Varicose veins of lower extremities: insights from the ... · The cumulative evidence from epidemiological, family, and genetic association studies 53. strongly indicates that there

1

Varicose veins of lower extremities: insights from the first large-scale genetic study 1

2

Alexandra S. Shadrina1,2*, Sodbo Z. Sharapov2,3, Tatiana I. Shashkova2,4,5, Yakov A. Tsepilov2,3* 3

4

1 Laboratory of Pharmacogenomics, Institute of Chemical Biology and Fundamental Medicine, 5

Novosibirsk, Russia 6

2 Laboratory of Theoretical and Applied Functional Genomics, Novosibirsk State University, 7


3 Laboratory of Recombination and Segregation Analysis, Institute of Cytology and Genetics, 9


4 Knomics LLC, Skolkovo Innovation Center, Moscow, Russia 11

5 Department of Biological and Medical Physics, Moscow Institute of Physics and Technology, 12

Moscow, Russia 13

14

15

16

* Corresponding authors: 17

Dr. Alexandra Shadrina, Laboratory of Pharmacogenomics, Institute of Chemical Biology and 18

Fundamental Medicine, 8 Lavrentjev Avenue, Novosibirsk 630090, Russia. Tel: +7 383 19

3635171, E-mail: [email protected] 20

ORCID 0000-0003-1384-3413 21

Dr. Yakov Tsepilov, Laboratory of Recombination and Segregation Analysis, Institute of 22

Cytology and Genetics, 10 Lavrentjev Avenue, Novosibirsk 630090, Russia. Tel: +7 383 23

3634980, E-mail: [email protected] 24

ORCID 0000-0002-4931-6052 25

26

not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which wasthis version posted July 13, 2018. . https://doi.org/10.1101/368365doi: bioRxiv preprint

mailto:[email protected]


https://doi.org/10.1101/368365

2

Abstract 27

Varicose veins of lower extremities (VVs) are a common multifactorial vascular disease. 28

Genetic factors underlying VVs development remain largely unknown. Here we report the first 29

large-scale study of VVs performed on a freely available genetic data of 408,455 European-30

ancestry individuals. We identified 7 reliably associated loci that explain 10% of the SNP-based 31

heritability, and prioritized the most likely causal genes CASZ1, PPP3R1, EBF1, STIM2, and 32

HFE. Genetic correlation analysis confirmed known epidemiological associations and found 33

genetic overlap with various traits including fluid intelligence score, educational attainment, 34

smoking, and pain. Finally, we observed causal effects of height, weight, both fat and fat-free 35

mass, and plasma levels of MICB and CD209 proteins. 36


https://doi.org/10.1101/368365

3

Varicose veins (VVs) are one of the clinical manifestations of chronic venous disease 37

posing both a cosmetic and medical problem. VVs can be found in different parts of the body, 38

but most commonly occur in the lower extremities. Prevalence estimates of this condition vary 39

across ethnic groups ranging from 2-4% in the Northern group of the Cook Islands to 50-60% in 40

some countries of the Western world1. Increased age, female sex, number of pregnancies, 41

obesity, history of deep venous thrombosis, and standing occupation are among other risk 42

factors2. VVs not related to the post-thrombotic syndrome or venous malformations are defined 43

as primary VVs. 44

Pathogenesis of VVs is still not fully clarified. According to current understanding, key 45

factors implicated in VVs development include changes in hemodynamic forces (decrease in 46

laminar shear stress and increase in venous filling pressure), endothelial activation, 47

inflammation, hypoxia, and dysregulation of matrix metalloproteinases and their tissue 48

inhibitors3-5. These alterations underlie pathological remodeling of the vascular wall and loss of 49

its tone. Questions remain about the order of events and the primary stimulus triggering the set of 50

disease-related changes. 51

The cumulative evidence from epidemiological, family, and genetic association studies 52

strongly indicates that there is a hereditary component in VVs etiology6-8. However, despite 53

progress in this field9-12, current knowledge of the genetic basis of this pathology is far from 54

being complete. Elucidating genes involved in susceptibility to VVs would help to identify key 55

molecular players in the disease initiation, provide deeper insights into its pathogenesis, and 56

eventually contribute to development of improved targeted therapy aimed at VVs treating and 57

preventing. 58

Large-scale biobanks linked to electronic health records open up unparalleled opportunities 59

to investigate the genetics of complex traits. Today, UK Biobank is the largest repository that 60

contains information on genotypes and phenotypes for half a million participating individuals13. 61

This resource is open to all bona fide researchers, and access to data is provided upon approval 62


https://doi.org/10.1101/368365

4

of their application and payment of necessary costs. However, the need to incur high costs 63

related to data access and computation can be an insurmountable obstacle for those who cannot 64

afford these expenses. The Neale Lab (http://www.nealelab.is/) and the Gene ATLAS database14 65

(http://geneatlas.roslin.ed.ac.uk/) are two independent projects, which intend to remove this 66

burden by generating genetic association data and sharing them with broader scientific 67

community. These resources provide free open access to “quick-and-dirty” genome-wide 68

association study (GWAS) summary statistics for a wide range of phenotypes measured in the 69

UK Biobank. In our study, we aimed to employ state-of-the-art bioinformatics approaches to 70

extract maximum possible information from these open resources with regard to the genetics of 71

VVs of lower extremities. Our objectives were to (1) identify genetic loci reliably associated 72

with VVs risk and prioritize the genes that account for the revealed associations, (2) elucidate 73

pleiotropic effects of identified loci, (3) investigate genetic overlap between VVs and other 74

complex traits, (4) gain etiological insights and explore cause-and-effect relationships by means 75

of Mendelian randomization analysis. 76

Results 77

Study design, assumptions, and limitations 78

Our study was designed as a classical GWAS involving the initial discovery screening 79

followed by the replication stage. All calculations were entirely based on the UK Biobank data 80

for white British individuals available in open access databases. The overall workflow of the 81

study is depicted in Fig.1. In the discovery stage, we used GWAS summary statistics provided 82

by the Neale Lab. To be able to perform a replication, we extracted the UK Biobank data not 83

covered by this project by means of reverse meta-analysis of two overlapping datasets: genetic 84

association data provided by the Gene ATLAS database and the GWAS data supplied by the 85

Neale Lab. The projects used different software and methods of analysis. The Gene ATLAS 86

project14 applied less stringent filtering criteria and therefore had larger sample size. This 87

difference in quality control approaches enabled us to generate a subset of “subtracted” 88


http://www.nealelab.is/

http://geneatlas.roslin.ed.ac.uk/

https://doi.org/10.1101/368365

5

individuals and use them as a replication cohort. Data obtained in both GWAS stages were used 89

for further downstream analyses. 90

In carrying out this study, we had to face a number of challenges and limitations that must 91

be acknowledged. The first limitation was intrinsic to the general approach to phenotype 92

definition based on the electronic medical records system, which was employed in the UK 93

Biobank study. Phenotype “VVs of lower extremities” was defined based on International 94

Classification of Disease (ICD-10) billing code “I83” present in the electronic patient record. 95

The Neale Lab reported the phenotype prevalence of 2.1%. It is much lower than VVs 96

prevalence estimated by European epidemiological studies. Despite the evidence of a “healthy 97

volunteer” selection bias in the UK Biobank study15, such a low rate indicates that a proportion 98

of individuals remained undiagnosed. This is in line with a recent primary healthcare register-99

based study reporting VVs prevalence rate of 3% in German general practice16. This 100

phenomenon could be explained by a non-life-threatening nature of varicose veins, which might 101

discourage patients from communication to the doctor. Given that individuals not diagnosed with 102

I83 served as controls in our study, we could therefore expect an overall decrease in the 103

statistical power to detect gene-disease associations. Another potential source of missing 104

associations was immensely strict criteria used by the Neale Lab for single nucleotide 105

polymorphisms (SNPs) quality control removing around 75% of SNPs initially provided by UK 106

Biobank. 107

The next important limitation arose from the lack of access to individual-level data 108

resulting in the inability to control a possible selection or sampling bias. We suggested that traits 109

related to VVs risk factors could potentially cause unequal representation of patients with 110

different characteristics in the case and the control groups, and thereby induce spurious 111

associations or effect modification. The Neale Lab analyses were adjusted for sex, but other 112

factors were beyond our control. In order to address this challenge, we performed an adjustment 113

for two potential confounders – body mass index (BMI) and deep venous thrombosis (DVT) – by 114


https://doi.org/10.1101/368365

6

implementation of the method based on GWAS summary statistics17,18 (Supplementary 115

Methods). 116

Last but not least, a noteworthy debatable point in our study is the approach to generating a 117

replication cohort. Due to our intention to use only free open resources, we had to bypass the 118

need to refer to individual patient data. Our replication study sample comprised individuals who 119

have passed the Gene ATLAS quality control criteria, but have been filtered out by the Neale 120

Lab. It was almost 5 times smaller than our discovery study sample. However, although this 121

“dirty” design limited our power to replicate associations, it still could not produce any false 122

positive findings. Moreover, we calculated genetic correlations between VVs in the replication 123

cohort and VVs in both public databases to ensure that these traits are genetically similar 124

(Supplementary table 1). VVs in our replication study was highly correlated with the same trait 125

in other datasets (rg ~ 0.9). Furthermore, it is noteworthy that VVs in the Gene ATLAS, the 126

Neale Lab as well as in BMI- and DVT-adjusted Neale Lab GWAS were almost genetically 127

equivalent (rg = 0.99 or rg = 1.00). 128

Summarizing the above, we can state that the limitations which we could not circumvent 129

were mainly related to the loss of statistical power. Nevertheless, given a large sample size and a 130

large number of associations tested, we assume that this obstacle could be at least partially 131

compensated by a huge scale of the UK Biobank study itself. 132

Genome-wide association study for VVs of lower extremities 133

We defined associated loci as regions within 250 kb from the lead SNP and reported only 134

the most significant SNP hits per locus. 135

In the discovery stage, we identified 18 loci that met a genome-wide level of statistical 136

significance of P < 5.0e-08 as provided by the Neale Lab (Table 1, Supplementary table 2). 137

