Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Abstract

Objective

To identify susceptibility genes associated with varicose veins (VVs) using a cross-tissue transcriptome-wide association study (TWAS) framework.

Methods

We performed a cross-tissue TWAS by integrating GWAS data from the FinnGen R12 dataset (38,467 VVs cases and 432,223 controls of European ancestry) with eQTL data from GTEx V8. The initial analysis was conducted using the Unified Test for Molecular Signatures (UTMOST), and subsequent validation incorporated multiple complementary methods, including Functional Summary-based Imputation (FUSION), Conditional and Joint Association Analysis (COJO), Fine-mapping Of CaUsal gene Sets (FOCUS), and Multi-marker Analysis of Genomic Annotation (MAGMA). Mendelian randomization (MR) and colocalization analyses were performed to explore potential genetically supported associations between candidate genes and VVs. Finally, Western blotting (WB) was conducted in venous tissue samples collected from patients undergoing surgery for VVs to provide preliminary experimental validation.

Results

This cross-tissue TWAS analysis identified four genes (MAP3K2, PDK1, TMEM87B, and POLR1B) as susceptibility genes associated with VVs risk. MR indicated that PDK1, TMEM87B, and POLR1B may be associated with the development of VVs. Colocalization analysis suggested that POLR1B was the primary candidate gene, showing strong colocalization with VVs in whole blood (PPH4 = 0.987), and preliminary WB results suggested that the protein level of POLR1B was significantly elevated in varicose tissue.

Conclusion

Our findings identify POLR1B as a promising candidate susceptibility gene potentially associated with VVs risk, providing a basis for future mechanistic and translational studies.

Keywords

varicose veins cross-tissue TWAS mendelian randomization colocalization

1.Introduction

Varicose veins (VVs) are common and frequently occurring clinical diseases, marked by venous valve dysfunction, venous wall dilation, and chronic venous hypertension. This condition leads to blood pooling, inflammation, and progressive remodeling of the venous system, causing symptoms such as swelling, pain, and chronic venous ulcers in severe cases.^1,2 The pathophysiology of VVs involves hemodynamic abnormalities, extracellular matrix degradation, endothelial dysfunction, and abnormal phenotypic switching of smooth muscle cells (SMC), which collectively aggravate venous insufficiency.^1,3 VVs affect about 30% of adults globally, with a higher prevalence observed in females, older populations, and individuals with prolonged standing or sedentary lifestyles.^4,5 As a chronic and debilitating disease with no completely effective non-surgical treatment options,⁶ VVs impose significant healthcare burdens, affecting patients’ quality of life and incurring substantial economic costs.

While environmental and occupational factors are well-established contributors to VVs, emerging evidence highlights the critical role of genetic predisposition. Evidence from familial aggregation and twin studies indicates that the heritability of VVs ranges from 20% to 50%, supporting a polygenic mode of inheritance.⁷ Genetic susceptibility to VVs likely arises from the cumulative effects of multiple common variants, each with modest individual effects.⁸ Although GWAS have identified more than 100 risk loci (e.g., PIEZO1, FOXC2), these findings account for only a limited proportion of the heritable risk. Genetic overlap with other vascular diseases indicates shared risk factors and suggests potential pleiotropic effects.^9,10

TWAS provide a robust framework for bridging this knowledge gap by integrating eQTL data with GWAS data, prioritizing candidate genes, and elucidating their roles in disease mechanisms. Cross-tissue TWAS approaches, such as UTMOST,¹¹ significantly improve detection power by using eQTL effects across multiple tissues without losing tissue-specific regulatory signals. These methods have proven effective in dissecting the genetic architecture of complex, multi-system disorders, such as diabetes and autoimmune conditions,^12,13 yet their application to venous disorders remains limited.

Mendelian randomization (MR) is a powerful analytical approach that leverages genetic variants as instrumental variables to infer the relationship between genetically predicted exposures and disease outcomes, thereby providing evidence that is less susceptible to confounding and reverse causation compared with traditional observational analyses. Therefore, integrating MR with transcriptome-wide association studies (TWAS) can help prioritize candidate genes with higher biological plausibility.

This investigation implemented a cross-tissue TWAS by harmonizing GWAS summary statistics from the FinnGen R12 cohorts with GTEx v8 multi-omics regulatory quantitative trait loci (eQTL) profiles. FUSION and COJO analyses were used to identify tissue-specific and shared gene associations,¹⁴ followed by validation using MAGMA and FOCUS analyses.¹⁵ Subsequently, two-sample MR analysis was used to explore the potential causal association between target genes and VVs. In addition, the coloc R package was used for colocalization analysis to further clarify whether gene expression and VVs shared a driver locus. Furthermore, preliminary experimental validation was performed to support the biological relevance of key findings. This integrative approach aimed to elucidate the genetic background and signaling pathways involved in VVs, deepen the understanding of its pathogenesis, and provide insights that may inform future investigations. Figure 1 introduces the research process.

Figure 1.

