Adrenomedullin and Arterial Stiffness Integrative Approach Combining Monocyte ADM Expression, Plasma MR-Pro-ADM, and Genome-Wide Association Study

David-Alexandre Trégouët, François Cambien, Stefan Blankenberg and Laurence TiretThomas Münzel, Gilles Montalescot, Gary F. Mitchell, Germaine C. Verwoert, Kirill V. Tarasov,

Farzin Beygui, Philipp S. Wild, Tanja Zeller, Marine Germain, Raphaele Castagné, Karl J. Lackner,Expression, Plasma MR-proADM and Genome-Wide Association Study

ADMAdrenomedullin and Arterial Stiffness: An Integrative Approach Combining Monocyte

Print ISSN: 1942-325X. Online ISSN: 1942-3268 Copyright © 2014 American Heart Association, Inc. All rights reserved.

TX 75231is published by the American Heart Association, 7272 Greenville Avenue, Dallas,Circulation: Cardiovascular Genetics

published online July 22, 2014;Circ Cardiovasc Genet.

http://circgenetics.ahajournals.org/content/early/2014/07/22/CIRCGENETICS.113.000456World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://circgenetics.ahajournals.org/content/suppl/2014/07/22/CIRCGENETICS.113.000456.DC1.htmlData Supplement (unedited) at:

http://circgenetics.ahajournals.org//subscriptions/

is online at: Circulation: Cardiovascular Genetics Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. Further information aboutnot the Editorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center,Circulation: Cardiovascular Genetics Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by DAVID HARRISON on July 30, 2014http://circgenetics.ahajournals.org/Downloaded from by DAVID HARRISON on July 30, 2014http://circgenetics.ahajournals.org/Downloaded from by DAVID HARRISON on July 30, 2014http://circgenetics.ahajournals.org/Downloaded from by DAVID HARRISON on July 30, 2014http://circgenetics.ahajournals.org/Downloaded from

http://circgenetics.ahajournals.org/content/early/2014/07/22/CIRCGENETICS.113.000456

http://circgenetics.ahajournals.org/content/suppl/2014/07/22/CIRCGENETICS.113.000456.DC1.html

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://circgenetics.ahajournals.org//subscriptions/

http://circgenetics.ahajournals.org/




DOI: 10.1161/CIRCGENETICS.113.000456

1

Adrenomedullin and Arterial Stiffness: An Integrative Approach CombiningMonocyte ADM Expression, Plasma MR-proADM and Genome-Wide

Association Study

Running title: Beygui et al.; Adrenomedullin and arterial stiffness

Farzin Beygui, MD, PhD1-4; Philipp S. Wild, MD, MsC5,6; Tanja Zeller, PhD7; Marine Germain,

MsC1-3; Raphaele Castagné, PhD1-3; Karl J. Lackner, MD8; Thomas Münzel, MD5; Gilles

Montalescot, MD, PhD1-3,9; Gary F. Mitchell, MD10; Germaine C. Verwoert, MsC11; Kirill V.

Tarasov, MD, PhD12; David-Alexandre Trégouët, PhD1-3; François Cambien, MD, PhD1-3;

Stefan Blankenberg, MD7*; Laurence Tiret, PhD1-3*

1Sorbonne Universités, UPMC Univ Paris 06, UMR_S 1166, Team Genomics & Pathophysiology of Cardiovascular Diseases; 2INSERM, UMR_S 1166, Team Genomics & Pathophysiology of Cardiovascular

Diseases; 3ICAN Institute for Cardiometabolism and Nutrition, Paris; 4Department of Cardiology, Caen University Hospital, Caen, France; 5Department of Medicine II, 6Center for Thrombosis and Hemostasis,

8Institute for Clinical Chemistry and Laboratory Medicine, University Medical Center of the Johannes Gutenberg-University Mainz, Mainz; 7University Heart Center Hamburg, Department of General and

Interventional Cardiology, Hamburg, Germany; 9Institut de Cardiologie, Centre Hospitalier Universitaire Pitié--HP, Université Paris 6), Paris, France; 10Cardiovascular Engineering, Inc.,

Norwood, MA; 11Department of Epidemiology, Rotterdam, the Netherlands; 12Laboratory of Cardiovascular Science, National Institute on Aging, National Institutes of Health, Baltimore, MD

*contributed equally as co-last authors

Correspondence:

Laurence Tiret, PhD Stefan Blankenberg, MD

INSERM UMR_S_1166 Dept of General & Interventional Cardiology

91 boulevard de l'Hôpital University Heart Center Hamburg

75634 Paris cedex 13, France Martinistr. 52, 20246 Hamburg, Germany

Tel: +33 1 40 77 96 70 Tel: +49 40 7410 53972

Fax: +33 1 40 77 97 28 Fax: +49 40 7410 53622

E-mail: [email protected] E-mail: [email protected]

Journal Subject Codes: [142] Gene expression, [89] Genetics of cardiovascular disease, [146] Genomics

, ,

DDDD1-1-1-1-3333****

nne Universités, UPMC Univ Paris 06, UMR_S 1166, Team Genomics & Path hysiolc s; Cy te ne aa r

M aScience National Institute on Aging National Institutes of Health Baltimore MD

nne UUUUnininiveveversrrsititititésééé , UPUUU MC Univ Paris 06, UMR_SSSS 111 66, Team Genommmmiciii s & Pathophysiology culllarrrr Diseaseeeessss; 2INNSERM, ,, UMR____S 1166,,, Teamm GGenomim cs &&&& PPPathophppp ysiololoogygygygy of CaC rddiovas; 3333ICCCCAN Institutete ffor Caarardidididiomoometabababa oloo iiisi mm annndd Nuutrritiooonn, PPPParris; 4DeDepaartrttmeentntnt of ff CaCaCaCardrdrddioooololologygygy,, Cy HHHoooso pital, Caenn, Fraanccce; 55DDeD partmeeentt of MMediiciine IIIIIIII, 6666CeCeCentereer for Thhhromommombosisss aana ddd Hemommoste fofoorr r r ClClCllininininicicicalalalal Chehehemmimim ssts ryyyy andnddd LLababababoroo attttorororo y y y y MeMeMeM didd ccicc nnene, UnUnUnUniviviviverrrsisss tytyty MMMMeddediccccal CCCCenenenenteteeer r r r oofof ttttheeee Joooho aana nerg-University Maiaiaiainzn , Maininnnzzz;; 777University Heaaartrr Center HaHH mburg, DDDepepepartment of General aal Cardiology, Hamamamambububuurgrgrgr , GeGeGeGermrmmmananananyyy; 9999InInInnstststtititititutut dddeee CaCaCaardrdrdiioi lololologigigie,e,e,e, CCCCenenenntrtrtrtre ee HoHoHoH spspsppititititaalaa ier Universitair

-HP, ,, Université Paris 6),),), Paris, ,, France; ; 101010Cardiovascular EngggineeringggMMMA; 11Depapartrtrtrtmememmentntntnt ooof f f f EpEEpEpidididdememememiooioiolololologygggy,,, RoRoRoRottttttttereererdadadadam,m,mm, ttttheheheh NNNNetetetetheheherlrlrlrlanaa dsdsdsds;;; 12LaLaLaLabobooborarararatototoryryryry ooof ff Cardiovaa

ScScieiencncee NaNatitiononalal IInsnstititututete oonn AgAginingg NNatatioionanall InInststititututeses ooff HeHealalthth BaBaltltimimororee MMDD

by DAVID HARRISON on July 30, 2014http://circgenetics.ahajournals.org/Downloaded from



2

Abstract:

Background - Adrenomedullin (ADM) is a circulating vasoactive peptide involved in vascular

homeostasis and endothelial function. Single nucleotide polymorphisms (SNPs) of the ADM

gene are associated with blood pressure variability, and elevated levels of plasma midregional

proadrenomedullin (MR-proADM) are associated with cardiovascular diseases.

Methods and Results - We investigated the sources of variability of ADM gene expression and

plasma MR-proADM concentrations in the general population, and their relationship to markers

of atherosclerosis. MR-proADM levels were assessed in 4155 individuals who underwent

evaluation of carotid intima-media thickness (IMT) and arterial rigidity (reflection index [RI]

and stiffness index [SI]). In a sub-sample of 1372 individuals, ADM gene expression was

assessed as part of a transcriptomic study of circulating monocytes. Non-genetic factors

explained 45.8% and 7.5% of MR-proADM and ADM expression variability, respectively. ADM

expression correlated to plasma C-reactive protein, interleukin-receptor 1A and

myeloperoxidase, while MR-proADM levels correlated to C-terminal pro-endothelin-1,

creatinine and N-terminal pro-B-type natriuretic peptide. Genome-wide association study of

ADM expression and MR-proADM levels both identified a single locus encompassing the ADM

gene. ADM expression was associated to a single SNP rs11042717 (p=2.36 10-12) while MR-

proADM was associated to 2 SNPs with additive effects, rs2957692 (p=1.54 10-13) and

rs2957717 (p=4.24 10-8). RI was independently associated to rs11042717 (p<10-4) and ADM

expression (p=0.0002) but not to MR-proADM. Weaker associations were observed for SI. IMT

was not related to ADM SNPs or expression.

