July 2, 2010 Original Article

Advance Publication Journal of Atherosclerosis and Thrombosis Journal of Atherosclerosis and ThrombosisAccepted Vol.17, forNo.● publication: May 7, 20101 Published online: July 2, 2010 Original Article

Identification of Evidence Suggestive of an Association with Peripheral Arterial Disease at the OSBPL10 Locus by Genome-Wide Investigation in the Japanese Population

Hiroshi Koriyama1, Hironori Nakagami2, Tomohiro Katsuya1, 3, Ken Sugimoto1, Hidetoshi Yamashita2, Yoichi Takami1, Shiro Maeda4, Michiaki Kubo5, Atsushi Takahashi6, Yusuke Nakamura7, Toshio Ogihara8, Hiromi Rakugi1,Yasufumi Kaneda2, and Ryuichi Morishita3

1Department of Geriatric Medicine and Nephrology, Osaka University Graduate School of Medicine, Suita, Japan 2Division of Gene Therapy Science, Osaka University Graduate School of Medicine, Suita, Japan 3Division of Clinical Gene Therapy, Osaka University Graduate School of Medicine, Suita, Japan 4Laboratory for Endocrinology and Metabolism, Center for Genomic Medicine, RIKEN, Tokyo, Japan 5Laboratory for Genotyping Development, Center for Genomic Medicine, RIKEN, Yokohama, Japan 6Laboratory for Statistical Analysis, Center for Genomic Medicine, RIKEN, Tokyo, Japan 7Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, University of Tokyo, Tokyo, Japan 8Osaka General Medical Center, Osaka Prefectural Hospital Organization, Osaka, Japan

Aim: Peripheral arterial disease (PAD) is a common cause of cardiovascular morbidity and an independent predictor of cardiovascular mortality. However, little is known about the genetic basis of PAD. To elucidate this, we performed a two-staged genome-wide association study in Japanese individuals. Methods: We initially tested 222,285 single-nucleotide polymorphisms (SNPs). After the first screening in a panel of 195 PAD cases and 1,358 controls, 2,696 SNPs (1.2%) were further genotyped in the second screening using another panel of 699 PAD cases and 1540 controls. In both screenings, controls were subjects affected with some diseases other than PAD. Results: When analyzed in the combined panel, the strongest signal of PAD association was observed at rs1902341 in the intron of OSBPL10 (p =4.7E-7 for trend test; OR=1.31, 95% CI 1.18−1.46). Also, PAD was modestly associated at several other loci such as rs2554503 in CSMD1 (p =5.7E-5; OR=1.32, 95% CI 1.15−1.51) or rs235243 in VSP13D (p =0.04; OR=1.18, 95% CI 1.01−1.37). Conclusion: Our genome-wide exploration identified suggestive evidence of PAD association at the OSBPL10 locus. Because the association has not reached a genome-wide significant level, further replication study is warranted for verification in the Japanese population.

J Atheroscler Thromb, 2010; 17:000-000.

Key words; Peripheral arterial disease, Genome-wide association study, OSBPL10

ties caused by blood flow dysfunction resulting from Introduction atherosclerosis. Known risk factors include hyperten- Peripheral arterial disease (PAD) is characterized sion, diabetes mellitus, dyslipidemia, smoking, and by pain, ulceration and necrosis of the lower extremi- ageing. PAD is a significant predictor of cardiovascular mortality and is associated with an increased risk of Address for correspondence: Tomohiro Katsuya, Division of 1-4) Clinical Gene Therapy, Osaka University Graduate School of stroke, myocardial infarction, and death . In partic- Medicine, 2-2 Yamada-oka, Suita 565-0871, Japan ular, the risk of cardiovascular death over a 10-year 2) E-mail: [email protected] period is increased 6-fold in patients with PAD . In Received: November 2, 2009 high-risk groups, such as subjects over the age of 50 Accepted for publication: May 7, 2010 years with a history of diabetes or smoking and sub- Advance Publication Journal of Atherosclerosis and Thrombosis 2 Koriyama et al. Accepted for publication: May 7, 2010 OSBPL10 is Associated with PAD 3 Published online: July 2, 2010