Flowchart of the TWAS study of VVs. This study employed a multi-step integrated genomics approach. First, we integrated genome-wide association study (GWAS) summary statistics for varicose veins from FinnGen Release 12 and expression quantitative trait loci (eQTL) data from GTEx V8. Transcriptome-wide association studies (TWAS) were performed using UTMOST (cross-tissue analysis) as the core method, supplemented by other validation analyses, including FUSION, conditional and joint analysis (COJO), FOCUS, and MAGMA. Four candidate genes (MAP3K2, PDK1, TMEM87B, and POLR1B) were identified through multi-method cross validation. Subsequently, Mendelian randomization (MR) was performed on the candidate genes to evaluate potential causal relationships, and colocalization analysis (posterior probability of hypothesis 4, PPH4 ≥ 0.8) was conducted using the COLOC package to assess whether shared genetic signals were present. Finally, preliminary experimental validation was performed on four pairs of samples using Western blotting. Taken together, this integrative analytical framework suggests that POLR1B may represent a novel susceptibility gene for varicose veins, warranting further investigation.

2.Materials and methods

2.1 GWAS data

The GWAS data for VVs were obtained from the FinnGen R12 cohorts. This dataset included 38,467 clinically confirmed cases and 432,223 ancestry-matched controls of European descent, offering a solid genetic foundation for subsequent analyses.¹⁶ Variants with missing information, low imputation quality (INFO score <0.8), or minor allele frequency (MAF) <0.01 were excluded. SNPs with ambiguous alleles (A/T or C/G) were removed. All variants were aligned to the GRCh37 (hg19) reference genome build to ensure consistency with the eQTL datasets.

2.2 eQTL data

The eQTL data were obtained from the GTEx v8 dataset,¹⁷ which provides RNA-seq and eQTL summary statistics across 49 distinct human tissues. These tissues include vascular-related tissues (e.g., artery aorta, artery tibial, and artery coronary), as well as whole blood and other systemic tissues. The inclusion of multiple tissues enables the identification of both tissue-specific and shared genetic regulatory effects. This is particularly relevant for VVs, a complex vascular disorder influenced by both local venous wall alterations and systemic factors, including inflammation and immune responses reflected in circulating blood cells. In this study, whole-blood eQTL data were primarily used because of their larger sample size and higher statistical power compared with other tissues in the GTEx dataset, while the biological relevance of this signal was interpreted in relation to possible systemic inflammatory and immune pathways rather than direct venous wall expression alone. Only high-quality cis-eQTLs were retained. SNPs with low minor allele frequency (MAF < 0.01) or poor imputation quality were excluded. Gene expression levels were normalized according to GTEx standard pipelines, and only significant eQTL signals passing multiple testing correction were included in the analysis.

2.3 Venous tissue sample collection

Fresh varicose vein tissues and paired adjacent normal great saphenous vein branch tissues were obtained from four patients undergoing surgical treatment for varicose veins. All tissue specimens were collected immediately after excision, snap-frozen in liquid nitrogen, and stored at –80 °C until further use. Given that this study represents an exploratory analysis using existing surgical specimens, no formal sample size calculation was performed for the tissue validation experiments.

This research protocol was approved by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (approval number: 2025-E0408; approved on May 21, 2025), tissue samples were collected during June 2025, and all participants signed written informed consent forms. The study was conducted in accordance with the Declaration of Helsinki of 1975, as revised in 2024.

2.4 TWAS analysis methods

The study first used the UTMOST algorithm^11,18 for cross-tissue association analysis. This algorithm integrates multi-tissue GTEx data to construct cross-tissue covariance matrices and identifies gene-phenotype associations by analyzing the combined effects of single-nucleotide polymorphisms located within linkage disequilibrium (LD) regions. To further validate the robustness of the results, three complementary methods—FUSION (integrated into COJO), FOCUS, and MAGMA—were applied. The FUSION toolkit¹⁹ enables transcriptome-wide (TWAS) and regulome-wide (RWAS) association analyses by developing predictive models of the genetically regulated components of molecular phenotypes (e.g., transcriptomic or regulatory traits) and linking these components to disease associations. FOCUS²⁰ employs probabilistic fine-mapping based on correlations between TWAS signals to prioritize candidate genes with causal evidence within risk loci. MAGMA^15,21 adopts a dimensionality reduction strategy, extracts principal components from gene-SNP matrices while discarding low-eigenvalue components, and uses these components as predictors in linear regression models to derive gene-level association statistics and characterize the polygenic architecture of complex traits. Linkage disequilibrium (LD) structure was estimated using the European reference panel from the 1000 Genomes Project Phase 3. Genes that reached a significance level (FDR < 0.05) in UTMOST, FUSION, FOCUS, and MAGMA analyses were intersected to obtain a high-confidence susceptible gene set, which was used for subsequent validation.

2.5 Conditional and joint analyses

In the FUSION framework, identifying multiple genetic features within a locus requires evaluating their statistical independence using conditional models. To address this, the conditional and joint (COJO) analytical module in FUSION was applied during post-processing to distinguish independent genetic associations. By explicitly modeling linkage disequilibrium among genetic markers, this method increases the resolution of trait-variation analyses and provides a more precise depiction of the underlying genetic architecture.^22,23 Following this analysis, genes with retained statistical significance were categorized as having joint significance, highlighting their robust independent contributions. In contrast, genes that lost significance were classified as marginally significant, indicating their conditional dependence on colocalized genetic variants.

