A genetic polymorphism in connexin 37 as a prognostic marker for atherosclerotic plaque development

Journal of Internal Medicine 1999; 246: 211±218

A genetic polymorphism in connexin 37 as a prognostic marker for atherosclerotic plaque development 1

1

1

1

2,3

3

M. BOERMA , L. FORSBERG , L. VAN ZEIJL , R. MORGENSTERN , U. DE FAIRE , C. LEMNE , D. ERLINGE , T. THULIN , Y. HONG & I. A. COTGREAVE 4

1

4

5

1

2

From the Division of Biochemical Toxicology; Division of Cardiovascular Epidemiology, Institute of Environmental Medicine, Karolinska Institute; Division of Cardiovascular Medicine, Department of Medicine, Karolinska Hospital, Stockholm; Department of Medicine, Lund University Hospital, Lund, Sweden; Division of Biostatistics, Washington University School of Medicine, St. Louis MO, USA 3

4

5

Abstract. Boerma M, Forsberg L, van Zeijl L, Morgenstern R, de Faire U, Lemne C, Erlinge D, Thulin T, Hong Y, Cotgreave IA (Karolinska Institute and Karolinska Hospital, Stockholm; Lund University, Lund, Sweden. Washington University School of Medicine, MO, USA). A genetic polymorphism in connexin 37 as a prognostic marker for atherosclerotic plaque development. J Intern Med 1999; 246: 211±218. Background and objectives. Atherosclerosis is a multifactorial disease, in part characterized by chronic inflammatory changes in the vessel wall and loss of normal physical and biochemical interactions between endothelial cells and smooth muscle cells. Previous studies [Hu J., Cotgreave IA. J Clin Invest; 99: 1±5] have provided molecular links between inflammation and myoendothelial communication via gap junctions, suggesting that these structures may be important in the development of the atherosclerotic vessel phenotype. In order to strengthen this premise, the aim of the present work was to probe for structural polymorphisms in connexin 37, a gap junctional protein uniquely expressed in endothelial cells, and to assess for

Introduction Atherosclerosis is one of the most common chronic diseases in the western world. The medical consequences of the disease are manifold and may lead to myocardial or cerebral infarctions, due to obstructions of blood vessels [1]. In addition, the socio-economic implications of atherosclerotic dis# 1999 Blackwell Science Ltd

potential genotypic segregation in individuals displaying atherosclerotic plaque. Methods and results. Computer-based comparisons of Expressed Sequence Tags (ESTs) predicted a polymorphism in the human gap junctional protein connexin 37 (cx37). The C1019-T mutation results in a proline to serine shift at codon 319 (cx37*1± cx37*2). A Restriction Fragment Length Polymorphism (RFLP) assay, involving the insertion of a novel Drd I cleavage site in the proline variant revealed a statistically significant over-representation of the cx37*1 allele in association with atherosclerotic plaque-bearing individuals (Oddsratio for the homozygote = 2.38, X2 = 7.693, P = 0.006), in comparison to individuals lacking plaque, irrespective of a history of hypertension. Conclusions. These data suggest that the C1019-T polymorphism in cx37 may provide `single gene marker', which could be useful in assessing atherosclerotic plaque development, particularly in cardiovascular risk groups such as those with borderline hypertension. Keywords: atherosclerosis, connexin 37, plaque, polymorphism, prognostic indicator.

eases are considerable. Taken together, these factors are fostering great efforts both in the development of treatments for the ongoing disease and, importantly, the identification of predisposing genetic and lifestyle factors, useful for the identification and prophylactic treatment of high-risk populations. Atherosclerosis is largely a disease of the human arterial vessel, involving both inflammatory cells 211

212

M . B O E R M A et al.

