Genetic Regulation of Serum Phytosterol Levels and … Regulation of Serum Phytosterol Levels and Risk of Coronary Artery Disease Running Title: Teupser et al.; Genetics of serum phytosterols
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Avenue, Dallas, TX 72514Circulation: Cardiovascular Genetics is published by the American Heart Association. 7272 Greenville
DOI: 10.1161/CIRCGENETICS.109.907873 published online Jun 7, 2010; Circ Cardiovasc Genet
Erdmann, H. -Erich Wichmann, Heribert Schunkert and Joachim Thiery Schreiber, Karl Werdan, Thomas Meitinger, Markus Löffler, Nilesh J. Samani, Jeanette Gert Matthes, Christian Wittekind, Christian Hengstenberg, Francois Cambien, Stefan
El Mokhtari, Diana Rubin, Arif B. Ekici, André Reis, Christoph Garlichs, Alistair S. Hall, Anika Großhennig, Inke R. König, Peter Lichtner, Iris M. Heid, Alexander Kluttig, Nour E.
Raaz-Schrauder, Georg M. Fiedler, Wolfgang Wilfert, Frank Beutner, Stephan Gielen, Linsel-Nitschke, Arne Schäfer, Peter S. Braund, Laurence Tiret, Klaus Stark, Dorette
Gieger, Lesca M. Holdt, Alexander Leichtler, Karin H. Greiser, Dominik Huster, Patrick Daniel Teupser, Ronny Baber, Uta Ceglarek, Markus Scholz, Thomas Illig, Christian
Genetic Regulation of Serum Phytosterol Levels and Risk of Coronary Artery Disease
http://www.lww.com/reprintsReprints: Information about reprints can be found online at
[email protected]. E-mail: Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Fax: Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters Kluwer
http://circgenetics.ahajournals.org/subscriptions/Subscriptions: Information about subscribing to Circulation: Cardiovascular Genetics is online at
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
Genetic Regulation of Serum Phytosterol Levels and Risk of Coronary Artery Disease
Running Title: Teupser et al.; Genetics of serum phytosterols and CAD risk
Daniel Teupser, MD; Ronny Baber, MSc; Uta Ceglarek, PhD; Markus Scholz, PhD; Thomas Illig, PhD; Christian Gieger, PhD; Lesca M. Holdt, MD; Alexander Leichtle, MD; Karin H. Greiser, MD; Dominik Huster, MD; Patrick Linsel-Nitschke, MD; Arne Schäfer, PhD; Peter S. Braund, MSc; Laurence Tiret, PhD; Klaus Stark, PhD; Dorette Raaz-Schrauder, MD; Georg M. Fiedler, MD; Wolfgang Wilfert, MSc; Frank Beutner, MD; Stephan Gielen, MD; Anika Großhennig, MSc; Inke R. König, PhD; Peter Lichtner, PhD; Iris M. Heid, PhD; Alexander Kluttig, PhD; Nour E. El Mokhtari, MD; Diana Rubin, MD; Arif B. Ekici, PhD; André Reis, MD; Christoph D. Garlichs, MD; Alistair S. Hall, MD; Gert Matthes, MD; Christian Wittekind, MD; Christian Hengstenberg, MD; Francois Cambien, MD, PhD; Stefan Schreiber, MD; Karl Werdan, MD; Thomas Meitinger, MD; Markus Löffler, MD; Nilesh J. Samani, FRCP; Jeanette Erdmann, PhD; H.-Erich Wichmann MD, PhD*; Heribert Schunkert, MD*; Joachim Thiery, MD*
* Contributed equally
Inst of Lab Med, Clin Chem & Molecular Diagnostics (DT, RB, UC, LMH, AL, GMF, WW, FB, JT), Inst for Med Informatics, Stats & Epidemiology (MS, ML), Dept of Med II (DH), Heart Ctr - Dept of Internal Med/Cardio (SG), Inst of Transfusion Med (GM), & Inst of Pathology (CW) Univ Leipzig, Germany; Instof Epidemiology (TI, CG, IMH, HEW), & Inst of Human Genetics (PL, TM) Helmholtz Zentrum München, German Research Ctr for Environmental Health, Neuherberg, Germany; Dept for Epidemiology & Preventive Med, Regensburg Univ Med Ctr, Regensburg, Germany (IMH); Inst of Med Informatics, Biometry & Epidemiology, Ludwig-Maximilians-Univ, Munich, Germany (HEW); Inst of Human Genetics, Klinikum rechts der Isar, Technical Univ, Munich, Germany (PL, TM); Inst of Med Epidemiology, Biostatistics, & Informatics, Martin-Luther-Univ Halle-Wittenberg (KHG, AK), & Dept of Med III (KW), Martin-Luther-Univ Halle-Wittenberg, Halle (Saale), Germany; Medizinische Klinik II (PL-N, AG, JE, HS), & Inst für Med Biometrie und Statistik (AG, IRK) Univ zu Lübeck, Lübeck, Germany; Inst für Klinische Molekularbiologie & Dept of Internal Med I, Universitätsklinikum Schleswig-Holstein, Kiel, Germany (AS, DR, SS); Dept of Cardiovascular Sciences, Univ of Leicester, Glenfield Hospital, Leicester, UK (PSB, NJS); Inst Nat de la Santé et de la Recherche Médicale (INSERM) Unité Mixte de Recherche UMR_S 525, Univ Pierre et Marie Curie (UPMC) Univ. Paris 06, Paris, France (LT, FC); Klinik und Poliklinik für Innere Med II, Univ Regensburg, Germany (KS, CH); Dept of Cardio & Angiology, Univ Hospital Erlangen, Germany (DR-S, CG); Klinik für Innere Medizin, Kreiskrankenhaus Rendsburg, Rendsburg, Germany (NEEM); Inst of Human Genetics, Univ of Erlangen-Nuremberg, Erlangen, Germany (ABE, AR); Leeds Inst of Gen, Health & Therapeutics, Univ of Leeds, Leeds, UK (ASH);
Correspondence: Dr. Daniel Teupser, University Leipzig, Liebigstr. 27, 04103 Leipzig,E-mail: [email protected]; Telephone +49-341-9722204; Fax +49-341-9722379 or Dr. Heribert Schunkert, Universität zu Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany. E-mail: [email protected]; Telephone +49-451-5002501; Fax: +49-451-5003060.