2.6 Gene analysis

Enrichment analyses for both tissue specificity and biological pathways were performed using the MAGMA framework.¹⁵ This approach uses gene expression datasets, such as those derived from the GTEx V8 project, to evaluate spatial expression profiles of genes across diverse tissues and explore their potential links to phenotypic traits. By aggregating association signals within biologically defined gene groups (e.g., genes involved in a shared pathway), MAGMA facilitates the systematic identification of functional modules or molecular pathways with significant associations with diseases or phenotypes. This methodology prioritizes collective gene-set contributions over individual gene effects, improving the detection of coordinated biological mechanisms underlying trait variability.

2.7 Mendelian randomization

To assess putative causal regulatory mechanisms underlying phenotypic variance, a two-sample MR analysis was performed using summary-level data. MR relies on three core instrumental variable assumptions: (1) relevance (genetic variants are associated with the exposure); (2) independence (genetic variants are independent of confounders); and (3) exclusion restriction (genetic variants influence the outcome only through the exposure). The exposure dataset comprised eQTL summary statistics for the target genes, identified by our multi-method TWAS analysis, from GTEx v8 whole-blood tissue, while the outcome dataset was the FinnGen R12 GWAS summary statistics for VVs. There was no sample overlap between the two datasets, as GTEx and FinnGen are independent cohorts with no known participant sharing. Both datasets were derived from populations of European ancestry, reducing the risk of population stratification bias. eQTL SNPs were used as exposure data and met three criteria: (1) p < 5 × 10^–824; (2) R²< 0.001, 10,000 kb window; (3) F-statistic ≥10 (F = β²/SE²), while VVs GWAS were used as outcome data. For genes with a single independent IV, the Wald ratio was used to estimate the causal effect. In cases where two or more independent IVs were available, the inverse-variance weighted (IVW) approach was used to derive the MR estimate. Sensitivity analyses (MR-Egger, Cochran’s Q) were not performed when IV count <3. p < 0.05 was considered nominally significant.

2.8 Colocalization analysis

The coloc R package was used to assess the colocalization of target genes (eQTL) and VVs (GWAS). The colocalization analysis included five scenarios: H0 (neither the eQTL nor the GWAS is within the given region), H₁/H₂ (either the eQTL or GWAS is within the given region), H₃ (both eQTL and GWAS are within the given region, but the driver locus is different), and H₄ (both eQTL and GWAS are within a given region and have the same driver locus). The prerequisite for colocalization is H₄(PPH4) ≥0.8; then, we concluded that the gene and VVs were driven by the same locus.^25,26

2.9 Western blotting

Total protein was extracted from collected venous tissue samples using RIPA lysis buffer supplemented with protease and phosphatase inhibitors, as previously described.²⁷ Tissue lysates were homogenized on ice and centrifuged at 12,000 ×g for 15 min at 4 °C to remove insoluble debris. Protein concentrations were determined using a bicinchoninic acid (BCA) assay according to the manufacturer’s instructions.

Equal amounts of protein (30–40 μg) from each sample were separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes. Membranes were blocked with 5% non-fat milk in Tris-buffered saline containing 0.1% Tween-20 (TBST) for 1 h at room temperature and then incubated overnight at 4 °C with primary antibodies against POLR1B (rabbit polyclonal, Catalog No. PA5-48383; Thermo Fisher Scientific, dilution 1:1000). After washing with TBST, membranes were incubated with appropriate horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature.

Protein bands were visualized using an enhanced chemiluminescence (ECL) detection system, as previously reported. GAPDH (rabbit recombinant antibody, Catalog No. 81640–5-RR, Proteintech, dilution 1:10000) was used as an internal loading control. Band intensities were quantified using ImageJ software, and relative protein expression levels were normalized to GAPDH. Western blot experiments were performed using 4 varicose vein tissue samples and 4 adjacent normal vein tissue samples. Statistical analysis was conducted using an unpaired Student’s t-test, and p < 0.05 was considered statistically significant.

3. Results

3.1 TWAS results of varicose veins

In the discovery phase of VVs research, cross-tissue TWAS identified 814 genes exhibiting nominal associations (P < 0.05) with VVs. Following multiple testing correction (FDR-adjusted α= 0.05), 344 of these genes retained statistical significance (P_FDR < 0.05), indicating robust associations between their predicted expression levels and VVs susceptibility (S1 Table). To validate these findings, single-tissue TWAS was conducted. Genes that were significant after FDR correction in at least one tissue type were retained, and 556 genes met this criterion (P_FDR < 0.05) (S2 Table). The single-tissue TWAS results revealed tissue-specific regulatory mechanisms in the pathogenesis of VVs. Subsequently, 29 candidate genes were consistently identified across both FUSION and UTMOST analyses, meeting stringent thresholds (S3 Table).

3.2 Conditional and joint association assessments

To assess whether the 29 overlapping candidate genes—predominantly mapped to chromosomes 2, 4, and 22—exerted independent effects on the phenotype, conditional and joint association analyses were performed. The analysis showed that 21 genes remained significantly and independently associated with VVs, suggesting their potential role in the pathogenesis of VVs (S4 Table). Owing to their TWAS significance being limited to a single tissue and potential confounding by linkage disequilibrium, eight genes were excluded from subsequent analyses; for example, POLR1B emerged as the predominant association signal at the 2q14.1 locus (Figure 2).