and the development of abnormal cellular phenotypes within the affected vessel. The proliferation of smooth muscle cells (SMCs), particularly during the formation of `fibro-fatty lesions' and their extensive secretion of fibrous material to form tertiary `fibrous plaques' is a major component in the development of atherosclerotic plaque [1]. Evidence is accumulating to suggest that the endothelial cells (ECs) and SMCs of the human arterial wall interact with each other intimately, both biochemically and physically, to control a variety of important physiological functions. It has also been suggested that ECs and SMCs control each others' proliferative behaviour in the vessel wall, as well as influencing the development of their differentiated cell phenotype and function [2]. Thus, SMC are relatively refractory to growth stimulatory signals in vivo and it seems that under normal conditions, EC exert a clear inhibitory influence on SMC proliferation [3]. In contrast, inappropriate interactions between EC and SMC are thought to contribute to a number of human pathophysiological processes, such as atherosclerosis, which is primarily characterized by the development of abnormal SMC phenotypes and increased proliferation [4]. It has been proposed that EC and SMC interact primarily by humoral mechanisms, through the synthesis and release of mediators as diverse as NO, prostacyclin, endothelin and TGF-b [5]. On the other hand, EC±SMC interactions may also be mediated by contact-dependent mechanisms such as those operating by gap junctions. Gap junctional intercellular communication (GJIC) is mediated through gap junctions; trans-membrane channels that connect neighbouring cells and allow passive diffusion of small molecules such as ions, amino-acids, but also second messengers such as cAMP, Ca2+ and inositol triphosphates [6, 7]. One of the functions of GJIC is to provide a contribution to tissue homeostasis. Inhibition of GJIC can cause proliferation of initiated cells as a result of a decreased growth-control of normal neighbouring cells [8±10]. Gap junctions occur in clusters, or plaques, in the cell membrane; one gap junction consists of two connexons, one from each neighbouring cell. In turn, a connexon consists of six connexin proteins. Connexins are members of a multigene family and the different connexins are named after the molecular weight that is predicted by their respective cDNAs [11]. Endothelial cells and SMCs are capable of GJIC in

vitro [12] and both human ECs and SMCs express cx40 and cx43, whilst ECs also uniquely express cx37 [13, 14]. As cx40 and cx43 cannot form functional gap junctions [15], myoendothelial gap junctions may be homotypic, comprising of either cx40/cx40 or cx43/cx43; or heterotypic comprising of either cx37/cx40 or cx37/cx43. The possible existence of heterotypic gap junctions between SMCs and ECs has not presently been studied. Little is known about the function of cx37, or the regulatory mechanisms controlling this protein. However, it was recently shown that the bacterial endotoxin Lipopolysaccharide (LPS) and the proinflammatory cytokines Tumour Necrosis Factor-a (TNF-a) and Interleukin-1b (IL-1b) all rapidly and selectively down-regulate myoendothelial GJIC, without affecting homologous communication between the respective cell types. The selectivity in this effect was proposed to be due to the unique presence of cx37 in myoendothelial gap junctions. In view of the refractory role of ECs on SMC proliferation, and the established relationship of chronic inflammation to the development of atherosclerosis, it was thus proposed that the inhibition of myoendothelial GJIC could function as a molecular link between inflammation and the enhanced proliferation of SMCs in the vessel wall [16]. Mutations in connexin proteins can have major effects on the rate of GJIC [17]. Polymorphisms in cx37 have been found in different human populations, but most lay outside the open reading frame of the protein [18]. Recently, Richard and co-workers detailed a variant form of cx37 with a C to T shift at codon 1019. This causes a shift from proline to serine at amino acid 303 in the original published protein sequence [19]. This mutation was simultaneously revealed and confirmed in the present work, using a strategy based on prediction in silico using sequence comparisons to human expressed sequence tags (ESTs). In view of the potential importance of cx37 in myoendothelial GJIC, and its role in linking inflammation to enhanced SMC proliferation, we have studied the allelic distribution in well defined human populations and could demonstrate that one allelic form of the cx37 gene (base-pair 1019 C of the coding region, cx37*1) is significantly overrepresented in human populations exhibiting well defined thickening of the carotid intima, irrespective of a history of hypertension. These results are discussed in terms of the development of the first

# 1999 Blackwell Science Ltd Journal of Internal Medicine 246: 211±218

TH E C ONNEXIN 3 7 (C1019-T)

213

single gene prognostic tool for the identification of individuals at high risk to develop atherosclerotic plaques.

tubes. The samples were frozen at ±70 8C until the extraction of DNA. All blood samples were coded and analysed by RFLP (below) in a blind manner.

