed LD clumping using a strict r² threshold of \u0026lt; 0.001 within a 10,000 kb window, based on the 1000 Genomes European reference panel.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003ePrimary and Sensitivity Analyses:\u003c/strong\u003e Our choice of MR method depended on the number of available instrumental variables for each immune phenotype. For phenotypes with multiple independent SNPs, the \u003cstrong\u003eInverse-Variance Weighted (IVW)\u003c/strong\u003e\u003cstrong\u003e[23]\u003c/strong\u003e method was used as the primary method to estimate the causal effect. To assess the robustness of these findings, we employed a suite of sensitivity analyses, including \u003cstrong\u003eMR-Egger regression\u003c/strong\u003e\u003cstrong\u003e[24]\u003c/strong\u003e, the \u003cstrong\u003eWeighted Median\u003c/strong\u003e method[25], the \u003cstrong\u003eWeighted Mode\u003c/strong\u003e method[26], and \u003cstrong\u003eBayesian Weighted Mendelian Randomization\u003c/strong\u003e\u003cstrong\u003e[27]\u003c/strong\u003e. In cases where only a single SNP remained as a valid instrument after clumping, the causal effect was estimated using the \u003cstrong\u003eWald Ratio\u003c/strong\u003e method.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eHeterogeneity and Pleiotropy Assessment:\u003c/strong\u003e For instruments comprising two or more SNPs, we conducted several diagnostic tests. Heterogeneity was assessed using \u003cstrong\u003eCochran's Q statistic\u003c/strong\u003e. Directional horizontal pleiotropy was evaluated using the \u003cstrong\u003eMR-Egger intercept test\u003c/strong\u003e, where a p-value \u0026lt; 0.05 indicates significant pleiotropy. We also generated scatter plots and leave-one-out plots to visually inspect the influence of individual SNPs.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003e\u003cem\u003eSingle-cell MR and colocalization analysis to identify cellular mechanisms of GNL3\u003c/em\u003e\u003c/strong\u003e :\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eTo dissect the cell-type-specific causal effects of \u003cem\u003eGNL3\u003c/em\u003e expression, we extended the MR framework to single-cell eQTL (sc-eQTL) data. Instrumental variables were selected and harmonized following the same procedure described for the immune cell screen. Causal effects were estimated using the IVW method for multi-SNP instruments and the Wald Ratio for single-SNP instruments.The robustness of our single-cell findings was confirmed using the sensitivity analyses described previously (e.g., MR-Egger intercept). Additionally, we applied the \u003cstrong\u003eSteiger test\u003c/strong\u003e to verify the causal direction from gene expression to OA risk and rule out reverse causation. To account for multiple testing across the numerous cellular contexts, we applied the \u003cstrong\u003eFalse Discovery Rate (FDR)\u003c/strong\u003e correction, with a corrected p-value (FDR \u0026lt; 0.05) considered statistically significant. Finally, to test for a shared causal variant between the sc-eQTL and OA GWAS signals, we performed a Bayesian colocalization analysis for each cellular context, defining strong evidence as PP.H4 ≥ 0.8.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch5\u003e2.2.3 Functional Annotation of Genetic Variants\u003c/h5\u003e\n\u003cp\u003eTo functionally characterize genetic variants of interest, we used the ANNOVAR software. Variants were annotated using a comprehensive set of databases, including gene-based annotations (RefSeq, Ensembl), population frequency data (gnomAD, 1000 Genomes), and data on clinical and functional significance (ClinVar, dbNSFP, GWAS Catalog).\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003e\u003cstrong\u003e3.1. Widespread eQTL and mQTL Colocalization Provides Robust Evidence for GNL3\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo identify the most likely effector gene at the 3q21 locus, we performed a comprehensive colocalization analysis across multiple osteoarthritis (OA) GWAS datasets and molecular QTLs. Among all genes in the region, GNL3 was uniquely distinguished by widespread, consistent, and robust evidence of colocalization, firmly establishing it as the lead candidate gene. A detailed summary of all colocalization results is provided in \u003cstrong\u003eSupplementary Materials 2\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;(Figure S1-Figure S5)\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur investigation began with the foundational OA GWAS by Zeggini et al. (2012; P01779), which revealed exceptionally strong colocalization signals between the OA risk locus and GNL3 expression QTLs (eQTLs). This colocalization was observed in a broad array of tissues, with 17 diverse tissues showing strong evidence of a shared causal variant (PP.H4 \u0026gt; 0.8), most notably in Cultured Fibroblasts (PP.H4 = 0.98), Testis (PP.H4 = 0.97), and Skeletal Muscle (PP.H4 = 0.97). The critical question was whether this signal was robust. We confirmed this by replicating the analysis using larger, more recent GWAS datasets. In the UK Biobank analysis by Tachmazidou et al. (2019), strong evidence of colocalization (PP.H4 \u0026gt; 0.8) for GNL3 eQTLs was consistently observed for both hip OA (P02060) and combined hip/knee OA (P02059) in many of the same key tissues, such as Breast Mammary Tissue and Testis\u003cstrong\u003e(Table 1)\u003c/strong\u003e. This pattern of shared signals across independent cohorts was further validated using the latest large-scale meta-analysis by Boer et al. (2021; P02070, P02071). While the posterior probabilities in this latter dataset did not always strictly exceed the 0.8 threshold in every tissue, the signals remained consistently high (e.g., PP.H4 ≈ 0.7), reinforcing the overall conclusion. This slight variation, likely due to differences in cohort composition and statistical power, demonstrates the stability of the signal rather than contradicting it. The remarkable consistency of this eQTL colocalization across different OA study populations and a diverse panel of tissues provides powerful evidence that this association is not a result of chance or confounding by linkage disequilibrium.\u003c/p\u003e\n\u003cp\u003eTo explore upstream regulatory mechanisms, we found that the OA risk signal also strongly and consistently colocalized with methylation QTLs (mQTLs) that control DNA methylation at CpG sites within the GNL3 locus. This finding was also robust across the different OA datasets. For instance, the mQTLs for probes cg18595196, cg00845626, and cg11041457 all robustly colocalized with hip and knee OA risk from both the Tachmazidou et al. (2019) and Boer et al. (2021) studies, with posterior probabilities consistently exceeding 0.85\u003cstrong\u003e(Table 2)\u003c/strong\u003e. This suggests that the genetic effect on OA is likely mediated, at least in part, through the epigenetic modification of the GNL3 gene, which in turn regulates its transcription.\u003c/p\u003e\n\u003cp\u003eCollectively, the highly consistent colocalization results across multiple independent OA datasets, diverse tissue eQTLs, and distinct regulatory layers (mQTLs) build a powerful and solid evidentiary foundation, prioritizing GNL3 as the primary functional gene at this locus.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 GNL3 Expression is Negatively Correlated with Promoter Methylation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further elucidate the regulatory mechanism linking DNA methylation to GNL3 expression, we performed a pan-cancer correlation analysis using data from The Cancer Genome Atlas (TCGA). We specifically examined the correlation between GNL3 mRNA expression and methylation levels (Beta-values). Our analysis focused on two key individual CpG sites, cg18595196 and cg11041457, which were identified as significant in our mQTL colocalization analysis, as well as the average methylation value across 17 CpG sites within the GNL3 locus. As shown in the gene schematic (\u003cstrong\u003eFigure S1, Figure S6\u003c/strong\u003e), these 17 probes are strategically clustered within key regulatory regions of the GNL3 gene, including its promoter, the 5' UTR, and the first exon. Many of these probes are also located within or adjacent to a CpG island, positioning them to directly influence its transcriptional activity.