jects over the age of 70 years, the prevalence of PAD is sent. The protocol was approved by the ethics com- over 25% in North America and Europe5). mittee of Osaka University and of each participating Despite its prevalence and the high social cost institution. attributed to PAD6), the genetic basis of PAD and the factors that determine its responsiveness to treatment SNP Genotyping Methods are yet to be understood. There are no genetic tests Using standard protocols, we extracted genomic available that reliably identify the subset of subjects DNA from peripheral blood leukocytes. In the first carrying inherited risk factors for developing PAD on stage, we genotyped 268,068 SNPs from autosomal a large scale7); however, three family studies to date chromosomes; these SNPs were selected as tagging have reported estimates of heritability of PAD and SNPs for the Japanese population from the JSNP 11) or thus its heritability could be suggested8-10). A logical HapMap database12) using high-density oligonucle- next step is to perform a more comprehensive and otide arrays (Perlegen Sciences) as described previ- unbiased scan of variations in all genes or in the entire ously. SNPs having a call rate ＞90% and no extreme genome using what is commonly referred to as a departure from Hardy-Weinberg Equilibrium (p＜10－6) genome-wide association study (GWAS). Thus, the were passed and used for the association study. In the aim of this study was to identify PAD-susceptible second study, genotyping was conducted using the genes through GWAS of SNPs. multiplex-PCR invader assay13) for PAD subjects and high-density oligonucleotide arrays (Perlegen Sciences) for control subjects. In the multiplex-PCR invader Methods assay, SNPs having a call rate ＞85% were passed and Study Participants used for the association study. Regarding the success BioBank Japan: For the GWAS, we selected case- rates of the multiplex-PCR invader assay ＞90% agree- control samples (cases 1 and 2, controls 1 and 2) from ment existed between the results of genotyping and subjects enrolled in BioBank Japan. The subjects were direct sequencing. recruited from several medical institutes in Japan, including Fukujuji Hospital, Iizuka Hospital, Iwate Statistical Analysis Medical University School of Medicine, Juntendo Statistical methods for determining associations University, National Hospital Organization Osaka and calculating LD coefficients (r 2) have been National Hospital, Nihon University, Nippon Medical described previously14). We performed the HWE test School, Osaka Medical Center for Cancer and Cardio- according to a method previously described15). The vascular Diseases, and Shiga University of Medical cut-off value for the HWE test in control groups was Science, The Cancer Institute Hospital of Japanese 0.000001 for the first stage and 0.01 for the second Foundation for Cancer Research, Tokushukai Hospi- stage; SNPs with p values less than the cut-off values tals and Tokyo Metropolitan Geriatric Hospital. of the HWE test were excluded from analysis. We ana- We selected peripheral arterial disease cases from lyzed the differences between case and control groups the individuals registered. We defined PAD as those with respect to the genotype distribution and allele with either stenosis on angiography of the lower limbs frequency in genome-wide screening (first or second or with ABI fulfilling the diagnostic criteria among stage) by Fisher’s exact test using dominant, recessive patients complaining of symptoms such as intermit- and allelic models with autosomal SNPs. tent claudication or pain in the lower limbs. (case 1, n =195; case 2, n =699). Control groups were 1,358 Results (control 1) and 1,540 individuals (control 2) registered as subjects without apparent clinical symptoms Genome-Wide Association Study of PAD and having diseases other than PAD, includ- To identify variants associated with susceptibility ing bronchial asthma, myocardial infarction, breast to PAD, we performed a genome-wide case-control cancer, Basedow’s disease, cerebral infarction, cerebral study in 195 individuals with PAD (case 1) and 1,358 aneurysm, osteoporosis, heart failure, unstable angina, controls (control 1) collected from the BioBank Japan, pollinosis, type 2 diabetes, emphysema, atopic derma- and genotyped 268,068 SNPs, which covered approxi- titis, stomach cancer or liver cirrhosis. The control mately 56% of common SNPs in Japanese (flow dia- subject group includes 194 cases of diabetes, 194 of gram of study is shown in Table 1). We compared the cerebral infarction (control 1), 188 of myocardial allele frequencies of 222,285 successfully genotyped infarction, and 189 of unstable angina (control 2). SNPs (which covered approximately 46% of common All participants provided written informed con- SNPs in Japanese), and selected 4,000 SNPs showing Advance Publication Journal of Atherosclerosis and Thrombosis 2 Koriyama et al. OSBPL10 is Associated withAccepted PAD for publication: May 7, 20103 Published online: July 2, 2010

the lowest p values between the two groups. In a sec- Table 1. Flow diagram of study ond round of screening, we attempted to genotype Entry: these 4,000 SNPs in 699 individuals with PAD (case 2) 300,000 BioBank Japan including PAD patients and 1,540 controls (control 2) (stage 2), and successfully obtained data for 2,696 SNPs. The results of ⬇ principal component analysis in stage 1 and 2 samples and HapMap samples revealed no evidence of popula- 1st screening: tion stratification between case and control groups 222,285 SNPs genotyped in 195 cases and 1358 controls throughout the present tests (Fig.1A, 1B, 1C). Base- line characteristics for the populations in this study ⬇ are shown in Table 2. SNPs that showed small p values (allelic model p＜0.005 and/or dominant model 2nd screening: p＜0.005 and/or recessive model p＜0.005) in the sec- 2,696 SNPs genotyped in remaining 699 cases and 1540 controls

A C

Fig.1. Samples in the 1st test (A), 2nd test (B) and in HapMap database were analyzed using the Smartpca program (Price et al. 2006), and the first (X axis) and 2nd (Y axis) principal components were plotted. Japanese samples were clearly plotted in a single cluster, except for one sample. One subject in the 1st test was halfway between Japanese and CEU. The Japanese cluster and Chinese cluster were very close, although it seems that some differences exist between these two populations. Advance Publication Journal of Atherosclerosis and Thrombosis 4 Koriyama et al. Accepted for publication: May 7, 2010 OSBPL10 is Associated with PAD 5 Published online: July 2, 2010