Conclusions - These results support an involvement of the ADM gene in the modulation of

peripheral vascular tone.

Key words: adrenomedullin, gene expression, genome-wide association scan, arterial stiffness

y ( [[[[

eeeenenenene eeeexpxpxpxprerereressssssssioioioionnnn wawawawassss

on-geeenenenenetitititicccc fafafafactctctctororororssss

4

x

a o

e A

M p g (p ) while MR

45.888%%% ananand d d 7.777 5% %%% of MR-proADM and ADDDDMMM expression variririabaaa ilityM , respectively.

ccorrrrelated to pplaaasmsma a a CCCC-r-r-reaeae ctcttctivivive protttteieiee n, iintterrleleleukukukkininin-rrececeppptttorrr 1AAAA andndndd

xiddddasasase, whilililile ee MRMM -prprroADMDMDMD leeeeveevelsss cooorrrelaateed toto CC-C ttermmminall ppprp oo-ee-endootothhehh lllinnn-1,,,,

and N-terminal pppproro-B-tytytyypepepe natriuruu eteticcc ppepepeptit dede. GeeGenon mee-wwwidide e e asaassososocic ation study o

ession and MR-prpp oAAADMDMM lllel velsll bbbob hthh iiidded ntifififfiiiei dd d a siiinggglllel llllocus encompapp ssing gg the A

M eexpxpreressssioioionn wawass asassosociciciiatatededed tttoo aa sisisingnglelelel SSSSNPNPNPNP rrs1s1s11010101042424227171717 77 7 (p(p(p( =2=2=22.3.3.3666 6 1010101M -12121212)))) whwhwhilililee MRMRMR




3

Adrenomedullin (ADM) is a potent vasodilator and angiogenic peptide involved in

cardiovascular and renal protection through autocrine, paracrine and endocrine pathways. It is

produced by a large variety of cells, especially endothelial and vascular smooth muscle cells, but

also inflammatory cells. Its expression is inducible by shear stress, oxidative stress and hypoxia.

The synthesis and secretion of ADM are stimulated by several pro-atherogenic and pro-

inflammatory factors such as angiotensin II, endothelin-1, aldosterone, interleukin- -

and tumor necrosis factor- and - .1,2

In humans, plasma concentrations of ADM are increased in various disease states

including acute myocardial infarction, hypertension, chronic renal failure and congestive heart

failure.1 Because the reliable quantification of ADM in plasma is hindered by several technical

difficulties, it has been proposed to use midregional proADM (MR-proADM), a stable peptide

present in an equimolar balance with ADM, as a surrogate marker in epidemiological studies.3

Increased MR-proADM plasma levels have been related to future cardiovascular events, 4-8 left

ventricular hypertrophy, 9,10 mortality in congestive heart failure 11 and vascular dysfunction. 12-

14

The human ADM gene is located on chromosome 11 and genetic polymorphisms in its

DNA sequence are associated to blood pressure levels and susceptibility to hypertension.15-17 The

contribution of the ADM locus to blood pressure variability in the general population is further

supported by the results of genome-wide association studies (GWASs).18,19 All these findings

concur to support a causal role of ADM in the etiology of cardiovascular diseases, although the

underlying molecular mechanisms are not yet fully elucidated.

Here we report the results of an integrative study conducted in a population-based cohort

combining information on ADM issued from several levels, including genome-wide genotype

rious disease stateseseses

ure andddd conge tstivvvveeee hhhhe

ecause the reliable quantification of ADM in plasma is hindered by several techn

, it has been proposed to use midregional proADM MR-proAD , a stable pept

a

M 8

h pertroph 9,10 mortalit in congesti e heart fail re 11 and asc lar d sf nctio

causususee ththththee rerereliabababable quantification of ADMMM iin plasma is hindddeeered by several techn

, iiti hhhas been prroooposseededed tooo o uusu e mmimm drddregioonnal pproAAADMDMDM (MRMMR-prorooADDDDMMM)M , a stabababble peept

an equiiuiuimomomom lllar babab lllalancncee wiw thhh AAAADDDDM, assas a aaa suurrrrogogatatateee mamamarkrkrkererer iiinin eepiipipidemimmimiolololloggicici lalal ssttutuddid

MR-prpp oADM ppplalll sma levelsll have bebb en r llellat ddedd to ffffuture ca drdddiiioi vascular events,,, 4-8

hh rt hph 99,1010 rt laliit iin stii hhe t ffailil 1111 dd ll dd ff iti




4

data, ADM gene expression in circulating monocytes and MR-proADM levels in plasma. The

aim of this study was to investigate the genetic and non-genetic sources of variability of ADM

gene expression and plasma MR-proADM levels and to assess their relationship to markers of

atherosclerosis.

Methods

Study Population

A detailed description of the study population is provided elsewhere.20 Study participants of both

sexes aged 35-74 years, were successively enrolled into the Gutenberg Health Study (GHS), a

community-based single-center cohort study in the Rhein-Main region in western mid-Germany.

Participants were of European origin. All subjects gave written informed consent. Ethical

approval was given by the local ethics committee and by the local and federal data safety

commissioners. Hypertension and diabetes were defined as previously.21

Biological Measurements

Plasma creatinine (Abbott), NT-pro-B-type natriuretic peptide (NT-proBNP) (Roche), C-reactive

protein (CRP) (Abbott), myeloperoxidase (MPO) (Prognostix), IL-18 (MBL), IL-1 Receptor

antagonist (IL-1Ra) (R&D Systems) and C-terminal pro-endothelin-1 (CT-proET-1) (BRAHMS)

levels were measured using commercially available assays according to manufacturer's

recommendations. Plasma MR-proADM levels were measured using the commercially available

immunoluminometric assay (B.R.A.H.M.S. AG, Henningsdorf/Berlin, Germany).3

Carotid Arteries Ultrasound

Ultrasound exploration of carotid arteries was performed using an 11 to 3MHz linear array

transducer of the i33 system (Philips, Bove, the Netherlands). 21 Intima-media thickness (IMT)

y p p

g Healalalalthththth SSSStutututudydydydy ((GHGHGHGHSS)S)S)

y-based single-center cohort study in the Rhein-Main region in western mid Germ

s

a

n

M

y-based dd sisis ngngnggle-cccenter cohort study in the RhRhein-Main region n n in western mid-Germ

s wweeere of Eurroppopeaann oroorigigigginininin. AlAlAlA l ll suuubbjecccts gavave wrwrititittteteten ininiinfofformrmr eeed ccooonseseentnn . EtEtEtEthihihicacacal

as ggggivivivi enenenen bbby yyy thththe lololol ccac llll ete hihihihicscscss ccccomomommimimim tttttttteeeeeee aaaandndndn bbbby y y y thththhe e ee lolooocacaalll anananandd d fefefefedededeerarararall l dadadadataaaa safafaafetetetety y y

ners. Hypertenssssioioioion nn anaana d ddd diddid abababbetetete esesess wwwerererereeee dedededefiffineneneed dd as pppprerereeviviviv ououuuslslsls y.y.y.212121

MMMeaeasusureremementntnttss




5

was measured as recommended 1 cm before carotid bulb, bilaterally at the far wall. The IMT

phenotype used in the present analysis was the average of right and left IMT measurements.

Evaluation of Vascular Tone and Arterial Stiffness

Reflection index (RI) reflecting vascular tone of small arteries, and stiffness index (SI) reflecting

large artery stiffness, were measured by Pulse Trace 2000 device (Micro Medical Ltd.,

Rochester, UK) as described elsewhere.22

Genotyping and Imputation

Genome-wide genotyping was performed using the Affymetrix Genome-Wide Human SNP Array

6.0 (http://www.affymetrix.com). Genotyping was performed separately in cohort GHS I

including 2990 individuals first enrolled in GHS, and in cohort GHS II including 1165

individuals subsequently enrolled. Calling of genotypes was performed separately in each

cohort.20 Genome-wide imputation of genotypes was performed in cohorts GHS I and GHS II

separately using MACH (v1.0.18.c) and minimac (release 2012-03-14) software with the 1000

Genomes release of February 2012 reference dataset (hg19/build37). Only SNPs with good

imputation quality (r2>0.3) were kept. Imputation-based GWAS was performed using mach2qtl

(V1.1.0, 2011-05-23).

