University of Groningen

Genetics of human cardiovascular traits Verweij, Niek

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2015

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Verweij, N. (2015). Genetics of human cardiovascular traits. University of Groningen.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 29-09-2021 2.

G e n o m e - w i d e a s s o c i a t i o n study on plasma levels of midregional- proadrenomedullin a n d

C-terminal-pro- endothelin-1

21 ABSTRACT

Endothelin-1 (ET-1) and adrenomedul- tively). Epistasis analyses showed a sig-

- lin (ADM) are circulating vasoactive nificant interaction between the senti- un peptides involved in vascular homeo- nel SNP of F12 and KLKB1 for both stasis and endothelial function. Ele- traits. In addition, a variant near the , Hongjuan

1 vated levels of plasma ET-1 and ADM, ADM (rs2957692) was associated and their biologically stable surrogates, with MR-proADM (P=1.05E-12) and a C-terminal-pro-Endothelin-1 (CT-pro- variant in EDN-1 (rs5370) was associat- ET-1) and midregional proadreno- ed with CT-proET-1 (P=1.49E-27). The illjé, Pim van der Harst

. S medullin (MR-proADM), are predic- total phenotypic variation explained by

rank P. Brouwersrank P. tors of cardiac death and heart failure. the genetic variants was 7.2% for MR- . Gansevoort, Patricia B. M We studied the association of common proADM and 14.6% for CT-proET-1.

on T on genetic variation with MR-proADM KLKB1 encodes plasma kallikrein, a and CT-proET-1 by genome-wide proteolytic enzyme known to cleave . de F Boer, association analyses in 3,444 partic- high-molecular-weight kininogen to ipants of European ancestry. We per- bradykinin and prorenin to renin. We . Bakker, R udolf A L . formed follow-up genotyping of SNPs cloned the precursors of ADM and ET-1 that showed suggestive or significant and demonstrate that purified plasma iek H. van Gilst, Herman H. W each, R association in the discovery stage in kallikrein can cleave these recombinant tephan J an additional 3,230 participants. The into multiple smaller peptides. ateo L minor variant in KLKB1 (rs4253238) The discovery of genetic variants in truck, S and F12 (rs2731672), both part of the kallikrein-kinin system and in the eldhuisen, W rene M the kallikrein-kinin system, were as- encoding pre-proET-1 and pre- sociated with higher MR-proADM proADM provides novel insights into oachim S Adapted from Hypertension, Mar;71(3):702–8 2013 . van V (P=4.46E-52 and P=5.90E-24, respec- the (co-) regulation of these vasoactive ahmud, I irk J tively) and higher CT-proET-1 levels peptides in the vascular system. (P=1.23E-122 and P=1.26E-67, respec- sselbergs, J . A . Hillege, D erweij, Hasan M olkert W iek V u, F