Adjustment for DVT and BMI only slightly changed the observed effects of SNPs, but yielded 138

two additional genome-wide significant loci tagged by rs11693897 and rs247749. However, both 139

these SNPs as well as polymorphism rs1061539 turned insignificant after the correction for the 140


https://doi.org/10.1101/368365

7

genomic inflation factor (LD Score regression intercept) λ = 1.0403. In the replication stage, 7 141

loci were replicated at P < 0.0029 (0.05/17). Manhattan plot of –log10(P) is presented in Fig.2. A 142

quantile-quantile plot for observed vs. expected distribution of P-values is shown in 143

Supplementary fig. 2. 144

Regional association plots for replicated loci are given in Supplementary fig. 1. Genes 145

located in these regions are listed in Supplementary table 2. We performed conditional and joint 146

(COJO) analysis and detected two independent signals in the locus on chromosome 16 at a 147

distance of 39 kb from each other (Supplementary table 3). Lead SNPs rs9972645 and rs2911463 148

were in low linkage disequilibrium (r2 = 0.15). 149

Functional annotation of the revealed signals 150

Literature-based annotation 151

We explored whether genes in close proximity to replicated hits could have biologically 152

plausible roles in VVs based on their functions or previously revealed associations with other 153

complex traits. The results of literature-based candidate gene prioritization are provided in Table 154

2. Furthermore, we compiled a list of traits associated at a genome-wide significance level with 155

either lead SNPs, or variants in linkage disequilibrium (LD) with these polymorphisms using 156

both the Pubmed and the PhenoScanner databases (Table 2, Supplementary table 4). Our analysis 157

showed that among the genes nearest to the top GWAS loci (±250 kb), 2 genes were related to 158

vascular development and remodeling (CASZ1 and PIEZO1) with one gene having putative role 159

in vascular remodeling (STIM2), 4 genes were implicated in blood pressure and hypertension 160

(CASZ1, PIEZO1, SLC12A2, and EBF1), 2 genes were linked to inflammation (PPP3R1 and 161

EBF1), and 2 genes were ion channels involved in regulation of both cell volume and vascular 162

tone (PIEZO1 and SLC12A2). Near the SLC12A2, we identified the FBN2 gene which encodes a 163

protein regulating elastogenesis, so we prioritized both genes. 164

Two signals overlapped with findings from a recent GWAS for VVs conducted by 165

“23andMe” company12: rs11121615 in the CASZ1 gene, which was among the top SNPs 166


https://doi.org/10.1101/368365

8

identified in that study, and rs2861819 near the PPP3R1, which was in complete LD 167

(D’ = 1.00, r2 = 0.98) with their another hit rs6712038. For 3 loci (tagged by rs2911463, 168

rs3101725, and rs7773004) a genome-wide significant association with red blood cell traits has 169

previously been demonstrated. One of these signals was near the HFE gene implicated in iron 170

absorption. The same locus showed association with height. 171

Changes in blood pressure, abnormal vascular wall remodeling, altered vascular tone, 172

deterioration of vein wall elastic properties as well as inflammation are known factors related to 173

VVs. Association with red blood cell traits is intriguing. Hypothetically, both erythrocyte 174

homeostasis and VVs could be linked to iron overload19,20. It is noteworthy that association 175

between hemochromatosis-related functional variant in the HFE gene and primary VVs was 176

revealed in our previous candidate-gene study in ethnic Russian individuals19. 177

Searching for functional variants 178

Ensembl Variant Effect Predictor (VEP)51 analysis identified one moderate-impact 179

missense variant rs2044693 in the PNO1 gene (in the locus tagged by rs2861819). SIFT and 180

PolyPhen tools predicted no damaging effect of this SNP on the protein function. Additionally, 181

we revealed 19 SNPs in regulatory regions (in loci tagged by rs9972656, rs2861819, 182

rs11135046, and rs7773004), and one synonymous variant rs1368298 in the EBF1 gene (in the 183

locus tagged by rs11135046). The full data is available in Supplementary table 5. 184

Testing for pleiotropy: VVs and gene expression 185

We used Summary data-based Mendelian Randomization (SMR) analysis followed by the 186

Heterogeneity in Dependent Instruments (HEIDI)52 test to identify genes whose expression level 187

is associated with the same causal SNPs that affect the risk of VVs. The results are given in 188

Supplementary table 6. SMR/HEIDI analysis provided evidence for pleiotropic effects of 189

variants within rs3101725 and rs2861819 regions. In the first region, association was revealed 190

with the expression of long intergenic non-protein coding RNA 1184 (LINC01184) located in 191

close proximity to the SLC12A2 on the reverse strand. Changes in the SLC12A2 expression were 192


https://doi.org/10.1101/368365

9

not associated with VVs in our tests. Interestingly, the previous study showed that disruption of a 193

bi-directional promoter between these genes altered the expression of both SLC12A2 and 194

LINC01184, while deletion of the third exon of LINC01184 affected only LINC01184 195

expression36. 196

In the locus tagged by rs2861819, our analysis identified four genes – PPP3R1, PLEK, 197

WDR92, and PNO1. It is likely that this region contains a strong eQTL that affects the 198

expression of nearby genes in different tissues and influences the risk of VVs. 199

It is important to note that SMR/HEIDI analysis does not distinguish pleiotropy from 200

causality. Therefore, we can hypothesize causal relationships between the expression level of 201

identified genes and the risk of VVs development. 202

Besides this, we found associations with expression levels of four genes within rs2911463 203

region (APRT, ZFPM1, RNF166, and PIEZO1) as well as with expression level of the TRIM38 204

gene located ~280 kb from rs7773004. However, statistically significant heterogeneity revealed 205

in the HEIDI test indicated that these associations were mediated by polymorphisms different 206

from those that alter VVs risk. 207

DEPICT analysis 208

Data-driven Expression Prioritized Integration for Complex Traits (DEPICT) framework53 209

was used to conduct a gene set and tissue enrichment analysis as well as to provide additional 210

evidence for gene prioritization. Results of DEPICT analysis for SNP sets associated with VVs 211

at P < 5.0e-08 and P < 1.0e-05 are given in Supplementary tables 7 and 8, respectively. For the 212

first set, we identified a significant enrichment of the Mammalian Phenotype (MP) Ontology 213

terms “abnormal vasculogenesis” and “abnormal vascular development”. When we relaxed 214

significance threshold of input SNPs to P < 1.0e-05, 22 MP terms were shown to be enriched 215

with additional 9 categories related to abnormal morphology of heart and vessels (“abnormal 216

blood vessel morphology”, “abnormal vascular smooth muscle morphology”, “failure of heart 217

looping”, etc.). The remaining MP terms were associated with aberrant morphology of other 218


https://doi.org/10.1101/368365

10

tissues and organs and disruption of embryonic development. Besides this, we observed 219

enrichment for several subnetworks of protein-protein interactions, including that for ENG 220

(major glycoprotein of the vascular endothelium, mutations in its gene cause multisystemic 221

vascular dysplasia), SMAD 2, 4, 6, 7 (transduce signals from TGF-beta family members), and 222

Notch4 (regulates arterial specification54). 223

Tissue enrichment analysis revealed no statistically significant categories. 224

DEPICT gene prioritization tool provided 2 and 47 prioritized genes depending on the 225

level of statistical significance of the input SNPs (P < 5.0e-08 and P < 1.0e-05, respectively; 226

Supplementary tables 7 and 8). 227

Summary of gene prioritization 228

Basing on cumulative evidence from different analyses, we drew up a list of genes which 229

most likely account for the revealed associations and seem to be functionally relevant to VVs 230

development (Table 3). For two loci on chromosomes 2 and 5, we were unable to prioritize any 231

candidate gene. However, for the locus on chromosome 2, we advocate for the PPP3R1 gene 232

since it displayed the lowest P and the greatest effect size in the SMR test (P = 1.3e-22 and β = 233

120.0, Supplementary table 6). Moreover, biological functions of PPP3R1 suggest its 234

involvement in VVs etiology (Table 2). In particular, it induces MCP-1 production, which was 235

shown to be enhanced in VVs55,56. Finally, a recent study demonstrated that expression of 236

PPP3R1 is increased in the venous tissue of VVs patients compared with veins of healthy 237

individuals (fold change = 1.5)57. 238

Testing for pleiotropy: VVs and other complex traits 239

SMR/HEIDI analysis revealed 31 traits that shared the same casual variants with VVs 240

(traits for this and further analyses were obtained from the GWAS-MAP database, see Methods 241

section for details). These SNPs were located in 5 out of 7 loci identified in our study (Fig. 3, 242

Supplementary table 10). Variants within loci tagged by rs11121615 and rs3101725 showed 243

positive SMR beta coefficients (the same direction of effect) with predicted and fat-free mass of 244


https://doi.org/10.1101/368365

11

both legs and a whole body as well as with basal metabolic rate, and negative coefficients 245

(opposed effects) – with leg fat percentage as well as impedance of both legs and a whole body 246

(parameter negatively correlated with fat-free mass). Loci tagged by rs3101725 and rs7773004 247

comprised SNPs with pleiotropic effects on red blood cell erythrocyte distribution width 248

(negative beta). Locus tagged by rs7773004 was also related to numerous blood traits, such as 249

mean corpuscular haemoglobin concentration, mean platelet thrombocyte volume, and monocyte 250

and reticulocyte count. Three loci (rs11135046, rs28558138, and rs28558138) were linked to 251

blood pressure and hypertension with different SMR beta signs. Another interesting finding was 252

identification of positive SMR beta for associations with vascular/heart problems (rs28558138 253

locus), cellulitis (rs3101725 locus), forced viral capacity, and height size at age 10 (rs7773004 254

locus). Overall, our analysis revealed four main groups of traits: one cluster related mainly to 255

mass, one cluster involving mostly fat- and blood pressure-related traits, one cluster of blood-256

related traits linked only to rs7773004 locus, and one independent trait “red blood cell 257

erythrocyte distribution width” (Fig. 3). 258

Genetic correlations 259

Genetic correlations between VVs in different datasets (discovery and replication cohorts, 260

etc.) are given in Supplementary table 1. A list of genetic correlation estimates (rg) between VVs 261

and 861 complex traits is presented in Supplementary table 11. Twenty five traits showed 262

statistically significant correlation with VVs with absolute values of rg ≥ 0.2. Correlation matrix 263

for this subset is displayed as a heatmap in Fig. 4. We observed 5 main clusters: traits related to 264

the type of job, intelligence, and qualification; traits related to height and mass (including 265

predicted and fat-free mass of both legs); thrombosis-related traits; traits related to operations; 266

and traits related to pain (including leg pain and gonarthrosis) and health satisfaction. VVs trait 267

was closest to the thrombosis-related cluster (positive correlation). It was also positively 268

correlated with mass, operations, and pain-related traits as well as with lower levels of 269

qualification and heavy manual/walking/standing job. For traits with less prominent correlation 270


https://doi.org/10.1101/368365

12

with VVs, we observed the same trend: pain and anthropometric traits (sitting and standing 271

height, BMI, mass, etc.) showed positive correlations, whilst higher levels of education - 272

negative. Interestingly, negative correlation was observed with usual walking pace, and positive 273