Subjects, materials and methods

Laboratory procedures

Clinical subjects

Chemicals and molecular biological materials

Clinical samples were obtained from several Swedish male study populations. The first study group was a case-control group of 80 borderline hypertensives and 80 pair-matched normotensive men (35± 55 years) from the same district of northern Stockholm. The second population consisted of 2 cohorts of established hypertensives (approximately 120 in each) from the Stockholm and Lund areas of Sweden, taking part in a multicentred Atherosclerosis study (The European Lacedepine Atherosclerosis Study, ELSA). In all cases, full clinical histories were available. Informed, written consent was obtained from each of the participating patients, and the different studies had been approved by the local Ethical Committee of the Karolinska Institute.

Ethidium bromide Taq polymerase and Tween 20 were obtained from Sigma Chemicals (St. Louis, USA). Two types of PCR-reaction buffer were used: 10 3 PCR-buffer containing MgCl2 (Sigma Chemicals, USA) and 10 3 Reaction buffer IV, lacking MgCl2 (AB, Epsom, UK). The deoxynucleotide triphosphate (dNTP)-mix (2.5 mM of each NTP) was obtained from Stratagene (La Jolla, USA). Primers (40 pmol mL21) were synthesized by and obtained from Kebo Lab, Stockholm, Sweden. The restriction enzyme Drd I (10 000 U mL21), and the 10 3 reaction buffer suitable for the use of this enzyme, were obtained from New England Biolabs Inc., USA. DNA Marker VI (Boerhinger Mannheim, Germany) was used as the molecular weight marker for agarose gel electrophoresis. The Terminator Ready Reaction Mix from Perkin Elmer was used for the dideoxy sequencing reactions. b-Agarase and the suitable 50 3 reaction buffer were obtained from Biolabs, New England, USA. The following buffers were routinely utilized: Phosphate Buffered Saline (PBS) (8.00 g NaCl, 0.20 g KCl, 1.44 g Na2HPO4.2H2O, 0.20 g KH2PO4, made up to 1 L in distilled water and pH adjusted to 7.4); Tris-EDTA (TE) (10 mM Tris-HCl and 1 mM EDTA and pH adjusted to 8.0); Tris-Borate-EDTA (TBE) (from a 10 3 stock solution: 0.9 M Tris-HCl, 0.9 M Boric Acid and 40 mL 0.5 M EDTA, pH 8.0). Ethidium bromide was used at 1 mg mL21 in dH2O, freshly diluted from a 10-mg mL21 stock solution. PBS-Tween consisted of Tween 20 and PBS (1 : 1000).

Ultrasound assessment of carotic atherosclerosis For assessment of carotid atherosclerosis, ultrasound measurements of the carotid arteries were performed with duplex scanners. Borderline hypertensives and normotensive controls were examined with an Acuson, 128 XP/5 machine, (Mountain View, CA, USA) and the established hypertensives in the ELSA trial with a biosound 2000 II s.a. (Biosound Inc., Indianapolis, IN, USA). All measurements were performed by trained sonographers who were unaware of the subjects' blood pressure levels and blood pressure status. Determinations of intima media-thickness were made according to established routines [20, 21]. Atherosclerotic plaque was defined as a localized intima media thickening with a thickness of .1 mm (borderline hypertensives and normotensive controls), and with a thickness of .1.2 mm in established hypertensives. The far wall of the common carotid artery was used throughout for the measurement of intima media thickness. Collection and treatment of human blood samples All blood samples were obtained from patients by venipuncture and collected into heparinzed blood

Expressed Sequence Tag (EST) data-base analysis At present most genetic polymorphisms are detected by performing sequencing reactions on genomic DNA. However, many expressed human gene sequences are already available in data-bases as Expressed Sequence Tags (ESTs). Using the Basic Linear Alignment Search Tool (BLAST) and the published cDNA of the protein of interest as a probe,

# 1999 Blackwell Science Ltd Journal of Internal Medicine 246: 211±218

214

M . B O E R M A et al.