ADM Gene Expression Measurement

Expression level of the ADM gene was assessed from circulating monocytes in 1372 individuals

of cohort GHS I as part of a genome-wide expression study using the Illumina HT-12 v3

Beadchip (http://www.Illumina.com). Detailed protocol for separation of monocytes, RNA

preparation and micro-array hybridization is reported elsewhere.20 ADM gene expression was

detected by a single probe located in the 3'UTR sequence of the gene (ILMN_1708934) and

which had a perfect quality score according to ReMOAT (http://remoat.sysbiol.cam.ac.uk/).

me-Wide Human SNPNPNPN

ely in c hhohhortttt GHGHGHGHS SSS IIII

e S

using MACH (v1.0.18.c) and minimac (release 2012-03-14) software with the 1

elease of Febr ar 2012 reference dataset (hg19/b ild37) Onl SNPs ith good

99000 iiindndnddivivivi idididduaalslslsls first enrolled in GHS, ananandd in cohort dd GHS IIIIIII including 1165

ssus bbbsequently ennnroolleleleledd. CaCCC llininining ofoof geneenotyypes wwasasasas pperfofformededed sepepepparattttelelelly ininn eaaachhh h

enommmeee-e wiwiwiw ddde iiimpmpmpututt tatatioioi n ofofof ggggenottttypypypesesss wasas ppererfofooformrmrmeddeded iiinnnn cocohhohortr s GHGHGHGHSSSS III aa dndnd GHGHGHG S

usinggg MACH (v(v11.11 000.0 18111 .c) )) and imiii iniimac (r((r( lellease 20101010 2-222 030303 1-1114)4)4)) sofffftware with the 1

lle fof FF beb 20201212 fef ddat et ((hhg1919/b/b iildld3737)) OO lnl SNSNPPs iithh dod




6

After pre-processing and quantile normalization, data were transformed using an arcsinh-

transformation.

Statistical Analysis

R software (V 2.13.0, The R Foundation for Statistical Computing) and SAS software (V 9.3,

SAS Institute Inc., Cary, North Carolina) were used for statistical analysis. Association of ADM

expression and plasma MR-proADM with cardiovascular risk factors (age, sex, body mass index

(BMI), type 2 diabetes, hypertension and current smoking) was assessed by linear regression

analysis. Association of ADM expression and plasma MR-proADM levels with circulating

biomarkers and vascular phenotypes was assessed by Pearson correlation coefficients. Partial

correlations were adjusted for age, sex, BMI and smoking. Independent predictors of ADM

expression and plasma MR-proADM levels were selected by stepwise regression analysis. All

biomarkers were log-transformed to remove positive skewness. In regression of vascular

phenotypes on SNPs, genotypes were introduced as categorical variables (2 df). Since the two

SNPs had a dominant effect on vascular phenotypes, their independent effect was tested by

jointly introducing them as binary variables in the regression model. SNPxSNP interaction was

tested by introducing a product term in the model (1 df).

The GWAS of plasma MR-proADM levels was performed separately in each cohort

using PLINK.23 After quality control filters, 629,283 SNPs were common to both cohorts. Age-

and sex-adjusted MR-proADM levels were analyzed assuming additive allele effects for each

SNP. The validity of this assumption was then formally tested for the SNPs found associated to

phenotype. The results of both cohorts were combined in a meta-GWAS using METAL24 with a

fixed-effect model relying on an inverse-variance weighting. To identify additional signals after

accounting for the effect of the hit SNP of the meta-GWAS, a conditional analysis adjusting for

vels with circulatiiiingngngng

ion coeffffffffiiici iieii ttntts. PPPPaarararti

s M

t

a dominant effect on asc lar phenot pes their independentr effect as tested b

s wwwererere e adadadadjujujuussstedededd for age, sex, BMI and smmomokik ng. Independenenent t predictors of ADM

aaandndnd plasma MRMRMR-prroAoAoADMDMDMD levevv lsss wwerrre sellectededed bbbyy yy ssteeepwwisee rrregggrrressioiooonnn anananaalyssiiisi .

wereee lololologggg-traansnsfffoformrm deded to rerereemmmmove ppposososositititiivivee kskskkewewwwnennenessssss. InInIn regegreresssioooonnnn ofofoff vavascscululllarar

on SNPs,,, gegg notytytypepp s were iiiintrodddud cedd d as categggoriciii lalll variablblblles (((2222 ddfdd )). Since the t

dd iinant fefffect ll hph ot hth iir ii dnd ddentr fefffect t tedd bb




7

the hit SNP was performed. Conditional analyses were repeated until no more significant

association was found. The GWAS of age- and sex-adjusted ADM expression levels was

performed with PLINK in the sub-sample of cohort GHS I with expression measurements.

The r2 coefficient of linkage disequilibrium (LD) between SNPs was estimated by

maximum-likelihood method. Regional association plots were obtained using SNAP.25 Power

was calculated using Quanto (http://hydra.usc.edu/gxe/ ). A p-value <0.05 was considered

significant, except for GWAS where the commonly used genome-wide significance level was

applied (p<5x10-8). The potential functionality of SNPs was investigated using the HaploReg

software (http://www.broadinstitute.org/mammals/haploreg/haploreg.php).

Results

Plasma MR-proADM levels were measured in 4155 individuals of whom 1372 also had a

measurement of monocyte ADM gene expression. Descriptive characteristics of the two samples

are shown in Table S1 and associations of MR-proADM levels and ADM gene expression with

risk factors are shown in Table S2 in the online-only Data Supplement.

Independent correlates of ADM gene expression were sex, BMI, age and smoking, which all

together explained 7.5% of the variability of expression. Independent correlates of plasma MR-

proADM were age, BMI, smoking, sex, hypertension and diabetes, which all together explained

45.8% of the variability of plasma levels. The correlation between monocyte ADM gene

expression and plasma MR-proADM levels was relatively modest (r=0.13 after adjustment for

risk factors).

GWAS of ADM Gene Expression

The GWAS of ADM gene expression was performed in 1372 individuals. At a genome-wide

significance level, 31 SNPs were associated to ADM gene expression (herein referred to as

ted using the HapplolololoReRRR

php).

R

e a

i p gene expression w

R-prrrroAoAoAoADMDMDMDM lllevevevellllssss weweweererr mmmmeaeaeaasusususureddd d inininn 444415151555 5 5 innnndidididiviviviidudududualalallss ofofoff wwwwhohohoom mmm 131313137272722 aaalsssoo o hahahahaddd d aa a

nt of monocyteeee ADADADADMMM gegegeg nenenee eeeexpxpxpprererer sssssssioioioion.n.n. DDDesesesscrccriptitititivevevee cccchahahararararactctctererere iiistststticicicicsss of the two sa

innn TaTablblee S1S1S11 anand dd aassssocociaiatittiononss ofoff MMMR-R-R prproAoAADMDMDM llevevelelss anandd d ADADADADMMMM ggenenee exexprpresessisionon wMMMM




8

eSNPs), all located on chromosome 11p15 in a 609 kb region overlapping the SBF2 and ADM

genes (Figure 1). The lead eSNP rs11042717 was located in an intron of SBF2, 23kb upstream

from the transcription start site of the ADM gene. The minor allele C (frequency [MAF] 0.45)

was associated with increased ADM gene expression in a co-dominant fashion (p = 2.36 10-12).

After conditioning on rs11042717, no other SNP was significantly associated with ADM

expression. Imputation analysis did not reveal any stronger hit than rs11042717. The rs11042717

effect was independent of cardiovascular risk factors and accounted for 3.2% of ADM expression

variability. We looked for potential interaction of genotype with age, sex, BMI or smoking but

none reached statistical significance. None of the ADM-associated SNPs was associated with

SBF2 expression (p>0.55) and SBF2 expression did not correlate to ADM expression (p=0.35).

The eSNP rs11042717 was associated with plasma MR-proADM levels (p=0.0002). The

minor allele C was associated with increased levels and explained 1% of variability. The effect

of rs11042717 on MR-proADM levels was weakened by adjustment for ADM expression but

remained significant (p=0.009).