Journal Subject Codes: [89] Genetics of Cardiovascular Disease, [90] Lipid and Lipoprotein Metabolism, [109] Clinical Genetics, [146] Genomics
y
C & ( , , W
zC zs tgensburg Univ Med Ctr Regensburg Germany (IMH); Inst of M
y
ChChChChCheem & MMMMolooolo ecececcculululululararararar DDDDDiaiaiaiaagngngngngnooostitititicscccsc (DTD , , , , RBRBRBRBRB,,, ,, UCUUCUCUC, LMLMLLML H,H,H,H,H, AA A AAL,L,L,LL GG G GGMFMFMFMFMF, Wtaaaaatststsss & Epipppp deddemimiooologygygygygy (MSMSMSMM , MLLL), DDepppppt t ttt ofof MMeeed IIII (DDDDDHHH)H , HeHeHHH ararrt CtCtC r offf TTTTTrararaaannnnsn fuusisiiionononn M MMMMedddd ( ( (((GGMGMGG ),)))) & &&&& InInInI ststst o oof ffff PaPaPaPaPathththththololololologogogogo yyyy (C(C(C((CW)W)W)W)W) UUUUUniiniiivvv LeLeLLeL ipiipii z
CG, IMH, HEHEEHEEW)W)W)W)W ,,, , & &&&& InInInInnststststst ofofofofof HHHHHumumumumumanannnn GG G GGenenennneteteteteticicicicics ssss (P(P(P(P(PL,L,L,LL, TTTTTM)M)M)M)M) Helmholtzseeearararchch C C Ctrtrtrt f fororor EE Envnvnvirironononmemementntntalal HH Heeealalthth, NeNeNeN uhuherererbebebbb rgrgrggg, ,, GeGeGermrmrmananany;y;y; DD Depeppeptgeggensnsbububurgrg UUUUUnininiiivv MeMeMeM ddddd CtCtCtC rr RRRRRegegenenssbubububb rgrg GeGeGeGG rmrmananyy (I(I(I(I(IMHMHMHMHMH);););) IIIIInsnstt ofofof MMMM
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
contrrrrrrrrrrrrrrrrrolololololo s.sssssssssss GGGGGGGGGGGGGGGGGGenenenenenneneneeneennenenenenenenneeeeeteeeeeeee i
e w
u
3
eeeee tt tttrrrarr nspopopooortrtrtrttereeee ABABABABABCGCGCGCGG88888 aaa aandndndndn aaaaat tt thhee blblblblblooooooooood d d d d grgrgrgrouououououp ABABABABABOOOOO lllocococococususususus w
ooooweweweririringngnggg ee effffffecececttt, thththererereee isisis ii incncncrerereasasasinining g g cococoncncncererernnn thththatatat ee elelelevavavateteteddd ssss
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
Human SNP Array 6.0 was employed in the German MI Family Study II. Genotyping of
individual SNPs was performed using iPlex single base primer extension and MALDI-TOF
(matrix assisted laser desorption/ionization time-of-flight) mass spectrometry (Sequenom, San
Diego, CA, U.S.A.),18 a melting curve based method with a single fluorescently labelled probe on
an ABI PRISM 7900 HT Sequence Detection System (Applied Biosystems, Darmstadt,
Germany),19 and TaqMan allelic discrimination (Applied Biosystems, Darmstadt, Germany).
Gene expression analysis
RNA was isolated from healthy appearing segments of liver samples using the monophasic Trizol
reagent (Invitrogen, Carlsbad, CA). Gene expression of ABCG5, ABCG8 and beta-actin was
determined in an ABI PRISM 7900 HT Sequence Detection System (Applied Biosystems,
Darmstadt, Germany) by TaqMan quantitative RT-PCR.19
stududddddddddddddddy y yy yyyyyyyyyyyyyyyyyyyy wewwewewewwewewewewewewewewwwererererererererererererereereree pp p ppp ppppp ppppppppper
x
I
a n
p®p®p®p®p® 50000KK KKK MaMaMaMaMappppppppppininining g g g g ArAAAA rarararr y yyy SeSeSeSS t,1666, 17 w w w w wheeeeerererer asasasasas t heeee AA AA Affffff ymymymymymeeeeetrtrtrtrrix
6.6666 000 00 wawawawawasss s emememmmplplplplployoyoyoyededededed iiiiin ttttthehehehehe G GGGGerereere mamamamaman nnnn MIMIMIMIMI FFFFFamamamamamililililily yyyy StStStStStududududu y yyyy IIII
as ss pepeperfrfrfororormememeddd usususinininggg g iPiPiPlelelexxx sisisingngnglelele bb basasaseee prprprimimimererer ee extxtxtenenensisisiononon a aa an
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
First stage genome-wide association study of KORA S3/F3
From 490,032 SNPs, a total of 390,130 were selected based on stringent quality criteria
(inclusion criteria for autosomal SNPs: call rate 95%, minor allele frequency (MAF) 1%, P-
values of exact HWE test 10-6). Campesterol, sitosterol, brassicasterol and corresponding ratios
normalized to total cholesterol concentrations as well as total cholesterol itself were log-
transformed prior to analysis. Models of additive genetic effects and recessive minor allele
effects were calculated adjusting for age, sex and log(BMI). For detection of population
stratification, we analysed QQ-Plots for all these test statistics. Inflation factors ranged between
1.00063 and 1.012, indicating no relevant inflation of test statistics (Supplementary Figure 1).
Adjustment for the first three principal components did not substantially change the identified
associations, supporting the absence of significant bias caused by population stratification
(Supplementary Table 1). In addition, we used multivariate analysis of variance (MANOVA) to
calculate a summary statistic for the combination of both the total phytosterol concentration and
the ratios of total phytosterol and total cholesterol concentration.
Second stage, validation in KORA S3/F3 stage 2
We selected 68 SNPs for further validation in remaining individuals of the KORA S3/F3 study
(n=1157). These included the 65 top SNPs of the list of SNPs ordered by the minimum of the p-
values of all univariate phenotype associations. In addition, three SNPs located in ABCG8 were
genotyped. These include SNP rs4245791, which had initially violated quality criteria (call rate,
HWE) on the 500K Array Set due to misgenotyping, and the two coding SNPs rs11887534
(D19H) and rs4148217 (T400K) not present on the 500k Array Set with known associations with
serum phytosterol levels.9 SNPs were genotyped using the Sequenome assay. From the 68
Inflation factttttororororororrorororroorororoororoooo s
isticccccccccs s s s ss s s s s sss s ssss sss (S(S(S(S(S(S((S(S(S(S(S(S(S(S(S(S(S(S(SS(S(Supupupupupupupupupupupupupupuupupupupupupupuuppu plplplplplplplplplpplplplpplplplpllpllpllemememememeemeeemememeemeememeeeeeem
f n
r t
l e
fiririririrssstss threeeeeeee p pp ppriiiiincncncncncipipipipipalalalalal ccc c comomomomompopopopoonenn nnts dddddididididid n nnnototototo sssssubububububstttttananananantititititialalalaa lylylylyly c cc cchahhhh n
rtinggggg t tttthehehehehe aaabsbsbsbsseeeencncncn e ofofofofof sigigigigignininininififififificaccacac ntntnt bbbbbiaiaiaiaiasss ss cacacacacausususu edeede bbbbby yyyy popopop pupupupulalalall tdd
lelle 1 1)). I In nnnn adadadadaddidididid titititionononnn, , , ,, wewwewew uu uusesesessed d d dd mumumumuultltltll ivivivii arararariaiaiaaiateteteee a a aaananananan lylylylylysisisisisis s s s ofofofoof v vv vvariiancce
7
initially selected SNPs, a total of 9 SNPs, including the 4 SNPs located in ABCG8 (rs41360247,
rs4245791, rs11887534 and rs4148217) and 5 additional SNPs showed p-values less than 0.01 in
at least one of the test statistics in the second stage and were selected for the final replication step.