N Y roe, Hans L

23 Introduction Methods

The endothelium plays a major role and midregional proadrenomedullin Study population the manufacturer’s instruction manu- in maintaining vascular homeostasis (MR-proADM) which are correlat- This study was performed in partic- als. The design of these assays is based by controlling blood fluidity, platelet ed with ET-1 and ADM in equimolar ipants of the Prevention of Renal and on immunoluminometric assays de- aggregation and (local) vascular tone. amounts 3,4. Increased plasma levels of Vascular End-Stage Disease (PRE- scribed previously 3,4. Assays were per- Control of vascular tone is mediated by CT-proET-1 and MR-proADM have VEND) study. Details of PREVEND formed in EDTA-plasma aliquots tak- the release of nitric oxide and various both been associated to worse vascular have been described elsewhere 10. This en at baseline. The samples were stored vasoactive peptides, including endothe- function, cardiac death and heart fail- study has been approved by the review at -80°C prior to analysis. All blood lin (ET-1) and adrenomedullin (ADM). ure 5-9. It is unknown whether increased board of the University Medical Center samples were processed by personnel ET-1 is a 21-amino acid peptide, secret- peptide levels are causal or consequence Groningen. All participants provided blinded from any patient data. A total ed by endothelial cells and is known to in disease and only limited knowledge informed consent. This study adheres of 3,444 samples were available for the be one of the most potent vasoconstric- exists on the factors regulating the re- to the principles expressed in the Dec- discovery analysis and 3,230 for repli- tors 1. Conversely, ADM, a 52-amino lease and subsequent bioactivation of laration of Helsinki. cation. acid peptide hormone, is secreted by a ET-1 and ADM. Genome wide associa- variety of different cells and is a potent tion (GWA) analyses have been proven Genotyping, quality control & impu- SNP/gene biology functional annota- vasodilator 2. In addition to their direct a powerful and unbiased tool to iden- tation tion effects on vascular tonus, both ADM tify novel mechanisms and pathways, Genotyping of 4,016 participants in For identifying the likely candidate and ET-1 are also involved in the ho- which might reveal novel targets for PREVEND was carried out using Illu- gene in each loci, we used the follow- meostasis of the sodium and water bal- therapy. Here we report the results of mina HumanCytoSNP-12 arrays. SNPs ing information: We considered the ance 1,2. GWA analyses in 3,444 participants were called using Illumina Genome nearest gene of the sentinel SNP, all reliable measurements of plas- of Prevend and follow-up genotyping Studio software and quality control nsSNPs in LD (r2 ≥0.8 in the HapMap ma ET-1 and ADM are highly chal- in 3,230 participants to identify com- was applied before and after imputa- phase II CEU or 1000 Genomes) with lenging due to their short half-life, mon genetic variants associated with tion (Online supplement). Replication the sentinel SNP, conservation among existence of binding proteins and oth- CT-proET-1 and MR-proADM levels, genotyping of 19 SNPs was performed species (GERP