Table 2. Clinical characteristics of study population in case-control study 1st Cases Controls p value 2nd Cases Controls p value N 195 1358 N 699 1540 Age (years) 70.2±8.4 56.9±15.4 Age (years) 70.4±8.8 61.7±16.0 Male (%) 87.2 47.8 ＜0.0001 Male (%) 81.9 58.7 ＜0.0001 Diabetes (%) 27.7 22.6 0.0975 Diabetes (%) 35.6 11.4 ＜0.0001 Hypertension (%) 70.3 53.5 ＜0.0001 Hypertension (%) 77.0 51.5 ＜0.0001 Dyslipidemia (%) 23.6 16.5 0.0111 Dyslipidemia (%) 29.6 12.0 ＜0.0001 Smoker (%) 90.8 52.2 ＜0.0001 Smoker (%) 79.7 59.2 ＜0.0001

Table 3. Association results for variants analyzed in Japanese individuals with PAD and control individuals Case Control dbSNPID Genes Chr. Genotype Genotype Alleles p value OR (95%CI) RAF RAF 11 12 22 Sum 11 12 22 Sum rs1902341 OSBPL10 3p22.3 Stage 1 46 89 60 195 0.464 254 708 588 1550 0.392 G/A 7.2×10－3 1.34 (1.09−1.66) Stage 2 153 333 202 688 0.464 264 712 564 1540 0.403 1.2×10－4 1.29 (1.13−1.46) Combined 199 422 262 883 0.464 518 1420 1152 3090 0.397 4.7×10－7 1.31 (1.18−1.46) rs6779621 OSBPL10 3p22.3 Stage 1 20 77 98 195 0.300 80 588 884 1552 0.241 G/T 0.01 1.35 (1.07−1.70) Stage 2 60 296 341 697 0.298 85 589 866 1540 0.246 2.8×10－4 1.30 (1.13−1.50) Combined 80 373 439 892 0.299 165 1177 1750 3092 0.244 2.7×10－6 1.32 (1.18−1.49) rs2168422 OSBPL10 3p22.3 Stage 1 98 75 21 194 0.302 848 615 88 1551 0.255 A/C 0.05 1.26 (1.00−1.59) Stage 2 328 298 64 690 0.309 843 599 97 1539 0.258 4.3×10－4 1.29 (1.12−1.48) Combined 426 373 85 884 0.307 1691 1214 185 3090 0.256 2.1×10－5 1.29 (1.15−1.44) rs2045298 OSBPL10 3p22.3 Stage 1 20 71 103 194 0.286 67 557 921 1545 0.224 T/C 6.1×10－3 1.39 (1.10−1.76) Stage 2 47 276 366 689 0.269 73 540 912 1525 0.225 1.8×10－3 1.26 (1.09−1.46) Combined 67 347 469 883 0.272 140 1097 1833 3070 0.224 2.7×10－5 1.29 (1.15−1.46) rs2554503 CSMD1 8p23 Stage 1 121 71 3 195 0.197 1121 395 36 1552 0.150 C/G 0.02 1.39 (1.06−1.82) Stage 2 443 229 20 692 0.194 1087 415 38 1540 0.159 4.5×10－3 1.27 (1.08−1.50) Combined 564 300 23 887 0.195 2208 810 74 3092 0.155 5.7×10－5 1.32 (1.15−1.51) rs1483466 LOC389676 8q21.3 Stage 1 88 89 17 194 0.317 863 587 102 1552 0.255 T/C 0.01 1.36 (1.08−1.71) Stage 2 350 287 59 696 0.291 863 581 94 1538 0.250 4.2×10－3 1.23 (1.07−1.42) Combined 438 376 76 890 0.297 1726 1168 196 3090 0.252 1.9×10－4 1.25 (1.11−1.40) rs431537 CDH18 5p15.1 Stage 1 117 59 19 195 0.249 863 617 73 1553 0.246 T/A 0.90 1.02 (0.80−1.30) Stage 2 367 277 47 691 0.268 929 529 82 1540 0.225 1.8×10－3 1.26 (1.09−1.46) Combined 484 336 66 886 0.264 1792 1146 155 3093 0.235 0.01 1.17 (1.03−1.32) rs235243 VPS13D 1p36.2 Stage 1 149 37 9 195 0.141 1177 360 16 1553 0.126 T/G 0.42 1.14 (0.84−1.54) Stage 2 510 163 19 692 0.145 1172 352 16 1540 0.125 0.06 1.19 (0.99−1.43) Combined 659 200 28 887 0.144 2349 712 32 3093 0.125 0.04 1.18 (1.01−1.37)

The results include data from stage 1 and stage 2 for SNPs that showed small p values (allelic model p＜0.005 and/or dominant model p＜0.005 and/or recessive model p＜0.005) in the 2nd screening. Risk allele frequency and genotype counts in individuals with PAD and control subjects are shown. 11, homozygous for major allele; 12, heterozygous; 22, homozygous for minor allele; Chr., chromosomal position; RAF, risk allele frequency. Allele indicates risk allele/non-risk allele. P values are for the allelic model. In OR of three SNPs there are significant differences in OSBPL10 (rs1902341, p =9.7×10－7, OR=1.32, 95%CI 1.18−1.47; rs6779621, p =1.1×10－5, OR=1.31, 95%CI 1.16−1.48; rs2168422, p =5.6 ×10－5, OR=1.28, 95%CI 1.14−1.44) after removing atherosclerotic disease subjects (n =765) from the controls. Advance Publication Journal of Atherosclerosis and Thrombosis 4 Koriyama et al. OSBPL10 is Associated withAccepted PAD for publication: May 7, 20105 Published online: July 2, 2010