Third stage, validation in CARLA
The 9 SNPs selected in stage 2 were genotyped in n=1760 individuals with full phenotype and
covariate information in the CARLA cohort. For association analyses data were additionally
adjusted for statin treatment. Five SNPs of the total of 9 SNPs selected in the second stage were
finally validated with significance levels below Bonferroni corrected thresholds in at least one of
the test statistics. The set of validated SNPs comprised again all four SNPs in ABCG8
(rs41360247, rs4245791, rs11887534, rs4148217) and one SNP in ABO (rs657152).
Fine mapping and haplotype analysis in CARLA
For fine mapping of the ABCG5/8 locus, we genotyped additional SNPs in the haplotype block
containing the four SNPs validated in the third stage from HapMap including flanking and known
coding SNPs in CARLA subjects. After phasing of the data,20 we determined the allelic
association for each of the haplotypes. Finally, we determined the genetic association for the
major haplotype variants determined by rs4952688 and rs11887534 using additive models.
Combined analysis
We calculated a combined effect for the validated SNPs (rs41360247, rs4245791 and rs657152)
which were genotyped in all three stages using regression models which additionally included
cohort assignment variables.
ected thresholdssdssdsdssdsss
gain aaaaaaaaaaaaaaalllllllllllll f ff fffffouououououououououououououuououououuuur rrrrr S
Genome-wide Study of Plasma Phytosterols and Replication
The initial genome-wide analysis using the Affymetrix 500k array identified one single
association at the ATP-binding cassette hemitransporter G8 (ABCG8) gene (rs41360247)
achieving genome-wide significance for phytosterol serum levels (Table 1). One additional SNP
(rs4245791), located 775 bp distal to rs41360247, was also highly significant but had to be
excluded in the initial analysis due to quality problems (Supplementary Table 3). This SNP also
achieved genome-wide significance after re-genotyping using the Sequenome assay (Table 1). A
total of 68 SNPs (Supplementary Table 3) were taken forward for validation in additional 1157
subjects of the KORA S3/F3 study (Supplementary Table 4) and 9 SNPs achieving nominal
significance of p<0.01 were taken forward for replication in 1760 individuals of the independent
CARLA study (Supplementary Table 5).
Fine-mapping and Haplotype Analysis of ABGC5/8
SNPs rs4245791 and rs41360247 at the ABCG8 locus were significantly associated in all three
studies (Table 1, Supplementary Table 6) and were independent of each other (r2=0.03,
Supplementary Table 7). Fine mapping of the haplotype block in CARLA (Figure 2,
Supplementary Table 8) revealed that rs41360247 was in close linkage disequilibrium (r2 = 0.93)
with coding SNP rs11887534 (D19H), which has been associated with phytosterol levels in
previous studies and is known to affect protein structure.9 In addition, SNP rs4952688 was
identified by fine-mapping as a proxy for rs4245791 (r2 = 0.89) with lowest p-values of
association of all SNPs used for fine-mapping. Haplotype analyses of the ABCG8 locus indicated
e Sequenome asaaaaaaaaaa
for vvvvvvvvvvvvvvvvvalalalaala ididididididididatatatatata ioioioioioooioioioioioioiooon n n nnn n nn iiin
R c
0 f
p
RAAAAA S3/F3F3F3F3F s ss sstudududududy y y yy (S(S(S(S(Supupupupupplplplpllememememmenenentaarry TTTTTababababablelelelele 4 4 44) ) ) )) anaaaa d d d d d 9 9 9 99 SNNNSNNPsPsPsPsP a c
01 wwwwwererererere eeee tatattakekekekeen nnnn fofofofoforwwwwwararararard dddd fofofofofor rrrr rerererereppppplililililicacacacacatititititiononononon ii iiin nnnn 17171717176060606060 iiiiindndndndndivivivivividididididuauauauau lslslslsls oo of
Association of ABCG5/8 SNP rs4952688 with mRNA Expression
One possible mechanism for the association between SNP rs4952688 and serum phytosterol
levels was by affecting expression levels of ABCG5 or ABCG8. To test this hypothesis, we
determined mRNA levels of these genes in 57 patient samples of normal human liver tissue and
observed significantly reduced mRNA expression levels of these two genes in association with
the T allele of rs4952688 (Figure 3) but not with rs41360247 or rs11887534 (D19H). Sequencing
of the putative intergenic promoter region revealed no SNPs that were associated with expression
levels, suggesting that the responsible variant resides outside this region (Supplementary Figure
3).
Association of Phytosterols with ABO Blood Groups
Another novel finding was that in addition to ABCG8, the ABO-gene locus was consistently
associated and also achieved genome-wide significance for association with phytosterol levels in
the combined analysis (Table 1, Supplementary Table 6). The effect of the ABO gene SNP
rs657152 on phytosterol levels was independent of the effects mediated by SNPs in the ABCG8
gene (Supplementary Table 7). The explained variance of serum phytosterols by ABO and
ABCG8 loci was ~10% (Supplementary Table 11). ABO codes for a polymorphic glycosyl-
transferring enzyme, responsible for the major blood groups. Our studies revealed that rs657152
was tightly linked with the blood group O1 allele (Supplementary Figure 4), coding for a protein
devoid of glycosyltransferase activity. Genetic analysis of blood groups in CARLA and
immunological determination in an independent cohort of blood donors confirmed that the non-
rs11887534 (D1DDDDDDDDDD
werererererererrererererrrrerre aaaaassssssssssssssssssssococococococococococcococcocccooo iaiaiaiaiaiaiaiaiaiaaiaiiai tttttetettttttttttt d
a e si s paaaaattt tt ttthtt e reeeeespspspspponononoonsisisisisiblblbblb e eeee vavavavavarirrrr ananananant tt rerererr sidedes ouououououtsttstst ididididde e e e thhhhhisiiii rrregegegegegioioioioon nn (S(S(S(S(Supupupupupp
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
functional O allele was associated with decreased phytosterol serum levels (Figure 4,
Supplementary Tables 12, 13).