ESTs similar to this cDNA can be identified and internal comparisons can be performed. Mismatches occurring in different ESTs could be the result of a polymorphism, providing that, in other ESTs, no mismatch at this particular position occurs. In the present work the originally published cx37 cDNA [11] was used to probe for mismatches in the open reading frame (ORF) of cx37. DNA purification DNA was purified from thawed whole blood (5 mL) using the Qiagen Blood DNA Kit (Quiagen Corp, CA, USA), essentially according to the manufacturers protocol (after purification the concentration of DNA was defined by the ratios of optical densities 260 nm: 280 nm. The DNA solutions were then diluted to 0.1 mg mL21 in TE). Restriction fragment length polymorphism (RFLP)assay In the present investigation primers were constructed that would amplify a 275 base-pairs part of the cx37-gene, coding for a part of the C-terminus of cx37. By mismatch of one base, the upper primer (59-CTGGACCCACCCCCTCAGAATGGCCAAAGA-39) inserted a unique restriction site for Drd I in the PCR-product of the published genotype, that will be referred to as cx37*1 (59-AGGAAGCCGTAGTGCCTGGTGG-39 lower primer). The main products of cx37*1 would be 240 and 35 base-pairs long and the main product of the other genotype, cx37*2, 275 base-pairs long. The restriction site was chosen to be present in cx37*1, in order to be able to see if the restriction site was actually built into the

PCR-product, in the event that the cx37*2 genotype was not present in the population (Fig. 1). Hot start PCR was performed by adding 5 mL of a 0.1-mg mL21 gDNA-solution to the following master-mix: 36 mL sterile dH2O, 5 mL 10 3 PCR-buffer, 3 mL dNTP and 0.5 mL of each primer, or: 33 mL sterile dH2O, 5 mL 10 3 PCR-buffer (AB), 3 mL MgCl2 (AB), 3 mL dNTP and 0.5 mL of each primer. After 1 min of heating the samples at 94 8C in the PCR-block, 0.5 mL of taq-polymerase was added to each sample and the following amplification program was run: 30 times (94 8C for 1 min, 60 8C for 1 min, 72 8C for 1 min and 30 s, 72 8C for 5 min). The PCR-products were stored at 4 8C until analysis. In order to cleave the PCR-products, 15 mL were then mixed with 0.5 mL Drd I and 1.6 mL 10 3 reaction buffer and incubated at 37 8C for about 3 h. After the incubation with the restriction-enzyme, products were separated by electrophoresis on a 2% agarose-gel, supplemented with 0.5 mg mL21 ethidium bromide. The DNA-bands were visualized using UV illumination at 302 nm. DNA sequencing A section of the DNA flanking and including the region of the 1019 base in the cx37-gene was amplified by PCR as follows. An upper-primer was constructed, which recognized a region of the template 45 bases up-stream of the upper-primer used for the RFLP-assays. Using this new upper primer, and the lower-primer used for the RFLPassays, we performed a new PCR-reaction on four cDNA-samples selected from two individuals each homozygote for the respective cx37 alleles, as described above. The new PCR-product, 319 basepairs long, was purified on a 2% low temperature

Fig. 1 The basis of the RFLP assay of the C to T conversion at base 1019 of the cx37 gDNA involving the insertion of a novel Drd I cleavage site by mismatch priming. A represents Part of the gDNA around the site of polymorphism and the upper PCR primer. Note that the primer contains the mismatch at *. B shows part of the PCR product derived from cx37*1 gDNA with the accompanying Drd I restriction site. C denotes part of the PCR product derived from cx37*2 gDNA and the product lacks a Drd I cleavage site. D denotes the consensus Drd I restriction sequence. # 1999 Blackwell Science Ltd Journal of Internal Medicine 246: 211±218