– with current smoking. 274

We calculated partial genetic correlations for the subset of 7 non-collinear traits with 275

|rg |≥ 0.2. Two traits – DVT and height size at age 10 – were shown to share common genetic 276

background with VVs after the adjustment for the remaining traits in the analyzed subset 277

(Supplementary fig. 3). In other words, their shared genetic components were demonstrated to be 278

at least partially independent of other complex traits. Additionally, we estimated partial genetic 279

correlations between VVs, standing height, and weight. Both standing height and weight had 280

independent genetic components shared with VVs (Supplementary fig. 4). 281

Hypothesis-free search for causal relationships 282

We applied a 2-sample Mendelian randomization (2SMR)58 strategy to infer causal 283

relationships between a broad range of “exposure” phenotypes and VVs as an outcome. In total, 284

39 complex traits were shown to be potential causative factors. Although only genome-wide 285

associated SNPs from exposure GWAS were selected as instrumental variables, we did not 286

require these loci to be replicated. Nevertheless, we checked the stability of our tests with regard 287

to instruments selection by performing the robustness analysis. This test along with the Steiger 288

test59 for the correct direction of effect underpinned the exclusion of 2 out of 39 traits 289

(Supplementary table 12). Further, we assessed violations of MR assumption of absence of 290

horizontal pleiotropy (influence of genetic instruments on the outcome only through the 291

exposure, also known as “exclusion-restriction criterion”) by means of sensitivity analyses58. We 292

did not observe a statistically significant intercept in MR-Egger regression for any trait. 293

However, only a small proportion of traits showed symmetry in Funnel plots and had no 294

heterogeneity in causal effects amongst instruments. This provides evidence that, for the 295

majority of traits, at least some of the selected instruments exhibit horizontal pleiotropic effects. 296


https://doi.org/10.1101/368365

13

Such traits mainly involved several hundred genome-wide SNPs that made leave-one-out 297

analysis also uninformative. In order to correct for horizontal pleiotropy, we applied a 298

straightforward approach having excluded all instrumental variables associated with VVs at the 299

level of statistical significance higher than 0.01. Our correction led to symmetry in 26 out of 33 300

asymmetrical Funnel plots and eliminated heterogeneity in causal effects for 28 out of 34 traits, 301

although 6 traits lost the statistical significance of 2SMR coefficients (Supplementary fig. 5-42, 302

Supplementary table 12). Removing potential sources of heterogeneity also reduced absolute 303

values of 2SMR beta for all the corrected phenotypes. A graphical representation of our results, 304

including 25 causal inferences that we consider the most reliable, is shown in Fig. 5. Twenty one 305

traits were related to anthropometry and included standing and sitting height, weight, hip and 306

waist circumference, fat-free, predicted, and fat mass of legs and arms, etc. One trait was a 307

spirometry measurement associated with pulmonary function. Nonetheless, since it is positively 308

correlated with height, we suppose that it has no independent effect on VVs development. 309

Similarly, the reverse association between malabsorption/coeliac disease and VVs could actually 310

be induced by a weight loss as one of the complications of these conditions. Moreover, although 311

this trait has passed all the necessary tests, we avoid making strong claims about its causality 312

since it was self-reported and involved only ~1,500 cases. 313

All the above mentioned traits were derived from the Neale Lab repository and therefore 314

had about 80% sample overlap with VVs dataset obtained from the Gene ATLAS. Shared 315

participants between the “exposure” and the “outcome” GWAS can cause bias in the Mendelian 316

randomization analysis when weak instruments are selected60. In our study, this limitation was 317

partially mitigated by using only strong instruments associated with exposure traits at a high 318

level of statistical significance (P < 1.0e-08). In order to fully overcome the problem and confirm 319

our results, we searched for independent GWAS and performed a replication analysis. Data were 320

obtained for height, weight, waist circumference, and body fat percentage. Only height reached 321

Bonferroni-corrected level of statistical significance and remained significant after the correction 322


https://doi.org/10.1101/368365

14

for pleiotropy (2SMR P < 1.0e-07, Supplementary table 13). However, null results for other 323

traits could be explained by a limited power of the analysis. For example, sample size for weight 324

was 4.5 times smaller than in the Neale Lab study. 325

Finally, the remaining two traits from the “reliable” set were plasma levels of MHC class I 326

polypeptide-related sequence B protein and CD209 antigen. These phenotypes were not UK 327

Biobank traits, therefore our results could not be confounded by a sample overlap. 328

Discussion 329

In the present study, we utilized freely available GWAS summary data to unravel the 330

genetic underpinnings of VVs of lower extremities. Lack of individual-level data made us be as 331

rigorous as possible to avoid false positive discoveries, so we skimmed off the most apparent 332

evidence. Nevertheless, as far as we are aware, our study is the largest and the most 333

comprehensive study of VVs genetics to date. Beyond that, the availability of GWAS results 334

makes our research fully reproducible. 335

We identified and replicated 7 susceptibility loci that explain 10% of the SNP-based 336

heritability, and prioritized the most likely causal genes (Fig.2, Tables 1-3). Overall, SNP 337

heritability on the liability scale was estimated at ~28 % assuming the disease prevalence of 20-338

30%. Almost 40% of the variance explained by our top SNPs was attributable to the 339

polymorphism in the CASZ1 gene involved in blood vessel development. This strong signal has 340

previously been revealed by “23andMe”12 and subsequently replicated in our own sample of 341

ethnic Russian individuals10. An especially interesting finding, in our opinion, is a novel 342

association of VVs with SNPs in a recently discovered PIEZO1 gene. PIEZO1 encodes a 343

pressure-activated ion channel which senses shear stress and controls vascular architecture27. 344

Mice embryos lacking functional Piezo1 exhibit defects in vascular remodeling and die at 345

midgestation61. Other prioritized genes were PPP3R1, EBF1, STIM2, and HFE. The latter gene 346

has been linked to the risk of VVs in our recent candidate-gene study19. Meanwhile, we could 347

not prioritize any gene in the locus tagged by rs3101725. On the one hand, this region contains 348


https://doi.org/10.1101/368365

15

two genes (SLC12A2 and FBN2) that could play a role in varicose transformation, and on the 349

other hand, the causal polymorphism was shown to be eQTL for a nearby non-protein coding 350

RNA LINC01184 with unknown function. The revealed genes can be considered as good 351

candidates for future follow-up functional studies. It is noteworthy that SLC12A2, FBN2, STIM2, 352

and HFE were included in the druggable gene set by Finan et al.62, and the PIEZO1 gene product 353

belongs to a “potential drug target” class according to the Human Protein Atlas 354

(https://www.proteinatlas.org/). 355

Gene set enrichment analysis involving both genome-wide and less strongly associated 356

signals detected gene categories related to abnormal vascular development and morphology. This 357

observation is consistent with the hypothesis that pathological changes in the vein wall are the 358

primary event preceding VVs formation63. Furthermore, our genetic correlations analysis 359

confirmed known epidemiological associations between VVs and DVT1,16 as well as standing 360

job1,64,65, rough labour65,66, and lower levels of education64 (Fig.4, Supplementary table 11). 361

Shared familial susceptibility with venous thromboembolism has already been shown by Zöller 362

et al67. Here we demonstrate that DVT and VVs share specific genetic components which are 363

independent from other factors such as obesity or number of operations (Supplementary fig. 3). 364

Since none of the top GWAS hits were related to thrombosis, we can conclude that SNPs with 365

less prominent associations are responsible for this genetic overlap. On the contrary, albeit 366

several identified loci were also associated with blood pressure/hypertension and red blood cell 367

traits, including mean corpuscular hemoglobin concentration (Table 2, Fig. 3), we found no 368

evidence for genetic correlation between VVs and these traits. We can therefore attribute these 369

effects only to pleiotropy. Intriguingly, we observed small, but significant genetic overlap with 370

smoking (rg = 0.16). Smoking is considered only as suggestive risk factor for VVs since 371

epidemiological studies have mainly shown no association with this habit1,64,66,68. Other novel 372

interesting findings include genetic links with prospective memory and fluid intelligence 373


https://www.proteinatlas.org/

https://doi.org/10.1101/368365

16

(negative correlation), pain (knee pain, pain all over the body, neck or shoulder pain, and leg 374

pain on walking), usual walking pace, and gonarthrosis. 375

Further, we obtained strong evidence for association between VVs and anthropometric 376

traits such as weight, height, waist and hip circumference. We not only observed genetic 377

correlations (Fig. 3, Fig. 4, Supplementary tables 10, 11), but also demonstrated these traits to be 378

causative factors for VVs development (Fig. 5, Supplementary table 12). It is important to note 379

that genetic overlap and causal relationships were inferred for both fat and fat-free mass. Thus, 380

we can speculate that increased weight is a risk factor for VVs regardless of whether it was 381

caused by excess body fat or a large mass of other tissues. According to some theories, 382

association between overweight and VVs can be explained by greater concentrations of 383

circulating estrogens or even by a confounding effect of parity1. Although this can be true, our 384

results suggest that the cause may also be an increase in mass per se. Moreover, we found that 385

height has common genetic component with VVs independent from weight and other traits 386