Table 3. Association results for variants analyzed in Japanese individuals with PAD and control individuals Case Control dbSNPID Genes Chr. Genotype Genotype Alleles p value OR (95%CI) RAF RAF 11 12 22 Sum 11 12 22 Sum rs3765337 VPS13D 1p36.2 Stage 1 147 36 9 192 0.141 1010 313 15 1338 0.128 C/A 0.52 1.11 (0.82−1.52) Stage 2 507 162 20 689 0.147 1158 354 15 1527 0.126 0.06 1.19 (0.99−1.44) Combined 654 198 29 881 0.145 2168 667 30 2865 0.127 0.05 1.17 (1.00−1.36) rs7659075 ARAP2 4p14 Stage 1 127 60 8 195 0.195 1019 475 15 1509 0.167 C/T 0.17 1.20 (0.92−1.57) Stage 2 485 179 31 695 0.173 1048 448 30 1526 0.166 0.57 1.05 (0.89−1.24) Combined 612 239 39 890 0.178 2067 923 45 3035 0.167 0.27 1.08 (0.94−1.24) rs16946196 non gene 18p11.3 Stage 1 103 70 22 195 0.292 838 625 89 1552 0.259 A/C 0.16 1.18 (0.94−1.49) Stage 2 375 240 72 687 0.279 813 622 105 1540 0.270 0.54 1.05 (0.91−1.21) Combined 478 310 94 882 0.282 1651 1247 194 3092 0.264 0.13 1.09 (0.97−1.23) rs17647070 non gene 18q12 Stage 1 169 26 0 195 0.067 1443 107 2 1552 0.036 C/T 4.4×10－3 1.93 (1.24−2.99) Stage 2 626 66 6 698 0.056 1413 125 1 1539 0.041 0.03 1.38 (1.03−1.84) Combined 795 92 6 893 0.058 2856 232 3 3091 0.038 2.9×10－4 1.54 (1.22−1.96) rs17832415 non gene 8p22 Stage 1 148 41 6 195 0.136 1223 317 11 1551 0.109 T/C 0.13 1.28 (0.94−1.75) Stage 2 530 141 20 691 0.131 1192 330 18 1540 0.119 0.26 1.12 (0.92−1.35) Combined 678 182 26 886 0.132 2415 647 29 3091 0.114 0.04 1.18 (1.01−1.38) rs1807019 non gene 18q21 Stage 1 150 38 7 195 0.133 1245 297 9 1551 0.102 C/A 0.05 1.36 (0.99−1.86) Stage 2 552 125 14 691 0.111 1242 286 10 1538 0.099 0.26 1.13 (0.92−1.38) Combined 702 163 21 886 0.116 2487 583 19 3089 0.101 0.07 1.17 (0.99−1.38) rs1847040 non gene 5q14 Stage 1 24 65 105 194 0.291 106 616 827 1549 0.267 T/G 0.33 1.13 (0.89−1.42) Stage 2 53 299 327 679 0.298 128 568 842 1538 0.268 0.04 1.16 (1.01−1.34) Combined 77 364 432 873 0.297 234 1184 1669 3087 0.268 0.02 1.15 (1.03−1.30) rs1916998 non gene 5q32 Stage 1 53 71 55 179 0.494 422 633 232 1287 0.426 A/C 5.3×10－3 1.38 (1.10−1.72) Stage 2 173 304 127 604 0.462 451 618 224 1293 0.412 4.3×10－3 1.22 (1.07−1.40) Combined 226 375 182 783 0.472 873 1251 456 2580 0.419 2.2×10－4 1.24 (1.11−1.39) rs2291016 non gene 8q21.3 Stage 1 43 104 44 191 0.497 249 757 522 1528 0.411 T/G 1.4×10－3 1.42 (1.15−1.76) Stage 2 140 314 228 682 0.435 227 759 539 1525 0.398 0.02 1.17 (1.03−1.33) Combined 183 418 272 873 0.449 476 1516 1061 3053 0.404 8.0×10－4 1.20 (1.08−1.34) rs2359536 non gene 10p12 Stage 1 168 25 1 194 0.070 1214 74 2 1290 0.030 C/T 2.8×10－4 2.40 (1.53−3.77) Stage 2 617 75 0 692 0.054 1213 84 2 1299 0.034 2.5×10－3 1.63 (1.19−2.24) Combined 785 100 1 886 0.058 2427 158 4 2589 0.032 1.5×10－6 1.84 (1.43−2.37) rs6481686 non gene 10p11.2 Stage 1 43 72 80 195 0.595 338 762 453 1553 0.537 G/C 0.03 1.27 (1.02−1.57) Stage 2 139 318 239 696 0.572 348 776 414 1538 0.521 1.9×10－3 1.23 (1.08−1.39) Combined 182 390 319 891 0.577 686 1538 867 3091 0.529 3.8×10－4 1.21 (1.09−1.35) rs7217914 non gene 17q24 Stage 1 31 66 98 195 0.328 119 645 771 1535 0.288 C/T 0.10 1.21 (0.97−1.52) Stage 2 80 308 310 698 0.335 134 620 783 1537 0.289 1.8×10－3 1.24 (1.08−1.42) Combined 111 374 408 893 0.334 253 1265 1554 3072 0.288 2.2×10－4 1.24 (1.10−1.38) rs994950 non gene 3p22 Stage 1 27 72 95 194 0.675 212 739 581 1532 0.620 A/G 0.04 1.27 (1.02−1.59) Stage 2 78 328 290 696 0.652 244 714 576 1534 0.608 5.1×10－3 1.21 (1.06−1.38) Combined 105 400 385 890 0.657 456 1453 1157 3066 0.614 9.8×10－4 1.20 (1.08−1.34) Advance Publication Journal of Atherosclerosis and Thrombosis 6 Koriyama et al. Accepted for publication: May 7, 2010 OSBPL10 is Associated with PAD 7 Published online: July 2, 2010

