Progression of Coronary Atherosclerosis Is Associated with a Common Genetic Variant of the Human Stromelysin-1 Promoter Which Results in Reduced Gene Expression*

There is a common polymorphism in the promoter sequence of the human stromelysin-1 gene, with one allele having a run of six adenosines (6A) and the other five adenosines (5A). We have previously reported, in a 3-year follow-up study of patients with coronary ather- osclerosis, that those patients who are homozygous for the 6A allele show a more rapid progression of the dis- ease. In this study, we have investigated whether the 5A/6A promoter polymorphism plays a role in the regu- lation of stromelysin-1 gene expression. In transient transfection experiments, a stromelysin-1 promoter con- struct with 6A at the polymorphic site was found to express less of the chloramphenicol acetyltransferase reporter gene than a construct containing 5A. Electrophoretic mobility shift assay and DNase I footprinting revealed the interaction of one or more nuclear protein(s) with the DNA sequence at the 5A/6A polymorphic site. The binding of one of the nucleoprotein factors was more readily detectable with an oligonucleotide probe corresponding to the 6A allele as compared with a probe corresponding to the 5A allele. Replacing the core bind- ing sequence with a random DNA sequence abolished the interaction between the nuclear protein(s) and Taken together, these results suggest that, in fibroblasts, the 5A/6A polymorphism may influence stromelysin-1 promoter activity in an allele-specific manner and that the effect is independent of the orientation and position of the regulatory sequence. Such allele-specific regulation was also observed in transient expression experiments on a rat vascular smooth muscle cell line (A10), in which the 5T-CAT transfectants expressed 1.5-fold more CAT activity than the 6T-CAT transfectants (125 (cid:54) 8 versus 81 (cid:54) 16; n (cid:53) 4, p (cid:44) 0.05).

Stromelysin-1 is a key member of the matrix metalloproteinase (MMP) 1 family, with a broad substrate specificity. It can degrade types II, IV, and IX collagen, proteoglycans, laminin, fibronectin, gelatins, and elastin (1)(2)(3). In addition, stromelysin-1 can also activate other MMPs such as collagenase, ma-trilysin, and gelatinase B, rendering stromelysin-1 crucial in connective tissue remodeling (4 -6). Expression of stromelysin-1 is primarily regulated at the level of transcription, where the promoter of the gene responds to various stimuli, including growth factors, cytokines, tumor promoters, and oncogene products (7)(8)(9)(10). The regulatory effects of such stimuli are mediated through a number of cis-elements located in the stromelysin-1 promoter. For instance, the activator protein-1 binding site at positions Ϫ63 to Ϫ70 is necessary for the basal expression of the gene and is also involved in interleukin-1 induction (11)(12)(13). A promoter element located between Ϫ1218 and Ϫ1202 is responsible for the induction of stromelysin-1 expression by platelet-derived growth factor B/B (14,15), whereas three sequences that share strong homology with the glucocorticoid-responsive consensus element are likely to be involved in the dexamethasone suppression (16).
Over the last few years, MMPs have been implicated in the connective tissue remodeling during atherogenesis (17)(18)(19)(20)(21)(22). By in situ mRNA hybridization, we originally demonstrated the presence of stromelysin-1 in coronary atherosclerotic plaques (18). Extensive expression of the stromelysin-1 gene was localized particularly to the regions considered prone to rupture, such as the cap and its adjacent tissues. These observations have been supported subsequently by the studies of Galis et al. (19,20), who demonstrated the gelatinolytic and caseinolytic activity in atherosclerotic but not in the uninvolved arterial tissues, using in situ zymography, a method that allows direct detection and microscopic localization of MMP activity in tissue sections. Additional reports of MMP expression in atheroma have also appeared (21,22).
Using single strand conformation polymorphism analysis, we recently identified a common variant in the promoter of the stromelysin-1 gene (23). This variant gives rise to one allele having a run of six adenosines (6A) and another having five adenosines (5A) at position Ϫ1171 according to the sequence published by Quinones et al. (16). The frequency of the two alleles (5A/6A) was found to be 0.51/0.49 in a sample of 354 healthy individuals from the United Kingdom. However, in a study of 72 patients with coronary heart disease defined by angiography, the 5A/6A polymorphism was associated with progression of coronary atherosclerosis, with those who were homozygous for the 6A allele showing more rapid progression of both global and focal atherosclerotic stenoses over the 3-year study period (17). This observation supports the findings by others that the MMPs are involved in the connective tissue remodeling during atherogenesis (18 -22).
It was, however, unclear whether the 5A/6A polymorphism plays a role in the regulation of stromelysin-1 expression, or if * This work was supported by Grants PG/92021, FS/95011, and RG16 from the British Heart Foundation, Grant 8691 from the Swedish Medical Research Council, and by the Swedish Heart-Lung Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom correspondence should be addressed. it is in linkage disequilibrium with variants elsewhere at the gene locus that are functional. Therefore, we carried out experiments to address this question and report here the results of these studies.