(Supplementary fig. 4) and is also causally related to VVs (Fig. 5). Causal inference was 387

additionally confirmed using an independent dataset (Supplementary table 13). This is in 388

agreement with the results of the Edinburgh Vein Study that has shown a significant relationship 389

between increasing height and VVs64. However, the majority of epidemiological studies to date 390

have been focused only on BMI and did not consider height or other body characteristics. 391

Notably, the formula for BMI contains the square of height in the denominator. It is possible that 392

the impact of height underlies the inconsistency in the results, when some studies show a 393

positive association between BMI and VVs, while others reveal no effect1. We therefore 394

recommend that future epidemiological studies collect and analyze data on height and weight 395

along with data on BMI. 396

Last but not least, we detected the causal effect of plasma levels of two proteins – MHC 397

class I polypeptide-related sequence B protein (MICB) and CD209 antigen. Both molecules are 398

involved in innate and adaptive immunity. MICB is a ligand for the activating receptor NKG2D 399


https://doi.org/10.1101/368365

17

present on the surface of natural killer and some other immune-related cells. CD209, also known 400

as DC-SIGN, is a C-type lectin receptor expressed on dendritic cells and macrophages. In our 401

study, Mendelian randomization analysis for CD209 was performed using two genetic 402

instruments – rs505922 and rs8106657 identified by Suhre et al69. The first SNP is located in the 403

ABO gene responsible for ABO blood group determination. Allele rs505922 C tags non-O group 404

and is in linkage disequilibrium with the allele A of the neighboring SNP rs507666 (D’ = 1.00, r2 405

= 0.39), which was found to be the top variant associated with VVs in the “23andMe” GWAS12 406

This association was validated in our previous study using data from UK Biobank10. At the same 407

time, rs505922 explains nearly 40% of the variability of the plasma CD209 concentration (our 408

estimate based on the published data69) with the allele C being linked to higher levels of CD209. 409

It is tempting to speculate that association between the ABO gene polymorphisms and VVs as 410

well as between VVs and blood group A70 is mediated by this protein. We conducted 411

SMR/HEIDI analysis for the locus containing these SNPs and demonstrated that both 412

associations with VVs and CD209 were related to the same causal polymorphism (PSMR = 3.3e-413

06, PHEIDI = 0.75), that supports our hypothesis. Nevertheless, the role of CD209 as well as 414

MICB in VVs etiology needs further experimental confirmation. Besides this, it should be taken 415

into account that the level of CD209 in plasma does not necessarily correlates with its level on 416

the surface of the cells, and circulating CD209 may possess its own specific function. 417

418


https://doi.org/10.1101/368365

18

Methods 419

Study sample, genotyping, and quality control 420

All cases and controls analyzed in our study were derived from UK Biobank13. 421

Sociodemographic, physical, lifestyle, and health-related characteristics of UK Biobank 422

participants have been reported elsewhere15. In brief, enrolled individuals were aged 40–69 423

years, were less likely to be obese, to smoke, to drink alcohol, and had fewer self-reported health 424

conditions compared with the general population. All study participants provided written 425

informed consent, and the study was approved by the North West Multi-centre Research Ethics 426

Committee. Genotyping was performed using the Affymetrix UK BiLEVE and the Affymetrix 427

UK Biobank Axiom arrays. Details on DNA extraction, genotyping, imputation, and quality 428

control (QC) procedures have been reported previously71. 429

The discovery stage of our GWAS was based on genetic association data provided by the 430

Neale Lab (http://www.nealelab.is/) for 337,199 QC positive individuals (6,958 patients 431

diagnosed with “I83: VVs of lower extremities” and 330,241 control individuals without this 432

diagnosis). Information on data processing is given on the Neale Lab website 433

(http://www.nealelab.is/blog/2017/9/11/details-and-considerations-of-the-uk-biobank-gwas). In 434

short, phenotypes were analyzed automatically by the PHEnome Scan ANalysis Tool. SNPs QC 435

criteria included minor allele frequency (MAF) > 0.1%, the Hardy-Weinberg equilibrium 436

P > 1.0e-10, and INFO score > 0.8. After filtration, 10,879,183 autosomal SNPs were left for the 437

analysis. Per-individual QC procedure included removing non-white British and closely related 438

individuals, patients with sex chromosome aneuploidies, and individuals who had withdrawn 439

consent from the UK Biobank study. Associations were adjusted for sex and the first 10 principal 440

components from the UK Biobank sample QC file. We also adjusted the data for body mass 441

index and deep venous thrombosis using GWAS summary statistics for these traits downloaded 442

from the same resource, and corrected the results for potential confounding using the intercept72 443

as a genomic control inflation factor λ. Details are given in Supplementary Methods. 444


http://www.nealelab.is/

http://www.nealelab.is/blog/2017/9/11/details-and-considerations-of-the-uk-biobank-gwas

https://doi.org/10.1101/368365

19

A replication stage was performed by a reverse inverse-variance weighted meta-analysis 445

that yielded genetic association data for 71,256 individuals (Fig. 1). For these purpose, we 446

derived summary statistics for 408,455 UK Biobank participants (10,861 cases and 397,594 447

controls) from the Gene ATLAS database14 (http://geneatlas.roslin.ed.ac.uk/) and used these data 448

as a “minuend” dataset. A detailed description of our approach with statistical processing is 449

presented in Supplementary Methods, in the study of Deng et al.18, and in our manuscript 450

deposited on bioRxiv17. Data provided by the Gene ATLAS project were obtained for a cohort of 451

both related and unrelated individuals of white British descent. The associations were calculated 452

using Mixed Linear Models with adjustment for age, age2, sex, array batch, UK Biobank 453

Assessment Center, and the leading 20 genomic principal components as computed by UK 454

Biobank. Details of this study including QC criteria have been reported previously14. In brief, 455

they excluded individuals with missing or conflicting data, non-biallelic variants, and SNPs with 456

MAF <0.01% and/or the Hardy-Weinberg equilibrium P < 1.0e-50. The Gene ATLAS contains 457

information for over 30 million variants. In our analysis, we focused on 10,829,469 SNPs that 458

overlap the Neale Lab SNP set. 459

Conditional analysis 460

Regional association plots were generated with LocusZoom tool (http://locuszoom.org/) 461

for regions within 250 kb from the lead SNP. Only replicated loci were analyzed. Conditional 462

and joint (COJO) analysis was carried out by a summary statistics-based method described by 463

Yang et al.73 Calculations were performed using the GCTA software74. LD matrix was computed 464

with PLINK 1.9 software (https://www.cog-genomics.org/plink2) using 1000 Genomes data for 465

503 European individuals (http://www.internationalgenome.org/data/). We claimed the presence 466

of one independent signal per locus if none of the polymorphisms except the lead SNP passed 467

the significance threshold of P = 5.0e-08. 468

469


http://geneatlas.roslin.ed.ac.uk/

http://locuszoom.org/

https://www.cog-genomics.org/plink2

http://www.internationalgenome.org/data/

https://doi.org/10.1101/368365

20

Literature-based functional annotation 470

We used regional association plots to identify genes located within associated loci. These 471

genes were further queried for potential involvement in the processes relevant to VVs 472

pathogenesis. For each gene, we scanned an Online Catalog of Human Genes and Genetic 473

Disorders (OMIM, https://www.omim.org/), the NCBI Gene 474

(https://www.ncbi.nlm.nih.gov/gene), and the Pubmed database to inquire into their biological 475

functions. Furthermore, we interrogated whether other hypothetical varicose-related traits were 476

associated with these genes according to previously published GWAS. The Pubmed search was 477

performed using the gene name as a keyword as well as combinations of the gene name and 478

“GWAS”, or “genome-wide association study”, or “varicose”, or “venous”, or “vascular”. The 479

information obtained was used for the literature-based gene prioritization. 480

Additionally, we used a PhenoScanner75 tool 481

(http://www.phenoscanner.medschl.cam.ac.uk/phenoscanner) to make a list of complex traits 482

associated at P < 5.0e-08 with lead SNPs or other polymorphisms in high LD (r2 ≥ 0.7) with lead 483

variants. 1000 Genomes-derived polymorphisms served as proxies for the analysis. 484

VEP analysis 485

For the VEP51 analysis, we used a set of SNPs within replicated regions associated with 486

VVs at P ≤ T, where log10(T) = log10(Pmin) + 1, and Pmin is a P-value for the strongest association 487

per locus (±250 kb from the independent hit). Analysis was performed using software available 488

online (https://www.ensembl.org/info/docs/tools/vep/index.html). 489

SMR/HEIDI analysis 490

SMR/HEIDI52 approach was used to test for potential pleiotropic effects of identified loci 491

on VVs and other complex traits including gene expression levels in certain tissues. SMR 492

analysis provides evidence for pleiotropy, but is unable to define whether both traits are affected 493

by the same underlying causal polymorphism. The latter is specified by a HEIDI test that 494

distinguishes pleiotropy from linkage disequilibrium. Summary statistics for gene expression 495


https://www.omim.org/

https://www.ncbi.nlm.nih.gov/gene

http://www.phenoscanner.medschl.cam.ac.uk/phenoscanner

https://www.ensembl.org/info/docs/tools/vep/index.html

https://doi.org/10.1101/368365

21

levels was obtained from Westra Blood eQTL76 (peripheral blood, 496

http://cnsgenomics.com/software/smr/#eQTLsummarydata) and the GTEx77 database (44 tissues, 497

https://gtexportal.org). VVs summary statistics was obtained from the Gene ATLAS database 498

since it enables the maximum statistical power due to the greatest sample size. At the same time, 499

VVs in this GWAS is completely correlated at a genetic level with VVs in our discovery dataset 500

(rg = 1.00, Supplementary table 1). Summary statistics for other complex traits was derived from 501

the GWAS-MAP database (see below). 502

Nominal P for SMR test was set at 2.63e-05 (0.05/1,898, where 1,898 is the total number 503

of probes for the gene expression analysis) and 1.89e-06 (0.05/7*3,787, where 7 is the number of 504

loci and 3,787 is the total number of non-VVs traits in GWAS-MAP). For HEIDI analysis in a 505

gene expression study, we used a conservative threshold of P = 0.01 (P < 0.01 corresponds to the 506

rejection of pleiotropy hypothesis). For HEIDI in a complex traits analysis, we implemented a 507

less conservative threshold of P = 0.001 since the number of independent test was much higher. 508