TH E C ONNEXIN 3 7 (C1019-T) melting agarose gel, supplemented with 0.5 mg mL21 ethidium bromide. After electrophoresis, the gel was visualized on a UV-board and the 319 base-pair products were excised from the gel. Fifty 3 Agarase-buffer was added to the pieces of gel in a ratio of 1 : 5 (v/w). After incubation at 70 8C for 15 min, the tubes were adjusted to 42 8C, 5 mL of b-Agarase was added and the mixture incubated at 42 8C for 5 h. The DNA was then precipitated and washed once with 70% ice-cold ethanol. The derived pellets were air-dried and dissolved in 10 mL sterile, distilled water. Four microlitres of Terminator Ready Reaction Mix were mixed with 6 mL of purified PCRproduct and 3.2 pmol sequencing primer was added. The following PCR program was applied: 26 times (96 8C for 30 s, 50 8C for 15 s, 60 8C for 4 min) and PCR-products were stored at 4 8C until analysis. The PCR-products were then precipitated, washed with 70% ethanol, air-dried and di-deoxy sequencing was subsequently performed by the Sequencing Service of the Department of Cell and Molecular Biology of the Medical Nobel Institute, Stockholm.

Statistical analysis Variations in the frequency distributions (chisquared analysis and the analysis of odds ratios) of the cx37*1 and cx37*2 alleles with respect to various clinical parameters of blood pressure and the existence of plaque were performed using version 6 of the EPI INFO statistical package as described previously [22].

215

Results A mutation in the regulatory domain of cx37 is predicted by EST-based comparisons Using an in silico-based prediction strategy for mutation from the published cx37 cDNA sequence, a single base-pair mismatch (C to T at base-pair 1019, codon 319) was detected in two independent human ESTs (entry numbers: AA090855 and AA249172), whilst the published sequence was detected in all other ESTs. This revealed a possible structural polymorphism at codon 319 of the protein involving a proline to serine shift within the carboxy-terminal terminus of the human cx37protein. A PCR-based RFLP assay detects the cx37 mutation in the Swedish population. Using an appropriate RFLP-assay (Fig. 1), the presence of this polymorphism was confirmed in various populations of normotensive and hypertensive Swedish men. The gels obtained by the RFLPassay of the cx37-gene showed three different PCR products: One each with either 275 or 240 base-pair products and one with both products in equal densities (Fig. 2). Thus, the polymorphism at basepair 1019, predicted by comparing ESTs, exists in the Swedish population with all expected allelic combinations detected. The 240 base-pairs product is predicted to be obtained from a cx37-gene, carrying a C at base-pair 1019, which was originally published as the sequence of the human cx37-gene [13]. We will refer to this allele as cx37*1. The 275 base-pairs product was predicted

Fig. 2 A typical gel obtained by an RFLPassay of cx37*1 and cx37*2 in 19 of the tested samples of human genomic DNA.The occurence of cx37*1 and cx37*2 in the DNA samples was determined by RFLP, as determined in the materials and methods section. The upper band is the larger 275 base-pair fragment. Lane A: molecular weight marker. Lane 11 is an example of a homozygote cx37*1, lane 12 of a homozygote cx37*2 and lane 14 of a heterozygote. # 1999 Blackwell Science Ltd Journal of Internal Medicine 246: 211±218

216

M . B O E R M A et al.