and 29 mammals) and er technical difficulties 3,4. Recently, two related vaso-active peptides that by KBiosciences (KBiosciences, Herts, biological function. We used GRAIL assays have been developed to measure have not been assessed by GWAS be- UK) utilizing the SNPline system in an analysis 11 to perform a text-based anal- the biologically stable surrogates, C-ter- fore. additional 3,230 independent partici- ysis in abstracts on PubMed prior to minal-pro-Endothelin-1 (CT-proET1) pants of the PREVEND study. Dec 2006 (to avoid confounding from GWAS results arising after that date). Biochemical measurements Two commercially available, fully au- Statistical analysis tomated sandwich immunoassays were We used linear regression on untrans- used for the measurement of CT-pro- formed MR-proADM measures using ET-1 and MR-proADM (BRAHMS an additive genetic model with age, sex CT-proET-1 KRYPTOR and MR- and body mass index (BMI) as covari- proADM KRYPTOR; BRAHMS AG, ates. Linear regression was performed Hennigsdorf, Germany) according to on untransformed CT-proET-1 mea-

2425 sures using an additive genetic model amino acids) was also PCR amplified enna, Austria) as previously described (50 mg/ml) for 0, 5, 30 and 60 minutes with age and sex as covariates. Gen- using a human cDNA clone (Source 13. D-Phe-Phe-Arg-chloromethylketone at 37°C with 5μg/mL activated purified otype-phenotype analyses were per- bioscience life sciences, Nottingham, (PPACKII, Calbiochem, Darmstadt, plasma kallikrein. The reactions were formed using PLINK (Version 1.07) 12. UK). For amplification, the following Germany) was used as kallikrein inhib- stopped by adding sample buffer and The most significant (P

(Source bioscience life sciences, Not- For purification of myc-proADM and control (λGC ) = 1.021 MR-proADM, evidence of association was calculated tingham, UK). For amplification, the myc-proET-1, anti-myc 9E10 agarose λGC = 1.039 CT-proET-1; Figure S1). using inverse-variance fixed-effect me- following primers were used: forward, resin was used (Santa Cruz Biotechnol- We observed significant associations at ta-analysis in METAL 15. The variants CGTGGATCCACCATGGACAC- ogy, Santa Cruz, USA). 3 loci for MR-proADM (P < 5 x 10-8) at the 4 loci with P < 5 x 10-8 in the dis- CGCTCGGTTGGATGTC and and at 3 loci for CT-proET-1. Two of covery phase were confirmed in further reverse, CGTGCGGCCGCCTA- In vitro assay of kallikrein the loci (SNPs in or near KLKB1 and genotyping (Table). The sentinel SNPs AAGAAAGTGGGGAGCACTTC. cleavage F12) were the same for both traits. at the four loci were included as covari- Pre-proET-1 cDNA without the signal Activated human plasma kallikrein was Each trait also had significant associ- ates in conditional analyses for each peptide (2,061 base pairs encoding 687 purified by Coachrom Diagnostica (Vi- ations at one additional trait-specific trait to determine whether there were

2627 fig.1a

55 klkb1 FXII 50 45 40 35 30 25 EDN-1

-log10(P) 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 16 18 20 22

fig.1b

25 klkb1 FXII 20

15 ADM

10 -log10(P) 5

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 16 18 20 22

Figure 1. Manhattan plot of CT-proET-1 (A), MR-proADM (B).