Details of data processing are given in Supplementary Methods. 509

DEPICT analysis 510

DEPICT analysis was performed using DEPICT53 software version 1.1, release 194 511

(https://data.broadinstitute.org/mpg/depict/) with default parameters. GWAS summary statistics 512

was obtained from the Gene ATLAS database. We employed DEPICT for both genome-wide 513

significant SNPs (P < 5.0e-08) as well as for loci associated with VVs at P < 1.0e-05. As in 514

previous analyzes, we defined locus as ±250 kb from the lead SNP. The major histocompatibility 515

complex (MHC) region was eliminated. Significance threshold was set at FDR < 0.05. 516

GWAS-MAP platform 517

GWAS-MAP platform integrates a database which was created to study cardiovascular 518

diseases and contains summary-level GWAS results for 123 metabolomics traits, 1,206 519

circulating proteins, 2,419 complex traits from the UK Biobank, and 8 traits related to coronary 520

artery disease, myocardial infarction and their risk factors. We additionally added 6 VVs-related 521


http://cnsgenomics.com/software/smr/#eQTLsummarydata

https://gtexportal.org/

https://data.broadinstitute.org/mpg/depict/

https://doi.org/10.1101/368365

22

traits analyzed in the present study (our discovery and replication datasets; the Gene ATLAS 522

data for VVs; and the Neale Lab data for VVs, BMI, and DVT) as well as 33 traits from the 523

Gene ATLAS database that we supposed to be biologically relevant to VVs. Description of all 524

3,795 traits is provided in Supplementary table 9. GWAS-MAP platform contains embedded 525

software for LD Score regression72, 2SMR analysis (MR-Base package58), and our 526

implementation of SMR/HEIDI analysis52. Further details are given in Supplementary methods. 527

Genetic correlations 528

Genetic correlations (rg) between VVs and other complex traits were calculated using 529

LDSC software (https://github.com/bulik/ldsc/). We applied a cross-trait LD Score regression 530

technique as previously described by Bulik-Sullivan et al.78 This method requires only GWAS 531

summary statistics and is not biased by a sample overlap. We analyzed 861 heritable non-VVs 532

traits from the GWAS-MAP database. Only traits with a total sample size of ≥ 10,000 and ≥ 1 533

million SNPs tested were included in the analysis. GWAS summary statistics for VVs was 534

obtained from the Gene ATLAS database. Statistical significance threshold was set at 1.16e-05 535

(0.01/861). For 25 traits that passed this threshold with |rg| ≥ 0.2, we calculated a matrix of 536

genetic correlations. Partial genetic correlations were estimated using the inverse of the genetic 537

correlation matrix. Significance threshold was set at 3.1e-05 (0.01/325, where 325 is the number 538

of pairwise combinations). For partial genetic correlations between VVs, standing height, and 539

weight, nominal P was set at 3.3e-03 (0.01/3). Clustering and visualization was performed by the 540

“corrplot” package for the R language (basic “hclust” function). Further details are provided in 541

Supplementary Methods. 542

2SMR analysis 543

Casual relationships between 2,221 non-VVs phenotypes (“exposures”) from the GWAS-544

MAP database and VVs (“outcome”) were assessed by 2-sample Mendelian randomization 545

(2SMR) as previously described by the MR-Base collaboration58 (http://www.mrbase.org/). All 546

the details of our protocol are given in Supplementary Methods. Binary traits with the number of 547


https://github.com/bulik/ldsc/

http://www.mrbase.org/

https://doi.org/10.1101/368365

23

cases or controls less than 1000 were not included in the study. GWAS summary statistics for 548

VVs was obtained from the Gene ATLAS database. Analysis was performed on the GWAS-549

MAP platform. Two 2SMR approaches were used: an inverse variance weighted meta-analysis 550

of Wald ratios (IVW) and MR-Egger regression. The nominal P was set at 1.1e-05 551

(0.05/2*2,221, where 2 is the number of approaches). For traits that passed either IVW or MR-552

Egger test, we performed the Steiger test59 for identifying the correct direction of effect, and 553

conducted the robustness analysis (our own approach). Traits that passed all the analyses were 554

then subjected to sensitivity tests58, that included assessing heterogeneity in causal effects 555

amongst instruments, horizontal pleiotropy test (based on the intercept in MR Egger regression), 556

leave-one-out analysis, and Funnel plots inspection. The nominal P for the Steiger test was set at 557

2.3e-05 (0.05/2,221), and for the robustness analysis, horizontal pleiotropy test, and 558

heterogeneity analysis – at 1.3e-03 (0.05/38, where 38 is the number of traits that passed both 559

2SMR analysis and the Steiger test). Leave-one-out and Funnel plots were examined manually. 560

Sensitivity analyses revealed the presence of horizontal pleiotropy for the majority of traits. To 561

correct for this confounder, we omitted all instrumental variables associated with VVs with 562

P < 0.01, and then repeated IVW 2SMR and sensitivity analyses. Additionally, for the resulting 563

set of traits, we searched for independent GWAS in the MR-Base database58 and conducted 564

confirmatory IVW 2SMR analysis (where appropriate) with the MR-Base default parameters. 565

The nominal P was set at 0.013 (0.05/4, where 4 is the number of traits). 566

567

568

569


https://doi.org/10.1101/368365

24

Acknowledgments 570

The work of ASS was supported by the Russian Science Foundation [Project No 17-75-20223]. 571

The work of YAT was supported by the Russian Ministry of Science and Education under the 5-572

100 Excellence Programme. 573

The work of SZS was supported by the Institute of Cytology and Genetics [Project No 0324-574

2018-0017]. 575

The development of GWAS-MAP platform was supported by grants from the Russian Ministry 576

of Science and Education under the 5-100 Excellence Programme, British Council's Institutional 577

Links Programme for Novosibirsk State University and University of Edinburgh (Project 578

reference No IL4277322879), and by PolyKnomics BV. We thank Maatschap PolyOmica for 579

providing free of charge access to computational services. 580

581

We express our gratitude to the Neale Lab and the Gene ATLAS projects for providing GWAS 582

summary statistics for UK Biobank traits. We sincerely thank Yurii Aulchenko for valuable 583

discussion. We gratefully acknowledge Denis Gorev, Eugene Pakhomov, and Anna Torgasheva 584

who contributed to the development of the GWAS-MAP platform. We acknowledge Eugene 585

Phakomov who provided computational support with SMR/HEIDI analysis. We also would like 586

to thank Mariya Smetanina and Maxim Filipenko for obtaining funding. 587

588

Author contributions 589

YAT and ASS conceived the study and provided its design. YAT carried out statistical analyses. 590

ASS worked with literature sources. ASS and YAT interpreted all data and wrote the paper. TIS, 591

SZS, and YAT contributed to the design and implementation of the GWAS-MAP database. SZS 592

contributed to statistical analysis. All authors reviewed and approved the final version of the 593

manuscript. 594


https://doi.org/10.1101/368365

25

Additional information 595

Competing interests 596

The authors declare no competing financial interests. 597

Materials & Correspondence 598

Correspondence and requests for materials should be addressed to Alexandra S. Shadrina (email: 599

[email protected]) or Yakov A. Tsepilov (email: [email protected]). 600




https://doi.org/10.1101/368365

26

References 601

1. Beebe-Dimmer, J. L., Pfeifer, J. R., Engle, J. S. & Schottenfeld, D. The epidemiology of 602

chronic venous insufficiency and varicose veins. Ann. Epidemiol. 15, 175–184 (2005). 603

2. Segiet, O. A. et al. Biomolecular mechanisms in varicose veins development. Ann. Vasc. 604

Surg. 29, 377–384 (2015). 605

3. Raffetto, J. D. Pathophysiology of chronic venous disease and venous ulcers. Surg. Clin. 606

North Am. 98, 337–347 (2018). 607

4. Pfisterer, L., König, G., Hecker, M. & Korff, T. Pathogenesis of varicose veins - lessons 608

from biomechanics. Vasa 43, 88–99 (2014). 609

5. Lim, C. S., Kiriakidis, S., Sandison, A., Paleolog, E. M. & Davies, A. H. Hypoxia-610

inducible factor pathway and diseases of the vascular wall. J. Vasc. Surg. 58, 219–230 611

(2013). 612

6. Ng, M. Y. M., Andrew, T., Spector, T. D. & Jeffery, S. Linkage to the FOXC2 region of 613

chromosome 16 for varicose veins in otherwise healthy, unselected sibling pairs. J. Med. 614

Genet. 42, 235–239 (2005). 615

7. Krysa, J., Jones, G. T. & van Rij, A. M. Evidence for a genetic role in varicose veins and 616

chronic venous insufficiency. Phlebology 27, 329–335 (2012). 617

8. Grant, Y., Onida, S. & Davies, A. Genetics in chronic venous disease. Phlebol. J. Venous 618

Dis. 32, 3–5 (2017). 619

9. Ellinghaus, E. et al. Genome-wide association analysis for chronic venous disease 620

identifies EFEMP1 and KCNH8 as susceptibility loci. Sci. Rep. 7, 45652 (2017). 621

10. Shadrina, A. et al. Polymorphisms of genes involved in inflammation and blood vessel 622

development influence the risk of varicose veins. Clin. Genet. 94, 191–199 (2018). 623

11. Shadrina, A. et al. Genome-wide association study in ethnic Russians suggests an 624

association of the MHC class III genomic region with the risk of primary varicose veins. 625

Gene 659, 93–99 (2018). 626

12. Bell R.K., Durand E.Y., McLean C.Y., Eriksson N., Tung J.Y., Hinds D. A large scale 627

genome wide association study of varicose veins in the 23andMe cohort. In: The 64th 628

Annual Meeting of The American Society of Human Genetics, San Diego, California, 629

USA, 18-22 October 2014, paper no. 2082M, p.487. San Diego: ASHG. 630

https://blog.23andme.com/wp-content/uploads/2014/10/Bell_ASHG2014_varicose.pdf 631