ond screening are shown in Table 3. A 3p22.3 100kb Through out this analysis, we identified several LOC728665 LOC391524 LOC344787 candidate SNPs, including an SNP with the lowest p STT3B OSBPL10 GPD1L value, located on chromosome 3p22.3. The combined p values of these SNPs located on chromosome 3p22.3 －7 －6 were 4.7×10 (rs1902341), 2.7×10 (rs6779621), 131211 10 9 8 7 6 5 4 3 2 1 exon and 2.1×10－5 (rs2168422), 2.7×10－5 (rs2045298), respectively, some of which were close to the p value * －7 －8 6 threshold of 5×10 to 5×10 , which is necessary in 20kb the setting of GWAS (Table 3). 5 Since the control subject group included cases 4 value)

of diabetes, cerebral infarction, myocardial infarction, P ( ( 3 and unstable angina, we performed additional associa- 10 2 tion analysis for three SNPs (rs1902341, rs6779621, -Log rs2168422) using control subjects without patients 1 with these atherosclerotic diseases. Allele frequency 0 in the control group changed little after subtracting B these subjects. The odds ratios of these 3 SNPs with- * out these subjects were also statistically significant (rs6779621, OR =1.31 (95% CI, 1.16−1.48); rs2168422, OR =1.28 (95% CI, 1.14−1.44); rs1902341, OR=1.32 (95% CI, 1.18−1.47)). We further performed age- and sex-matched re-analysis of stage 2 samples in 645 individuals with PAD and 645 controls. These three SNPs were still associated (p =0.009, 0.012, 0.018, respectively).

LD Block We subsequently attempted to construct a link- Fig.2. Genomic context of OSBPL10 region on chromosome age-disequilibrium (LD) block at chromosome 3p22.3 3. (A) Exon-intron structure of OSBPL10 (exons shown to define the region showing a strong association with as vertical bars). Asterisk indicates SNP rs1902341 at PAD. As shown in Fig.2, significant SNPs were intron 5. (B) Pairwise correlation structure in 400 kb located within intron 5 of OSBPL10 (oxysterol bind- intervals (3.163-3.203 Mb on chromosome 3, NCBI Build 36). The plot includes pairwise r 2 values from ing protein like 10). Since the total length of the HapMap release 27 for the JPT population. Associa- OSBPL10 gene is about 300 kb, and SNPs on both tion test scores are shown as －log10 (p value), where sides of the 3 SNPs have high p values (Table 4), it is p values are taken from Table 2 (set 1＋set 2) for unlikely that they influence adjacent genes other than rs6779621, rs2168422 and rs1902341 and from Table OSBPL10. 3 (set 1) for other SNPs.

Discussion lesterol, they cause conformational change, and induce Identification of PAD susceptibility genes could signal transduction to regulate cholesterol homeostasis. have an important impact on the diagnosis and treat- In this study, we found that SNPs of OSBPL10 ment of this disease. In this study, we utilized the were associated with PAD. Recently, Perttila et al. GWAS approach and identified a susceptible gene for reported that OSBPL10 is associated with a high tri- PAD, OSBPL10. glyceride trait in Finnish subjects and regulates cellular Oxisterol binding protein (OSBP) has been iden- lipid metabolism 19). Because dyslipidemia is one of the tified in eukaryotic species from yeast to humans16). risk factors of PAD, it can be speculated that The human OSBP family consists of 12 members. OSBPL10 associates with PAD through the dysregula- OSBP family proteins have an oxysterol binding tion of serum lipid homeostasis. OSBPL10 may be domain, which binds oxysterol, and a PH domain, associated with cholesterol metabolism or transport, which attaches to non-ER membranes via binding of suggesting that it may be associated with coronary phosphatidylinositol17, 18). By binding oxysterol or cho- artery disease and/or ischemic stroke. However, the Advance Publication Journal of Atherosclerosis and Thrombosis 6 Koriyama et al. OSBPL10 is Associated withAccepted PAD for publication: May 7, 20107 Published online: July 2, 2010