from a cx37-gene carrying a T at base-pair 1019. We will refer to this allele as cx37*2. When the two bands are visible in equal densities, the tested individual is a heterozygote with respect to this polymorphism. The fidelity of the RFLP assay in detecting the base alteration at codon 1019 was confirmed by sequencing of a region of the cx37 gene approximately 100 base pairs up and down stream from the Drd I cleavage site in four randomly chosen gDNAs, two each homozygous for the cx37*1 or the cx37*2 genotypes. The resulting sequence data confirmed the observations of the RFLP-assay. However, it will be noted that at base-pairs 966 and 967 in codon 301, all four sequences differed from the published cx37 sequence (G and C, as opposed to C and G, respectively) [13], suggesting an alanine to glycine shift in the protein in the individuals tested. A cx37 allele is significantly over-represented in individuals exhibiting atherosclerotic plaques in the common carotid artery

cx37*1 homozygote. The Chi-squared value for linear trend is 6.154, P = 0.014. Similarly, comparison of all individuals with plaque against all individuals without plaque yielded a 13% overrepresentation of the cx37*1 homozygotes and a 5% under-representation in the cx37*2 homozygotes. Again, assuming an OR of 1.00 for the cx37*2 homoygous genotype, the corresponding ORs for the heterozygote and the cx37*1 homozygote are 1.49 and 2.38, respectively (chi-squared = 7.693, P = 0.006). When comparing the corresponding allele distributions for the cx37 gene in the populations under study, the data in Table 1 reveal overrepresentation of the cx37*1 allele in the plaquebearing hypertensives compared to the normotensives without plaque (0.69 vs. 0.60), which was further polarized when comparing all individuals with and without plaque (0.69 vs. 0.58).

Discussion

When the RFLP assay was applied to a case-control study group of normotensive and borderline hypertensives, as well as to two cohorts of established hypertensives, Table 1 shows the occurrence of the cx37*1 and cx37*2 homozygotes and the cx37*1/ cx37*2 heterozygotes. In comparing all hypertensive patients with plaque (n = 267) with all normotensive individuals without plaque (n = 66) there was a 12% over-representation of the homozygote cx37*1 and a 5% under-representation of cx37*2. Assuming the odds ratio (OR) for the cx37*2 homozygous genotype to be 1.00, then the OR for the heterozygote is 1.87 and 2.96 for the

The use of EST-based prediction of mutations has been a possibility for a number of genes for some time. However, to our knowledge, the prediction of the cx37 polymorphism at codon 1019 is one of the first instances where such an in silico prediction results in the detection of a novel polymorphism. It will be noted however, that during the course of this work, this polymorphism was reported by Richard and co-workers, who isolated the mutation by conventional sequencing of a human satellite DNA co-segregating with a human dermatological disorder in a mixed population of North American subjects [19]. Thus, our results confirm the presence of this allelic variation in the Swedish male

Table 1 Frequencies of cx37*1 and cx37*2 genotypes in normotensive and hypertensive Swedish men studied in relation to the presence or absence of carotid intimal thickening (plaque)

All samples Hypertensives with plaques Normotensives without plaques All individuals with plaques All individuals without plaques

Number of samples

Homozygote cx37*1

Homozygote cx37*2

Homozygote cx37*1/*2

396 267 66 275 121

171 (43%) 126 (47%)b 23 (35%) 130 (47%)d 41 (34%)

49 26 10 28 21

176 (44%) 115 (43%)a 33 (50%) 117 (43%)c 59 (49%)

(13%) (10%) (15%) (10%) (17%)

The originally published sequence is termed cx37*1, the sequence differing at base 1019 is termed cx37*2. Genotypes are given as number (percentage) in the tested population. Assuming that the odds ratio (OR) for homozygous cx37*2 is 1.00, a indicates an odds ratio of 1.87 and b an OR of 2.96, chi-square 6.154, P = 0.014. Similarly c indicates an OR of 1.49 and d an OR of 2.38, chi-square 7.693, P = 0.006. # 1999 Blackwell Science Ltd Journal of Internal Medicine 246: 211±218