\u003c/p\u003e\n\u003cp\u003eThe analysis revealed a consistent and statistically significant negative correlation between GNL3 expression and the methylation levels of both the individual CpG sites and their aggregate average across a wide spectrum of cancer types (\u003cstrong\u003eFigure\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e2,\u0026nbsp;\u003c/strong\u003e \u003cstrong\u003eFigure S7\u003c/strong\u003e). This inverse relationship, where higher methylation is associated with lower gene expression, was observed for all three methylation metrics and was particularly pronounced in numerous cancers, including Liver Hepatocellular Carcinoma (LIHC), Stomach Adenocarcinoma (STAD), Lung Adenocarcinoma (LUAD), and Lung Squamous Cell Carcinoma (LUSC). This pan-cancer evidence provides strong, independent support for the hypothesis derived from our mQTL analysis: that DNA methylation within the GNL3 promoter acts as a repressive regulatory mechanism, contributing to the silencing of its expression. This finding mechanistically links the genetic variants (mQTLs) to a functional outcome (GNL3 expression levels), strengthening the overall causal argument.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3 Mendelian Randomization Confirms a Causal, Protective Role for GNL3 Expression\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHaving established GNL3 as the likely causal gene through colocalization, we performed a two-sample Mendelian Randomization (MR) analysis to formally test for a causal relationship and determine the direction of its effect on OA risk. Using genetically predicted GNL3 expression in whole blood as the exposure, our analysis revealed a significant and consistent protective effect against major forms of osteoarthritis. Specifically, a genetically predicted one-standard-deviation increase in GNL3 expression was associated with a significantly lower risk of knee osteoarthritis (OR = 0.85, 95% CI [0.78, 0.93], p = 4.4e-4) and hospital-diagnosed hip osteoarthritis (OR = 0.63, 95% CI [0.46, 0.86], p = 3.8e-3). Furthermore, a significant protective effect was also observed for a broader, localized OA phenotype (PheCode 740.1; OR = 0.81, 95% CI [0.70, 0.93], p = 3.5e-3). While not all tested OA-related traits reached statistical significance, these consistent findings across distinct, large-scale OA datasets provide strong statistical support for the hypothesis that maintaining higher levels of GNL3 expression is causally protective against the development of osteoarthritis(\u003cstrong\u003eTable 3\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4 Exploratory Analysis of Splicing QTLs Suggests Tissue-Specific Effects\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe next conducted an exploratory analysis to investigate if alternative splicing could constitute an additional regulatory mechanism. A broad colocalization analysis between OA risk and GNL3 splicing QTLs (sQTLs) did not reveal a widespread pattern of shared causal variants comparable to the eQTL results (\u003cstrong\u003eSupplementary\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eMaterials\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;3\u003c/strong\u003e). However, a more sensitive, secondary analysis using SMR did identify significant, albeit highly tissue-specific, associations (\u003cstrong\u003eTable 4\u003c/strong\u003e). For example, a distinct splicing event in the Thyroid was causally associated with a protective effect on OA risk, while a different splicing event in the same tissue was associated with increased risk. These preliminary, tissue-specific findings suggest that while the regulation of total gene expression appears to be the primary mechanism at this locus, alternative splicing of GNL3 may represent a secondary, context-dependent regulatory layer, representing a potential avenue for future investigation into its tissue-specific functions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5 Systematic single-cell analysis pinpoints causal mechanisms to specific immune, neural, and progenitor cells\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo dissect the precise cellular context of \u003cem\u003eGNL3\u003c/em\u003e's causal effect, we performed a systematic colocalization and Mendelian randomization analysis using single-cell eQTL data. Our systematic MR analysis of 47 available single-cell contexts provided robust and consistent evidence of causality. After correcting for multiple testing, all significant associations revealed a protective effect of higher \u003cem\u003eGNL3\u003c/em\u003e expression on knee OA risk, with the effects concentrated in specific, state-dependent cellular contexts (\u003cstrong\u003eTable 5\u003c/strong\u003e\u003cstrong\u003e, Supplementary Materials 4\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThis causal effect was particularly strong in the immune system. For instance, genetically predicted \u003cem\u003eGNL3\u003c/em\u003e expression in \u003cstrong\u003eactivated CD4+ Memory T-cells\u003c/strong\u003e showed a robust protective signal \u003cstrong\u003e(OR = 0.81, 95% CI [0.73, 0.90], FDR = 0.003)\u003c/strong\u003e. We also uncovered a novel and significant link to neural lineages under cellular stress. \u003cem\u003eGNL3\u003c/em\u003e expression in \u003cstrong\u003estressed dopaminergic neurons\u003c/strong\u003e was strongly protective \u003cstrong\u003e(OR = 0.73, 95% CI [0.60, 0.88], FDR = 0.007)\u003c/strong\u003e and showed highly suggestive evidence of a shared genetic architecture with OA risk \u003cstrong\u003e(PP.H4 = 0.71)\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eCrucially, the most powerful evidence for a specific causal pathway emerged from a clear convergence of our analytical approaches in developmental progenitors. The signal in \u003cstrong\u003efloor plate progenitors\u003c/strong\u003e not only demonstrated a strong and significant causal protective effect \u003cstrong\u003e(OR = 0.89, 95% CI [0.84, 0.95], FDR = 0.003)\u003c/strong\u003e but was also the only context to surpass the high-confidence threshold for a shared genetic variant \u003cstrong\u003e(PP.H4 = 0.81)\u003c/strong\u003e. This dual, quantitative evidence provides a definitive link between a shared causal variant, the regulation of \u003cem\u003eGNL3\u003c/em\u003e in this specific cell type, and protection against osteoarthritis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.6 GNL3 Locus is a Pleiotropic Hub for Systemic Metabolic, Hematological, and Neurological Traits\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo understand the broader systemic impact of the GNL3 locus, we performed a GWAS-GWAS colocalization analysis using a body composition trait, impedance of the leg (a proxy for fat-free mass), as the anchor. This revealed that the same causal variant influencing GNL3 and OA risk is highly pleiotropic, significantly affecting a range of systemic traits and established OA risk factors (\u003cstrong\u003eTable 6\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThe most compelling evidence of pleiotropy was observed with key metabolic and endocrine traits. We found exceptionally strong colocalization with Body Mass Index (BMI; PP.H4 \u0026gt; 0.97) and waist-to-hip ratio (WHR; PP.H4 \u0026gt; 0.99), two critical indicators of metabolic health. A similarly robust signal was identified for 25-hydroxyvitamin D levels (PP.H4 \u0026gt; 0.98), a crucial factor in bone and joint health, with the evidence remaining strong even after conditioning on BMI.\u003c/p\u003e\n\u003cp\u003eIntriguingly, the locus also showed significant, albeit more moderate, evidence of colocalization with other systemic phenotypes. These included hematological traits such as hematocrit (PP.H4 ≈ 0.80) and several neuro-behavioral traits, most notably neuroticism (PP.H4 \u0026gt; 0.96) and its related \"worry\" subcluster (PP.H4 ≈ 0.79). This pattern of shared genetic architecture across metabolic, endocrine, hematological, and neurological domains positions GNL3 as a central genetic hub. These findings strongly reinforce the view of OA as a disease with significant systemic components, driven by genes that operate far beyond the local joint environment. Key pleiotropic colocalization results are summarized in Table 6, with a comprehensive list provided in \u003cstrong\u003eSupplementary Materials\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e5\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.7 Mendelian Randomization Reveals Causal Roles of Specific Immune Cell Subsets in OA Pathogenesis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo dissect the systemic inflammatory component of OA pathogenesis, we conducted a comprehensive two-sample MR analysis to systematically evaluate the causal effects of 729 genetically predicted immune cell phenotypes on four distinct OA outcomes. This analysis identified a complex landscape of both protective and risk-increasing roles for specific innate and adaptive immune cell populations, providing strong causal evidence for their involvement in the disease (\u003cstrong\u003eTable 7; full results in Supplementary Materials\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e6\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eWithin the adaptive immune system, we found compelling evidence for the involvement of both T and B cell lineages. Notably, higher genetically predicted levels of naive CD4+ T-cells were causally associated with a reduced risk of localized OA (OR = 0.78, 95% CI [0.69, 0.89], p = 1.5e-4). A strong protective signal was also observed for IgD- CD27- B-cells, which significantly lowered the risk of hip OA (OR = 0.66, 95% CI [0.54, 0.80], p = 3.0e-5), an association that remained robust after multiple testing correction. In contrast, higher levels of effector memory CD8+ T-cells were associated with an increased risk of knee and hip OA (OR = 1.33, 95% CI [1.05, 1.68], p = 1.7e-02).