Accessed July 12, 2018. 632

13. Sudlow, C. et al. UK Biobank: an open access resource for identifying the causes of a 633

wide range of complex diseases of middle and old age. PLOS Med. 12, e1001779 (2015). 634

14. Canela-Xandri, O., Rawlik, K. & Tenesa, A. An atlas of genetic associations in UK 635

Biobank. Preprint at bioRxiv 176834 (2017). doi:10.1101/176834 636

15. Fry, A. et al. Comparison of sociodemographic and health-related characteristics of UK 637

Biobank participants with those of the general population. Am. J. Epidemiol. 186, 1026–638

1034 (2017). 639

16. Müller-Bühl, U. et al. Varicose veins are a risk factor for deep venous thrombosis in 640

general practice patients. Vasa 41, 360–365 (2012). 641

17. Tsepilov, Y. A. et al. Network based conditional genome wide association analysis of 642

human metabolomics. Preprint at bioRxiv 096982 (2016). doi:10.1101/096982 643

18. Deng, Y. & Pan, W. Conditional analysis of multiple quantitative traits based on marginal 644

GWAS summary statistics. Genet. Epidemiol. 41, 427–436 (2017). 645

19. Sokolova, E. A. et al. HFE p.C282Y gene variant is associated with varicose veins in 646

Russian population. Clin. Exp. Med. 16, 463–470 (2016). 647

20. Zamboni, P. et al. The overlapping of local iron overload and HFE mutation in venous leg 648

ulcer pathogenesis. Free Radic. Biol. Med. 40, 1869–1873 (2006). 649

21. Charpentier, M. S. et al. CASZ1 promotes vascular assembly and morphogenesis through 650

the direct regulation of an EGFL7/RhoA-mediated pathway. Dev. Cell 25, 132–143 651


https://blog.23andme.com/wp-content/uploads/2014/10/Bell_ASHG2014_varicose.pdf

https://doi.org/10.1101/368365

27

(2013). 652

22. Charpentier, M. S., Dorr, K. M. & Conlon, F. L. Transcriptional regulation of blood vessel 653

formation. Cell Cycle 12, 2165–2166 (2013). 654

23. Levy, D. et al. Genome-wide association study of blood pressure and hypertension. Nat. 655

Genet. 41, 677–687 (2009). 656

24. Kato, N. et al. Meta-analysis of genome-wide association studies identifies common 657

variants associated with blood pressure variation in east Asians. Nat. Genet. 43, 531–538 658

(2011). 659

25. Lu, X. et al. Genome-wide association study in Chinese identifies novel loci for blood 660

pressure and hypertension. Hum. Mol. Genet. 24, 865–874 (2015). 661

26. Takeuchi, F. et al. Blood pressure and hypertension are associated with 7 loci in the 662

Japanese population. Circulation 121, 2302–2309 (2010). 663

27. Li, J. et al. Piezo1 integration of vascular architecture with physiological force. Nature 664

515, 279–282 (2014). 665

28. Gudipaty, S. A. et al. Mechanical stretch triggers rapid epithelial cell division through 666

Piezo1. Nature 543, 118–121 (2017). 667

29. Retailleau, K. et al. Piezo1 in smooth muscle cells is involved in hypertension-dependent 668

arterial remodeling. Cell Rep. 13, 1161–1171 (2015). 669

30. Rode, B. et al. Piezo1 channels sense whole body physical activity to reset cardiovascular 670

homeostasis and enhance performance. Nat. Commun. 8, 350 (2017). 671

31. Wang, S. et al. Endothelial cation channel PIEZO1 controls blood pressure by mediating 672

flow-induced ATP release. J. Clin. Invest. 126, 4527–4536 (2016). 673

32. Ilkan, Z. et al. Evidence for shear-mediated Ca2+entry through mechanosensitive cation 674

channels in human platelets and a megakaryocytic cell line. J. Biol. Chem. 292, 9204–675

9217 (2017). 676

33. Faucherre, A., Kissa, K., Nargeot, J., Mangoni, M. E. & Jopling, C. Piezo1 plays a role in 677

erythrocyte volume homeostasis. Haematologica 99, 70–5 (2014). 678

34. Cahalan, S. M. et al. Piezo1 links mechanical forces to red blood cell volume. Elife 4, 679

e0737 (2015). 680

35. van der Harst, P. et al. Seventy-five genetic loci influencing the human red blood cell. 681

Nature 492, 369–375 (2012). 682

36. Hodonsky, C. J. et al. Genome-wide association study of red blood cell traits in 683

Hispanics/Latinos: The Hispanic Community Health Study/Study of Latinos. PLOS 684

Genet. 13, e1006760 (2017). 685

37. Wu, W. et al. Calcineurin B stimulates cytokine production through a CD14-independent 686

Toll-like receptor 4 pathway. Immunol. Cell Biol. 94, 285–292 (2016). 687

38. Satonaka, H. et al. Calcineurin promotes the expression of monocyte chemoattractant 688

protein-1 in vascular myocytes and mediates vascular inflammation. Circ. Res. 94, 693-689

700 (2004). 690

39. Orlov, S. N., Tremblay, J. & Hamet, P. NKCC1 and hypertension: a novel therapeutic 691

target involved in the regulation of vascular tone and renal function. Curr. Opin. Nephrol. 692

Hypertens. 19, 163–168 (2010). 693

40. Olivieri, J., Smaldone, S. & Ramirez, F. Fibrillin assemblies: extracellular determinants of 694

tissue formation and fibrosis. Fibrogenesis Tissue Repair 3, 24 (2010). 695

41. Carta, L. et al. Fibrillins 1 and 2 perform partially overlapping functions during aortic 696

development. J. Biol. Chem. 281, 8016–8023 (2006). 697

42. Chami, N. et al. Exome genotyping identifies pleiotropic variants associated with red 698

blood cell traits. Am. J. Hum. Genet. 99, 8–21 (2016). 699

43. Gisler, R., Jacobsen, E. W. & Sigvardsson, M. Cloning of human early B-cell factor and 700

identification of target genes suggest a conserved role in B-cell development in man and 701

mouse. Blood. 96, 1457-1464 (2000). 702

44. Griffin, M. J. et al. Early B-cell factor-1 (EBF1) is a key regulator of metabolic and 703


https://doi.org/10.1101/368365

28

inflammatory signaling pathways in mature adipocytes. J. Biol. Chem. 288, 35925–35939 704

(2013). 705

45. Ehret, G. B. et al. Genetic variants in novel pathways influence blood pressure and 706

cardiovascular disease risk. Nature 478, 103–109 (2011). 707

46. Singh, A. et al. Gene by stress genome-wide interaction analysis and path analysis 708

identify EBF1 as a cardiovascular and metabolic risk gene. Eur. J. Hum. Genet. 23, 854–709

862 (2015). 710

47. Brandman, O., Liou, J., Park, W. S. & Meyer, T. STIM2 is a feedback regulator that 711

stabilizes basal cytosolic and endoplasmic reticulum Ca2+ levels. Cell 131, 1327–1339 712

(2007). 713

48. Fernandez, R. A. et al. Upregulated expression of STIM2, TRPC6, and Orai2 contributes 714

to the transition of pulmonary arterial smooth muscle cells from a contractile to 715

proliferative phenotype. Am. J. Physiol. Physiol. 308, C581–C593 (2015). 716

49. van der Harst, P. et al. Seventy-five genetic loci influencing the human red blood cell. 717

Nature 492, 369–75 (2012). 718

50. Wood, A. R. et al. Defining the role of common variation in the genomic and biological 719

architecture of adult human height. Nat. Genet. 46, 1173–1186 (2014). 720

51. McLaren, W. et al. The Ensembl Variant Effect Predictor. Genome Biol. 17, 122 (2016). 721

52. Zhu, Z. et al. Integration of summary data from GWAS and eQTL studies predicts 722

complex trait gene targets. Nat. Genet. 48, 481–487 (2016). 723

53. Pers, T. H. et al. Biological interpretation of genome-wide association studies using 724

predicted gene functions. Nat. Commun. 6, 5890 (2015). 725

54. Fish, J. E. & Wythe, J. D. The molecular regulation of arteriovenous specification and 726

maintenance. Dev. Dyn. 244, 391–409 (2015). 727

55. del Rio Solá L. et al. Varicose veins show enhanced chemokine expression. Eur. J. Vasc. 728

Endovasc. Surg. 38, 635–641 (2009). 729

56. Lattimer, C. R., Kalodiki, E., Geroulakos, G., Hoppensteadt, D. & Fareed, J. Are 730

inflammatory biomarkers increased in varicose vein blood? Clin. Appl. Thromb. Hemost. 731

22, 656–664 (2016). 732

57. Smetanina, M. A. et al. DNA methylation and gene expression profiling reveal MFAP5 as 733

a regulatory driver of extracellular matrix remodeling in varicose vein disease. 734

Epigenomics (2018). doi:10.2217/epi-2018-0001 [Epub ahead of print]. 735

58. Hemani, G. et al. The MR-Base platform supports systematic causal inference across the 736

human phenome. Elife 7, e34408 (2018). 737

59. Hemani, G., Tilling, K. & Davey Smith, G. Orienting the causal relationship between 738

imprecisely measured traits using GWAS summary data. PLOS Genet. 13, e1007081 739

(2017). 740

60. Burgess, S., Davies, N. M. & Thompson, S. G. Bias due to participant overlap in two-741

sample Mendelian randomization. Genet. Epidemiol. 40, 597–608 (2016). 742

61. Ranade, S. S. et al. Piezo1, a mechanically activated ion channel, is required for vascular 743

development in mice. Proc. Natl. Acad. Sci. 111, 10347–10352 (2014). 744

62. Finan, C. et al. The druggable genome and support for target identification and validation 745

in drug development. Sci. Transl. Med. 9, eaag1166 (2017). 746

63. Lim, C. S. & Davies, A. H. Pathogenesis of primary varicose veins. Br. J. Surg. 96, 1231–747

1242 (2009). 748

64. Lee, A. J., Evans, C. J., Allan, P. L., Ruckley, C. V. & Fowkes, F. G. R. Lifestyle factors 749

and the risk of varicose veins: Edinburgh Vein Study. J. Clin. Epidemiol. 56, 171–179 750