GWAS of Plasma MR-proADM Levels

The GWAS of MR-proADM levels was performed in 4155 individuals. At a genome-wide

significance level, 34 SNPs were associated with plasma MR-proADM levels (herein referred to

as pSNPs), all located at the 11p15 locus already associated to ADM expression. The lead pSNP

rs2957692 was located in the 3' region of the ADM gene (Figure 2) and its minor allele G (MAF

0.39) was associated with increased MR-proADM levels in a co-dominant fashion (p = 1.54 10-

13). Conditional analysis on rs2957692 revealed 3 additional pSNPs associated with MR-

proADM levels, all at the 11p15 locus, the strongest signal being observed with the pSNP

rs2957717 (p = 4.24 10-8). The minor allele T of rs2957717 (MAF 0.30) was associated with

sex, BMI or smokiiiingngnn

Ps was associiii tttatedddd wwwwi

e

e )

e C was associated with increased levels and explained 1% of variability. The ef

7 b

ignificant (p 0 009)

essioioioion n (p(p(p(p>0>00>0.5.. 5)5)5)) and SBF2 expression did ddd nnot correlate to ADADADADM expression (p=0.M

e eeeSSNNP rs11042727271777 wwwassss aasa socicicic atttedd wwwitth ppplaasmama MMMRRR-ppprroADDMMM leleleevelssss ((((p=p=p=000.0 0000000020 )

e C wwwasasasas aaassocciaiai ttteted dd iwiwith iiincncnccrrrreasedededd llllevevevev llelss anand ddd exexexxplpplplaiaiaiineddd 1%1%1%1% oof fff vaarirririabababbilittty. TTThehehh eef

717 on MR-prpp oAAAADMDMM lllev lells was weakkkeneddd bbbby yy addjujjuj stment fffor ADADADDMMM expppression bM

ig inififi nt (( 00 00009)9)




9

decreased MR-proADM levels. After conditioning on both rs2957692 and rs2957717, no other

SNP was significantly associated with MR-proADM levels. The two pSNPs rs2957717 and

rs2957692, located 14kb apart from each other, were not in LD (r2=0.0006). They had additive

effects on MR-proADM levels and accounted together for 1.5% of the variability of MR-

proADM levels. There was no interaction of either SNP with age, sex, BMI or smoking.

The lead pSNP rs2957692 was in strong LD with the lead eSNP rs11042717 (r2=0.67). In

a regression analysis including rs11042717 and the two pSNPs rs2957692 and rs2957717, only

the two pSNPs remained significantly associated with plasma MR-proADM.

The imputation-based GWAS identified the same region as the original one. The

strongest hit was observed with an imputed SNP rs7925720 (MAF 0.37, p = 8.28 10-15). The

rs7925720 was in nearly complete association with the typed lead pSNP rs2957692 (r2=0.95) and

explained almost the same proportion of the variability of MR-proADM (1.05% vs 0.96%).

Because of the nearly complete association between the two pSNPs, it was impossible to

disentangle their respective roles.

Association of ADM Gene Expression and Plasma MR-proADM Levels with Vascular

Phenotypes

ADM gene expression negatively correlated to RI (r = -0.13 and -0.08 before and after

adjustment for risk factors) whereas it was not associated to SI and IMT (Table 1). Plasma MR-

proADM levels positively correlated to SI and IMT in raw analyses, but these correlations

disappeared after adjustment for risk factors (Table 1).

Association of Vascular Phenotypes with SNPs at the ADM Locus

In 4155 individuals, RI was significantly associated to the 3 ADM SNPs identified by GWAS of

plasma and expression levels (Table 2). When the 3 SNPs were introduced simultaneously in a

ADM.

e origigiiginallll one. TTTThehehehe

i 15 h

w 9

a

e their respecti e roles

it wawawas s obobobobseseservrr edededd with an imputed SNP rs7s7s79292925720 (MAF 0.3737377, p = 8.28 10-15). Th

wwwaw ssss in nearly cooomppllelel te aaassoccciiai tiiioon wwwiith ththe tyypepepepedd d llel adaad pSNNNP rsrsrs22292 5777766696 222 (((r( 2=0=0=00.9

almoststst tttheheheh sammee prprpropoporo tionnn ooooffff the vavavariririr bababbililililititity ofofof MMMMRRR-prrroAAoAoADMDMDMM (1.1 0500505%%%% vsvs 0000.9996666%%%%)

the nearly yy compppllel te associaiii tiiion bbbetween thhheh two pppSNSNSNSNPPs,,, iiti was iiiimpppossible to

hth iei iti lle




10

regression model, rs2957717 and rs11042717 showed independent effects on RI while

rs2957692 was no longer significant. The rs11042717/TT and rs2957717/TT genotypes were

both independently associated with reduced RI (p<10-4 and p=0.009, respectively; p=0.18 for

interaction between the 2 SNPs). Further adjustment on plasma MR-proADM and ADM gene

expression maintained the association of RI with rs11042717 while the effect of rs2957717 was

no longer significant. However, it should be noted that this last analysis had a reduced statistical

power because it was restricted to the 1372 individuals with ADM expression. In the final model

including simultaneously cardiovascular risk factors, plasma MR-proADM levels, ADM

expression and rs11042717/TT genotype, ADM expression was negatively associated to RI

independently of genotype whereas plasma MR-proADM levels was not associated to RI (Table

3).

The 3 SNPs were associated with SI in a similar fashion (Table 2). When including in the

same model MR-proADM levels, ADM expression and rs11042717/TT genotype, only the

genotype remained associated to SI (Table 4). The association, however, was weaker than with

RI. Both indexes reflect vascular rigidity and they were highly correlated (r=0.52, p<10-4 after

adjustment for risk factors). In multivariate regression, RI remained significantly associated to

rs11042717 after adjustment for SI, while the association of SI to SNP did not persist after

adjustment for RI. The IMT phenotype was not associated to any of the 3 SNPs (data not

shown).

For purpose of replication, we interrogated the data of a meta-GWAS of carotid-femoral

pulse wave velocity (CFPWV), a standard measure of arterial stiffness.26 None of the ADM

SNPs were associated to CFPWV (all p-values > 0.20). Conversely, we examined whether the

ADM levels, ADMMMM

vely associiii tttat ddedd to ooo RRRRI

n T

e

emained associated to SI (Table 4) The association ho e er as eaker than

ntly y y ofofofo gggenenenotototyppeee whereas plasma MR-proAoAoADM levels was nnnooto associated to RI (T

e 3 SNSNSNSNPsPsPsPs werree asassoso iiciciaateddd wwwwitititithhhh SIIII iiinnn a aaa iisimimiillalarr fafafashshshshioioion (T(T(TT bbabablelel 2)22 . WhWWhWhen iiincnclulul dididdingng

l MR-ppproADM lllel vellsl ,,, ADADADA MMM exprpp essiiion andddd rs110004242422717171777/T/T/TTT TT gegg notyyypepp ,,, onlyy the

iai dd iciat ded t SISI ((TTablbl 4)4) ThTh iciatiio hh kak thha




11

SNP rs7152623 found associated with CFPWV in the meta-GWAS26 was related to RI and SI

GHS, but neither association was significant (p=0.85 and 0.84, respectively).

Association of ADM Gene Expression and Plasma MR-proADM Levels with Circulating

Biomarkers

All biomarkers except IL-18 were significantly associated with ADM gene expression (Table 5).

The strongest correlations were observed with CRP, IL1-Ra and MPO, which were the only

biomarkers associated to ADM expression after adjustment for cardiovascular risk factors. In a

stepwise regression including cardiovascular risk factors (age, sex, BMI and smoking) and

circulating biomarkers, the independent predictors of ADM gene expression were CRP, sex, IL1-

Ra, MPO and MR-proADM levels which all together explained 15.4% of the variability of

expression.

Associations of circulating biomarkers with plasma MR-proADM levels were radically

different from those of ADM expression (Table 5). After adjustment for risk factors, the strongest

correlations were observed with CT-proET-1, creatinine and NT-proBNP, none being associated

to ADM expression. By contrast, plasma MR-proADM levels did not correlate to MPO. The

independent predictors of MR-proADM levels selected by stepwise regression were CT-proET-

1, BMI, age, NT-proBNP, IL1-Ra, creatinine, ADM expression, smoking and sex. These factors

accounted for 71.7% of the variability of plasma MR-proADM levels, the major part being

attributable to CT-proET-1.

Discussion

The present study is the first large-scale investigation of ADM integrating data from multiple

sources, i.e. DNA sequence variability, gene expression in monocytes and plasma protein levels.

This integrative analysis shed new light on the role of ADM in cardiovascular pathophysiology.