2829 fig.2

ProADM Plasma kallikrein Inhibitor Incubation (min)

Full length

17kDa

10kDa

proET-1 Plasma kallikrein Inhibitor Incubation (min) Full length

17kDa

10kDa

Figure 2. In-vitro cleavage of recombinant proADM and proET-1 by plasma kallikrein. Recombinant proADM (A), proET-1 (B) and Luciferase (C) were incubated with plasma kallikrein Luciferase for 0, 5, 30 and 60 minutes with and without the presence of PPACK II, a specific plasma kallikrein inhibitor. The full Plasma kallikrein length proteins are indicated. After 5 minutes of incubation Inhibitor the amount of full length proADM and proET-1 decreases Incubation (min) while several smaller products are formed. Luciferase (DNA supplied with the in vitro translation assay as positive control) Full length is not cleaved by plasma kallikrein suggesting that kallikrein is a relatively specific protease for proADM and proET-1.

3031 independent associations at these loci. associated with SNPs at a locus harbor- Novel role of KLKB1 on proADM and into a pcDNA3.1 expression vector These analyses identified no secondary ing 13 genes within 1 MB of the senti- proET-1 cleavage and generated [35S]-methionine labeled signals. The total variance explained by nel SNP (rs2957692), this SNP is clos- We tested the hypothesis that KLKB1 recombinant proteins by in-vitro tran- the genome wide significant variants est to the ADM encoding gene (ADM). is the causal gene related to the ob- scription/translation. We incubated was 6.7% for MR-proADM and 14.3% we examined all nsSNPs served association with MR-proADM recombinant proADM and proET-1 for CT-pro-ET-1 levels (Table 3). Ad- that are in LD (r2 ≥0.8) with one or and CT-proET1 trough cleavage be- proteins with active purified plasma ditional epistasis analyses were carried more of the sentinel SNPs in the Hap- cause 1.) KLKB1 encodes a cleavage kallikrein (human) and observed a out by testing pairwise combinations Map phase II or 1000 genome CEU protein with cleavage sites that overlap time-dependent cleavage of proADM between the 3 sentinel SNPs of CT-pro- datasets. We identified one nsSNP with cleavage sites of the precursors of and proET-1 to multiple smaller pep- ET-1 and between the 3 sentinel SNPs (rs3733402) in the KLKB1 gene in full MR-proADM and CT-proET-13,4,17,18, tides. This reaction was completely in- of MR-proADM. We observed a signif- LD (r2 = 1.00) with the sentinel SNP. 2.) previous GWA publications associ- hibited by a kallikrein inhibitor (Figure icant interaction between the sentinel We performed wet-lab genotyping of ated bradykinin 19 and renin levels 20 to 2). Cleavage kinetics was also depen- SNPs of the KLKB1 and F12 loci (Ta- the rs3733402 variant in 3,230 partic- the same variant, and 3.) the sentinel dent on the kallikrein concentration ble S2). The total phenotypic variation ipants and confirmed its strong associ- SNP is in full LD with a coding SNP (data not shown). In contrast, luciferase explained slightly increased to 7.2% for ation with both CT-proET-1 and MR- 21 affecting the proteolytic activity of was used as a positive control DNA for MR-proADM and 14.6% for CT-pro- proADM levels (Table). The nsSNP plasma kallikrein. For this purpose we the in-vitro translation assay was not ET-1 by including the interaction term rs3733402 is located in KLKB1 provid- designed an in-vitro assay. The ADM cleaved by plasma kallikrein (Figure 2). of the KLKB1 and F12 SNPs. ing a potential biological mechanism. cDNA and EDN-1 cDNA were cloned we also brought 12 loci for- The nsSNP (rs3733402) in KLKB1 by ward for replication that showed sug- itself explains approximately 4.7% of gestive evidence for association (P > variance of MR-proADM and 8.6% of 5 x 10-8 and P < 1 x 10-5) with MR- CT-pro-ET-1 plasma levels. rs2731672 Discussion proADM or CT-pro-ET-1. However lays 5.8kb from F12 and is in LD (r2 none of these loci became significantly = 1) with rs1801020 that is located in associated after genotyping in the addi- a highly conserved region within the tional 3,230 samples of the PREVEND 5’UTR part of F12, making it a poten- Using GWA analyses and function- ic variants was high; 7.2% for MR- cohort (Table S3). tial candidate. F12 is also a biological al in vitro follow-up we identified a proADM and 14.6% for CT-proET-1. plausible candidate when considering novel role for plasma kallikrein in the KLKB1 encodes plasma Identification of candidate genes and the KLKB1 locus as both genes are part regulation of both MR-proADM and prekallikrein and is activated to plasma putative causal genetic variants of the kallikrein-kinin system (KKS) CT-proET-1 (Figure 1). We identified kallikrein by factor XIIa. Factor XIIa The peptides CT-proET-1 and MR- and have interactions with each other two loci harboring components of the is the active form of factor XII and is proADM were both associated to SNPs on a molecular level16. Grail literature KKS; KLKB1 on and encoded by F12. Plasma kallikrein located in the locus that contained the mining tool based on publications pri- F12 on chromosome 5. In addition, and factor XII co-localize to endothe- gene encoding for the precursor of the or 2006 also suggested ADM, EDN-1, we observed a strong epistatic effect lial cells for their activation 16. Plasma peptides. CT-proET-1 was associat- KLKB1 and F12 as candidate genes between the genetic variants at these kallikrein plays a key role in the pre- ed with SNPs at a locus harboring 5 (P<0.01). loci. We also found an association with cursor maturation of other cardiovas- genes within 1 MB of the sentinel SNP a SNP near ADM with MR-proADM cular peptides such as bradykinin and (rs5370). rs5370 is a non-synonymous and a nsSNP in EDN-1 with CT-pro- renin by proteolysis 22. The KLKB1 lo- SNP (nsSNP) located in the endothe- ET-1. The total phenotypic variation cus has also been associated with circu- lin-1 (EDN-1) gene. MR-proADM was explained by all the significant genet- lating bradykinin 19 and renin levels 20