\u003c/p\u003e\n\u003cp\u003eIn the innate immune compartment, we identified striking and opposing effects. A genetically predicted increase in classical monocytes (CD14+ CD16-) demonstrated a strong protective effect against knee OA (OR = 0.44, 95% CI [0.27, 0.70], p = 6.7e-4). Conversely, a robust risk-increasing signal was observed for several markers of granulocyte activation. The most significant of these was the expression of CD86 on granulocytes, which was causally associated with a more than two-fold increase in the risk of knee OA (OR = 2.07, 95% CI [1.36, 3.14], p = 6.7e-04). Collectively, these MR results provide strong causal evidence that the balance between specific immune cell subsets is a critical determinant of OA risk, highlighting distinct cell-mediated pathways for joint-specific pathology.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eOur systematic, multi-modal genomic analysis has definitively established \u003cstrong\u003e\u003cem\u003eGNL3\u003c/em\u003e\u003c/strong\u003e \u003cstrong\u003eas the causal effector gene at the\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e3q21\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;osteoarthritis (OA) risk locus\u003c/strong\u003e. Moving beyond this primary identification, our findings converge to propose a unifying mechanistic model: \u003cem\u003eGNL3\u003c/em\u003e functions as a \u003cstrong\u003emaster regulator of systemic homeostasis\u003c/strong\u003e. We posit that genetically determined \u003cem\u003eGNL3\u003c/em\u003e expression levels act as a systemic set point, modulating the response of immune, neural, and progenitor cells to metabolic and inflammatory stressors, thereby dictating an individual\u0026apos;s susceptibility to developing OA.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.1 A multi-layered evidence chain establishes \u003cem\u003eGNL3\u003c/em\u003e as the causal gene\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe foundation of our argument is the exceptionally consistent evidence of shared causal variants between OA risk and the regulation of \u003cem\u003eGNL3\u003c/em\u003e. The genetic signal for OA strongly and repeatedly colocalized with \u003cem\u003eGNL3\u003c/em\u003e expression QTLs (eQTLs) across a diverse panel of 17 tissues and was robustly replicated across independent, large-scale GWAS cohorts (Zeggini et al. 2012, Tachmazidou et al. 2019, and Boer et al. 2021), providing powerful evidence against confounding by chance or linkage disequilibrium ( Section 3.1 ; Supplementary Materials S2 ). We further strengthened this link by demonstrating that the OA risk locus also consistently colocalized with methylation QTLs (mQTLs) controlling CpG sites within the \u003cem\u003eGNL3\u003c/em\u003e promoter, and functionally validated this by showing a significant negative correlation between promoter methylation and \u003cem\u003eGNL3\u003c/em\u003e expression ( Section 3.2 ). While our exploratory analysis of splicing QTLs did not suggest a widespread role ( Section 3.4 ), our Mendelian Randomization analysis provided formal proof of causality, demonstrating that genetically predicted higher expression of \u003cem\u003eGNL3\u003c/em\u003e is significantly protective against both knee and hip OA ( Section 3.3 ). This multi-layered evidence chain forms an unshakable foundation for \u003cem\u003eGNL3\u003c/em\u003e\u0026apos;s causal role.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.2 A unified mechanistic hypothesis: \u003cem\u003eGNL3\u003c/em\u003e as a master regulator of cellular stress response in OA\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA critical insight from our study is that the function of \u003cem\u003eGNL3\u003c/em\u003e in OA is not confined to the joint but is deeply embedded in systemic biology. Our pleiotropy analysis firmly establishes the \u003cem\u003eGNL3\u003c/em\u003e locus as a hub sharing causal variants with key metabolic and endocrine mediators, including BMI and 25-hydroxyvitamin D ( Section 3.6 ). This provides a direct genetic underpinning for the modern paradigm of OA as a systemic, metabolic disease characterized by \u0026quot;metaflammation\u0026quot;[28-30].\u0026nbsp;Similarly, Vitamin D is essential for bone health and has known immunomodulatory functions[31, 32]. Our systematic single-cell analyses provide the crucial next step, dissecting precisely \u003cem\u003ehow\u003c/em\u003e these systemic risks are executed at a cellular level through the action of \u003cem\u003eGNL3\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003e4.2.1 A Novel Neuro-Skeletal Axis: Linking Central Nervous System Integrity to Joint Pathogenesis\u003c/p\u003e\n\u003cp\u003eOne of the most striking and novel findings of our study is the strong causal protective effect of GNL3 in neural cells, particularly under conditions of cellular stress. This discovery provides compelling genetic evidence for the emerging concept of a \u0026quot;neuro-skeletal axis\u0026quot; in OA, moving beyond the traditional view of pain as a mere consequence of joint damage, and positioning the nervous system as an active participant in the disease process[33].\u003c/p\u003e\n\u003cp\u003eOur analysis revealed a significant protective signal for GNL3 expression in rotenone-treated dopaminergic neurons and astrocytes, suggesting that individuals with genetically lower GNL3 expression may have a reduced capacity to buffer their neural systems against metabolic or oxidative stress. We propose this genetic vulnerability could manifest in two distinct, yet interconnected, pathways relevant to OA:\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003ePromotion of Neurogenic Inflammation and Structural Joint Damage: The nervous system is not a passive bystander to joint pathology. Peripheral sensory neurons that innervate the joint can, when stressed or damaged, release pro-inflammatory neuropeptides such as Calcitonin Gene-Related Peptide (CGRP) and Substance P directly into the synovial space[34]. This process, known as neurogenic inflammation, can directly drive OA pathogenesis by promoting synovitis, increasing vascular permeability, and stimulating the production of matrix-degrading enzymes by chondrocytes and synoviocytes. Our findings suggest a novel genetic basis for this mechanism: genetically determined lower GNL3 expression may render these sensory neurons more susceptible to stress-induced dysfunction, leading to a heightened state of neurogenic inflammation that actively contributes to the structural degradation of the joint. This provides a direct, mechanistic link from a genetic variant to the cellular processes that cause cartilage breakdown.\u003c/li\u003e\n \u003cli\u003eModulation of Central Sensitization and the Pain Experience: A well-known paradox in OA is the often-poor correlation between the degree of joint damage visible on X-rays and the level of pain experienced by the patient[35]. This is largely explained by central sensitization, a process where the central nervous system becomes hyperexcitable, amplifying pain signals[36]. Our findings provide a potential genetic explanation for an individual\u0026apos;s predisposition to central sensitization. The causal effect of GNL3 in central nervous system cells (astrocytes, dopaminergic neurons) suggests that its expression level influences the overall \u0026quot;tone\u0026quot; and resilience of these circuits. Lower GNL3 expression could lead to a state of heightened neural excitability or impaired glial cell function, creating a permissive environment for the development of chronic, amplified pain. This aligns intriguingly with our pleiotropy finding linking the GNL3 locus to neuroticism, suggesting that this single genetic variant may simultaneously influence the structural integrity of the joint (via neurogenic inflammation) and the central nervous system\u0026apos;s processing of pain and stress.