Figure 2.

Conditional and joint analysis. Green markers indicate genes (such as POLR1B) that were significantly associated in the joint analysis.

3.3 MAGMA analysis

The MAGMA framework identified 2,688 genes with significant genome-wide associations (P_FDR < 0.05) in multi-tissue transcriptome-wide analyses (S5 Table). Following false-discovery rate correction, the significant gene set showed enrichment in pathways encompassing the Nikolsky breast cancer 16q24 amplicon, circulatory system development, the PID AP-1 signaling cascade, cellular growth, and vasculature development (Supplementary Figure 1). Tissue-specific MAGMA evaluations further highlighted that VVs-related SNPs exhibited preferential localization in the uterus, coronary artery, tibial artery, fallopian tube, endocervix, aorta, and bladder, which are rich in epithelial/endothelial cells and smooth muscle cells (Supplementary Figure 2).

3.4 Statistical fine mapping

To identify potential genetic features associated with VVs, statistical fine-mapping was performed using FOCUS, which estimates posterior inclusion probabilities (PIPs) for prioritized variants. Notably, MAP3K2, PDK1, TMEM87B, and POLR1B exhibited robust PIP values exceeding the 0.5 threshold (0.983, 0.886, 0.87, and 0.975, respectively), indicating a high likelihood of association with VVs (S6 Table and Supplementary Figures 3–5).

3.5 Integration of the four TWAS methods

After false discovery rate (FDR) adjustment, integrative multi-method analysis—including UTMOST, FUSION, FOCUS, and MAGMA—consistently identified four high-confidence genes, MAP3K2, PDK1, TMEM87B, and POLR1B, associated with VVs. This consensus across methodologies is summarized in the Venn diagram (Figure 3).

Figure 3.

Venn diagram of genes identified by four methods shows that a total of four candi-date genes were identified (MAP3K2, PDK1, TMEM87B, and POLR1B).

3.6 Mendelian randomization analysis

The advantage of MR analysis is that SNPs are used as instrumental variables to reduce potential confounding and infer genetically supported associations between genes and VVs (GWAS) at the genetic level. eQTLs for the four genes, derived from whole-blood tissue datasets, were integrated with GWAS data for VVs traits. After filtering based on SNP thresholds (p < 5 × 10^–8; R²< 0.001, 10-Mb window; F-statistic ≥10), the MR results showed that MAP3K2 lacked a valid instrumental variable. For POLR1B (1 SNP) and PDK1 (1 SNP), the Wald ratio was applied; for TMEM87B (2 SNPs), IVW was used. Due to the limited number of IVs, pleiotropy tests and sensitivity analyses could not be reliably conducted. The MR analysis revealed a protective association for PDK1. Conversely, TMEM87B and POLR1B showed detrimental associations (Figure 4 and S7 Table).

Figure 4.

Results of Mendelian randomization analysis.

3.7 Colocalization analysis

Subsequently, colocalization analysis was performed using a 300 kb regional window to clarify whether the eQTL and VVs were driven by the same variant locus. The results showed that among the three candidate genes (PDK1, TMEM87B, and POLR1B), only POLR1B, located at the 2q14.1 locus, showed strong colocalization with VVs, with a PPH4 value of 0.987. Notably, rs62158643 was identified as a shared variant locus between POLR1B gene expression in whole-blood tissue and VVs (Figure 5 and S8 Table). These findings suggest that POLR1B expression may be genetically associated with VVs susceptibility through a shared regulatory locus.

Figure 5.

Colocalization analysis results of POLR1B and varicose veins. The common locus for regulating gene expression and varicose veins is rs62158643 , PPH4 = 0.987, indicating that there is a strong colocalization relationship between POLR1B and varicose veins.

3.8 Increased POLR1B protein expression in varicose vein tissues

To further validate the transcriptome-wide association and genetically supported association results, POLR1B protein expression was examined in varicose vein tissues and paired adjacent normal vein tissues by Western blot analysis. As shown in Figure 6 (Supplementary Figure 6), POLR1B protein levels appeared elevated in varicose vein tissues compared with paired adjacent normal vein tissues in this small exploratory cohort. Densitometric analysis showed increased POLR1B expression in varicose vein tissues after normalization to GAPDH.

Figure 6.

Representative Western blot images showing POLR1B protein expression in varicose vein (VV) tissues and paired adjacent normal vein tissues obtained from patients undergoing surgery for varicose veins. GAPDH was used as a loading control. The right panel shows densitometric quantification of POLR1B protein levels normalized to GAPDH. Data are presented as mean ± standard deviation from four paired samples (n = 4). Statistical significance was assessed using an unpaired Student’s t-test. P < 0.01.

These findings provide preliminary experimental support for the genetic and transcriptomic analyses, indicating that increased POLR1B expression may be associated with varicose vein pathology.