3233 fig.3a fig.3b

endothelin converting enzyme (ECE)

NH2- -COOH NH2- -COOH

Pre-proET-1 Pre-proADM

Figure 3a / 3b. Structural features of pre-proET-1 and pre-proADM. Numbers indicate amino acids.

3435 by GWA studies (r2=1.00 with our lead kallikrein cleaves proADM and pro- is independent requires further condi- in vitro functional follow-up we report SNP). We found the sentinel SNP at the ET-1 into smaller fragments (Figure 2). tional analyses. Our sentinel SNP at the involvement of the KKS system in KLKB1 locus to be in complete LD (r2 = The exact cleavage products remain to the EDN-1 locus (rs5370) is a nsSNP the regulation of MR-proADM and 1.00) with a nsSNP (rs3733402) caus- be identified and future studies are war- located in the CT-proET-1 part of the CT-proET-1 plasma levels. Preliminary ing an asparagine to serine amino acid ranted to clarify putative physiological endothelin-1 coding gene. We cannot functional data supports the hypoth- substitution (at position 124), located mechanisms between plasma kallikrein exclude this is a false positive associa- esis that the precursors are ligands for within the functional catalytic domain and these peptides. Luciferase was not tion as the nsSNP is located two amino plasma kallikrein. Future functional of plasma kallikrein. This amino acid cleaved by plasma kallikrein indicating acids downstream from the epitope of studies should further characterize the substitution is known to modify the plasma kallikrein substrate specificity. the CT-proET-1-tracer, this may affect potentially complex regulations of these proteolytic activity of kallikrein 21 sug- The KKS is tightly coordinated the antibody-epitope affinity 4. rs5370 peptides by the kallikrein-kinin system. gesting that the KLKB1 locus is associ- through multiple complex interactions has previously been associated to idio- We also found associations between ated to MR-proADM and CT-proET-1 with the renin-angiotensin-aldoste- pathic pulmonary arterial hyperten- MR-proADM and CT-proET-1 plasma due to differences in cleavage, similar to rone system (RAAS)22,23. Interestingly, sion29 and risk for hypertension 30. levels and the genes encoding their pre- bradykinin and renin. ADM and ET-1 are also suspected to in summary, using GWAS and cursors. Human pre-proET1, consist- interact with the RAAS 24. The KKS ing of 212 amino acids17 and is cleaved and RAAS are involved in a multitude into various peptides: pre-proET-17-53, of physiologic and pathophysiologic ac- Big-endothelin (pre-proET43-92), tions that impacts salt sensitivity, blood Perspectives pre-proET90-168 and CT-proET-1 flow and vascular reactivity, similar to (pre-proET-168-203). Pre-proADM the actions of ADM and ET-1. It is sug- consists of 185 amino acids, which can gested that the plasma KKS operates at be cleaved into four known peptides: the level of individual tissues and is de- Endothelin-1 (ET-1) and adreno- how MR-proADM and CT-proET-1 PAMP (proadrenomedullin N-termi- pendent on the local production of its medullin (ADM) are circulating va- levels are related to disease and whether nal 20 peptide; pre-proADM22–41), substrates, which is also in concordance soactive peptides involved in vascular they should be considered a target for MR-proADM (pre-proADM45-91), with the local production and action of homeostasis and endothelial function. therapy. In addition, the variants of the ADM (preproADM94-143) and ADT ADM and ET-1 1,2,22,25,26. In this study we identified 4 common KKS (KLKB1 and F12) could be used (adrenotensin; pre-proADM150–185). in addition to the KLKB1 and variants to influence plasma levels of in a mendelian randomization study to The cleavages sites involved in the bi- F12 loci, we identified genetic variants MR-proADM and CT-proET-1. The assess the potential of plasma kallikrein oprocessing of pre-proET1 and pre- at the ADM locus to be associated with common variants in ADM and EDN- as a novel therapeutic target, as suggest- proADM derived peptides include arg- MR-proADM levels and SNPs at the 1 should be evaluated further to assess ed recently 31. arg and lys-arg sequences (see Figure 3a EDN-1 locus with CT-proET-1 levels. / 3b for a schematic overview), which The variant (rs2957692) identified near are among the specific recognition the ADM locus is not in LD (r2 = 0.00) sites of plasma kallikrein18. On these with a previously reported variant in funding grounds, we hypothesized that plasma ADM (rs4910118) identified in an ear- kallikrein is involved in the bioprocess- lier candidate gene study 27. Recently, ing of pre-proET1 and pre-proADM the ADM locus was reported to be as- derived peptides. Using a custom de- sociated with systolic blood pressure, This work was supported by the fol- dation (Grant E033), the National Insti- signed ‘in vitro’ assay we provide pre- our sentinel SNP is low LD (r2<0.2) lowing grants: PREVEND genetics is tutes of Health (grant LM010098), The liminary data that purified plasma with this Locus28. Whether this locus supported by the Dutch Kidney Foun- Netherlands organisation for health re-

3637 search and development (NWO VENI supported by the Medical Research Literature grant 916.761.70), and the Dutch Inter Council of Great Britain and the Brit- University Cardiology Institute Neth- ish Heart Foundation. This work forms erlands (ICIN). N. Verweij is supported part of the research themes contribut- by the Netherlands Heart Foundation ing to the translational research port- (grant NHS2010B280). S.J.L. Bakker folio for Barts and the London Car- is supported by the Netherlands Heart diovascular Biomedical Research Unit, Foundation, Dutch Diabetes Research which is supported and funded by the Foundation and Dutch Kidney Foun- National Institute for Health Research. dation, together participating in the The funders had no role in study de- framework of the Center for Transla- sign, data collection and analysis, de- tional Molecular Medicine (CTMM) cision to publish, or preparation of the (www.ctmm.nl) project PREDICCt manuscript. (grant 01C-104-07). PB Munroe was

Disclosures

J Struck is employed by BRAHMS GmbH, a company manufacturing and holding patent rights on the MR- proADM and CT-proET-1 assays. The authors have declared that no compet- ing interests exist.