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eIn summary, our data reframe the role of the nervous system in OA from a simple pain reporter to a critical, genetically-influenced modulator of the disease. The GNL3 locus appears to be a key factor governing the resilience of the neuro-skeletal axis, where its protective, high-expression alleles may help maintain both peripheral neural homeostasis and central pain circuit stability, thereby protecting against both the structural and symptomatic progression of osteoarthritis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.2.2 Modulating Immune Homeostasis and Cellular Senescence\u003c/strong\u003e Our systematic analysis confirmed a robust causal role for \u003cem\u003eGNL3\u003c/em\u003e in activated T-cells ( Section 3.5 ), providing a mechanistic anchor for the well-documented immune component of OA, which our broader immune cell MR screen showed was defined by a delicate balance between protective (e.g., naive T-cells) and pathogenic (e.g., activated effector cells) populations ( Section 3.7 ). We hypothesize that \u003cem\u003eGNL3\u003c/em\u003e, a known regulator of the p53-mediated stress response[37, 38], acts as a critical checkpoint for maintaining immune cell homeostasis and preventing premature exhaustion. In individuals with risk-associated, low-expression alleles, T-cells, upon activation by systemic inflammatory triggers (such as those from adipose tissue in obesity), may have a lower threshold for entering a state of activation-induced senescence. These senescent T-cells are known to adopt a pro-inflammatory senescence-associated secretory phenotype (SASP), releasing a cocktail of cytokines and chemokines that can directly contribute to synovitis and cartilage degradation in the joint[39].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.2.3 Impaired Regenerative Capacity in Progenitor Cells\u003c/strong\u003e The strongest convergence of our causal and colocalization evidence was found in developmental progenitor cells ( Section 3.5 ). This finding is profound because it connects the OA risk locus directly to \u003cem\u003eGNL3\u003c/em\u003e\u0026apos;s fundamental biological role in ribosome biogenesis, cell cycle control, and stem cell maintenance. While the specific progenitor lineage remains to be identified, this result supports a compelling hypothesis: genetically determined lower \u003cem\u003eGNL3\u003c/em\u003e expression could impair the regenerative potential of certain cell populations crucial for joint homeostasis. This could include mesenchymal stem cells responsible for cartilage repair, or even hematopoietic progenitors that give rise to balanced immune cell populations. Under conditions of chronic micro-trauma or inflammatory stress, this reduced regenerative capacity could lead to a net loss of tissue integrity over time, providing a plausible, cell-intrinsic mechanism for the initiation and progression of OA.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.3 Strengths, Limitations, and Future Directions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe primary strength of this study is its rigorous, multi-modal design, which constructs a comprehensive evidence chain from a GWAS locus to a causal gene and its specific, systemic, mechanistic context. By systematically integrating multiple layers of evidence\u0026mdash;colocalization, MR, pan-cancer analysis, and, crucially, systematic single-cell causal inference\u0026mdash;we have built a robust and multifaceted case for the role of \u003cem\u003eGNL3\u003c/em\u003e in OA. The convergence of our MR and colocalization results in a specific cell type is a major strength that provides a high-confidence starting point for experimental validation. However, certain limitations must be acknowledged. Our study relies on publicly available summary statistics, and while we have used the largest and most well-powered datasets available, the exposure QTL data are not always derived from the most disease-relevant tissue (i.e., synovium or cartilage). While the strong signals in immune cells and fibroblasts provide a compelling systemic link,\u0026nbsp;direct validation in joint tissues is a critical next step. Future research should be directed by the specific, testable hypotheses generated by our study. First, \u003cem\u003ein vitro\u003c/em\u003e studies using iPSC-derived neurons, T-cells, and mesenchymal stem cells with CRISPR-mediated modulation of \u003cem\u003eGNL3\u003c/em\u003e expression are needed to functionally validate its role in buffering against metabolic and inflammatory stressors. Second, the development of cell-type-specific conditional \u003cem\u003eGNL3\u003c/em\u003e knockout mouse models (e.g., in sensory neurons or T-lymphocytes) will be essential to dissect its contribution to OA pathogenesis in vivo, assessing not only joint degradation but also pain behaviors. Finally, from a translational perspective, exploring whether pharmacological agents can safely and effectively upregulate \u003cem\u003eGNL3\u003c/em\u003e expression could represent a novel therapeutic avenue, targeting the fundamental cellular stress pathways that appear to underpin the genetic risk of OA.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eIn summary, our rigorous, multi-modal integrative genomic analysis establishes \u003cem\u003eGNL3\u003c/em\u003e as a causal gene for osteoarthritis. The weight of the evidence moves beyond a simple genetic association to propose a coherent mechanistic model where \u003cem\u003eGNL3\u003c/em\u003e acts as a systemic regulatory hub. Genetically determined lower expression of \u003cem\u003eGNL3\u003c/em\u003e appears to sensitize the host to metabolic and inflammatory insults, creating a permissive environment for immune, neural, and progenitor cell dysregulation that culminates in joint degradation. This work validates \u003cem\u003eGNL3\u003c/em\u003e as a high-confidence therapeutic target and reframes our understanding of OA, highlighting the potential for novel therapeutic strategies aimed at modulating systemic homeostatic and cellular stress-response pathways to preserve joint health.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor Disclosure Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics and Consent to Participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Publish\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot Applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of Competing Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTianhao Qu:\u003c/strong\u003e Conceptualization, Methodology, Formal Analysis, Data Curation, Writing – Original Draft, Visualization. \u003cstrong\u003eYan Zhong:\u003c/strong\u003e Methodology, Software, Validation. \u003cstrong\u003eYonghuan Zhou:\u003c/strong\u003e Investigation, Resources, Supervision. \u003cstrong\u003eLin Liu:\u003c/strong\u003e Software, Visualization. \u003cstrong\u003eZheng Ye:\u003c/strong\u003e Conceptualization, Project Administration, Funding Acquisition, Writing – Review \u0026amp; Editing, Supervision.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data analyzed in this study are publicly available, summary-level data obtained from the sources as described in the Materials and Methods section of the manuscript. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eJ.D. Steinmetz, G.T. Culbreth, L.M. Haile, Q. Rafferty, J. Lo, K.G. Fukutaki, J.A. Cruz, A.E. Smith, S.E. Vollset, P.M. Brooks, Global, regional, and national burden of osteoarthritis, 1990\u0026ndash;2020 and projections to 2050: a systematic analysis for the Global Burden of Disease Study 2021, The Lancet Rheumatology, 5 (2023) e508-e522.\u003c/li\u003e\n\u003cli\u003eS.a. Tang, C. Zhang, W.M. Oo, K. Fu, M.A. Risberg, S.M. Bierma-Zeinstra, T. Neogi, I. Atukorala, A.-M. Malfait, C. Ding, Osteoarthritis, Nature Reviews Disease Primers, 11 (2025) 1-22.\u003c/li\u003e\n\u003cli\u003eJ. Zhu, W. Chen, Y. Hu, Y. Qu, H. Yang, Y. Zeng, C. Hou, F. Ge, Z. Zhou, H. Song, Physical activity patterns, genetic susceptibility, and risk of hip/knee osteoarthritis: a prospective cohort study based on the UK Biobank, Osteoarthritis and cartilage, 30 (2022) 1079-1090.\u003c/li\u003e\n\u003cli\u003eB.G. Faber, M. Frysz, C.G. Boer, D.S. Evans, R. Ebsim, K.A. Flynn, M. Lundberg, L. Southam, A. Hartley, F.R. Saunders, The identification of distinct protective and susceptibility mechanisms for hip osteoarthritis: findings from a genome-wide association study meta-analysis of minimum joint space width and Mendelian randomisation cluster analyses, EBioMedicine, 95 (2023).\u003c/li\u003e\n\u003cli\u003eF. Gee, C.F. Clubbs, E.V. Raine, L.N. Reynard, J. Loughlin, Allelic expression analysis of the osteoarthritis susceptibility locus that maps to chromosome 3p21 reveals cis-acting eQTLs at GNL3 and SPCS1, BMC medical genetics, 15 (2014) 1-7.\u003c/li\u003e\n\u003cli\u003eM.D. Gallagher, A.S. Chen-Plotkin, The post-GWAS era: from association to function, The American Journal of Human Genetics, 102 (2018) 717-730.\u003c/li\u003e\n\u003cli\u003eG. Quiroga-Artigas, D. de Jong, C.E. Schnitzler, GNL3 is an evolutionarily conserved stem cell gene influencing cell proliferation, animal growth and regeneration in the hydrozoan Hydractinia, Open biology, 12 (2022) 220120.