Meta-analysis of Association of Identified Phytosterol SNPs with CAD Risk
Given the evidence suggesting that elevated phytosterol levels may increase the risk of
atherosclerosis, we next tested the association of variants in ABGC8 (rs41360247, rs4245791)
and ABO (rs657152) with CAD. This was done in a metaanaylsis of 11 different studies
comprising a total of 13,764 CAD cases and 13,630 healthy controls (Figure 5). Detailed results
for each study are presented in Supplementary Figure 5 and Supplementary Tables 14-16. We
found that alleles associated with increased phytosterol levels were positively associated with
increased probability of CAD, while alleles associated with reduced phytosterols were associated
with reduced probability of CAD (Figure 5). We also tested the effect of identified genetic
variants on LDL-cholesterol levels, since recent studies have shown an association with SNPs in
ABCG5/8 (Aulchenko et al, Kathiresan et al Nat Genet 2009). The latter could be confirmed for
ABCG8 rs41360247 and rs4245791. We also found an association of ABO (rs657152) with LDL-
cholesterol (Supplementary Figure 6).
Discussion
Our genome-wide analysis and functional studies revealed that ~10% of the variability of serum
phytosterol levels in the normal population is explained by three variants found at the ABCG8
and ABO gene loci. Using this information, we investigated whether genetic variants affecting
phytosterol levels also modulate the risk of CAD. We found that all three polymorphisms
identified to display association with phytosterols were independently associated with CAD.
upplementaryyyyyyyyyyyy TTTTTTT
werererererererereereerrrrerr p ppp pppppppppppppppppososososoooososososossossititititititittititititititi ivivivivivivivivivivivivvivivveleelelelelllee y
y
b i
o t
y ofofofofof CADDDDD, ,,, whwhwhwhwhililillile ee e e alaaaa leleleleleleleleleles asasasa sosososoocicc aateed wwwwwitititiith hh hh rerereredududuuducecccc d d d d phphphphphytytytyy ososososo teteteteterorrrr l
bilililililitytytytyty o o o o of ffff CACACACAAD DDDD (F(F(F(FFigigigigigurururururee 5)5)5)5)5). WeWeWeWeWe a aaaalslslslslso oooo tetetetetestststststededededed tt ttthehehehehe efefefefeffefefeff ctctctctct ooooofff ff i
olllesesesteteterororolll lelelevevevelslsls, sisisincncnceee rererecececentntnt ss stttudududieieiesss hahahaveveve ss shohohownwnwn aa annn asasassososociciciatattatttt
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
Polymorphisms associated with increased phytosterol serum levels were associated with an
increased risk of CAD, whereas a polymorphism associated with decreased phytosterols was
associated with decreased CAD risk. Thus, our approach using genome-wide analysis of the
intermediate phenotype of serum phytosterols, which is as a maker of cholesterol homeostasis led
to the identification of 3 novel genetic variants modulating CAD risk.
ABCG8 is a plausible candidate for affecting the inherited variability of serum phytosterol levels,
given that the gene encodes the ATP-binding cassette hemitransporter that carries phytosterols
into the bile.1, 2, 5 Indeed, smaller studies previously reported an association between the coding
variant D19H in this gene and serum phytosterols,9 a finding that was confirmed by our data.
D19H was also found to affect the susceptibility for cholesterol gall stone disease.22 It was
speculated that the 19H variant may increase the efficiency of sterol excretion into the bile
lumen, causing hypersaturation of the bile, subsequently leading to gall stone formation.23
Indeed, there is published data about an association between the D19H variant and serum
cholesterol levels.24, 25 Moreover, recent genome wide studies identified an association of LDL-
cholesterol with proxies to D19H and the other ABCG8 variant, rs4245791.26, 27 This effect could
be confirmed in the present study (Supplementary Figure 6), albeit the association of D19H – and
the other variants we identified – with serum cholesterol levels was only weak and effects on
phytosterols remained highly significant after normalization to cholesterol (Table 1) or
adjustment to LDL-cholesterol (Supplementary Table 17).
A novel finding of the present study is that a second genomic effect at the ABCG8 locus adds
independently to the association with serum phytosterol levels. This one was tagged by SNPs
rs4952688/rs4245791 and related to increased liver expression of ABCG5 and of ABCG8
mRNAs. The parallel regulation of two genes suggested that rs4952688/rs4245791 might be
n association beeeeeeeeeet
that wawwawawawawwawawawawawawaaww s ssss cococococococoococoococoocoooccconfnffnfnfnfnfnfnffnfnfnffnfffnfnffnfiiiriiiririiiiiiiiiii m
n
t
t
nnnnnd dd dd to afffffff ececececect thththththeee ee susususususcscscscscepepepepptittt bibbibib lill tyy fororororor c c c cchooooolelelell ststststererereerolll g g g g galalaalall stststststononononone
199999HHH HH vavavavavarirrrrianananannt ttt mamamamam y ininininincrcccc eaeaeaeaeasesesesese ttttthehehehehe e eeeefffffffffficicicicicieieieieencncncncncy yyyy ofofofoof sssssteteteteterororororol exexexeexcrcrcrccreeetee
linked to a variant, which affects transcriptional activation. However, sequencing of 6kb around
the intergenic region revealed no obvious causative mutations, indicating that other factors
outside this region might be responsible (Supplementary Figure 3).
An unexpected finding was that the ABO blood group gene locus also affects serum phytosterol
levels. The O-allele, which leads to dysfunctional mutations devoid of glycosyltransferase
activity, was associated with significantly reduced phytosterol concentrations. One may speculate
that addition of carbohydrate groups to oligosaccharide chains of proteins might either reduce the
activity of proteins responsible for eliminating sterols or induce the activity of proteins
responsible for sterol uptake. In this regard, it is of interest that both ABCG5 and ABCG8
undergo N-linked glycosylation.28 However, the specific biological mechanism by which ABO
alters phytosterol levels is unclear.29 Interestingly, it has been previously reported that serum
cholesterol levels are slightly but consistently elevated in non-O subjects.30, 31 In this regard, it is
of interest that ABO also showed an association with serum total and LDL-cholesterol levels in
our analyses (Table 1, Supplementary Figure 6).