4. Discussion

After obtaining VVs GWAS data and eQTL data from FinnGen R12 and GTEx v8, respectively, candidate susceptibility genes for VVs risk were systematically evaluated. Using UTMOST, FUSION, FOCUS, and MAGMA analyses, followed by validation with MR and colocalization analyses, several candidate susceptibility genes for VVs were identified, among which POLR1B emerged as the most consistently supported gene. Furthermore, preliminary experimental validation suggested elevated POLR1B protein expression in varicose vein tissues, providing initial supportive evidence for its involvement in VVs.

Previous GWAS studies have identified more than 100 VVs-associated risk loci, although many lack functional insights.^28–31 By performing SMR (summary data-based Mendelian randomization) analysis, Cui et al.²⁸ identified four genes (PLEKHA5, CRIM1, CBWD1, and KRTAP5-AS1) significantly associated with VVs risk. Lin et al.²⁹ conducted MR analysis and showed that several plasma proteins have causal associations with VVs and may become potential targets for the treatment of this disease. Helkkula et al.³⁰ performed a GWAS analysis on VVs and identified a novel locus, GJD3, which was exclusively associated with a reduced risk of VVs. Similarly, Lee et al.³¹ reported six genome-wide significant SNPs, each mapping to a distinct gene, as being associated with VVs. Unlike single-tissue TWAS, which examines gene-trait associations within individual tissues, cross-tissue TWAS leverages eQTL signals across multiple tissues to improve power and uncover additional susceptibility genes expressed in diverse biological contexts. In this study, by leveraging cross-tissue regulatory information, the prioritization of candidate genes was improved, and POLR1B was highlighted as a novel gene potentially involved in VVs pathogenesis.

An important consideration in this study is that the strongest colocalization signal for POLR1B was detected in whole blood rather than vascular tissues. Although VVs are primarily characterized by venous wall remodeling, endothelial dysfunction, vascular smooth muscle cell (VSMC) phenotypic switching, and extracellular matrix (ECM) remodeling, increasing evidence suggests that systemic inflammatory and immune regulatory mechanisms also contribute to disease progression. Whole-blood eQTL signals reflect the transcriptional status of circulating blood cells and may therefore serve as surrogate indicators of systemic inflammatory or immune activity rather than direct vascular wall expression.³² In this context, circulating immune cells and inflammatory mediators have been shown to participate in endothelial dysfunction, VSMC remodeling, and ECM alterations, which are central pathological features of VVs.³³ Furthermore, a recent Mendelian randomization study identified genetically supported associations between multiple immune cell phenotypes—including myeloid dendritic cells, monocytes, and regulatory T cells—and VVs risk.³⁴ In addition, POLR1B-mediated ribosome biogenesis has been implicated in inflammatory responses and immune regulation,³⁵ suggesting that altered POLR1B expression in circulating cells may reflect systemic inflammatory states that indirectly contribute to vascular remodeling. Therefore, blood-based POLR1B transcriptomic signals may act as proxies for systemic inflammatory or immune processes associated with venous wall degeneration, rather than directly representing structural gene expression within vascular tissues. Future studies examining POLR1B expression specifically in venous endothelial cells and VSMCs will help clarify its tissue-specific role in VVs pathogenesis.

POLR1B encodes a core subunit of RNA polymerase I, which is responsible for ribosomal RNA transcription, a key rate-limiting step in ribosome biogenesis. Ribosome biogenesis is essential for maintaining cellular protein synthesis capacity and is closely linked to cell growth and proliferation.³⁶ Studies by Huang et al.³⁷ have demonstrated that enhanced ribosome biosynthesis can promote abnormal proliferation and activation of vascular cells, leading to cardiovascular disease, while Holdt et al.³⁸ found that controlling ribosome biogenesis can inhibit the proliferation of VSMCs, thereby preventing atherosclerosis. These findings suggest that POLR1B-mediated ribosome biosynthesis is crucial for vascular homeostasis and may contribute to the development of varicose veins.

In addition, ribosome biosynthesis has been shown to be involved in the fibrotic process. A study by Hu et al.³⁹ indicated that apigenin attenuates myocardial fibrosis and endothelial-to-mesenchymal transition (EMT) by inhibiting ribosome biogenesis within coronary endothelial cells. Enhanced ribosomal activity supports increased protein synthesis and ECM production, which are hallmarks of fibroblast activation. Given that excessive ECM deposition and venous wall thickening are key pathological features of VVs, altered POLR1B expression may contribute to disease progression by facilitating fibroblast activation and matrix remodeling. Moreover, ribosome biogenesis is closely linked to inflammatory signaling and cellular stress responses.⁴⁰,⁴¹ Dysregulation of ribosome biogenesis can trigger inflammatory pathways and influence vascular homeostasis. In this context, increased POLR1B expression in circulating cells may reflect a systemic pro-inflammatory state that promotes endothelial dysfunction and venous wall remodeling. Together, these findings provide a plausible biological framework linking POLR1B-mediated ribosome biogenesis to vascular remodeling in VVs, while the exact molecular mechanism by which POLR1B contributes to VVs requires further clarification.