\u003c/li\u003e\n\u003cli\u003eC.G. Boer, K. Hatzikotoulas, L. Southam, L. Stef\u0026aacute;nsd\u0026oacute;ttir, Y. Zhang, R.C. de Almeida, T.T. Wu, J. Zheng, A. Hartley, M. Teder-Laving, Deciphering osteoarthritis genetics across 826,690 individuals from 9 populations, Cell, 184 (2021) 4784-4818. e4717.\u003c/li\u003e\n\u003cli\u003eI. Tachmazidou, K. Hatzikotoulas, L. Southam, J. Esparza-Gordillo, V. Haberland, J. Zheng, T. Johnson, M. Koprulu, E. Zengini, J. Steinberg, Identification of new therapeutic targets for osteoarthritis through genome-wide analyses of UK Biobank data, Nature genetics, 51 (2019) 230-236.\u003c/li\u003e\n\u003cli\u003eE. Zeggini, K. Panoutsopoulou, L. Southam, N. Rayner, A. Day-Williams, M. Lopes, V. Boraska, T. Esko, E. Evangelou, A. Hoffman, Identification of new susceptibility loci for osteoarthritis (arcOGEN): a genome-wide association study, Lancet (London, England), 380 (2012).\u003c/li\u003e\n\u003cli\u003eN.R. Gay, M. Gloudemans, M.L. Antonio, N.S. Abell, B. Balliu, Y. Park, A.R. Martin, S. Musharoff, A.S. Rao, F. Aguet, Impact of admixture and ancestry on eQTL analysis and GWAS colocalization in GTEx, Genome biology, 21 (2020) 1-20.\u003c/li\u003e\n\u003cli\u003eU. V\u0026otilde;sa, A. Claringbould, H.-J. Westra, M.J. Bonder, P. Deelen, B. Zeng, H. Kirsten, A. Saha, R. Kreuzhuber, S. Yazar, Large-scale cis-and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression, Nature genetics, 53 (2021) 1300-1310.\u003c/li\u003e\n\u003cli\u003eT. Qi, Y. Wu, H. Fang, F. Zhang, S. Liu, J. Zeng, J. Yang, Genetic control of RNA splicing and its distinct role in complex trait variation, Nature genetics, 54 (2022) 1355-1363.\u003c/li\u003e\n\u003cli\u003eA. Jourdon, S. Scuderi, D. Capauto, A. Abyzov, F.M. Vaccarino, PsychENCODE and beyond: transcriptomics and epigenomics of brain development and organoids, Neuropsychopharmacology, 46 (2021) 70-85.\u003c/li\u003e\n\u003cli\u003eD.A. Bennett, A.S. Buchman, P.A. Boyle, L.L. Barnes, R.S. Wilson, J.A. Schneider, Religious orders study and rush memory and aging project, Journal of Alzheimer\u0026rsquo;s disease, 64 (2018) S161-S189.\u003c/li\u003e\n\u003cli\u003eR. Ding, Q. Wang, L. Gong, T. Zhang, X. Zou, K. Xiong, Q. Liao, M. Plass, L. Li, scQTLbase: an integrated human single-cell eQTL database, Nucleic Acids Research, 52 (2024) D1010-D1017.\u003c/li\u003e\n\u003cli\u003eV. Orr\u0026ugrave;, M. Steri, C. Sidore, M. Marongiu, V. Serra, S. Olla, G. Sole, S. Lai, M. Dei, A. Mulas, Complex genetic signatures in immune cells underlie autoimmunity and inform therapy, Nature genetics, 52 (2020) 1036-1045.\u003c/li\u003e\n\u003cli\u003eJ.N. Weinstein, E.A. Collisson, G.B. Mills, K.R. Shaw, B.A. Ozenberger, K. Ellrott, I. Shmulevich, C. Sander, J.M. Stuart, The cancer genome atlas pan-cancer analysis project, Nature genetics, 45 (2013) 1113-1120.\u003c/li\u003e\n\u003cli\u003eZ. Ye, Y. Gao, J. Yuan, F. Chen, P. Xu, W. Liu, The Role of Gut Microbiota in Modulating Brain Structure and Psychiatric Disorders: A Mendelian Randomization Study, NeuroImage, (2025) 121292.\u003c/li\u003e\n\u003cli\u003eC. Giambartolomei, D. Vukcevic, E.E. Schadt, L. Franke, A.D. Hingorani, C. Wallace, V. Plagnol, Bayesian test for colocalisation between pairs of genetic association studies using summary statistics, PLoS genetics, 10 (2014) e1004383.\u003c/li\u003e\n\u003cli\u003eY. Wu, J. Zeng, F. Zhang, Z. Zhu, T. Qi, Z. Zheng, L.R. Lloyd-Jones, R.E. Marioni, N.G. Martin, G.W. Montgomery, Integrative analysis of omics summary data reveals putative mechanisms underlying complex traits, Nature communications, 9 (2018) 918.\u003c/li\u003e\n\u003cli\u003eD.A. Lawlor, R.M. Harbord, J.A. Sterne, N. Timpson, G. Davey Smith, Mendelian randomization: using genes as instruments for making causal inferences in epidemiology, Statistics in medicine, 27 (2008) 1133-1163.\u003c/li\u003e\n\u003cli\u003eS. Burgess, A. Butterworth, S.G. Thompson, Mendelian randomization analysis with multiple genetic variants using summarized data, Genetic epidemiology, 37 (2013) 658-665.\u003c/li\u003e\n\u003cli\u003eJ. Bowden, G. Davey Smith, S. Burgess, Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression, International journal of epidemiology, 44 (2015) 512-525.\u003c/li\u003e\n\u003cli\u003eJ. Bowden, G. Davey Smith, P.C. Haycock, S. Burgess, Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator, Genetic epidemiology, 40 (2016) 304-314.\u003c/li\u003e\n\u003cli\u003eF.P. Hartwig, G. Davey Smith, J. Bowden, Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption, International journal of epidemiology, 46 (2017) 1985-1998.\u003c/li\u003e\n\u003cli\u003eJ. Zhao, J. Ming, X. Hu, G. Chen, J. Liu, C. Yang, Bayesian weighted Mendelian randomization for causal inference based on summary statistics, Bioinformatics, 36 (2020) 1501-1508.\u003c/li\u003e\n\u003cli\u003eY. Jiang, Osteoarthritis year in review 2021: biology, Osteoarthritis and cartilage, 30 (2022) 207-215.\u003c/li\u003e\n\u003cli\u003eC.R. Scanzello, Role of low-grade inflammation in osteoarthritis, Current opinion in rheumatology, 29 (2017) 79-85.\u003c/li\u003e\n\u003cli\u003eD. Azamar-Llamas, G. Hernandez-Molina, B. Ramos-Avalos, J. Furuzawa-Carballeda, Adipokine contribution to the pathogenesis of osteoarthritis, Mediators of inflammation, 2017 (2017) 5468023.\u003c/li\u003e\n\u003cli\u003eX.-X. Zhang, S.-H. He, X. Liang, W. Li, T.-F. Li, D.-F. Li, Aging, cell senescence, the pathogenesis and targeted therapies of osteoarthritis, Frontiers in Pharmacology, 12 (2021) 728100.\u003c/li\u003e\n\u003cli\u003eM. Aribi, F.J. Mennechet, C. Touil-Boukoffa, The role of vitamin D as an immunomodulator, in, Frontiers Media SA, 2023, pp. 1186635.\u003c/li\u003e\n\u003cli\u003eJ. Yu, L. Ji, Y. Liu, X. Wang, J. Wang, C. Liu, Bone-brain interaction: mechanisms and potential intervention strategies of biomaterials, Bone Research, 13 (2025) 38.\u003c/li\u003e\n\u003cli\u003eJ.D. Richardson, M.R. Vasko, Cellular mechanisms of neurogenic inflammation, The Journal of pharmacology and experimental therapeutics, 302 (2002) 839-845.\u003c/li\u003e\n\u003cli\u003eJ. Bedson, P.R. Croft, The discordance between clinical and radiographic knee osteoarthritis: a systematic search and summary of the literature, BMC musculoskeletal disorders, 9 (2008) 116.\u003c/li\u003e\n\u003cli\u003eC.J. Woolf, Central sensitization: implications for the diagnosis and treatment of pain, pain, 152 (2011) S2-S15.\u003c/li\u003e\n\u003cli\u003eR.Y. Tsai, R.D. McKay, A nucleolar mechanism controlling cell proliferation in stem cells and cancer cells, Genes \u0026amp; development, 16 (2002) 2991-3003.\u003c/li\u003e\n\u003cli\u003eH. Ma, T. Pederson, Nucleostemin: a multiplex regulator of cell-cycle progression, Trends in cell biology, 18 (2008) 575-579.\u003c/li\u003e\n\u003cli\u003eO.H. Jeon, C. Kim, R.-M. Laberge, M. Demaria, S. Rathod, A.P. Vasserot, J.W. Chung, D.H. Kim, Y. Poon, N. David, Local clearance of senescent cells attenuates the development of post-traumatic osteoarthritis and creates a pro-regenerative environment, Nature medicine, 23 (2017) 775-781.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003ch4\u003eTable 1: Widespread Colocalization of GNL3 Expression QTLs (eQTLs) with Osteoarthritis Risk across Diverse Human Tissues.\u003c/h4\u003e\n\u003cp\u003eLegend: Colocalization analysis of GWAS signals for Osteoarthritis (P01779) and eQTLs for \u003cem\u003eGNL3\u003c/em\u003e gene expression across 44 tissues from the GTEx v8 project. The table lists tissues with a posterior probability of a shared causal variant (PP.H4) greater than 0.8, sorted in descending order of PP.H4. This widespread colocalization across multiple organ systems suggests a systemic role for GNL3 in OA pathogenesis.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTissue\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eQTL Type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eColocalization Probability (PP.H4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGWAS Top SNP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL Top SNP\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCells Cultured Fibroblasts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.981\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers6976\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTestis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.975\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2028216\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMuscle Skeletal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.970\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers11714419\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBreast Mammary Tissue\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.967\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers8906\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAdipose Subcutaneous\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.950\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2028216\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHeart Left Ventricle\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.944\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2028216\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eEsophagus Gastroesophageal J.