(2003). 751

65. Bihari, I., Tornoci, L. & Bihari, P. Epidemiological study on varicose veins in Budapest. 752

Phlebol. J. Venous Dis. 27, 77–81 (2012). 753

66. Zolotukhin, I. et al. Prevalence and risk factors for chronic venous disease in Russian 754

general population. Eur. J. Vasc. Endovasc. Surg. 54, 752-758 (2017). 755


https://doi.org/10.1101/368365

29

67. Zöller, B., Ji, J., Sundquist, J. & Sundquist, K. Venous thromboembolism and varicose 756

veins share familial susceptibility: a nationwide family study in Sweden. J. Am. Heart 757

Assoc. 3, e000850 (2014). 758

68. Brand, F. N., Dannenberg, A. L., Abbott, R. D. & Kannel, W. B. The epidemiology of 759

varicose veins: the Framingham Study. Am. J. Prev. Med. 4, 96–101 (1988). 760

69. Suhre, K. et al. Connecting genetic risk to disease end points through the human blood 761

plasma proteome. Nat. Commun. 8, 14357 (2017). 762

70. Cornu-Thenard, A., Dab, W., de Vicenzi, I. & Valty, J. Relationship between blood 763

groups (ABO) and varicose veins of the lower limbs. A case-control study. Phlebol. J. 764

Venous Dis. 4, 37–40 (1989). 765

71. Bycroft, C. et al. Genome-wide genetic data on ~500,000 UK Biobank participants. 766

Preprint at bioRxiv 166298 (2017). doi:10.1101/166298 767

72. Bulik-Sullivan, B. K. et al. LD Score regression distinguishes confounding from 768

polygenicity in genome-wide association studies. Nat. Genet. 47, 291–295 (2015). 769

73. Yang, J. et al. Conditional and joint multiple-SNP analysis of GWAS summary statistics 770

identifies additional variants influencing complex traits. Nat. Genet. 44, 369–375 (2012). 771

74. Yang, J., Lee, S. H., Goddard, M. E. & Visscher, P. M. GCTA: A Tool for Genome-wide 772

Complex Trait Analysis. Am. J. Hum. Genet. 88, 76–82 (2011). 773

75. Staley, J. R. et al. PhenoScanner: a database of human genotype–phenotype associations. 774

Bioinformatics 32, 3207–3209 (2016). 775

76. Westra, H.-J. et al. Systematic identification of trans eQTLs as putative drivers of known 776

disease associations. Nat. Genet. 45, 1238–1243 (2013). 777

77. GTEx Consortium. The Genotype-Tissue Expression (GTEx) project. Nat. Genet. 45, 778

580–585 (2013). 779

78. Bulik-Sullivan, B. et al. An atlas of genetic correlations across human diseases and traits. 780

Nat. Genet. 47, 1236–1241 (2015). 781

782


https://doi.org/10.1101/368365

30

Figures 783

784

785

Figure 1. Scheme depicting the overall workflow of our study. 786

In the discovery stage, we used GWAS summary statistics for VVs provided by the Neale Lab 787

and adjusted it for two potential confounders – body mass index (BMI) and deep venous 788

thrombosis (DVT). GWAS data for these traits was also obtained from the Neale Lab database. 789

In order to validate the revealed association signals, we generated a replication cohort by means 790

of reverse meta-analysis using data from by the Gene ATLAS (larger sample size) and the Neale 791

Lab (smaller sample size) databases. VVs traits in all datasets including our simulated replication 792

sample were highly genetically correlated with each other (Supplementary table 1). For 7 793

replicated loci, we performed a functional annotation analysis. Other analyses were conducted 794

using the Gene ATLAS data since it enables to achieve the highest statistical power and is 795

genetically equivalent to our discovery dataset (rg = 1.00). 796

797


https://doi.org/10.1101/368365

31

798 Figure 2. Graphical summary of the discovery GWAS stage after the adjustment for deep 799

venous thrombosis and body mass index and the correction for genomic control. Red line 800

corresponds to the genome-wide significance threshold of P = 5.0e-08. Only associations with 801

P < 1.0e-02 are presented. Points with -log10(P) >15 are depicted as points with -log10(P) =15. 802

803


https://doi.org/10.1101/368365

32

804 Figure 3. Pleiotropic effects of variants within revealed loci on other complex traits. 805

Color depicts the sign and the magnitude of SMR beta coefficient. Negative sign (red) means 806

opposed effects, and positive sign (blue) means the same direction of effect. Traits that passed 807

both SMR and HEIDI test are marked with an asterisk. Prioritized genes in loci are shown in 808

brackets. Full names of traits as indicated in the Neale lab and the Gene ATLAS databases are 809

given in Supplementary table 14. 810

RBC, red blood cell. 811

812


https://doi.org/10.1101/368365

33

813 Figure 4. Matrix of genetic correlations between VVs and other complex traits. 814

Color depicts the sign and the absolute value of genetic correlation coefficient (rg). Only the 815

traits having |rg| ≥ 2 with VVs are shown. Combinations with statistically significant rg are 816

marked with points (P < 3.1e-05). O levels of qualification correspond to the General Certificate 817

of Education Ordinary Level/the General Certificate of Secondary Education; A or AS level 818

correspond to Advanced or Advanced Subsidiary levels. “Qualifications: NN” means no 819

qualification. Full names of traits as indicated in the Neale lab and the Gene ATLAS databases 820

are given in Supplementary table 15. For “Prospective memory result” and “Health satisfaction”, 821

high scores correspond to poor performance. 822

DVT, deep venous thrombosis; VTE, venous thromboembolism. 823

824


https://doi.org/10.1101/368365

34

825 Figure 5. Volcano plot depicting the results of hypothesis-free 2SMR analysis (IVW approach). 826

The x-axis shows 2SMR beta, and the y-axis denotes the level of statistical significance on a 827

logarithmic scale. Grey color indicates the traits that did not pass either 2SMR, or the Steiger 828

test, or additional robustness analysis. Small colored circles depict the traits that passed these 829

test, but either exhibited heterogeneity in causal effects amongst instruments after removing 830

potential outlier SNPs, or did not passed 2SMR after this procedure. Large colored circles 831

represent the traits that both passed 2SMR and showed no heterogeneity after the outliers 832

removal (“the most reliable traits”). 2SMR coefficients are shown as calculated before the 833

correction for pleiotropy. 834

“Pulmonary function” traits include different estimates of forced vital capacity and forced 835

expiratory volume in 1-second. Further details are provided in Supplementary table 12. 836

MICB, MHC class I polypeptide-related sequence B. 837

838


https://doi.org/10.1101/368365

35

Supplementary tables 839

Supplementary table 1. Genetic correlations between VVs in different datasets included in the 840

analysis. 841

Supplementary table 2. Top loci associated with VVs of lower extremities. 842

Supplementary table 3. Results of conditional and joint analysis for replicated loci. 843

Supplementary table 4. Complex traits associated with the revealed hits (found in a 844

PhenoScanner catalogue). 845

Supplementary table 5. Results of the VEP analysis. 846

Supplementary table 6. Results of SMR/HEIDI analysis. Searching for pleiotropic effects on 847

VVs and gene expression. 848

Supplementary table 7. Results of DEPICT analysis for SNPs with P < 5e-08. 849

Supplementary table 8. Results of DEPICT analysis for SNPs with P < 1e-05. 850

Supplementary table 9. Description of traits included in the GWAS-MAP database. 851

Supplementary table 10. Results of SMR/HEIDI analysis. Searching for pleiotropic effects on 852

VVs and other complex traits. Associations that passed both SMR and HEIDI analyses. 853

Supplementary table 11. Genetic correlations between VVs and other complex traits. 854

Supplementary table 12. Results of hypothesis-free 2SMR analysis. 855

Supplementary table 13. Results of 2SMR analysis performed on independent datasets. 856

Supplementary table 14. Trait abbreviations, full names of traits, and their sources 857

(SMR/HEIDI). 858

Supplementary table 15. Trait abbreviations, full names of traits, and their sources (genetic 859

correlations). 860

861


https://doi.org/10.1101/368365

36

Supplementary figure captions 862

863 Supplementary figure 1. Regional association plot of –log10 (P) for SNPs located at the 864

distance of ≤ 250 kb from the index SNP. Color of circles indicates the strength of linkage 865

disequilibrium with the lead SNP based on the squared correlation coefficient (r2). Blue line 866

indicates recombination rate (cM/Mb). Genes are indicated as blue bars under the plot. 867

868

Supplementary figure 2. A quantile-quantile plot for observed vs. expected distribution of P-869

values for χ2 statistics. 870

871

Supplementary figure 3. Matrix of partial genetic correlations between VVs and a subset of 872

non-collinear complex traits. Color depicts the sign and the absolute value of genetic correlation 873

coefficient (rg). Combinations with statistically significant correlations are marked with points. 874

Full names of traits as indicated in the Neale lab and the Gene ATLAS databases are given in 875

Supplementary table 15. 876

877

Supplementary figure 4. Matrix of partial genetic correlations between VVs, standing height, 878

and weight. Color depicts the sign and the absolute value of genetic correlation coefficient (rg). 879

Combinations with statistically significant correlations are marked with points. 880

881

Supplementary figures 5-42. Leave-one-out and Funnel plots before and after removal of 882

instrumental variables associated with VVs with P < 0.01. 883

884


https://doi.org/10.1101/368365

37

Table 1. Top SNPs associated with varicose veins of lower extremities. Replicated associations are shown in bold. SNP Chr: position* EffA/RefA† EAF Discovery stage Replication stage,