In addition to POLR1B, this study identified two other potential genes associated with VVs risk, including PDK1 and TMEM87B. The results suggest that dysregulation of these genes may directly or indirectly contribute to the pathogenesis of VVs. PDK1 plays a crucial role in cellular adaptation to hypoxic conditions by inhibiting pyruvate dehydrogenase activity, which in turn regulates metabolic pathways. Additionally, PDK1 promotes cell proliferation in response to hypoxia, a process that may contribute to endothelial cell dysfunction and the development of varicose veins.⁴²,⁴³ Dysregulation of PDK1 may disrupt mitochondrial function and dynamics, leading to vascular endothelial cell injury, a pathological feature of VVs. TMEM87B encodes a transmembrane protein that has been implicated in congenital heart disease,⁴⁴,⁴⁵ suggesting its potential involvement in vascular development. Its dysregulation in varicose veins may reflect abnormal vascular remodeling and contribute to disease progression.

Although this comprehensive study yielded relatively robust results regarding the genetic susceptibility to VVs, several limitations should be acknowledged. First, in the context of TWAS analysis, due to the effects of linkage disequilibrium, genetic variants may influence the expression of multiple neighboring genes, thereby making it difficult to distinguish the true pathogenic gene from associated signals. Furthermore, co-regulation among genes situated within the same regulatory network may further confound the results of TWAS analysis. Second, although Mendelian randomization (MR) and colocalization analyses suggested a potential genetically supported association between POLR1B expression and VVs, the MR estimates for some genes were derived from single genetic instruments, precluding formal pleiotropy assessment. Consequently, these findings should be interpreted cautiously and require replication using more instrumental variables or functional validation. Moreover, the analyses were based on available genetic data and may not fully capture the complexity of gene-environment interactions. Third, the Western blot results, while supportive, were derived from a small cohort (n = 4). This limited sample size substantially restricts statistical power and may not capture population-level variability. Therefore, these findings should be considered preliminary and hypothesis-generating, rather than definitive validation. Independent replication in larger cohorts is essential before any firm conclusions can be drawn. Fourth, this study was based on the FinnGen cohort, which predominantly includes individuals of European ancestry, potentially limiting the generalizability of the findings to other populations. Future studies using animal models or longitudinal human cohorts will be crucial to validate the causal role of POLR1B in VVs development and to explore potential environmental factors that may interact with genetic predisposition.

5. Conclusion

In conclusion, this study identified POLR1B as a potential genetic susceptibility gene for VVs risk through cross-tissue TWAS analysis. MR analysis suggested a genetically supported association between POLR1B expression and VVs risk. Colocalization analysis identified shared SNPs between POLR1B and VVs. The Western blot results provide preliminary experimental support for the research findings. However, given the preliminary nature and limitations of the present study, future research should further explore the specific molecular mechanisms through which POLR1B—and the ribosome biogenesis it mediates—influences the structure and function of the venous wall. Furthermore, investigating the relationship between Pol I activity and the risk of varicose veins (VVs) may offer new perspectives on the understanding of this disease.

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Supplemental material for Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study by Quanxing Kuang, MM, Qingfeng Zhu, MM, Xiaocheng Li, MM, Han Yang, MD, Wenhong Jiang, MM, Youfu Wang, MM, Xiao Qin, MD in Science Progress

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Supplemental material

Supplemental material - Novel susceptibility genes for varicose veins revealed by a cross-tissue transcriptome-wide association study

Footnotes

Acknowledgments

We thank the participants and researchers of FinnGen R12 and GTEx V8, as well as the patients who participated in this research study. During the preparation of this work, the authors used ChatGPT and Grammarly to improve the language and readability of the manuscript. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

ORCID iDs

Youfu Wang

Xiao Qin

Ethical considerations

The study was approved by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (protocol code 2025-E0408; May 21, 2025).

Consent to participate

Written informed consent was obtained from all subjects involved in the study.

Author contributions

Q.K.: Conceptualization, Data curation, Writing–original draft. Q.Z.: Software, Formal analysis.

X.L.:Writing–original draft, Formal analysis.

H.Y.:Writing–original draft, Writing – review & editing.

W.J.: Data curation, Writing–original draft.

Y.W.: Software, Writing–original draft.

X.Q.: Writing – review & editing, Funding acquisition. All authors reviewed the manuscript and agreed to the published version of the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by Guangxi Science and Technology Plan Project (2017AB45033).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this articleThe authors declare no conflicts of interest.

Data Availability Statement

The GWAS summary statistics for varicose veins are available from the FinnGen consortium (https://www.finngen.fi/en). The eQTL data are available from the GTEx Portal (). All other data supporting the findings of this study are available within the article and its supplementary materials or from the corresponding author upon reasonable request.

Supplemental material

Supplemental material for this article is available online.