Importantly, the genetic variants associated with serum phytosterols were also associated with
risk of CAD. It should be emphasised that we only tested the associations of these variants with
CAD after their strong association with phytosterol levels became apparent. Therefore, the
significance levels achieved for the association of the variants with CAD can be considered to be
reasonably definitive. Hence the present study adds three additional variants to the evolving list
of genetic markers of this common disease.16, 32, 33 However, our data fall short to prove that these
two associations are causally linked, i.e. that the increase in CAD risk is functionally mediated by
higher phytosterol serum levels, since the identified variants also had a concomitant effect on
nduce the actitttttttttt
that bobobobobobothththhhhhh AA AA AAAAA AAAAAAAAAAAABCBCBCBBCBCBCBCBBCBBCBBBBB
m
v p
e n
yyyyycococococosylatitititit ononononon.28 HHH HHowowowowoweveveveveverererrr, thththhhe ee sppecififififificicicicic b b bb bioioioioiololoologigigiggicaaaallll l memmmem chchchchchanananananisiisisi m
velelelelelssss s isisisisis uuuuuncncncnclelelellearararara ...29 IIIIIntntntntntereeee esesesesestititititingngngngnglylylylyly,,, ititittt h hhhhasasasasas bbbbbeeeeeeeeeen nnnn prrprprprevevevevevioioioioioususussslylylyyy rrrrrepepepee
eeee ss slililighghghtltltly y y yy bububuttt cococonsnsnsisisistetetentntntlylylyyy ee elelelevavavateteteddd ininin nn nononon-O-O-O ss subububjejejectctctsss.30, 31 II Innnn
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
regard to the frequent use of phytosterol food supplements, since a substantial number of
individuals with certain genotypes may respond with relatively high phytosterol serum levels
after intake of these additives.3
In summary, this is the first genome-wide association study investigating genetic variability of
serum phytosterol levels in the general population. We identified significant associations of
serum plant sterols with three functional genetic variants. Particularly, our data suggest novel
additive mechanisms for ABCG8 and ABO in regulating serum phytosterol levels which also
impact serum LDL-cholesterol levels. Moreover, we show that common genetic variants
associated with serum phytosterol levels affect CAD risk in a concordant fashion. These data
show for the first time that genetic variants affecting sterol homeostasis play a role in susceptibly
to CAD.
Funding Sources: The KORA research platform was initiated and financed by the Helmholtz Center Munich, which is funded by the German Federal Ministry of Education and Research and by the State of Bavaria. The KORA GWAS was supported by the German Ministry of Education and Research through the National Genome Research Network (NGFN). Members of the KORA Study Group are listed in the online supplement. The CARLA Study was funded in part by a grant from the German Research Foundation. The German MI Study was supported by the Deutsche Forschungsgemeinschaft and the German Federal Ministry of Education and Research (BMBF) in the context of the German National Genome Research Network (NGFN-2 and NGFN-plus). We are grateful to the WTCCC and the Cardiogenics Consortium for allowing us to use data from their CAD genome-wide association scans. Cardiogenics is an EU funded integrated project (LSHM-CT-2006-037593). The Leipzig Heart Study was funded in part by a grant from the Roland-Ernst-Foundation to D.T.. N.J.S. holds a Chair funded by the British Heart Foundation. Part of the study was funded by a grant from the German Ministry of Education and Research through the National Genome Research Network (NGFNplus) to D.T. and J.T. M.S. was funded by the German Federal Ministry for Education and Research 01KN0702. Part of the study was funded by a grant of the Medical Faculty, University Leipzig to A.L.
en JC,CC,C,C,C,C,C,C,CC,C,C,C,C,C,C,C,,,, H HHH H Hobobobobobobbbbobobbobbobbsbsbsbsbsbsbsbsbsbsbsbssbsbbsb HHHH HH
i
e
is , ai e es et , i
L A
l bi h d d h l i d l Gl b l
ingngngngng cc casasasasassssess ttttttttttee eee proteins ABCG55555 and ABCBCBCCBCG8G8G8G8G8 permits their
e. JJJ JJ Clin Invnvest. 200000002000 ;111111000:659559-666969999....
is ALALALALAL, LLe M M MMMouo lllllllacacacacc-V-V-V-V-Vaidydydydye B,BB,BB, DD Despipipipipiauauaauau SSSSS, ,,,, RoRoRoR uuubu iniii et FFF, BBoBB vi
Table 1: Validation and replication of major genetic associations of serum phytosterol levels
Genome wide association in KORA S3, validation in the remaining individuals of KORA S3, replication in the CARLA cohort and combined analysis of the three
SNPs with best p-values of association with phytosterols. CA, campesterol; SI, sitosterol; BR, brassicasterol; CH, cholesterol; MANOVA, multivariate analysis of
CA, SI, BR; CA/CH, campesterol normalized to cholesterol; SI/CH, sitosterol normalized to cholesterol; BR/CH, brassicasterol normalized to cholesterol;
MANOVA/CH, multivariate analysis of CA/CH, SI/CH, BR/CH; bp position refers to NCBI build 36. Alleles, major allele > minor allele; MAF, minor allele
frequency; CR, call rate; HWE, P value of deviation from Hardy-Weinberg equilibrium; p-values of association are given for the additive model for rs41360247 and
rs4245791 and for the recessive model for rs657152. Effects on plasma phytosterol concentrations are shown in Supplementary Table 6.
Allelic effect and p value of association
Cohort SNPGene Chr bp position Alleles
MAF CR HWE CA SI BR MANOVA CA/CH SI/CH BR/CH MANOVA/CH CH
Supplemental Material Teupser et al., Genetic regulation of serum phytosterol levels and risk of coronary artery disease The supplemental materials have the following sections in order:
1. Study cohorts…………………………………………………………………....2 2. Genotyping and gene expression analysis……………………….………..5 3. Statistical analysis……………………………………………………………...7 4. References……………………………………………………………….……..13 5. Supplementary Tables………………………………………………………..16 6. Supplementary Figures…………………………………………………........35 7. Members of the KORA Study Group………………………………………..41
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
Effects and p-values of association of first stage genome-wide association study after adjustment for the first three principle components
SNP CA SI BR MANOVA CA/CH SI/CH BR/CH MANOVA/CH CH rs41360247
ABCG8 -15%
2.0x10-10 -24%
8.8x10-16 -17%
6.0x10-12 5.2x10-15
-14%
1.8x10-12 -24%
2.8x10-18 -16%
8.0x10-14 1.2x10-16
-0.5% 0.71
rs4245791 ABCG8
13% 4.8x10-18
21% 3.2x10-24
15% 2.9x10-20
2.2x10-22
11%
4.5x10-19 20%
6.4x10-26 13%
7.6x10-21 9.2x10-24
1.2% 0.11
rs657152 ABO
8% 9.1x10-5
10% 1.6x10-4
7% 1.6x10-3
5.4x10-4
8%
6.6x10-6 10%
2.8x10-5 7%
3.8x10-4 2.3x10-5
-0.04%
0.97 CA, campesterol; SI, sitosterol; BR, brassicasterol; CH, cholesterol; MANOVA, multivariate analysis of CA, SI, BR; CA/CH, campesterol normalized to cholesterol; SI/CH, sitosterol normalized to cholesterol; BR/CH, brassicasterol normalized to cholesterol; MANOVA/CH, multivariate analysis of CA/CH, SI/CH, BR/CH; p-values of association are given for the additive model for rs41360247 and rs4245791 and for the recessive model for rs657152.