Table 4. List of SNPs in and around OSBPL10 taining fusion protein24); however, the function of Chromosomal Position in Allele human genes (VPS13A-D) has not been analyzed dbSNP 25) position OSBPL10 p value yet . Some SNPs in intergenic regions also showed relatively small p values, such as rs2359536 (p =1.5× rs7650117 31700742 intron 0.134 10－6). At present, the function and significance of rs3792547 31698465 intron 0.260 these regions are unknown. Further studies are neces- rs3749404 31685390 intron 0.300 sary to confirm the significance of these associations. rs9849345 31682329 intron 0.025 PAD is a complex trait whose development is rs4955200 31767319 intron 0.088 probably modified to some degree by unchanged rs6779621 31747517 intron 0.012 genetic determinants that are independent of tradi- rs2168422 31744846 intron 0.050 tional risk factors for atherosclerosis. Although we rs2045298 31739258 intron 0.006 applied the GWAS approach to PAD, the large num- rs2100543 31718616 intron 0.066 bers of association tests that result from GWAS scans rs11921928 31674993 intron 0.028 require that an even lower probability value threshold rs1869849 31673676 intron 0.005 be used to identify promising associations between rs2278959 31665320 intron 0.169 variants and disease. It has been reported that a p value rs17027930 31651978 intron 1.000 threshold of 5×10－7 to 5×10－8 is necessary in the rs9866738 31648747 intron 0.691 setting of GWAS26, 27). This study is still in the screen- rs3792554 31645284 intron 0.424 ing stage and needs a replication study using an inde- rs11058 31643468 intron 0.154 pendent panel of cases (PAD patients) and controls in rs2305610 31645653 intron 0.813 the future. To date, several genes susceptible to PAD rs7633447 31825873 intron 0.897 have been reported in a case-control study with sam- rs6780930 31830855 intron 0.561 ple sizes ranging from ~100 to 1300 subjects7), but we rs6767086 31894102 intron 0.558 could not confirm these genes as candidate genes in rs6788326 31856091 intron 0.149 our study. Thus, we speculate that our genome-wide rs3749403 31692204 intron 0.101 association study still cannot cover all the genes for rs2290531 31729288 intron 0.675 PAD. Further studies are necessary to elucidate the rs1902341 31735574 intron 0.007 underlying molecular mechanism of the genes and gene polymorphisms conferring susceptibility to PAD. In conclusion, our genome-wide exploration function of oxysterol binding proteins and the identified evidence suggestive of PAD association at OSBPL10 gene have not been clarified yet, and no the OSBPL10 locus. study has examined the association with coronary artery disease and/or ischemic stroke. In the present study, several genes other than Acknowledgments OSBPL10 that showed borderline association were We express our gratitude to Kazuko Iwasa and identified (Table 3). SNP in the CSMD1 gene Eriko Nagata for their continuous support of our (rs2554503) showed a relatively small p value (p = investigations. The DNA samples used for this research 5.7×10－5). CSMD1 is a complement regulatory pro- were provided by the Leading Project for Personalized tein that blocks the classical but not alternate comple- Medicine of the Ministry of Education, Culture, ment pathway20) and is associated with Kawasaki dis- Sports, Science and Technology, Japan. ease21). Interestingly, recently, it was reported that SNPs in the CSMD1 gene are also associated with metabolic syndrome22) and hypertension23). SNPs in Sources of Funding VPS13D genes showed high combined p values This work was partly supported by a Grant-in- (p =0.04－0.05). This gene encodes a protein belong- Aid from the Ministry of Education, Culture, Sports, ing to the vacuolar-protein-sorting-13 gene family. In Science and Technology of Japan (21390223), the yeast, vacuolar-protein-sorting-13 proteins are Leading Project for Personalized Medicine of the Min- involved in trafficking membrane proteins between istry of Education, Culture, Sports, Science and Tech- the trans-Golgi network and the prevacuolar compart- nology, Japan, and research grants from Takeda Sci- ment. It was reported that in wild-type yeast, insulin ence Foundation. is ultimately delivered to the vacuole, whereas vps mutants secrete primarily unprocessed insulin-con- Advance Publication Journal of Atherosclerosis and Thrombosis 8 Koriyama et al. Accepted for publication: May 7, 2010 OSBPL10 is Associated with PAD 9 Published online: July 2, 2010