MI and smoking) anananandd dd

ession were CCCCRPRRR , sesesesex

n f

c

o

s ere obser ed ith CT proET 1 creatinine and NT proBNP none being assoc

nd MRMRRMR-pp-prororoADADADDM MMM levels which all togetheheheher explained 15.4% % % % of the variability of

ociatattiooioionsnsnsns of ciciircrculull tatatiiining g bibibiommomomarkeeersrsrs wwwwititithh hh plplplaasmamamaa MMMRRRR-prprprpr AAoAoADMDMDMDM llllevevevevelelells wewerere rr dadadicii

om those of ADDMMM exprpp ession (((TaTT blblble 5)5)5). AfAfAfA ter adjujjustment fffor risii kk kk factors,,, the str

bob ded iithh CTCT ETET 11 iti ini dnd NNTT BoBNPNP bbeiin




12

The GWAS of ADM expression and plasma MR-proADM revealed a strong effect of the ADM

locus on both phenotypes. No other locus was detected at a genome-wide level for either

phenotype, which does not exclude the possibility of more modest genetic contributors that

would not have passed the GWAS stringent level of significance. A recent GWAS in 3444

Europeans reported two loci associated to plasma MR-proADM, KLKB1 encoding prekallikrein

and F12 encoding factor XII, these two loci being even more strongly associated to MR-

proADM than the ADM locus itself.27 In our study, the lead SNP rs4253238 at the KLKB1 locus

had a nominal p-value of 0.065. The lead SNP rs2731672 at the F12 locus was not correctly

imputed in GHS but there was no significant SNP in the region (all p-values > 0.10). Yet, the

power for replicating these signals in the 4155 individuals of GHS was almost 100%.

Confirmation of these loci in further studies is warranted.

ADM expression was associated to a single cis eSNP rs11042717 located upstream from

the ADM gene sequence in an intron of the nearby SBF2 gene. The rs11042717 is in almost

complete association (r2=0.87) with the rs11042725, a SNP not genotyped in the present study

and whose functional role on ADM expression has been suggested by reporter gene assays.28 The

rs11042725 SNP is in the 5' region of the ADM gene and according to the HaploReg software, it

is located within a strong enhancer when tested in human umbilical vein endothelial cells, which

supports a role in expression regulation. However, in the imputation-based GWAS, the

rs11042725 did not yield a stronger signal than rs11042717. The rs11042717 is also in almost

complete association with the lead pSNP of plasma MR-proADM, rs2957692. The HaploReg

software did not predict any functional role for either SNP. Because of the strong LD between all

these SNPs, it is impossible to disentangle their effect. Experimental studies are required to

understand the underlying functional mechanisms.

ocus was not corrececececttltt y

values >>>> 0000.11110)0)0)0). YeYeYeYett,tt th

o

M

e t

ssociation (r2 0 87) ith the rs11042725 a SNP not genot ped in the present st

epllllicicicicatattatininining g ththththessseee signals in the 4155 indivivivivididuals of GHS waasss almost 100%.

onnnn oooof these locci in fuuuurtheeheerr r studududu ies is wwwaarraannteddd.

M exprprprreseseses isisionn wasas aassssoco iateteteedddd to a sssiniininglglglg e ciiis eeSNSNSNNPPPP rsrsrs111104404042727272 17171717 lococccatatated upspstrtrt eaeam m

ene sequqq ence in an iiiintron offff thheh nearbbby yy SBSBSBF2F2F2 gggene. TTThehh r 1s1110101010424242277177 7 is in almost

iia iti (( 22 00 887)7) iithh hth 1111040427272525 SNSNPP t t dd iin thhe nt t




13

The GWAS of MR-proADM identified a second pSNP rs2957717 which was not in LD

with the other SNPs and had independent additive effect on MR-proADM. Unlike the lead

pSNP, rs2957717 was not associated with monocyte ADM expression, emphasizing the existence

of a likely different regulation mechanism in monocytes and endothelial cells. Other possible

mechanisms of genetic control consistent with a location of the SNP in the 3'UTR may be those

involved in protein degradation, stability or clearance. According to HaploReg, the rs2957717

SNP is located in an intron of the nearby AMPD3 gene which codes for an erythrocyte-specific

AMP deaminase. Given the erythroid specificity of this isoform, the functional role of AMPD3 in

endothelial cells is unlikely.

The association of ADM SNPs with RI and SI suggests a causal involvement of ADM in

the modulation of vascular tone. Since ADM SNPs were associated to monocyte ADM expression

and plasma MR-proADM levels, a hypothesis would be that the SNP effect is mediated by one

or both of these proximal phenotypes. However, the association was not modified by adjustment

on plasma levels and even strengthened by adjustment on monocyte gene expression, indicating

that the hypothesis of mediation by these intermediate traits was not supported. A likely

explanation is that genetic variants exert their effect through modulation of ADM expression in

other sites, especially the vascular endothelium. It is also known that the colocalization of cis

eSNPs and disease-associated variants at the same locus are not sufficient to infer a causal

mechanism as there could be two distinct causal variants in LD with the SNP.29 On the other

hand, circulating MR-proADM was not associated to RI and its association with SI disappeared

after adjustment for cardiovascular risk factors. In the Framingham Heart Study, which is

comparable to GHS in terms of sample size and study population, plasma adrenomedullin was

not associated with several measures of arterial stiffness, including CFPWV.30 The only

unctional role of AMAMAMAMP

e D

t r

a

h t

le els and e en strengthened b adj stment on monoc te gene e pression indic

e assssococociaiaiaiatititit ononon oof f f ADAAA M SNPs with RI and SSSIII suggests a causasasal lll involvement of AD

tiiiionnnn of vasculararr tonenee. SiiSiinncn e ADADADMMM SNNNPPs wweere aasssssssococoo iatetted too mmmonononocytttte ADDDMMM eeexxprM

a MRRRR ppp-prorororoAADA MMM lllelevevellsls, a hyhyhypopopopothesssisisis wwwwououldldldld bbbbee thththhaaaat ttthhhehe SNSNSNSNP PPP ffefeffefff ctttt iiiisss medididi tatatededdd bbby

hese prpp oximal phhphhenotypypypes. However,,, thhheh associatiioi n was not moddidd fied by yy adjuj st

l lel dd st hth ded bb djdj tm t te isi iindidi




14

correlation was found with pulse pressure, and exclusively in men.30 This suggests that

circulating adrenomedullin, despite being predominantly of endothelial origin, does not exactly

reflect the in situ vascular synthesis. Functional experiments in endothelial cells are required to

understand the molecular mechanisms involved in the genetic control of vascular tonus.

At variance with the present study, the ADM locus was not found in a large-scale meta-

GWAS of CFPWV, a standard measure of arterial stiffness.26 On the other hand, we failed to

replicate the BCL11B locus identified in the meta-GWAS26 despite a 92% power in the GHS

study. These discrepancies are likely to be explained by differences in phenotypes, since

CFPWV measures aortic stiffness whereas RI is a measure of peripheral vasodilation and relative

wave reflection and SI has been shown to poorly correlate to pulse wave velocity.31

There are limitations of our study that should be considered. ADM expression was

assessed in monocytes and MR-proADM levels were assessed in plasma whereas circulating

ADM levels mainly originate from endothelial synthesis. The circulating levels of MR-proADM

may not reflect the autocrine and/or paracrine activities of ADM that may play a major role

within the vascular wall. ADM expression and MR-proADM levels were assessed by a single

measure and did not account for time-variability. SNPs included in GWAS were biased towards

high allele frequency and rare variants of interest that were not correctly imputed may have been

missed. The RI and SI scores are surrogate measures of arterial stiffness that may lack

sensitivity/specificity.

In conclusion, we showed that genetic variants at the ADM locus influence ADM gene

expression in monocytes and circulating levels of MR-proADM. These variants are also

associated with modulation of vascular tone, mainly of peripheral arteries. The molecular

mechanisms mediating this association need to be elucidated in further experimental work.

pphenotypep s, sinceeee

ral vasodidididilalll tititition andndndnd re

c 31

e

n

A

flect the a tocrine and/or paracrine acti ities of ADM that ma pla a major role

ctioon n n ananand ddd SISIISI hasasas been shown to poorly cooorrrrele ate to pulse wavavave velocity.31

ereee aaaare limitatioonnns oof fff ourrr r ssts uddddyy thhhaat shohhouldd bbe ccononnnsisisidererred. ADAADMMM eeexprrresesese siiononn wwwaas

monononocycycycytetetes ananddd MMMMRRR-prpproAAADMDMDMDM levvvellelels wewerere aassssessesesseeedddd inn pppplllalasmsmaa whwhwhwhererere eas ciciircrculull tatatinii

s mainlyy origig nate ffffrom endddod thhhh lleliaii l syyynthehhh iisiis. Thhhhe ciiiircullllatinii g gg lell vels of MR-prpp oA

flflfl t hth to iin d/d/ iri ctii iitiie fof AADMDM thhat lpl jaj lol




15

Further studies will be required to evaluate if these variants are related to the occurrence of

cardiovascular diseases and whether ADM should be considered as a target for therapy.