N = 71,256

h2 (%)§ Without adjustment for BMI and DVT,

N = 337,199

Adjusted for BMI and DVT,

N = 336,107

β SE P β SE P P‡(GC) Z P

rs11121615 1: 10825577 T/C 0.69 -0.0056 0.0004 1.1e-49 -0.0056 0.0004 7.6e-50 6.0e-48 -8.55 1.2e-17 3.7

rs2911463 16: 88835545 A/G 0.69 -0.0041 0.0004 1.0e-27 -0.0041 0.0004 9.1e-28 9.7e-27 -4.05 5.1e-05 1.6

rs2861819 2: 68489221 C/G 0.66 0.0030 0.0004 8.5e-17 0.0030 0.0004 7.4e-17 3.0e-16 5.34 9.5e-08 1.2

rs3101725 5: 127524018 C/T 0.76 -0.0030 0.0004 1.5e-13 -0.0030 0.0004 2.9e-13 8.4e-13 -4.02 5.7e-05 0.9

rs11135046 5: 158230013 T/G 0.54 -0.0025 0.0003 5.9e-13 -0.0025 0.0003 3.7e-13 1.1e-12 -4.13 3.6e-05 0.9

rs28558138 4: 26818080 C/G 0.42 -0.0024 0.0004 1.8e-11 -0.0024 0.0004 1.3e-11 3.2e-11 -6.39 1.7e-10 1.1

rs7773004 6: 26267755 G/A 0.49 -0.0023 0.0003 5.2e-11 -0.0023 0.0003 6.4e-11 1.5e-10 -3.05 0.002 0.6

rs7856039 9: 118385117 C/T 0.67 0.0024 0.0004 5.7e-11 0.0024 0.0004 8.1e-11 1.9e-10 0.10 0.92 ̶

rs9880192 3: 128297569 C/G 0.41 -0.0023 0.0004 8.0e-11 -0.0023 0.0004 8.7e-11 2.0e-10 -2.03 0.04 ̶

rs192647746 4: 182760366 G/A 0.001 0.0336 0.0054 4.5e-10 0.0333 0.0052 1.2e-10 2.9e-10 -1.01 0.31 ̶

rs12625547 20: 50154647 G/T 0.17 -0.0029 0.0005 1.2e-10 -0.0029 0.0005 1.3e-10 2.9e-10 -2.18 0.03 ̶

rs236530 17: 68217471 T/C 0.75 -0.0026 0.0004 1.8e-10 -0.0026 0.0004 1.4e-10 3.2e-10 -2.84 0.005 ̶

rs2241173 17: 70028445 G/A 0.58 -0.0022 0.0004 4.5e-10 -0.0022 0.0004 4.4e-10 9.7e-10 -1.84 0.07 ̶

rs73107980 12: 48187601 T/C 0.24 0.0025 0.0004 1.7e-09 0.0025 0.0004 1.4e-09 2.8e-09 1.65 0.10 ̶

rs10084824 4: 89515141 A/C 0.01 0.0100 0.0018 4.8e-08 0.0101 0.0018 1.2e-08 2.3e-08 -1.75 0.08 ̶

rs6062619 20: 62683002 G/A 0.27 -0.0022 0.0004 2.0e-08 -0.0022 0.0004 1.4e-08 2.6e-08 -2.80 0.005 ̶

rs1936800 6: 127436064 T/C 0.53 -0.0019 0.0003 2.0e-08 -0.0020 0.0003 1.9e-08 3.6e-08 -0.31 0.76 ̶

rs11693897 2: 45933993 T/C 0.004 0.0148 0.0029 3.3e-07 0.0149 0.0027 3.3e-08 6.2e-08 -0.94 0.34 ̶

rs247749 5: 76296735 C/T 0.31 0.0021 0.0004 5.5e-07 0.0021 0.0004 3.9e-08 7.3e-08 1.78 0.07 ̶

rs1061539 6: 29937555 C/T 0.83 -0.0025 0.0005 4.0e-08 -0.0025 0.0005 4.8e-08 8.9e-08 -2.69 0.007 ̶

BMI, body mass index; DVT, deep venous thrombosis; EAF, effect allele frequency; SE, standard error; SNP, single nucleotide polymorphism *Chromosome: position on chromosome according to GRCh37.p13 assembly †Effect allele/reference allele ‡P-value corrected for the genomic inflation factor §Proportion of SNP-based heritability explained by the variant


https://doi.org/10.1101/368365

38

Table 2. Gene prioritization based on a literature review.

Lead SNP Candidate

gene

OMIM*

code

Location

Functional effects of the encoded proteins Traits† associated with SNPs in the revealed

loci according to previous GWAS (lead SNP,

D’ and r2 for LD‡ with our hit)

rs11121615

CASZ1 609895 Intronic Castor zinc finger 1 transcription factor

Development of heart and blood vessels: blood vessel assembly,

lumen formation, and sprouting21,22

Blood pressure and arterial hypertension23-26

(rs880315 in CASZ1, D’ = 0.12, r2 = 0.01)

VVs of lower extremities12 (rs11121615 in

CASZ1, the same SNP)

rs2911463§ PIEZO1 611184 Intronic Piezo type mechanosensitive ion channel component 1

Converts mechanical forces into biological signals: senses shear

stress in vascular endothelium, smooth muscle cells, red blood

cells, and other cells, and mediates stretch-activated currents

Determinant of vascular architecture in both development and

adult physiology27

Controls epithelial cell number: triggers cell division in response

to mechanical stretch and induces cell extrusion in response to

crowding 28

Mediates hypertension-dependent arterial remodeling through

smooth muscle cell-related mechanisms29

Regulation of blood pressure, vascular tone, NO formation, and

blood flow redistribution30,31

Regulation of platelet function: contributes to the shear-dependent

thrombus formation under arterial flow (suggestive evidence)32

Red blood cell volume regulation33,34

Mean corpuscular hemoglobin

concentration35,36

(rs10445033 in PIEZO1, D’ = 0.88,

r2 = 0.55 with rs2911463, and D’ = 0.84,

r2 = 0.12 with rs9972645;

rs9932423 in PIEZO1, D’ = 0.93,

r2 = 0.86 with rs2911463 and D’ = 0.85,

r2 = 0.17 with rs9972645;

rs551118 5’near PIEZO1, D’ = 0.96,

r2 = 0.557 with rs2911463 and D’ = 0.83,

r2 = 0.10 with rs9972645)

rs2861819 PPP3R1 601302 6 kb from

the 5’ end

Protein phosphatase 3 regulatory subunit B, alpha

(calcineurin subunit B, type 1)

Pro-inflammatory effect: induction of cytokine and chemokine

VVs of lower extremities12

(rs6712038 5’near PPP3R1, D’ = 1.00,

r2 = 0.98)


https://doi.org/10.1101/368365

39

LD, linkage disequilibrium; UTR, untranslated region *Online Mendelian Inheritance in Man® database (https://www.omim.org/)

(MCP-1) production37,38

rs3101725

SLC12A2 600840 3' UTR Solute carrier family 12 member 2

Membrane ion-transport protein involved in cell-volume

regulation

Regulation of vascular tone39

Regulation of blood pressure through vascular and renal effects39

Red blood cell distribution width36,42

(rs17764730 5’near SLC12A2, D’ = 0.95,

r2 = 0.84;

rs10063647 5’near SLC12A2,

D’ = 0.93, r2 = 0.30;

rs10089 in SLC12A2, D’ = 1.00, r2 = 0.06)

FBN2 612570 70 kb from

the 3’ end Fibrillin 2

Extracellular matrix protein, component of connective tissue

microfibrils involved in elastic fiber assembly40

Is implicated into the assembly of the aortic matrix41

rs11135046 EBF1 164343 Intronic Early B cell factor 1

Transcription factor

Regulation of development and differentiation of B lymphocytes43

Regulator of metabolic and inflammatory signaling pathways in

mature adipocytes44

Blood pressure and hypertension14,45

(rs11135046 in EBF1, the same SNP, and

rs11953630 3’near EBF1, D’ = 0.13,

r2 = 0.01)

Cardiometabolic traits46

(rs4704963 in EBF1, D’ = 1.00, r2 = 0.09)

rs28558138 STIM2 610841 44 kb from

the 5’ end Stromal interaction molecule 2

Regulation of Ca2+ concentration in cytosol and endoplasmic

reticulum47

Potential role in vascular remodeling due to smooth muscle cell

phenotype alteration48

N/A

rs7773004 HFE 613609 171 kb

from the 5’

end

Homeostatic iron regulator

Regulation of iron absorption

Mean corpuscular hemoglobin concentration

and mean corpuscular volume49

(rs7773004 3’near HFE,, the same SNP)

Height50

(rs7773004 3’near HFE, the same SNP)


https://www.omim.org/

https://doi.org/10.1101/368365

40

†The full list of traits is given in Supplementary table 2 ‡LD were determined for European populations using LDlink online tool (https://analysistools.nci.nih.gov/LDlink/) §The second independent signal in this region (tagged by rs9972645) was also located in the PIEZO1 gene


https://analysistools.nci.nih.gov/LDlink/

https://doi.org/10.1101/368365

41

Table 3. Summary of gene prioritization.

Lead SNP Locus* Number of

genes per

locus†

Prioritized

gene

Nearest

gene, yes/no

Evidence for

prioritization

rs11121615 1: 11075577–10575577 4 CASZ1 yes L, D

rs2911463 16: 89085545–88585545‡ 20 PIEZO1 yes L, D

rs2861819 2: 68739221–68239221 8 PPP3R1 yes L, E, S

PLEK no S

WDR92 no S

PNO1 no V, S

rs3101725 5: 127774018–127274018 3 SLC12A2 yes L

FBN2 no L

LINC01184 no S

rs11135046 5: 158480013–157980013 1 EBF1 yes L

rs28558138 4: 27068080–26568080 2 STIM2 no L, D

rs7773004 6: 26517755–26017755 41 HFE no L

D, DEPICT analysis; L, literature-based prioritization; E, expression change in VVs; S, SMR/HEIDI analysis; V, Variant Effect Predictor analysis

*Chromosome: position on chromosome according to GRCh37.p13 assembly †Calculated based on regional association plots generated with LocusZoom tool (http://locuszoom.org/, Supplementary fig.1) ‡This locus contains two independent association signals. SNP rs9972645 that tags the second signal is also located in the PIEZO1 gene


http://locuszoom.org/

https://doi.org/10.1101/368365

Documents

Varicose veins of lower extremities: insights from the ... · The cumulative evidence from epidemiological, family, and genetic association studies 53. strongly indicates that there