Geometric mean and standard error per genotype for phytosterols and cholesterol (in mg/L)
SNP Allele CA SI BR CA/CH SI/CH BR/CH CH
rs41360247 Hom. major T/T
5.9 (1.01)
2.3 (1.01)
0.60 (1.01)
2.7 (1.005)
1.1 (1.01)
0.28 (1.01)
2162 (1.003)
Het. T/C
5.1 (1.01)
1.8 (1.02)
0.50 (1.02)
2.4 (1.01)
0.85 (1.02)
0.23 (1.02)
2132 (1.01)
Hom. minor C/C
4.6 (1.05)
1.4 (1.12)
0.46 (1.06)
2.3 (1.05)
0.68 (1.11)
0.22 (1.06)
2038 (1.02)
rs4245791 Hom. major T/T
5.4 (1.01)
2.0 (1.01)
0.54 (1.01)
2.5 (1.01)
0.92 (1.01)
0.25 (1.01)
2141 (1.004)
Het. T/C
6.1 (1.01)
2.4 (1.01)
0.62 (1.01)
2.8 (1.01)
1.11 (1.01)
0.29 (1.01)
2175 (1.005)
Hom. minor C/C
6.7 (1.02)
2.8 (1.02)
0.71 (1.02)
3.1 (1.01)
1.31 (1.02)
0.33 (1.02)
2166 (1.01)
rs657152 Hom. major G/G
5.6 (1.01)
2.1 (1.01)
0.57 (1.01)
2.6 (1.01)
0.99 (1.01)
0.27 (1.01)
2140 (1.005)
Het. G/T
6.0 (1.01)
2.3 (1.01)
0.60 (1.01)
2.8 (1.01)
1.06 (1.01)
0.28 (1.01)
2172 (1.004)
Hom. minor T/T
5.9 (1.01)
2.3 (1.02)
0.59 (1.02)
2.7 (1.01)
1.04 (1.02)
0.27 (1.01)
2172 (1.01)
Values were adjusted for age, sex, log(BMI), statin treatment and study. CA, campesterol; SI, sitosterol; BR, brassicasterol; CH, cholesterol; CA/CH, campesterol normalized to cholesterol; SI/CH, sitosterol normalized to cholesterol; BR/CH, brassicasterol normalized to cholesterol.
Analysis of all significant SNPs of serum phytosterol levels within one regression model
Allelic effects and
p-value of association
SNP CA SI BR MANOVA CA/CH SI/CH BR/CH MANOVA/CH CH
rs41360247 -10% 4.4 x 10-14
-17% 3.0 x 10-18
-13% 1.1 x 10-15 1.3 x 10-29 -9%
4.4 x 10-13 -16%
1.5 x 10-18 -11%
2.1 x 10-14 2.6 x 10-31 -1.6% 0.065
rs4245791 11% 2.3 x 10-39
19% 1.7 x 10-52
13% 4.1 x 10-43 2.4 x 10-55 10%
1.5 x 10-41 18%
1.3 x 10-56 12%
1.6 x 10-43 5.1 x 10-62 0.8% 0.068
rs657152 8% 9.4 x 10-13
9% 2.4 x 10-8
6% 4.9 x 10-7 2.8 x 10-10 6%
2.2 x 10-10 7%
5.1 x 10-7 5%
5.0 x 10-5 3.5 x 10-9 1.5% 0.012
Allelic effects relative to the major allele and corresponding p-values. Analysis is based on the combined data sets of KORA S3 500k, KORA S3 Stage 2 and CARLA replication. Data were adjusted for age, sex, log(BMI), statin treatment and study. The regression model simultaneously included the additive effects of rs41360247 and rs4245791 and the recessive effect of rs657152. For rs657152 and phytosterol phenotypes, the results are identical with Table 1 (combined analysis). CA, campesterol; SI, sitosterol; BR, brassicasterol; CH, cholesterol; MANOVA, multivariate analysis of CA, SI, BR; CA/CH, campesterol normalized to cholesterol; SI/CH, sitosterol normalized to cholesterol; BR/CH, brassicasterol normalized to cholesterol; MANOVA/CH, multivariate analysis of CA/CH, SI/CH, BR/CH.
Yellow: P<0.01; Red: P<0.001. *indicates SNPs which were used for haplotype analysis (Supplementary Table 9). Otherwise see legend to Supplementary Table 3.
The variants at rs4952688 and rs11887534 are marked red in the haplotype column. Haplotypes containing C/T variant showed high phytosterol concentrations throughout while haplotypes
containing the G/A variant showed low concentrations. Results of association analysis for these haplotypes are shown in Supplementary Table 10. ca, mean and standard error of serum
campesterol (mg/L); si, mean and standard error of serum sitosterol (mg/L); br, mean and standard error of serum brassicasterol (mg/L); ca/ch, mean and standard error of serum campesterol
normalized to cholesterol (x103); si/ch, mean and standard error of serum sitosterol normalized to cholesterol (x103); br/ch, mean and standard error of serum brassicasterol normalized to
cholesterol (x103); ch, mean and standard error of serum cholesterol.
Geometric mean and standard error of phytosterol and cholesterol concentrations in the CARLA cohort for all genotypes of the three haplotype varaints of rs4952688 and rs11887534 (in mg/L)
Dose effect C/T 0.11 0.19 0.14 0.11 0.19 0.13 0.0045 p-value C/T 2.7 x 10-20 2.2 x 10-24 1.5 x 10-18 1.9 x 10-22 3.7 x 10-27 2.0 x 10-19 0.54 Dose effect G/A -0.11 -0.20 -0.17 -0.097 -0.19 -0.16 -0.013 p-value G/A 2.8 x 10-06 1.4 x 10-7 4.0 x 10-08 5.2 x 10-06 4.8 x 10-08 7.7 x 10-08 0.36 Positive dose effect for haplotype C/T and a negative dose effect for haplotype G/A for all phytosterol traits but not for cholesterol. Effects and p-values are shown for the additive model. Date
were adjusted for age, sex, log(BMI) and statin treatment status. ca, mean and standard error of serum campesterol (mg/L); si, mean and standard error of serum sitosterol (mg/L); br, mean and
standard error of serum brassicasterol (mg/L); ca/ch, mean and standard error of serum campesterol normalized to cholesterol (x103); si/ch, mean and standard error of serum sitosterol
normalized to cholesterol (x103); br/ch, mean and standard error of serum brassicasterol normalized to cholesterol (x103); ch, mean and standard error of serum cholesterol.