Funding Sources: The Gutenberg Health Study is funded through the government of Rheinland-Pfalz (http://www.rlp.de/) (“Stiftung Rheinland Pfalz für Innovation”, contract AZ 961–386261/733), the research programs “Wissen schafft Zukunft” and “Schwerpunkt Vaskuläre Prävention” of the Johannes Gutenberg-University of Mainz (http://www.uni-mainz.de/), and its contract with Boehringer Ingelheim (http://www.boehringer-ingelheim.de/) and PHILIPS Medical Systems (http://www.healthcare.philips.com/), including an unrestricted grant for the Gutenberg Health Study. Specifically, the research reported in this article was supported by the National Genome Network “NGFNplus” (http://www.ngfn.de/en/start.html) (contract 01GS0833 and 01GS0831) by the BMBF, and a joint funding grant from the BMBF, and the Agence Nationale de la Recherche, France (http://www.agence-nationale-recherche.fr/) (contract BMBF 01KU0908A and ANR 09 GENO 106 01). F.B. was supported by a grant from the Fondation pour la Recherche Médicale (contract ING20091218112).

Conflict of Interest Disclosures: Pr Blankenberg is member of the BRAHMS GmbH medical advisory board. Assays for measuring MR-proADM were provided by BRAHMS for free. Dr Mitchell is owner of Cardiovascular Engineering, Inc, a company that designs and manufactures devices that measure vascular stiffness. The company uses these devices in clinical trials that evaluate the effects of diseases and interventions on vascular stiffness. The remaining authors report no conflicts.

References:

1. Jougasaki M, Burnett JC, Jr. Adrenomedullin: potential in physiology and pathophysiology. Life Sci. 2000;66:855-872.

2. Tsuruda T, Burnett JC, Jr. Adrenomedullin: an autocrine/paracrine factor for cardiorenal protection. Circ Res. 2002;90:625-627.

3. Morgenthaler NG, Struck J, Alonso C, Bergmann A. Measurement of midregional proadrenomedullin in plasma with an immunoluminometric assay. Clin Chem. 2005;51:1823-1829.

4. Brouwers FP, de Boer RA, van der Harst P, Struck J, de Jong PE, de Zeeuw D, et al. Influence of age on the prognostic value of mid-regional pro-adrenomedullin in the general population. Heart. 2012;98:1348-1353.

5. Khan SQ, O'Brien RJ, Struck J, Quinn P, Morgenthaler N, Squire I, et al. Prognostic value of midregional pro-adrenomedullin in patients with acute myocardial infarction: the LAMP (Leicester Acute Myocardial Infarction Peptide) study. J Am Coll Cardiol. 2007;49:1525-1532.

BF, and the Agencececeerererchchche.e.e.frfrfr///) ) ) (c(c(cononontrtrtraca t t t BBBant frfff om tttthhhhe FFFFondadadadatttitio

f doowner of Cardiovascular Engineering, Inc, a company that designs and manufact measure vascular stiffness. The company uses these devices in clinical trials the oo

f IIInttterest Discscloloosuureeess:: PrPrPrP BlBlBllanankkekk nbnbbeeergg iss mmmembmbmbereer of f f f thhhe BRBRBRAHHHHMSMSMSMS GGGGmbmbmbmbH HHH mmmedoaaaardrdrdd. Assays fororr meaeaasurrrir nngn MR---prroADAADMM wwererere pppprorovidedded byy BRARARAAHMMMMSS S fofofor frrree. owwwwnenenen rrr r ofofofo CCCCarrrdididiovovovo aasa cucucuc laaaar rr r EnEnEnEngigigg neeeeerererininining,g,g,g IInncnn ,,, a a a cococoompmpmpm ananana y y y ththththatatat ddddesesese igigigignsnsnns aaaandndnd mmmmananannufuffacaat measure vascuuulalallar stiffnfnfnfnesee s. Thee compap nynyny uses ththththesese dedeviv cecees ss in clinical trials the effects of diseaeaeaeaseseseesss s anannnd ddd ininintetetetervrvrvenenenentititionononons sss ononono vvvasasasa cculaaaarrr r ststststififififfnfnnesesesess.s.s. TTThehehee rerereemaining authoooonfnfnflililictctctsss.




16

6. Schnabel RB, Schulz A, Messow CM, Lubos E, Wild PS, Zeller T, et al. Multiple marker approach to risk stratification in patients with stable coronary artery disease. Eur Heart J.2010;31:3024-3031.

7. Tzikas S, Keller T, Ojeda FM, Zeller T, Wild PS, Lubos E, et al. MR-proANP and MR-proADM for risk stratification of patients with acute chest pain. Heart. 2013; 99:388-395.

8. Wild PS, Schnabel RB, Lubos E, Zeller T, Sinning CR, Keller T, et al. Midregional proadrenomedullin for prediction of cardiovascular events in coronary artery disease: results from the AtheroGene study. Clin Chem. 2012;58:226-236.

9. Bhandari SS, Davies JE, Struck J, Ng LL. The midregional portion of proadrenomedullin is an independent predictor of left ventricular mass index in hypertension. Metabolism. 2010;59:7-13.

10. Coutinho T, Al-Omari M, Mosley TH, Jr., Kullo IJ. Biomarkers of left ventricular hypertrophy and remodeling in blacks. Hypertension. 2011;58:920-925.

11. Adlbrecht C, Hulsmann M, Strunk G, Berger R, Mortl D, Struck J, et al. Prognostic value of plasma midregional pro-adrenomedullin and C-terminal-pro-endothelin-1 in chronic heart failure outpatients. Eur J Heart Fail. 2009;11:361-366.

12. Coutinho T, Turner ST, Mosley TH, Kullo IJ. Biomarkers associated with pulse pressure in African-Americans and non-Hispanic whites. Am J Hypertens. 2012;25:145-151.

13. Kita T, Kitamura K, Hashida S, Morishita K, Eto T. Plasma adrenomedullin is closely correlated with pulse wave velocity in middle-aged and elderly patients. Hypertens Res.2003;26:887-893.

14. Schnabel RB, Wild PS, Schulz A, Zeller T, Sinning CR, Wilde S, Kunde J, et al. Multiple endothelial biomarkers and noninvasive vascular function in the general population: the Gutenberg Health Study. Hypertension. 2012;60:288-295.

15. Chen S, Lu X, Zhao Q, Wang L, Li H, Huang J. Association of adrenomedullin gene polymorphisms and blood pressure in a Chinese population. Hypertens Res. 2013 ;36:74-78.

16. Ishimitsu T, Tsukada K, Minami J, Ono H, Matsuoka H. Variations of human adrenomedullin gene and its relation to cardiovascular diseases. Hypertens Res. 2003;26 Suppl:S129-134.

17. Li Y, Staessen JA, Li LH, Gao PJ, Thijs L, Brand E, et al. Blood pressure and urinary sodium excretion in relation to the A-1984G adrenomedullin polymorphism in a Chinese population. Kidney Int. 2006;69:1153-1158.

f fff leleleleftftftft vvvvenenenentrtrtrtriciciciculululularararar 5.

c ld f.

h rm

ith p lse a e elocit in middle aged and elderl patients H t R

cht C,C,C, HHHHulululu smsmsmananannnn M, Strunk G, Berger R,RR MMMortl D, Struck JJJ,,, et al. Prognostic valdreeegigigigional prorororo-aaaadrddd ennnnomomoomedeee ululullilililin n n n anananand dd C-C-CC tetetermminnal-prprprproooo-e-e--endndndndototothehehelililin-nnn 1 inininin ccchrhhh onnnicicicic hhheaeaeaeart f. EuEEur J Heart FaFaFail. 222009999;11:363361-3366.

ho TTTT, TuTuTuTurnrnrnr er SSST,T,T, MMMososlell y yy THHTHTH, Kulllllllooo IJIJIJI . BiBiBBiomomarararkekekekersrsrs assssooocociaiaii tteted ddd wiwiwiwithtthth pulllsese ppreressssurumericans and noooon-H-H-H-Hisisisi papapaaninnic c c whwhwhw ititittesesese .. AmAmAmAm JJJ HHHypypypypeere teeeensnsnsn .. 2020202 12121212;2;2;25:5:5:5 14141445-5-5-5-151.