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.937\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers11714419\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNerve Tibial\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.932\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers6976\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eColon Transverse\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.919\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2028216\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePancreas\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.918\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2028216\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eArtery Aorta\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.888\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2028216\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eArtery Coronary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.873\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers8906\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHeart Atrial Appendage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.864\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers1866268\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBrain Cerebellum\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.839\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers13063160\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eArtery Tibial\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.834\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers12488461\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers3796353\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePsychENCODE (Prefrontal Cortex)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.898\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eP00954_rs2336147\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2336147\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eROSMAP (Brain Cortex)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eeQTL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.823\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eP00954_rs7646741\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers7646741\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch4\u003eTable 2: Strong Evidence of Colocalization between GNL3 Methylation QTLs (mQTLs) and Hip Osteoarthritis.\u003c/h4\u003e\n\u003cp\u003eLegend: Colocalization analysis of GWAS signals for hip osteoarthritis and total joint replacement with methylation QTLs (mQTLs) for CpG sites at the \u003cem\u003eGNL3\u003c/em\u003e locus. The results from different brain-related datasets (ROSMAP, FB_Brain) and blood provide evidence for an epigenetic regulatory mechanism linking genetic risk to OA.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTrait Description\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePMID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTissue/Source\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMethylation Probe\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eColocalization Probability (PP.H4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHip osteoarthritis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e30664745\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eROSMAP (Brain Cortex)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ecg18595196\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.948\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTotal joint replacement\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e34822786\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eROSMAP (Brain Cortex)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ecg18595196\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.948\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHip osteoarthritis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e30664745\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eROSMAP (Brain Cortex)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ecg00845626\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.926\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTotal joint replacement\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e34822786\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eROSMAP (Brain Cortex)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ecg00845626\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.916\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHip osteoarthritis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e30664745\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFB_Brain (Brain)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ecg11041457\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.896\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOsteoarthritis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e22763110\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eUS_Blood (Whole Blood)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ecg08332332\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.943\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch4\u003eTable 3: Mendelian Randomization Analysis of GNL3 Expression on Osteoarthritis and Related Traits.\u003c/h4\u003e\n\u003cp\u003eLegend:\u0026nbsp;Results from two-sample Mendelian randomization (MR) analysis assessing the causal effect of genetically predicted GNL3 expression on various OA-related outcomes. A negative beta and an Odds Ratio (OR) less than 1 indicate a protective effect. Statistically significant results (p \u0026lt; 0.05) are highlighted in bold.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 157px;\"\u003e\n \u003cp\u003eOutcome\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eGWAS ID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003eBeta\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp\u003eStd. Error\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eP-value\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 134px;\"\u003e\n \u003cp\u003eOdds Ratio (95% CI)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 157px;\"\u003e\n \u003cp\u003eKnee osteoarthritis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eebi-a-GCST007090\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e-0.159\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp\u003e0.045\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e4.38E-04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 134px;\"\u003e\n \u003cp\u003e0.85 [0.78, 0.93]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 157px;\"\u003e\n \u003cp\u003eHip osteoarthritis (hospital diagnosed)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eebi-a-GCST005810\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e-0.457\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp\u003e0.158\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.0038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 134px;\"\u003e\n \u003cp\u003e0.63 [0.46, 0.86]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 157px;\"\u003e\n \u003cp\u003eOsteoarthritis; localized (PheCode 740.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eGCST90044591\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e-0.216\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp\u003e0.074\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.0035\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 134px;\"\u003e\n \u003cp\u003e0.81 [0.70, 0.93]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 157px;\"\u003e\n \u003cp\u003eOsteoarthritis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eGCST90134288\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e-0.567\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp\u003e0.273\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.0380\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 134px;\"\u003e\n \u003cp\u003e0.57 [0.33, 0.98]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 157px;\"\u003e\n \u003cp\u003eOsteoarthritis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eGCST90134279\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e-0.055\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.1038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 134px;\"\u003e\n \u003cp\u003e0.95 [0.88, 1.01]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 157px;\"\u003e\n \u003cp\u003eOsteoarthritis of the hip or knee\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eebi-a-GCST007092\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e-0.092\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp\u003e0.095\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.3357\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 134px;\"\u003e\n \u003cp\u003e0.91 [0.76, 1.10]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 157px;\"\u003e\n \u003cp\u003eOsteoarthritis (self-reported)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eebi-a-GCST005811\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e+0.056\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp\u003e0.069\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.4125\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 134px;\"\u003e\n \u003cp\u003e1.06 [0.92, 1.21]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eNote:\u003c/strong\u003e Data was sourced from the IEU OpenGWAS database for studies with an \u0026apos;ebi-a-\u0026apos; prefix in the GWAS ID, and from the EBI GWAS Catalog FTP server (http://ftp.ebi.ac.uk/pub/databases/gwas/summary_statistics/) for all others.