TH E C ONNEXIN 3 7 (C1019-T) population. It is of interest to note that the absolute allelic frequencies (cx37*1: 0.72; cx37*2: 0.28) in this North American population were close to the frequencies observed in the plaque-bearing Swedish individuals (cx37*1: 069; cx37*2 : 0.31). The reasons underlying the differences in frequency distributions of the cx37 alleles between these two populations are obscure, but may be related to undefined interethnic variations in the American population. The polymorphism in the cx37 gene results in a rather nonconservative amino-acid alteration in the cx37-protein, with the cx37*1 allele yielding a proline-residue at codon 319 and the cx37*2 allele a serine-residue in its place. It may be speculated that this amino-acid alteration has effects on the tertiary structure of the protein and therefore on the functionality of the gap-junctions that these proteins partake in. Proline-residue are well known to cause bends or `kinks' in nascent polypeptides, which would be lost by replacement by a serine-residue. Moreover, the serine-residue of cx37*2 at codon 319 is present in the carboxy-terminal tail of the cx37protein, an area already rich in serines, which are known to be involved in reversible phosphorylation reactions in the case of other connexin proteins. This might result in altered regulation of the protein. Thus, it is interesting to note that a mutation in cx43, causing a serine to proline shift at codon 364, has been found to be associated with viscero-atrial heterotaxia. Protein kinase A increases GJIC between cells expressing wild type cx43 but has no effect on this mutant. Protein kinase C, however, increases GJIC in the mutants, but has no effect on the wild type [17]. It remains to be determined whether the proline to serine shift in the present cx37 polymorphism results in alterations to the function and/or regulation of the protein, particularly with respect to myoendothelial GJIC. Here, it may be a useful strategy to study the coupling behaviour of different populations of ECs, exhibiting defined cx37*1 or cx37*2 genotypes. From the data presented in Table 1 it is clear that the cx37*1 and cx37* homozygous genotypes are, respectively, over- and under-represented in the plaque bearing individuals, in comparison with nonplaque bearers. In extension from the calculations of the appropriate linear trends, and assuming that the homozygous cx37*1 genotype is proatherosclerotic and that it is under recessive

217

inheretance, the OR for the homozygous cx37*1 genotype in association with plaque in the hypertensive group is 1.84 (95% Cl: 0.96±3.55), with a chi-square of 3.86, P = 0.05. This reveals a significant correlation of the cx37*1 homozygous genotype with the occurrence of atherosclerotic plaque in hypertensive patients. Further, if the entire population is analysed, and a recessive inheritance assumed, then the OR for the cx37*1 homozygous in association with the development of plaque is 1.75 (95% Cl: 1.10±2.80), with a chisquare of 6.14, P = 0.013. This reveals an even stronger association with the exsistence of atherosclerotic plaque. These data indicate a strong and relatively unprecedented single gene allele correlation to the development of a multifactorial disease state as complex as atherosclerosis. The clinical ramifications of the above observations are clear. The use of the RFLP-based analysis of the cx37*1 and cx37*2 alleles may provide, for the first time, a simple and reliable single gene polymorphism marker for the assessment of a susceptibility to the development of atherosclerotic plaques. This may provide considerable clinical and socio-economic benefits in the early prediction of individuals at high risk for the development of atherosclerosis. Here, one particular group of interest are young, borderline hypertensives, exhibiting early stages of vascular pathology.

Acknowledgements The authors would like to acknowledge financial support from the Swedish Medical Research Council (IC: 07140 and UDF: 09533) and the Swedish Heart-Lung Foundation (UDF).

References

# 1999 Blackwell Science Ltd Journal of Internal Medicine 246: 211±218

1 In: Gallin JI, Goldstein IM, Snyderman R, eds. Atherogenesis Inflammation: Basic Principles and Clinical Correlates, 2nd edn. 1992: New York: Raven Press, 1051±9. 2 Powell RJ, Cronenwett JL, Fillinger MF, Wagner RJ, Sampson LN. Endothelial cell modulation of smooth muscle cell morphology and organizational growth pattern. Ann Vasc Surg 1996; 10: 4±10. 3 Griendling KK, Alexander RW. Cellular biology of blood vessels. In: Hurst's the Heart. Schlant RC, Alexander RW, eds. New York: McGraw-Hill, 1993: 31±45. 4 Sosa-Melgarejo JA, Berry CL. Myoendothelial contacts in arteriolosclerosis. J Pathol 1992; 167: 235±9. 5 Griendling KK, Alexander RW. Endothelial control of the