KKitamuraa KKKK, HaHaHaHashshshshidididida aa S,S,SS, MMMMororororisisishihihihitatata KKKK,, EtEtEtEto oo o T.TTT PPPlalalalasmssmsma aa a addadadrerererenononoomemmemedudududullllllininnin iiisss closely iithh lls lel iit iin ididdldl ded dnd ldld ll atiient HH t RR



https://www.researchgate.net/publication/26775958_The_midregional_portion_of_proadrenomedullin_is_an_independent_predictor_of_left_ventricular_mass_index_in_hypertension?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/26775958_The_midregional_portion_of_proadrenomedullin_is_an_independent_predictor_of_left_ventricular_mass_index_in_hypertension?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/7169839_Blood_pressure_and_urinary_sodium_excretion_in_relation_to_the_A-1984G_adrenomedullin_polymorphism_in_a_Chinese_population?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==



https://www.researchgate.net/publication/23974297_Prognostic_value_of_plasma_midregional_pro-adrenomedullin_and_C-terminal-pro-endothelin-1_in_chronic_heart_failure_outpatients?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==



https://www.researchgate.net/publication/230762693_Association_of_adrenomedullin_gene_polymorphisms_and_blood_pressure_in_a_Chinese_population?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/230762693_Association_of_adrenomedullin_gene_polymorphisms_and_blood_pressure_in_a_Chinese_population?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/51731047_Biomarkers_Associated_With_Pulse_Pressure_in_African-Americans_and_Non-Hispanic_Whites?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/51731047_Biomarkers_Associated_With_Pulse_Pressure_in_African-Americans_and_Non-Hispanic_Whites?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/51706128_Biomarkers_of_Left_Ventricular_Hypertrophy_and_Remodeling_in_Blacks?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/51706128_Biomarkers_of_Left_Ventricular_Hypertrophy_and_Remodeling_in_Blacks?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/8925553_Plasma_Adrenomedullin_Is_Closely_Correlated_with_Pulse_Wave_Velocity_in_Middle-Aged_and_Elderly_Patients?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==



https://www.researchgate.net/publication/10861378_Variations_of_Human_Adrenomedullin_Gene_and_Its_Relation_to_Cardiovascular_Diseases?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/10861378_Variations_of_Human_Adrenomedullin_Gene_and_Its_Relation_to_Cardiovascular_Diseases?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==


17

18. Ehret GB, Munroe PB, Rice KM, Bochud M, Johnson AD, Chasman DI, et al. Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature.2011;478:103-109.

19. Wain LV, Verwoert GC, O'Reilly PF, Shi G, Johnson T, Johnson AD, et al. Genome-wide association study identifies six new loci influencing pulse pressure and mean arterial pressure. Nat Genet. 2011;43:1005-1011.

20. Zeller T, Wild P, Szymczak S, Rotival M, Schillert A, Castagne R, et al. Genetics and beyond--the transcriptome of human monocytes and disease susceptibility. PLoS One.2010;5:e10693.

21. Neumann JT, Tzikas S, Funke-Kaiser A, Wilde S, Appelbaum S, Keller T, et al. Association of MR-proadrenomedullin with cardiovascular risk factors and subclinical cardiovascular disease. Atherosclerosis. 2013;228:451-459.

22. Schnabel RB, Schulz A, Wild PS, Sinning CR, Wilde S, Eleftheriadis M, et al. Noninvasive vascular function measurement in the community: cross-sectional relations and comparison of methods. Circ Cardiovasc Imaging. 2011;4:371-380.

23. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet.2007;81:559-575.

24. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics. 2010;26:2190-2191.

25. Johnson AD, Handsaker RE, Pulit SL, Nizzari MM, O'Donnell CJ, de Bakker PI. SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap. Bioinformatics.2008;24:2938-2939.

26. Mitchell GF, Verwoert GC, Tarasov KV, Isaacs A, Smith AV, Yasmin, et al. Common genetic variation in the 3'-BCL11B gene desert is associated with carotid-femoral pulse wave velocity and excess cardiovascular disease risk: the AortaGen Consortium. Circ Cardiovasc Genet. 2012;5:81-90.

27. Verweij N, Mahmud H, Mateo Leach I, de Boer RA, Brouwers FP, Yu H, et al. Genome-wide association study on plasma levels of midregional-proadrenomedullin and C-terminal-pro-endothelin-1. Hypertension. 2013;61:602-608.

28. Glubb DM, McHugh PC, Deng X, Joyce PR, Kennedy MA. Association of a functional polymorphism in the adrenomedullin gene (ADM) with response to paroxetine. Pharmacogenomics J. 2010;10:126-133.

adis MMMM, tettt allll. NNNoN nininininnnvnvations aaaandndndnd ccccomomomompapapapaririririsososon

C

at

C d

Circ CCCararardididid ovovovassc c c Imaging. 2011;4:371-3808080.

SS,S NNNeale B, Toodddd-BBBBrownwnwnw K, TThT omoomasss LL, FeFerreieieiraraaa MMMAA, BBenndeeer D,DDD et aaaallll. PLPLPLLINNNNKKK:K ale-e-gegegeg nome aaasssooociaation aaandnd pppopopuuulaatiooonn-baaseed lililinknkagge annalyysy eees. AAAmA JJJJ HHHHumumum Geneenet9-57777555.

CJ,,, Li Y,,, Abecasis GRGRGRG . MEMEMETAAAL:LL fffast and ddd ffefffififif cient meta-an llallysyy iiisi of gegg nomewidscans. Biioioioioinfnfnfnfororrormamamamatitititicscscs.. 202202010101010;2;2;26::6:6:212122190909090-2-22-2191919191.1.1.




18

29. Wallace C, Rotival M, Cooper JD, Rice CM, Yang JH, McNeill M, et al. Statistical colocalization of monocyte gene expression and genetic risk variants for type 1 diabetes. Hum Mol Genet. 2012;21:2815-2824.

30. Levy D, Hwang SJ, Kayalar A, Benjamin EJ, Vasan RS, Parise H, et al. Associations of plasma natriuretic peptide, adrenomedullin, and homocysteine levels with alterations in arterial stiffness: the Framingham Heart Study. Circulation. 2007;115:3079-3085.

31. Salvi P, Magnani E, Valbusa F, Agnoletti D, Alecu C, Joly L, et al. Comparative study of methodologies for pulse wave velocity estimation. J Hum Hypertens. 2008;22:669-677.



https://www.researchgate.net/publication/5328242_Comparative_study_of_methodologies_for_pulse_wave_velocity_estimation?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==

https://www.researchgate.net/publication/5328242_Comparative_study_of_methodologies_for_pulse_wave_velocity_estimation?el=1_x_8&enrichId=rgreq-0cee0e74-ee6a-47bb-bcae-82658b91f96f&enrichSource=Y292ZXJQYWdlOzI2NDE2MzI3NjtBUzoxMjQ4Njc3MTg0MjI1MzNAMTQwNjc4MjE5MjU2MQ==


19

Table 1. Correlation of ADM gene expression and plasma MR-proADM levels with vascular parameters (n=1372)

*Adjusted for age, sex, BMI and smoking. Data are Pearson correlation coefficients.

ADM gene expression Plasma MR-proADM

Raw correlation p-value Adjusted

correlation* p-value Raw correlation p-value Adjusted

correlation* p-value

Reflection index (RI) -0.134 <.0001 -0.082 0.003 0.023 0.412 -0.023 0.398

Stiffness index (SI) -0.040 0.135 -0.033 0.234 0.251 <.0001 -0.005 0.856

Intima-media thickness (IMT) -0.009 0.726 -0.045 0.093 0.376 <.0001 0.026 0.344

0.0.0.0.020202023333

040 0 135 0 033 0 234 0 251 <

0 <

. Data are Peaeaeaearsrsrsrsononon coccocorrrrrrrrelelelelatatatatioioioion nnn cocococoefefefe fififificicicieene tstststs.