\u003c/p\u003e\n\u003ch4\u003eTable 4: Bidirectional Causal Effects of GNL3 Splicing Events on Osteoarthritis Risk from SMR Analysis.\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003eLegend:\u003c/strong\u003e Summary-data-based Mendelian Randomization (SMR) analysis using splicing QTLs (sQTLs) as instruments. The results reveal distinct splicing events with opposing causal effects on OA risk. Results are considered robust when the HEIDI test for pleiotropy is non-significant (p_heidi \u0026gt; 0.05).\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTrait Description\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTissue\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTop sQTL SNP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEffect Size (b_smr)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCausal P-value (p_smr)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePleiotropy Test (p_heidi)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eConclusion\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHip or Knee OA (P02059)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eThyroid\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2710323\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.125\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6.65e-05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.134\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSignificant Protective Effect\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOsteoarthritis (P01779)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eThyroid\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers2710323\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.303\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.85e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.300\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSignificant Protective Effect\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOsteoarthritis (P01779)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eThyroid\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers9853056\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e+0.229\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6.93e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.061\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSignificant Risk Effect\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOsteoarthritis (P01779)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLung\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ers1010554\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e+0.240\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e7.42e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.031\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eAssociation likely due to pleiotropy\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 5: Systematic single-cell analysis identifies causal cellular contexts for \u003cem\u003eGNL3\u003c/em\u003e in osteoarthritis.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLegend:\u0026nbsp;\u003c/strong\u003eTo identify the precise cellular drivers of GNL3\u0026apos;s effect on osteoarthritis, we performed a systematic Mendelian randomization (MR) and colocalization analysis across all available single-cell eQTL datasets. The table shows all cellular contexts with a significant causal protective effect (Odds Ratio \u0026lt; 1) of GNL3 expression on knee OA after False Discovery Rate (FDR) correction. The corresponding Posterior Probability of a shared causal variant (PP.H4) is also shown. The convergence of a significant MR result and a high-confidence colocalization signal (PP.H4 \u0026gt; 0.8) in floor plate progenitors provides the strongest evidence for a specific causal pathway.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eStudy Source\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBiological Domain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCellular Context\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCausal Effect on Knee OA (MR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFDR-corr. P\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eColoc. (PP.H4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eJerber et al. 2021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eProgenitor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFloor Plate Progenitors\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.89 (0.84 - 0.95)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSoskic et al. 2022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eImmune\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCD4+ Memory T-Cell (Activated)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.81 (0.73 - 0.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.41\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSoskic et al. 2022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eImmune\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCD4+ Naive T-Cell (Activated)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.57 (0.41 - 0.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eJerber et al. 2021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNeural\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDopaminergic Neurons (Stress)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.73 (0.60 - 0.88)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.71\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eJerber et al. 2021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNeural\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAstrocyte-like Cells (Stress)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.80 (0.68 - 0.93)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.024\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.34\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 6: Key Pleiotropic Effects of the \u003cem\u003eGNL3\u003c/em\u003e Locus Identified Through GWAS-GWAS Colocalization.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLegend:\u003c/strong\u003e Representative results from GWAS-GWAS colocalization analysis showing shared genetic signals between body composition traits (a proxy for the OA-associated locus) and key systemic traits. Only the most significant findings for major metabolic, endocrine, and neurological traits are presented. PP.H4 indicates the posterior probability of a shared causal variant.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAnchor Trait (PMID)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eColocalizing Trait\u0026nbsp;(PMID)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eColocalization Probability (PP.H4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImpedance of leg (right) (31768069)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBody mass index (BMI)\u0026nbsp;(30124842)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.998\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImpedance of leg (left) (31768069)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWaist-to-hip ratio adjusted for BMI\u0026nbsp;(25673412)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.990\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImpedance of leg (right) (31768069)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e25-hydroxyvitamin D\u0026nbsp;(32242144)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.998\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImpedance of leg (right) (31768069)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNeuroticism\u0026nbsp;(30867560)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.979\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImpedance of leg (left) (31768069)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHematocrit\u0026nbsp;(32888493)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.796\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 7: Causal Effects of Genetically Predicted Immune Cell Traits on Osteoarthritis Risk from Mendelian Randomization.