218

6 7 8 9 10 11

12 13

14

15

M . B O E R M A et al.

cardiovascular system: recent advances. FASEB J 1996; 10: 283±92. Budunova IV. Alteration of gap junctional intercellular communication during carcinogenesis. Cancer J 1994; 7: 228±37. Goodenough DA, Goliger JA, Paul DL. Connexins, connexons, and intercellular communication. Ann Rev Biochem 1996; 65: 475±502. Yamasaki H. Gap junctional intercellular communicationn and carcinogenesis. Carcinogenesis 1990; 11: 1051±8. Yotti LP, Chang CC, Trosko JE. Elimenation of metabolic cooperation in chinese hamster cells by tumour promoter. Science 1979; 206: 1089±91. Murray AW, Fitzgerald DJ. Tumour promoters inhibit metabolic cooperation in co-cultures of epidermal and 3T3 cells. Biochem Biophys Res Comms 1979; 91: 395±401. SaÂez JC, Berthoud VM, Moreno AP, Spray DC. Multiplicity of controls in differentiated and undifferentiated cells and possible functional implications. Adv Second Messenger Phosphoprotein Research 1993; 27: 163±98. Larson DM, Haudenschild CC, Beyer EC. Gap junction messenger RNA expression by vascular wall cells. Circulation Res 1990; 66: 1074±80. Reed KE, Westphale EH, Larson, DM, Wang, H, Veenstra, RD, Beyer, EC. Molecular cloning and functional expression of human connexin 37 and endothelial cell gap junction protein. J Clin Investigations 1993; 91: 997±1004. Van Rijen HVM, Kempen MJAV, Analbers LJS et al. Gap junctions in human umbilical cord endothelial cells contain multiple connexins. Am J Physiol 1997, 272 (Cell Physiol); 41: C117±C130. Bruzzone R, Haeflinger J, Gimlich RL, Paul DL. Connexin40, a component of gap junctions in vascular endothelium, is restricted in its ability to interact with other connexins. Mol Biol Cell 1993; 4: 7±20.

16 Hu J, Cotgreave IA. Differential regulation of gap junctions by proinflammatory mediators in vitro. J Clin Investigation 1997; 99: 1±5. 17 Britz-Cunningham SH, Shah MM, Zuppan CW, Fletcher WH. Mutations of the connexin43 gap-junction gene in patients with heart malformations and defects of laterality. New England J Med 1995; 332: 1323±9. 18 Krutovskikh V, Mironov N, Yamasaki H. Human connexin 37 is polymorphic but not mutated in tumours. Carcinogenesis 1996; 17: 1761±3. 19 Richard G, Lin JP, Smith L et al. Linkage studies in erythrokeratodermias: fine mapping, genetic heterogeneity, and analysis of candidate genes. J Invest Dermatol 1997; 109: 666±71. 20 Lemne C, Jogestrand T, de Faire U. Carotid intima-media thickness and plaque in borderline hypertension. Stroke 1995; 26: 34±9. 21 Mercuri M, Tang R, Phillips M, Bond MG. Ultrasound protocol and quality control procedures in the European Lacidipine Study on Atherosclerosis (ELSA). Blood Pressure 1996; 4 s:20±3. 22 Dean AG, Dean JA, Columbier D et al. Info, version 6.0: A word processing, Database and Statistics program for public health on IBM-compatible microcomputers. Center for Disease Control and Prevention, Atlanta, Georgia, USA. 1995. Received 22 January 1999; accepted 20 April 1999. Correspondence: Ian A Cotgreave Ph.D, Division of Biochemical Toxicology, Institute of Environmental Medicine, Karolinska Institute, Box 210, S-17177 Stockholm, Sweden (tel.: + 87287654; fax: + 8334467; e-mail: [email protected]).

# 1999 Blackwell Science Ltd Journal of Internal Medicine 246: 211±218