3839 1 kohan, D. E., Rossi, N. F., Inscho, E. 8 maisel, A. et al. Midregion prohor- METAL: fast and efficient meta-analysis of ge- iology - Regulatory, Integrative and Comparative W. & Pollock, D. M. Regulation of blood pres- mone adrenomedullin and prognosis in patients nomewide association scans. Bioinformatics 26, Physiology 285, R1-R13 (2003). sure and salt homeostasis by endothelin. Physio- presenting with acute dyspnea: results from the 2190-2191 (2010). logical reviews 91, 1-77 (2011). BACH (Biomarkers in Acute Heart Failure) trial. 23 shen, B. & El-Dahr, S. S. Cross-talk Journal of the American College of Cardiology 58, 16 shariat-Madar, Z., Mahdi, F. & of the renin-angiotensin and kallikrein-ki- 2 samson, W. K. Adrenomedullin and 1057-1067 (2011). Schmaier, A. H. Assembly and activation of the nin systems. Biological chemistry 387, 145-150 the control of fluid and electrolyte homeostasis. plasma kallikrein/kinin system: a new interpre- (2006). Annual review of physiology 61, 363-389 (1999). 9 schnabel, R. B. et al. Multiple En- tation. International immunopharmacology 2, dothelial Biomarkers and Noninvasive Vascular 1841-1849 (2002). 24 vogt, S. et al. Endothelin-1 and Ad- 3 morgenthaler, N. G., Struck, J., Alon- Function in the General Population: The Guten- renomedullin Plasma Levels After Exposure to so, C. & Bergmann, A. Measurement of Mid- berg Health Study. Hypertension 60, 288-295 17 inoue, A. et al. The human preproen- Fludrocortisone, Dexamethasone, and Spirono- regional Proadrenomedullin in Plasma with an (2012). dothelin-1 gene. Complete nucleotide sequence lactone. Clin Exp Hypertens Early Online, 1-6 Immunoluminometric Assay. Clin Chem 51, and regulation of expression. The Journal of bio- (2012). 1823-1829 (2005). 10 Hillege, H. L. et al. Urinary albumin logical chemistry 264, 14954-14959 (1989). excretion predicts cardiovascular and noncardio- 25 yayama, K., Kunimatsu, N., Ter- 4 Papassotiriou, J., Morgenthaler, N. vascular mortality in general population. Circu- 18 metters, K. M., Rossier, J., Paquin, J., anishi, Y., Takano, M. & Okamoto, H. Tissue G., Struck, J., Alonso, C. & Bergmann, A. Im- lation 106, 1777-1782 (2002). Chretien, M. & Seidah, N. G. Selective cleav- kallikrein is synthesized and secreted by human munoluminometric Assay for Measurement of age of proenkephalin-derived peptides (less than vascular endothelial cells. Biochim Biophys Acta the C-Terminal Endothelin-1 Precursor Frag- 11 raychaudhuri, S. et al. Identifying 23,300 daltons) by plasma kallikrein. The Journal 1593, 231-238 (2003). ment in Human Plasma. Clin Chem 52, 1144- Relationships among Genomic Disease Re- of biological chemistry 263, 12543-12553 (1988). 1151 (2006). gions: Predicting Genes at Pathogenic SNP As- 26 lipke, D. W. & Olson, K. R. A specif- sociations and Rare Deletions. PLoS Genet 5, 19 suhre, K. et al. Human metabolic ic inhibitor of mammalian kallikrein, Phe-Phe- 5 dhillon, O. S. et al. Prognostic value e1000534 (2009). individuality in biomedical and pharmaceutical Arg-chloromethyl ketone, inhibits the produc- of mid-regional pro-adrenomedullin levels taken research. Nature 477, 54-60 (2011). tion of vasoactive substances from trout plasma on admission and discharge in non-ST-elevation 12 Purcell, S. et al. PLINK: a tool set for by kallikrein and blocks endogenous kallikrein- myocardial infarction: the LAMP (Leicester whole-genome association and population-based 20 shah, S. et al. Four genetic loci influ- like activity in trout gills. Fish Physiology and Bio- Acute Myocardial Infarction Peptide) II study. linkage analyses. American journal of human ge- encing electrocardiographic indices of left ven- chemistry 10, 339-346 (1992). Journal of the American College of Cardiology 56, netics 81, 559-575 (2007). tricular hypertrophy. Circulation. Cardiovascular 125-133 (2010). genetics 4, 626-635 (2011). 27 cheung, B. M. et al. Plasma adreno- 13 Gallimore, M. J., Fareid, E. & Stor- medullin level is related to a single nucleotide 6 jankowska, E. A. et al. Identification morken, H. The purification of a human plasma 21 katsuda, I., Maruyama, F., Ezaki, polymorphism in the adrenomedullin gene. Eu- of Chronic Heart Failure Patients with a High kallikrein with weak plasminogen activator ac- K., Sawamura, T. & Ichihara, Y. A new type of ropean journal of endocrinology / European Feder- 12-Month Mortality Risk Using Biomarkers In- tivity. Thrombosis Research 12, 409-420 (1978). plasma prekallikrein deficiency associated with ation of Endocrine Societies 165, 571-577 (2011). cluding Plasma C-Terminal Pro-Endothelin-1. homozygosity for Gly104Arg and Asn124Ser in PLoS ONE 6, e14506 (2011). 14 Hayashi, J., Salomon, D. R. & Hugli, apple domain 2 of the heavy-chain region. Euro- 28 ehret, G. B. et al. Genetic variants T. E. Elevated kallikrein activity in plasma from pean Journal of Haematology 79, 59-68 (2007). in novel pathways influence blood pressure and 7 von Haehling, S. et al. Mid-regional stable liver transplant recipients. International cardiovascular disease risk. Nature 478, 103-109 pro-adrenomedullin as a novel predictor of mor- immunopharmacology 2, 1667-1680 (2002). 22 schmaier, A. H. The kallikrein-kinin (2011). tality in patients with chronic heart failure. Euro- and the renin-angiotensin systems have a mul- pean journal of heart failure 12, 484-491 (2010). 15 willer, C. J., Li, Y. & Abecasis, G. R. tilayered interaction. American Journal of Phys- 29 vadapalli, S., Rani, H. S., Sastry, B.