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLegend: Representative results from two-sample MR analysis showing causal associations between immune phenotypes and OA risk. Results with p \u0026lt; 0.01 are shown. OR (Odds Ratio) \u0026lt; 1 indicates a protective effect; OR \u0026gt; 1 indicates a risk-increasing effect.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eExposure (Immune Cell Trait)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eOutcome\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMethod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eOR (95% CI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eP-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eProtective Associations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCD14+ CD16- monocyte Absolute Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eKnee OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.44 (0.27-0.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6.7e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNaive CD4+ T cell Absolute Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLocalized OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.78 (0.69-0.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.5e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eIgD- CD27- B cell %lymphocyte\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eHip OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.66 (0.54-0.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.0e-05\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCD25++ CD4+ T cell Absolute Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLocalized OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.89 (0.82-0.96)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.8e-03\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCentral Memory CD4+ T cell %CD4+ T cell\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLocalized OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.83 (0.73-0.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.6e-03\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eRisk-Increasing Associations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCD86 on granulocyte\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eKnee OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.07 (1.36-3.14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6.7e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCD39 on granulocyte\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eKnee OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2.05 (1.36-3.11)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6.7e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCCR2 on granulocyte\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eKnee OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.71 (1.25-2.32)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6.7e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCD19 on switched memory B cell\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eKnee OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.54 (1.12-2.13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e8.2e-03\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eEffector Memory CD8+ T cell Abs. Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eKnee \u0026amp; Hip OA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.33 (1.05-1.68)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.7e-02\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"human-genomics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"hugm","sideBox":"Learn more about [Human Genomics](http://humgenomics.biomedcentral.com/)","snPcode":"40246","submissionUrl":"https://submission.nature.com/new-submission/40246/3","title":"Human Genomics","twitterHandle":"@OAgenetics","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Osteoarthritis, GNL3, Mendelian Randomization, Integrative Genomics, Integrative Genomics, Pleiotropy","lastPublishedDoi":"10.21203/rs.3.rs-7630649/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7630649/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e Osteoarthritis (OA) is a leading cause of disability, yet there are no approved disease-modifying therapies (DMOADs) capable of halting its progression. While genome-wide association studies (GWAS) have robustly linked the 3q21locus to OA, the causal effecator gene and its underlying mechanism have remained elusive, hindering translational progress.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e We implemented a multi-tiered, systematic integrative genomics strategy to proceed from genetic association to causal mechanism. By integrating summary statistics from large-scale OA GWAS with multi-tissue molecular quantitative trait loci (QTL) and single-cell expression QTL (sc-eQTL) data, we employed a combination of Bayesian colocalization and Mendelian randomization (MR) to establish a robust chain of causal evidence.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Our analysis definitively identifies \u003cstrong\u003eGNL3\u003c/strong\u003e as the causal effector gene at the 3q21 locus. MR analysis formally demonstrated that genetically predicted higher GNL3 expression is causally protective against both knee and hip OA (e.g., Knee OA: Odds Ratio = 0.85, p = 4.4e-4). Further single-cell causal inference pinpointed this protective effect to specific cellular contexts, most prominently in \u003cstrong\u003eactivated T-cells, neural cells under cellular stress, and developmental progenitors\u003c/strong\u003e. Moreover, we establish this locus as a pleiotropic hub, showing that its causal variants are shared with systemic OA risk factors, including Body Mass Index (BMI) and vitamin D levels.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003e This study establishes, for the first time, a complete chain of evidence from a shared causal variant to the regulation of \u003cem\u003eGNL3\u003c/em\u003e expression in specific cell types and, ultimately, to a causal impact on OA risk. Our findings converge on a novel mechanistic model: GNL3 acts as a master regulator of systemic homeostasis, where its genetically determined expression level modulates immune and neural cell responses to stress, thereby dictating an individual's susceptibility to OA. This work validates GNL3 as a high-confidence therapeutic target and provides a new framework for developing DMOADs aimed at reinforcing systemic homeostatic pathways.\u003c/p\u003e","manuscriptTitle":"Integrative Genomics Establishes GNL3 as a Pleiotropic Hub and Causal Gene for Osteoarthritis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-29 11:06:12","doi":"10.21203/rs.3.rs-7630649/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-12-19T12:19:12+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-06T11:10:21+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-18T12:03:14+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"247544256744619671311298205380984206296","date":"2025-11-17T12:19:34+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"245183356425791525907049730236754808267","date":"2025-11-17T11:23:09+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"293288106422423980711937137677161372837","date":"2025-10-08T17:13:50+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-18T11:12:38+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-17T07:30:00+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-17T07:29:44+00:00","index":"","fulltext":""},{"type":"submitted","content":"Human Genomics","date":"2025-09-16T12:46:54+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"human-genomics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"hugm","sideBox":"Learn more about [Human Genomics](http://humgenomics.biomedcentral.com/)","snPcode":"40246","submissionUrl":"https://submission.nature.com/new-submission/40246/3","title":"Human Genomics","twitterHandle":"@OAgenetics","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2d218ac2-0a4b-452c-bd42-8fcdf22089fc","owner":[],"postedDate":"September 29th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-02-23T16:04:47+00:00","versionOfRecord":{"articleIdentity":"rs-7630649","link":"https://doi.org/10.1186/s40246-026-00920-3","journal":{"identity":"human-genomics","isVorOnly":false,"title":"Human Genomics"},"publishedOn":"2026-02-16 15:57:32","publishedOnDateReadable":"February 16th, 2026"},"versionCreatedAt":"2025-09-29 11:06:12","video":"","vorDoi":"10.1186/s40246-026-00920-3","vorDoiUrl":"https://doi.org/10.1186/s40246-026-00920-3","workflowStages":[]},"version":"v1","identity":"rs-7630649","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7630649","identity":"rs-7630649","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}