References

Chen

Liu

Zhang

, et al. Pathophysiology and Genetic Associations of Varicose Veins: A Narrative Review. Angiology 2024; 76: 33197241227598–33197241227628. https://doi.org/10.1177/00033197241227598

Lim

Davies

. Pathogenesis of primary varicose veins. Br J Surg 2009; 96: 1231–1242. https://doi.org/10.1002/bjs.6798

Ortega

Romero

Asúnsolo

, et al. Behavior of Smooth Muscle Cells under Hypoxic Conditions: Possible Implications on the Varicose Vein Endothelium. Biomed Res Int 2018; 2018: 7156150. https://doi.org/10.1155/2018/7156150

Aslam

Muhammad Asif

Ahmad

, et al. Global impact and contributing factors in varicose vein disease development. SAGE Open Med 2022; 10: 20503121221118992. https://doi.org/10.1177/20503121221118992

Evans

Fowkes

Ruckley

, et al. Prevalence of varicose veins and chronic venous insufficiency in men and women in the general population: Edinburgh Vein Study. J Epidemiol Community Health 1999; 53: 149–153. https://doi.org/10.1136/jech.53.3.149

Belramman

Bootun

Lane

TRA

, et al. Endovenous Management of Varicose Veins. Angiology 2019; 70: 388–396. https://doi.org/10.1177/0003319718780049

MYM

Andrew

Spector

, et al. Linkage to the FOXC2 region of chromosome 16 for varicose veins in otherwise healthy, unselected sibling pairs. J Med Genet 2005; 42: 235–239. https://doi.org/10.1136/jmg.2004.024075

Lyons

Saha

Seet

, et al. Human venous valve disease caused by mutations in FOXC2 and GJC2. J Exp Med 2017; 214: 2437–2452. https://doi.org/10.1084/jem.20160875

Levin

Huffman

Verma

, et al. Genetics of varicose veins reveals polygenic architecture and genetic overlap with arterial and venous disease. Nat Cardiovasc Res 2023; 2: 44–57. https://doi.org/10.1038/s44161-022-00196-5

10.

Shadrina

Smetanina

Sokolova

, et al. Association of polymorphisms near the FOXC2 gene with the risk of varicose veins in ethnic Russians. Phlebology 2016; 31: 640–648. https://doi.org/10.1177/0268355515607404

11.

, et al. A statistical framework for cross-tissue transcriptome-wide association analysis. Nat Genet 2019; 51: 568–576. https://doi.org/10.1038/s41588-019-0345-7

12.

Wang

Khunsriraksakul

Markus

, et al. Integrating single cell expression quantitative trait loci summary statistics to understand complex trait risk genes. Nat Commun 2024; 15: 4260. https://doi.org/10.1038/s41467-024-48143-1

13.

Dou

Liu

Kong

, et al. Shared genetic architecture of type 2 diabetes with muscle mass and function and frailty reveals comorbidity etiology and pleiotropic druggable targets. Metabolism 2025; 164: 156112. https://doi.org/10.1016/j.metabol.2024.156112

14.

Mai

Gao

, et al. Transcriptome-wide association studies: recent advances in methods, applications and available databases. Commun Biol 2023; 6: 899. https://doi.org/10.1038/s42003-023-05279-y

15.

de Leeuw

Mooij

Heskes

, et al. MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput Biol 2015; 11: e1004219. https://doi.org/10.1371/journal.pcbi.1004219

16.

Kurki

Karjalainen

Palta

, et al. Author Correction: FinnGen provides genetic insights from a well-phenotyped isolated population. Nature 2023; 615: E19. https://doi.org/10.1038/s41586-023-05837-8

17.

The Genotype-Tissue Expression (GTEx) project . Nat Genet 2013; 45: 580–585. https://doi.org/10.1038/ng.2653

18.

Rodriguez-Fontenla

Carracedo

. UTMOST, a single and cross-tissue TWAS (Transcriptome Wide Association Study), reveals new ASD (Autism Spectrum Disorder) associated genes. Transl Psychiatry 2021; 11: 256. https://doi.org/10.1038/s41398-021-01378-8

19.

Gusev

Shi

, et al. Integrative approaches for large-scale transcriptome-wide association studies. Nat Genet 2016; 48: 245–252. https://doi.org/10.1038/ng.3506

20.

Mancuso

Freund

Johnson

, et al. Probabilistic fine-mapping of transcriptome-wide association studies. Nat Genet 2019; 51: 675–682. https://doi.org/10.1038/s41588-019-0367-1

21.

Yang

Huang

. A cross-tissue transcriptome-wide association study identified susceptibility genes for age-related macular degeneration. Sci Rep 2025; 15: 4788. https://doi.org/10.1038/s41598-025-89246-z

22.

Deng

Pan

. Improved Use of Small Reference Panels for Conditional and Joint Analysis with GWAS Summary Statistics. Genetics 2018; 209: 401–408. https://doi.org/10.1534/genetics.118.300813

23.

Chen

Jin

Chu

, et al. A cross-tissue transcriptome-wide association study reveals GRK4 as a novel susceptibility gene for COPD. Sci Rep 2024; 14: 28438. https://doi.org/10.1038/s41598-024-80122-w

24.

Panagiotou

Ioannidis

JPA

. What should the genome-wide significance threshold be? Empirical replication of borderline genetic associations. Int J Epidemiol 2012; 41: 273–286. https://doi.org/10.1093/ije/dyr178

25.

Giambartolomei

Vukcevic

Schadt

, et al. Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genet 2014; 10: e1004383. https://doi.org/10.1371/journal.pgen.1004383

26.

Liu

, et al. Identification of novel therapeutic targets for chronic kidney disease and kidney function by integrating multi-omics proteome with transcriptome. Genome Med 2024; 16: 84. https://doi.org/10.1186/s13073-024-01356-x

27.