Phytosterol and cholesterol concentrations (mg/L) in relation to blood groups in CARLA (genetic determination)
Blood group ca si br ca/ch si/ch br/ch ch O (N=623) 5.44 (1.01) 1.91 (1.02) 0.58 (1.02) 2.71 (1.01) 0.95 (1.02) 0.296 (1.02) 2006 (1.01)A (N=777) 5.87 (1.01) 2.07 (1.02) 0.63 (1.02) 2.85 (1.01) 1.00 (1.02) 0.311 (1.01) 2067 (1.01)B (N=237) 5.72 (1.02) 1.98 (1.03) 0.61 (1.03) 2.84 (1.02) 0.98 (1.03) 0.311 (1.03) 2012 (1.01)AB (N=102) 5.77 (1.03) 1.98 (1.05) 0.60 (1.04) 2.82 (1.03) 0.97 (1.05) 0.298 (1.04) 2034 (1.02)p-value O vs. A,AB,B 7.6x10-5 0.015 0.0051 0.0020 0.056 0.042 0.056 Geometric mean and standard error of age, sex, log(BMI) and statin treatment status adjusted traits. Phytosterols were also adjusted for rs4245791 and rs41360247. Blood group O showed
reduced phytosterol concentrations while cholesterol concentrations are equal. P-values were calculated for the comparison of blood group O with the pooled blood groups A,B and AB. ca, mean
and standard error of serum campesterol (mg/L); si, mean and standard error of serum sitosterol (mg/L); br, mean and standard error of serum brassicasterol (mg/L); ca/ch, mean and standard
error of serum campesterol normalized to cholesterol (x103); si/ch, mean and standard error of serum sitosterol normalized to cholesterol (x103); br/ch, mean and standard error of serum
brassicasterol normalized to cholesterol (x103); ch, mean and standard error of serum cholesterol.
Phytosterol and cholesterol concentrations (mg/L) in blood donors (immunological determination)
Blood group ca si br ca/ch si/ch br/ch ch O (N=301) 5.17 (1.02) 2.26 (1.02) 0.67 (1.02) 2.84 (1.02) 1.25 (1.02) 0.370 (1.02) 1819 (1.02)A (N=296) 5.49 (1.02) 2.40 (1.02) 0.71 (1.02) 3.04 (1.02) 1.33 (1.02) 0.393 (1.02) 1810 (1.02)B (N=111) 5.48 (1.03) 2.38 (1.04) 0.72 (1.03) 3.02 (1.03) 1.32 (1.04) 0.395 (1.03) 1815 (1.03)AB (N=52) 5.49 (1.04) 2.39 (1.05) 0.69 (1.05) 3.01 (1.04) 1.32 (1.05) 0.378 (1.04) 1823 (1.04)p-value O vs. A,AB,B 0.011 0.044 0.031 0.014 0.030 0.021 0.874 Geometric mean and standard error of age, sex and log(BMI) adjusted traits. Blood group O showed reduced phytosterol concentrations while cholesterol concentrations are equal. P-values
were calculated to compare blood group O with the pooled blood groups A,B and AB. ca, mean and standard error of serum campesterol (mg/L); si, mean and standard error of serum sitosterol
(mg/L); br, mean and standard error of serum brassicasterol (mg/L); ca/ch, mean and standard error of serum campesterol normalized to cholesterol (x103); si/ch, mean and standard error of
serum sitosterol normalized to cholesterol (x103); br/ch, mean and standard error of serum brassicasterol normalized to cholesterol (x103); ch, mean and standard error of serum cholesterol.
Metaanalysis of association of ABCG8 SNP rs41360247 with CAD
Cohorts are described in supplementary methods; MAF, minor allele frequency (minor allele C, major allele T); OR (95% CI), odds ratio and 95% confidence interval using additive and recessive
Metaanalysis of association of ABCG8 SNP rs4245791 with CAD
Cohorts are described in supplementary methods; MAF, minor allele frequency (minor allele C, major allele T); OR (95% CI), odds ratio and 95% confidence interval using additive and recessive
Metaanalysis of association of ABO SNP rs657152 with CAD
Cohorts are described in supplementary methods; MAF, minor allele frequency (minor allele A, major allele C); OR (95% CI), odds ratio and 95% confidence interval using additive and recessive
Replication of major genetic associations of serum phytosterol levels in CARLA with additional adjustment to LDL-cholesterol levels
CA, campesterol; SI, sitosterol; BR, brassicasterol; CH, cholesterol; bp position refers to NCBI build 36. MAF, minor allele frequency; CR, call rate; HWE, P value of deviation from Hardy-Weinberg equilibrium; P values of association are given for the additive model for rs41360247 and rs4245791 and for the recessive model for rs657152 after additional adjustment to LDL-cholesterol.
Allelic effect and P value of association
Cohort SNP Gene Chr bp position MAF CR HWE CA SI BR
Supplementary Figure 2: Geometric mean and standard error of campesterol for functionally relevant haplotypes of the ABCG8 locus in the CARLA cohort (n=1760). Haplotype analysis showed that the variation of phytosterol levels at this locus could be best explained by haplotypes defined by rs11887534 (C/G) and rs4952688 (A/T). These SNPs were tightly linked with rs41360247 and rs4245791 (see Figure 1), respectively but showed improved P-values of association. The CT haplotype was associated with elevated phytosterols (dose effect 0.11, P = 2.7 x 10-20), whereas the GA haplotype was associated with decreased phytosterols (dose effect -0.11, P = 2.8 x 10-6).
CA/CA CA/CT CT/CT CA/GA GA/GA CT/GA
7.06.05.04.03.0
0
rs11887534, rs4952688 haplotype
2.01.0C
ampe
ster
ol (m
g/L)
698 662 162 135 6 54
Supplementary Figure 2
Supplemental Material, Teupser et al 36
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
Supplementary Figure 3: DNA sequence analysis of 6kb of the intergenic region of ABCG5 and ABCG8 in 17 human liver samples. The region was selected based on a high level of conservation determined with the Vista browser. SNPs in the region failed (left) to show significant linkage with SNP rs4245791 or rs4962688 (right). Homozygousity for major alleles is indicated in blue and homozygousity for minor alleles in orange. “H” stands for heterozygous samples.