retti V, Leboeuf M, Olivier JF, Phillips MS, Roumy S, References Sallee C, Verner A, Hudson TJ, Kwok PY, Cai D, Koboldt 1) Hooi JD, Kester AD, Stoffers HE, Rinkens PE, Knottnerus DC, Miller RD, Pawlikowska L, Taillon-Miller P, Xiao M, JA, van Ree JW: Asymptomatic peripheral arterial occlu- Tsui LC, Mak W, Song YQ, Tam PK, Nakamura Y, Kawa- sive disease predicted cardiovascular morbidity and mor- guchi T, Kitamoto T, Morizono T, Nagashima A, Ohnishi tality in a 7-year follow-up study. J Clin Epidemiol, 2004; Y, Sekine A, Tanaka T, Tsunoda T, Deloukas P, Bird CP, 57: 294-300 Delgado M, Dermitzakis ET, Gwilliam R, Hunt S, Mor- 2) Criqui MH, Langer RD, Fronek A, Feigelson HS, Klau- rison J, Powell D, Stranger BE, Whittaker P, Bentley DR, ber MR, McCann TJ, Browner D: Mortality over a period Daly MJ, de Bakker PI, Barrett J, Chretien YR, Maller J, of 10 years in patients with peripheral arterial disease. N McCarroll S, Patterson N, Pe’er I, Price A, Purcell S, Engl J Med, 1992; 326: 381-386 Richter DJ, Sabeti P, Saxena R, Schaffner SF, Sham PC, 3) Mohler ER, 3rd: Peripheral arterial disease: identification Varilly P, Stein LD, Krishnan L, Smith AV, Tello-Ruiz and implications. Arch Intern Med, 2003; 163: 2306-2314 MK, Thorisson GA, Chakravarti A, Chen PE, Cutler DJ, 4) Kannel WB: The Framingham Study: ITS 50-year legacy Kashuk CS, Lin S, Abecasis GR, Guan W, Li Y, Munro and future promise. J Atheroscler Thromb, 2000; 6: 60- HM, Qin ZS, Thomas DJ, McVean G, Auton A, Bottolo 66 L, Cardin N, Eyheramendy S, Freeman C, Marchini J, 5) Hirsch AT, Criqui MH, Treat-Jacobson D, Regensteiner Myers S, Spencer C, Stephens M, Donnelly P, Cardon JG, Creager MA, Olin JW, Krook SH, Hunninghake DB, LR, Clarke G, Evans DM, Morris AP, Weir BS, Mullikin Comerota AJ, Walsh ME, McDermott MM, Hiatt WR: JC, Sherry ST, Feolo M, Skol A, Zhang H, Matsuda I, Peripheral arterial disease detection, awareness, and treat- Fukushima Y, Macer DR, Suda E, Rotimi CN, Ade- ment in primary care. JAMA, 2001; 286: 1317-1324 bamowo CA, Ajayi I, Aniagwu T, Marshall PA, 6) Caro J, Migliaccio-Walle K, Ishak KJ, Proskorovsky I: The Nkwodimmah C, Royal CD, Leppert MF, Dixon M, Pei- morbidity and mortality following a diagnosis of periph- ffer A, Qiu R, Kent A, Kato K, Niikawa N, Adewole IF, eral arterial disease: long-term follow-up of a large data- Knoppers BM, Foster MW, Clayton EW, Watkin J, base. BMC Cardiovasc Disord, 2005; 5: 14 Muzny D, Nazareth L, Sodergren E, Weinstock GM, 7) Knowles JW, Assimes TL, Li J, Quertermous T, Cooke JP: Yakub I, Birren BW, Wilson RK, Fulton LL, Rogers J, Genetic susceptibility to peripheral arterial disease: a dark Burton J, Carter NP, Clee CM, Griffiths M, Jones MC, corner in vascular biology. Arterioscler Thromb Vasc Biol, McLay K, Plumb RW, Ross MT, Sims SK, Willey DL, 2007; 27: 2068-2078 Chen Z, Han H, Kang L, Godbout M, Wallenburg JC, L’ 8) Carmelli D, Fabsitz RR, Swan GE, Reed T, Miller B, Wolf Archeveque P, Bellemare G, Saeki K, An D, Fu H, Li Q, PA: Contribution of genetic and environmental influ- Wang Z, Wang R, Holden AL, Brooks LD, McEwen JE, ences to ankle-brachial blood pressure index in the Guyer MS, Wang VO, Peterson JL, Shi M, Spiegel J, NHLBI Twin Study. National Heart, Lung, and Blood Sung LM, Zacharia LF, Collins FS, Kennedy K, Jamieson Institute. Am J Epidemiol, 2000; 151: 452-458 R, Stewart J: A second generation human haplotype map 9) Kullo IJ, Turner ST, Kardia SL, Mosley TH Jr, Boerwinkle of over 3.1 million SNPs. Nature, 2007; 449: 851-861 E, de Andrade M: A genome-wide linkage scan for ankle- 13) Ohnishi Y, Tanaka T, Ozaki K, Yamada R, Suzuki H, brachial index in African American and non-Hispanic Nakamura Y: A high-throughput SNP typing system for white subjects participating in the GENOA study. Ath- genome-wide association studies. J Hum Genet, 2001; erosclerosis, 2006; 187: 433-438 46: 471-477 10) Murabito JM, Guo CY, Fox CS, D’Agostino RB: Herita- 14) Devlin B, Risch N: A comparison of linkage disequilib- bility of the ankle-brachial index: the Framingham Off- rium measures for fine-scale mapping. Genomics, 1995; spring study. Am J Epidemiol, 2006; 164: 963-968 29: 311-322 11) Haga H, Yamada R, Ohnishi Y, Nakamura Y, Tanaka T: 15) Wigginton JE, Cutler DJ, Abecasis GR: A note on exact Gene-based SNP discovery as part of the Japanese Millen- tests of Hardy-Weinberg equilibrium. Am J Hum Genet, nium Genome Project: identification of 190,562 genetic 2005; 76: 887-893 variations in the human genome. Single-nucleotide poly- 16) Yan D, Olkkonen VM: Characteristics of oxysterol bind- morphism. J Hum Genet, 2002; 47: 605-610 ing proteins. Int Rev Cytol, 2008; 265: 253-285 12) Frazer KA, Ballinger DG, Cox DR, Hinds DA, Stuve LL, 17) Olkkonen VM, Johansson M, Suchanek M, Yan D, Gibbs RA, Belmont JW, Boudreau A, Hardenbol P, Leal Hynynen R, Ehnholm C, Jauhiainen M, Thiele C, Lehto SM, Pasternak S, Wheeler DA, Willis TD, Yu F, Yang H, M: The OSBP-related proteins (ORPs): global sterol sen- Zeng C, Gao Y, Hu H, Hu W, Li C, Lin W, Liu S, Pan H, sors for co-ordination of cellular lipid metabolism, mem- Tang X, Wang J, Wang W, Yu J, Zhang B, Zhang Q, brane trafficking and signalling processes? Biochem Soc Zhao H, Zhou J, Gabriel SB, Barry R, Blumenstiel B, Trans, 2006; 34: 389-391 Camargo A, Defelice M, Faggart M, Goyette M, Gupta S, 18) Loewen CJ, Roy A, Levine TP: A conserved ER targeting Moore J, Nguyen H, Onofrio RC, Parkin M, Roy J, Stahl motif in three families of lipid binding proteins and in E, Winchester E, Ziaugra L, Altshuler D, Shen Y, Yao Z, Opi1p binds VAP. EMBO J, 2003; 22: 2025-2035 Huang W, Chu X, He Y, Jin L, Liu Y, Sun W, Wang H, 19) Perttila J, Merikanto K, Naukkarinen J, Surakka I, Martin Wang Y, Xiong X, Xu L, Waye MM, Tsui SK, Xue H, NW, Tanhuanpaa K, Grimard V, Taskinen MR, Thiele C, Wong JT, Galver LM, Fan JB, Gunderson K, Murray SS, Salomaa V, Jula A, Perola M, Virtanen I, Peltonen L, Olk- Oliphant AR, Chee MS, Montpetit A, Chagnon F, Fer- konen VM: OSBPL10, a novel candidate gene for high Advance Publication Journal of Atherosclerosis and Thrombosis 8 Koriyama et al. OSBPL10 is Associated withAccepted PAD for publication: May 7, 20109 Published online: July 2, 2010