4041 & Nallari, P. Endothelin-1 and endothelial ni- tric oxide polymorphisms in idiopathic pulmo- nary arterial hypertension. International journal of molecular epidemiology and genetics 1, 208-213 (2010).

30 rankinen, T. et al. Effect of endothe- lin 1 genotype on blood pressure is dependent on physical activity or fitness levels.Hypertension 50, 1120-1125 (2007).

31 liu, J. et al. Hyperglycemia-induced cerebral hematoma expansion is mediated by plasma kallikrein. Nature medicine 17, 206-210 (2011). les b T a

4243 44

Table: Summary of the loci associated with MR-proADM and CT-proET-1 levels, discovery, replication and combined. Beta values estimate the difference in concentrations CT( -proET-1 in pmol/l, MR-proADM in nmol/l) per copy of the coded allele, adjusted for the covariates in the model.

Discovery, N = 3,230 Replication, N = 3,444 Combined, N = 6,674

Locus Trait CHR SNP wA1/A2 Beta (se) P Beta(se) P FRQ P-value Gene

1 ADM 4 rs4253238 C/T 0.027(0.003) 6.93E-24 0.034(0.003) 2.76E-34 0.46 4.46E-52 KLKB1

ET 4 rs4253238 C/T 4.811(0.305) 4.07E-54 5.476(0.312) 5.49E-66 0.46 1.23E-122 KLKB1

ADM 4 rs3733402* G/A NA NA 0.035(0.003) 1.88E-35 0.46 2.71E-36 KLKB1

ET 4 rs3733402* G/A NA NA 5.501(0.314) 1.37E-65 0.46 1.41E-68 KLKB1

2 ADM 5 rs2731672 T/C 0.024(0.003) 7.07E-14 0.020(0.003) 6.19E-10 0.24 5.90E-24 F12

ET 5 rs2731672 T/C 5.041(0.375) 3.85E-40 4.178(0.375) 2.92E-28 0.24 1.26E-67 F12

3 ET 6 rs5370* T/G 2.928(0.379) 1.38E-14 2.983(0.390) 2.49E-14 0.22 1.49E-27 EDN1

4 ADM 11 rs2957692 G/A 0.017(0.003) 2.46E-08 0.013(0.003) 9.87E-06 0.40 1.05E-12 ADM

*non sysnonymous SNP

45