040000 0....13111 5 -0.033 0.234 0.251 <

000009999 0.727 66 -00.04455 0.0.00 00933 0.3737373766 <




20

Table 2. Association of ADM SNPs with reflection index (RI) and stiffness index (SI) in the total sample (n=4155)

SNP RI (%) SI (m/s)

Genotype n Adjusted mean* SE p-value Adjusted

mean* SE p-value

rs11042717 CC 874 67.4 0.47 <10-4 9.42 0.09 0.0009CT 1998 67.6 0.31 9.42 0.06

TT 1201 65.5 0.40 9.08 0.08

rs2957717 CC 1945 66.8 0.32 0.006 9.27 0.06 0.004

TC 1726 67.5 0.34 9.45 0.06

TT 357 65.0 0.74 8.98 0.14

rs2957692 AA 1498 66.2 0.36 0.025 9.13 0.07 0.002

GA 1895 67.4 0.32 9.43 0.06

GG 645 67.5 0.55 9.45 0.10

*Adjusted for age, sex, BMI, and smoking

65 0 0 74

66666666.8. 0.0.0.0.32323232 0.0.0.0.000000006666

6767667.5555 0.000 34343434

6565 00 00 7474




21

Table 3. Results of linear regression analysis of reflection index (RI) on ADM genotype, plasma MR-proADM and ADM expression (n=1372)

Model rs11042717 (TT vs CT+CC) MR-proADM (log) ADM expression

Estimate* SE p-value Estimate* SE p-value Estimate* SE p-value

Genotype only -3.75 0.85 <10-4 - - - - - -

Genotype + MR-proADM -3.86 0.86 <10-4 -2.88 2.13 0.18 - - -

Genotype + ADM expression -4.33 0.86 <10-4 - - - -4.03 1.05 10-4

Genotype + MR-proADM + ADM expression

-4.39 0.86 <10-4 -2.05 2.13 0.34 -3.95 1.06 2x10-4

*Estimate of linear regression coefficients adjusted for age, sex, BMI and smoking

3 000.0 18181818

3

3

s adjusted for age sex BMI and smoking

33333 0.0.0.0.8666 <10-4 - - -

39999 0.0.0.0.868666 <11110000-4 -2-2-22.00005 555 2.2.2.2.13 0.0.0.0 34343434

s adadadjujujustststededed fffororor aaagegege sesesexxx BBBMIMIMI aaandndnd sssmomomokikikinnnggg




22

Table 4. Results of linear regression analysis of stiffness index (SI) on ADM genotype, plasma MR-proADM and ADM expression (n=1372)

*Estimate of linear regression coefficients adjusted for age, sex, BMI and smoking

Model rs11042717 (TT vs CT+CC) MR-proADM (log) ADM expression

Estimate* SE p-value Estimate* SE p-value Estimate* SE p-value

Genotype only -0.36 0.16 0.02 - - - - - -

Genotype + MR-proADM -0.38 0.16 0.02 -0.18 0.40 0.66 - - -

Genotype + ADM expression -0.41 0.16 0.01 - - - -0.32 0.20 0.12

Genotype + MR-proADM + ADM expression

-0.42 0.16 0.01 -0.11 0.41 0.78 -0.32 0.20 0.11

-

8 0 16 0 02 0 18 0 40 0 66

2

8 0.0.0.0 16666 0.02 ---0.000 18 0.40404040 0.66

1111 0.1666 0.0101010 - - -

2 0.16 0.0.0.0.01010101 -0-0-00.1..1. 111 0.0.0.0 4141414 0.78




23

Table 5. Correlation of ADM gene expression and plasma MR-proADM levels with circulating biomarkers (n=1372)

*Adjusted for age, sex, BMI and smoking. Data are Pearson correlation coefficients.

ADM gene expression Plasma MR-proADM

Raw correlation

p-value Adjusted correlation*

p-value Raw correlation

p-value Adjusted correlation*

p-value

Creatinine -0.103 0.0001 0.007 0.811 0.313 <.0001 0.405 <.0001

CRP 0.254 <.0001 0.205 <.0001 0.288 <.0001 0.128 <.0001

NT-proBNP 0.073 0.009 0.010 0.715 0.423 <.0001 0.325 <.0001

MPO 0.180 <.0001 0.160 <.0001 0.073 0.007 0.026 0.343

IL-18 0.012 0.651 0.041 0.146 0.156 <.0001 0.077 0.006

IL-1Ra 0.260 <.0001 0.188 <.0001 0.359 <.0001 0.242 <.0001

CT-proET-1 0.065 0.016 0.042 0.137 0.675 <.0001 0.622 <.0001

MR-proADM 0.176 <.0001 0.128 <.0001 - - - -

0.0.0.0.313131313333 <<<<

< 0001 0 205 < 0001 0 288 <

<

0

<000.0 141414146666 0.000 151515156666

<...000000010010 0.205 <.0001 0.288 <

0.009 0.0 010100 00 0.0.0.0.7171717 5 0.4242424233 <

<.0001 0.0.0.0.161616160000 <.<<< 0000000001010101 0.0.00 073 0

0.655551111 0.000 040404041111 <<<




24

Figure Legends:

Figure 1. Regional association plot of monocyte ADM gene expression with SNPs of the 11p15

locus encompassing the SBF2 and ADM genes. The left vertical axis shows the p-value (-log10)

of association, the right vertical axis shows the recombination rate in the region (cM/Mb). The

best eSNP is shown by a red diamond. Other eSNPs significantly associated to expression are

shown by squares whose color goes from deep red to light pink according to decreasing LD (r2)

with the best eSNP.

Figure 2. Regional association plot of plasma MR-proADM with SNPs of the 11p15 locus

encompassing the SBF2 and ADM genes (see legend of figure 1).

R

i

Regigigigionononalalalal aassssssocciaiaiaiation plot of plasma MR-p-p-prooADM with SNPNPNPPss of the 11p15 locus

inngn the SBF2 aannnd ADADDMM gggeg nesss (sseee leeegeendd oof fffigigigururururee 1))).







SUPPLEMENTAL MATERIAL

Table S1. Baseline characteristics of the study population

Characteristics Total sample Sub-sample

with ADM expression

n 4155 1372

Female sex, n (%) 2031 (48.9) 667 (48.6)

Age, yrs 55.7 ± 10.9 55.2 ± 10.9

Body mass index, kg.m-2 27.2 ± 4.8 26.9 ± 4.5

Current smoker, n (%) 789 (19.0%) 257 (18.8%)

Diabetes, n (%) 303 (7.3%) 101 (7.4%)

Hypertension, n (%) 2140 (51.5%) 681 (49.6%)

Biomarkers

Creatinine (mg/dL) 0.88 (0.79 – 0.98) 0.88 (0.79 – 0.97)

CRP, mg/L 1.60 (0.94 – 3.20) 1.70 (0.97 – 3.20)

NT-proBNP (pg/mL) 69.1 (34.9 – 132.6) 65.5 (34.4 – 128.4)

MPO (pmol/L) 298.3 (236.9 – 372.7) 310.2 (242.1 – 385.3)

IL-18 (pg/mL) 217.4 (168.2 – 282.7) 216.7 (169.5 – 284.7)

IL-1Ra (pg/mL) 320.0 (241.0 – 427.9) 305.0 (224.5 – 415.3)

CT-proET-1 (pmol/L) 59.1 (50.8 – 68.2) 60.0 (52.8 – 68.3)

MR-proADM, nmol/L 0.46 (0.40 – 0.55) 0.45 (0.39 – 0.53)

Vascular parameters

Reflection index, % 67.0 ± 15.0 66.9 ± 15.1

Stiffness index, m/s 9.32 ± 2.99 9.29 ± 3.06

Mean intima-media thickness, mm 0.65 ± 0.13 0.65 ± 0.13

ADM gene expression - 8.85 ± 0.38

Values are mean ± SD, median (25th – 75th interquartile range) or n (%).

Table S2. Association of monocyte ADM gene expression and plasma MR-proADM levels with cardiovascular risk factors (n = 1372)

ADM gene expression Plasma MR-proADM (log)

Risk factor Coefficient SE p-value R2 (%) Coefficient SE p-value R2 (%)

Age (yrs) 0.002 0.001 0.035 0.32 0.013 0.001 <.0001 33.08

Sex (male vs female) -0.166 0.020 <.0001 4.65 0.020 0.013 0.127 0.17

BMI (kg/m2) 0.009 0.002 <.0001 1.16 0.024 0.001 <.0001 18.93

Diabetes (yes vs no) 0.044 0.040 0.271 0.09 0.210 0.025 <.0001 4.91

Hypertension (yes vs no) 0.006 0.007 0.765 0.02 0.163 0.013 <.0001 10.75

Current smoking (yes vs no) 0.057 0.026 0.032 0.33 -0.026 0.017 0.134 0.16

Data are coefficients from univariate regression analysis of ADM expression and log-transformed MR-proADM levels, respectively,

on each single risk factor

Documents

Adrenomedullin and Arterial Stiffness Integrative Approach Combining Monocyte ADM Expression, Plasma MR-Pro-ADM, and Genome-Wide Association Study