Mahmood

Yang

P-C

. Western blot: technique, theory, and trouble shooting. N Am J Med Sci 2012; 4: 429–434. https://doi.org/10.4103/1947-2714.100998

28.

Cui

Zhou

, et al. Identifying potential drug targets for varicose veins through integration of GWAS and eQTL summary data. Front Genet 2024; 15: 1385293. https://doi.org/10.3389/fgene.2024.1385293

29.

Lin

Zhou

Liu

, et al. Identification of potential drug targets for varicose veins: a Mendelian randomization analysis. Front Cardiovasc Med 2023; 10: 1126208. https://doi.org/10.3389/fcvm.2023.1126208

30.

Helkkula

Hassan

Saarentaus

, et al. Genome-wide association study of varicose veins identifies a protective missense variant in GJD3 enriched in the Finnish population. Commun Biol 2023; 6: 71. https://doi.org/10.1038/s42003-022-04285-w

31.

Lee

M-L

Liang

Chuang

C-H

, et al. A genome-wide association study for varicose veins. Phlebology 2022; 37: 267–278. https://doi.org/10.1177/02683555211069248

32.

Liefferinckx

Stern

Perée

, et al. The identification of blood-derived response eQTLs reveals complex effects of regulatory variants on inflammatory and infectious disease risk. PLoS Genet 2025; 21: e1011599. https://doi.org/10.1371/journal.pgen.1011599

33.

Xie

Chen

, et al. The crosstalks between vascular endothelial cells, vascular smooth muscle cells, and adventitial fibroblasts in vascular remodeling. Life Sci 2024; 361: 123319. https://doi.org/10.1016/j.lfs.2024.123319

34.

Zhang

Q-F

Liu

Q-L

. Causal role of immune cells in varicose veins: insights from a Mendelian randomization study. Sci Rep 2025; 15: 19527. https://doi.org/10.1038/s41598-025-03602-7

35.

Wang

Zhang

X-Y

Liao

N-Q

, et al. Identification of ribosome biogenesis genes and subgroups in ischaemic stroke. Front Immunol 2024; 15: 1449158. https://doi.org/10.3389/fimmu.2024.1449158

36.

Jiao

Liu

X-Y

, et al. Ribosome biogenesis in disease: new players and therapeutic targets. Signal Transduct Target Ther 2023; 8: 15. https://doi.org/10.1038/s41392-022-01285-4

37.

Huang

Wei

Jia

, et al. CircRNA-RBAC induces cardiac repair by promoting ribosome biogenesis and cell cycle progression in cardiomyocytes. Int J Biol Macromol 2024; 287: 138406. https://doi.org/10.1016/j.ijbiomac.2024.138406

38.

Holdt

Stahringer

Sass

, et al. Circular non-coding RNA ANRIL modulates ribosomal RNA maturation and atherosclerosis in humans. Nat Commun 2016; 7: 12429. https://doi.org/10.1038/ncomms12429

39.

Yang

Tang

, et al. Apigenin inhibits endothelial-to-mesenchymal transition of coronary artery endothelial cells and myocardial fibrosis by regulating ribosome biogenesis through GTPBP4 modulation. Hum Cell 2025; 38: 157. https://doi.org/10.1007/s13577-025-01291-y

40.

Shen

Peng

Zhang

, et al. Eukaryotic translation initiation factor 6-mediated ribosome biogenesis promotes synovial aggression and inflammation by increasing the translation of SP1 in rheumatoid arthritis. Int Immunopharmacol 2024; 142: 113164. https://doi.org/10.1016/j.intimp.2024.113164

41.

Cerezo

Henry

Henras

, et al. Ribosome biogenesis and homeostasis: in the front line to cope with cellular stress. Biochimie 2025; 238: 43–56. https://doi.org/10.1016/j.biochi.2025.08.016

42.

McFate

Mohyeldin

, et al. Pyruvate dehydrogenase complex activity controls metabolic and malignant phenotype in cancer cells. J Biol Chem 2008; 283: 22700–22708. https://doi.org/10.1074/jbc.M801765200

43.

Hitosugi

Fan

Chung

T-W

, et al. Tyrosine phosphorylation of mitochondrial pyruvate dehydrogenase kinase 1 is important for cancer metabolism. Mol Cell 2011; 44: 864–877. https://doi.org/10.1016/j.molcel.2011.10.015

44.

H-C

Coughlin

Geiger

, et al. Discovery of a potentially deleterious variant in TMEM87B in a patient with a hemizygous 2q13 microdeletion suggests a recessive condition characterized by congenital heart disease and restrictive cardiomyopathy. Cold Spring Harb Mol Case Stud 2016; 2: a000844. https://doi.org/10.1101/mcs.a000844

45.

Digilio

Dentici

Loddo

, et al. Congenital heart defects in the recurrent 2q13 deletion syndrome. Eur J Med Genet 2022; 65: 104381. https://doi.org/10.1016/j.ejmg.2021.104381

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.10 MB

0.01 MB

0.61 MB

0.02 MB

0.01 MB

0.21 MB

0.04 MB