triglyceride trait in dyslipidemic Finnish subjects, regu- tion modeling methods. BMC Proc, 2009; 3 Suppl 7: S50 lates cellular lipid metabolism. J Mol Med, 2009; 87: 23) Hong KW, Go MJ, Jin HS, Lim JE, Lee JY, Han BG, 825-835 Hwang SY, Lee SH, Park HK, Cho YS, Oh B: Genetic 20) Kraus DM, Elliott GS, Chute H, Horan T, Pfenninger variations in ATP2B1, CSK, ARSG and CSMD1 loci are KH, Sanford SD, Foster S, Scully S, Welcher AA, Holers related to blood pressure and/or hypertension in two VM: CSMD1 is a novel multiple domain complement- Korean cohorts. J Hum Hypertens, 2009 regulatory protein highly expressed in the central nervous 24) Zhang B, Chang A, Kjeldsen TB, Arvan P: Intracellular system and epithelial tissues. J Immunol, 2006; 176: retention of newly synthesized insulin in yeast is caused 4419-4430 by endoproteolytic processing in the Golgi complex. J 21) Burgner D, Davila S, Breunis WB, Ng SB, Li Y, Bonnard Cell Biol, 2001; 153: 1187-1198 C, Ling L, Wright VJ, Thalamuthu A, Odam M, Shimizu 25) Velayos-Baeza A, Vettori A, Copley RR, Dobson-Stone C, C, Burns JC, Levin M, Kuijpers TW, Hibberd ML: A Monaco AP: Analysis of the human VPS13 gene family. genome-wide association study identifies novel and func- Genomics, 2004; 84: 536-549 tionally related susceptibility Loci for Kawasaki disease. 26) Risch N, Merikangas K: The future of genetic studies of PLoS Genet, 2009; 5: e1000319 complex human diseases. Science, 1996; 273: 1516-1517 22) Nock NL, Wang X, Thompson CL, Song Y, Baechle D, 27) Hirschhorn JN, Daly MJ: Genome-wide association stud- Raska P, Stein CM, Gray-McGuire C: Defining genetic ies for common diseases and complex traits. Nat Rev determinants of the Metabolic Syndrome in the Framing- Genet, 2005; 6: 95-108 ham Heart Study using association and structural equa-