Supplementary Figure 3
Supplemental Material, Teupser et al 37 Sa
mpl
e ID
rs49
5302
1rs
6712
582
rs13
4252
60U
1-1
rs67
2805
3rs
6741
243
rs17
0317
00rs
1342
5681
U1A
-1rs
3452
0479
rs35
6362
39rs
3438
1269
U2-
1U
2-2
rs67
1054
4rs
1049
5909
rs67
5662
9rs
3806
471
rs11
8875
34 (D
19H
)rs
4245
789
rs38
0647
0rs
4148
202
rs10
1772
00
Pos
ition
[bp]
4391
6686
4391
6704
4391
6819
4391
6850
4391
6857
4391
6861
4391
7139
4391
7191
4391
7235
4391
7566
4391
7601
4391
7677
4391
7688
4391
7693
4391
7855
4391
8288
4391
8594
4391
9678
4391
9751
4392
0380
4392
0673
4392
1323
4392
1355
Major T C A C T A T T C C G A G G T G G T G G G T CMinor G T G T C G C A T T A C A A C A A G C A A G T# 1 H H H C H H T H C H H H H H T G G H G G H G C# 2 H H H H H H T H C H H H H H T G G H G G H H H# 3 T C A C T A T T C C G A G G H G G T G G H H C# 4 T C A C T A T T C C G A G G T G G T G G G T C# 5 H H H C H H T H C H H H H H T G G H G G H H H# 6 H H H C H H T H C H H H H H T G G H G G H H H# 7 T C A C T A T T C C G A G G T G G T G G G T C# 8 H H H H H H T H C H H H H H T G G H G G H H C# 9 T C A C T A T T C C G A G G T G G T G G G T C# 10 T C H C H G T H H H H A G G H G H T H G G H C# 11 H H H C H H T H C H H H H H H G G H G G H G T# 12 T C A C T A T T C C G A G G T G G T G G G T C# 13 G T G C C G T A C T A C A A H G G G G G A G C# 14 H H H C H H T H H H H A G G H G G T G G G H C# 15 H H H C H H T H C H H H H H H G G H G G H H C# 16 H H G C C G T A C T A C A A H G G G G G A G C# 17 H H H C H H T H C H H H H H T G G H G G H H C
rs41
4820
3rs
1017
9921
U10
-1U
11-1
4392
1386
4392
1795
4392
1845
4392
2480
C C C CT T G AC C C CC H C CC C C CC C C CC H C CC H C CC C C CC C C CC C C CC C H HC H C CC C C CC C C CC C H HC C C CC C C CC C C C
Sam
ple
ID
rs41
3602
47
rs42
4579
1
rs49
5268
8
rs41
4821
7 (T
400K
)
Pos
ition
[bp]
4392
7160
4392
7935
4395
0274
4395
2937
Major T T A CMinor C C T A# 1 T T A H# 2 T H H C# 3 T H H C# 4 T H H H# 5 T T A C# 6 T T A H# 7 T H H H# 8 T H H C# 9 T T A H# 10 H H H C# 11 T T A C# 12 T H H C# 13 T T A C# 14 T C T C# 15 H T A C# 16 T T A C# 17 T T H C
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
Supplementary Figure 4: (A) Haplotype structure of the ABO locus using the lead SNP rs657152 (75) and neighboring SNPs from the 500K Array Set. (B) SNPs determining blood groups and rs657152 and their corresponding LD-plot in CARLA (r2). Position and effect of the 5 SNPs on amino acids of ABO and blood groups are shown. Minor alleles of SNP 1 and 3 lead to blood group B (blue). Minor alleles of SNP 2 and 4 lead to blood group O2 and O1, respectively.
Supplementary Figure 5: Odds ratio and 95% CI in 11 studies of CAD and meta-effects using fixed effects and random effects models. (A) ABCG8, rs41360247 (B) ABCG8, rs4245791 (C) ABO, rs657152
Supplemental Material, Teupser et al 39
Supplementary Figure 5
A
B
C
0.6 0.8 1.0 1.2 1.6
Angio-LuebCARLA
ECTIM
Erlangen
GerMIFS II
GoKard
KORA-B
KORA-MI
LE-Heart
PopGen
WTCCC
Fixed effect
Random effects
0.8 1.0 1.2 1.4 1.6 2.0
Angio-Lueb
CARLA
ECTIM
Erlangen
GerMIFS II
GoKard
KORA-B
KORA-MI
LE-Heart
PopGen
WTCCC
Fixed effect
Random effects
0.8 1.0 1.2 1.4 1.6 2.0
Angio-Lueb
CARLA
ECTIM
Erlangen
GerMIFS II
GoKard
KORA-B
KORA-MI
LE-Heart
PopGen
WTCCC
Fixed effect
Random effects
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
Supplementary Figure 6: Fold change and 95% CI of LDL-cholesterol in CARLA for SNPslocated in ABCG8 and ABO genes using additive and recessive models, respectively.
0.8 0.9 1.0 1.1 1.2
rs41360247ABCG8
rs4245791
rs657152ABO
ABCG8
lower LDL-cholesterol higher LDL-cholesterol
Supplementary Figure 6
Supplemental Material, Teupser et al 40
P = 0.039
P = 0.078
P = 0.012
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from
Members of KORA Study Group Cooperative health research in the Region of Augsburg (KORA)
KORA study group consists of H.-Erich Wichmann1,2 (speaker), Rolf Holle3, Jürgen
John3, Thomas Illig2, Christa Meisinger1, Annette Peters1, and their coworkers, who are
responsible for the design and conduct of the KORA studies. The KORA S3/F3 500K
study was conducted by Christian Gieger1,2, Guido Fischer1, Iris M. Heid1,2, Susana
Eyheramendy1,2, Norman Klopp1,2, Peter Lichtner4, Gertrud Eckstein4, Thomas Illig2, H.-
Erich Wichmann1,2, and Thomas Meitinger4,5
1Institute of Epidemiology, GSF - National Research Center for Environment and Health,
85764 Neuherberg, Germany. 2Chair of Epidemiology, IBE, University of Munich, 81377 Munich, Germany. 3Institute of Health Economics and Health Care Management, GSF-National Research
Centre for Environment and Health, 85764 Neuherberg, Germany. 4Institute of Human Genetics, GSF National Research Center for Environment and
Health, 85764 Neuherberg, Germany 5Institute of Human Genetics, Technical University, 81765 Munich, Germany
by on June 8, 2010 circgenetics.ahajournals.orgDownloaded from