EXPERIMENTAL PROCEDURES
Cell Cultures-Human fetal foreskin fibroblasts (HFFF2) were purchased from ECACC. Primary human vascular smooth muscle cells (VSMCs) were kindly provided by Dr. Jan Nilsson, Karolinska Hospital, Stockholm, Sweden. A rat vascular smooth muscle cell line (A10) was a kind gift from Dr. Ö zkan Yalkinoglu, Bayer AG, Wuppertal, Germany. Cells were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% (v/v) fetal calf serum (Life Technologies, Inc.), penicillin (100 units/ml), and streptomycin (100 g/ml) in standard tissue culture flasks in a humidified air/CO 2 (19:1) incubator at 37°C. Culture medium for A10 cells was also supplemented with 1 mM L-glutamine.
Reporter Gene Constructs-The 1.3-kilobase pair stromelysin-1 gene promoter (16) containing six adenosines at the 5A/6A polymorphic site was cloned into a p-ALTER vector (Promega). Site-directed mutagenesis was used to replace the 6As at the polymorphic site with 5As, using an Altered Sites in vitro mutagenesis system (Promega). The p-ALTER plasmids containing the stromelysin-1 promoter were digested with BamHI and blunt-ended by filling in with Klenow enzyme, followed by HindIII cleavage. The blunt-end HindIII fragments of stromelysin-1 promoter, with 5As or 6As at the polymorphic site, was cloned upstream a chloramphenicol acetyltransferase (CAT) reporter gene (pCAT-basic vector, Promega). The resultant plasmids were named 5A-CAT and 6A-CAT, respectively.
Two further plasmids, 5T-CAT and 6T-CAT, with the Ϫ480 to Ϫ1304 fragment in the stromelysin-1 promoter in the opposite orientation as compared to 5A-CAT and 6A-CAT, respectively, were created by the following strategy. The plasmids 5A-CAT and 6A-CAT were linearized by XbaI and converted into blunt-ends by filling in with dNTP and Klenow enzyme. These linearized plasmids were then cut into two fragments by HindIII. The smaller blunt-end HindIII fragment (from Ϫ480 to Ϫ1304 in the stromelysin-1 promoter) was partially filled in with dATP and dGTP. In separate tubes, the plasmids 5A-CAT and 6A-CAT were linearized by HindIII, and blunt-ended with dNTP. The linearized plasmids were cut by XbaI to generate two fragments. The larger blunt-end XbaI fragment was then partially filled in with dCTP and dTTP. Finally, the smaller fragment (containing one blunt end and one partial HindIII end) and the larger fragment (containing one blunt end and one partial XbaI end) were ligated.
A mutated construct (named M-CAT) in which the 8 base pairs of (A) 6 (C) 2 at the polymorphic site in the stromelysin-1 promoter are replaced by a sequence of ATACTTAG was made by in vitro mutagenesis using 6T-CAT as a template.
Correct sequences and orientation of the stromelysin-1 promoter in all constructs were confirmed by restriction enzyme analysis, polymerase chain reaction, and DNA sequencing.
Transfection and Transient Gene Expression Assays-The plasmid DNAs used for transfection were purified using a Plasmid Mega Kit (Qiagen). Twenty-four hours before transfection, cells were seeded at a density of 7 ϫ 10 5 cells/100-mm dish and maintained in Dulbecco's modified Eagle's medium with 10% fetal calf serum and antibiotics. The next day, the cells (approximately 50% confluent) were transfected by a lipofection method; 500 l of Opti-MEM I reduced serum medium (Life Technologies, Inc.) containing 5 g of CAT constructs and 5 g of pSV-␤-galactosidase control vector (Promega) was mixed with 500 l of Opti-MEM containing 30 g of Lipofectin reagent (Life Technologies, Inc.). The Lipofectin-DNA mixture was incubated at room temperature for 15 min, during which time cells were washed once with serum-free medium. Four ml of Opti-MEM was then added to the Lipofectin-DNA complex and the solution overlaid onto the cells. At 16 h post-transfection, the Lipofectin-DNA complex was aspirated and replaced with Dulbecco's modified Eagle's medium containing fetal calf serum and antibiotics. Twenty-four hours later, cells were harvested and CAT enzyme activities in the cell extracts measured using a Quan-T-CAT assay system (Amersham Corp.). ␤-Galactosidase activity was measured using a ␤-galactosidase enzyme assay system (Promega). CAT levels were expressed as arbitrary units after standardization against ␤-galactosidase levels.
Nuclear Extract Preparation-Human fetal foreskin fibroblasts and human primary VSMCs were grown to subconfluence, and nuclear extracts prepared as described by Alksnis et al. (24). All buffers were supplemented with 0.7 g/ml leupeptin, 16.7 g/ml aprotinin, 0.5 mM phenylmethanesulfonyl fluoride, and 0.33 g/ml 2-mercaptoethanol. Protein concentration in the extracts was measured according to the method described by Kalb and Bernlohr (25).
The oligonucleotide probes were 5Ј-end-labeled with T4 polynucleotide kinase and [␥-32 p]ATP. Three g of nuclear extracts was preincubated for 10 min on ice in a solution composed of 1 l of 10 mM dithiothreitol, 1 l of 10 mM EDTA, 3.6 l of 0.1 M Hepes, pH 7.9, 3.2 l of 50% Ficoll, and 2 l of 1 mg/ml poly(dI-dC). Then 60,000 cpm of probes, with or without unlabeled oligonucleotide competitors in 50 -100-fold molar excess, was added and incubated for another 20 min at room temperature. The unbound free probes were separated from the DNA-protein complexes by electrophoresis through a 7% nondenaturing polyacrylamide gel (the ratio of acrylamide/bisacrylamide ϭ 80:1) in 0.25 ϫ TBE buffer for 2 h at 4°C, 200 V. The gels were then exposed to Hyperfilm-MP (Amersham) at room temperature without intensifying screen for 1 week.
DNase I Footprinting Assays-Two double-stranded 35/36-mer oligonucleotides (CCTTTGATGGGGGGAAAAA(A)CCATGTCTTGTCCTGA) were 5Ј-end-labeled on either the top or bottom strand. The 32 P-labeled probes were then incubated with nuclear extracts from HFFF2 in conditions identical to those for the EMSA described above. This was followed by partial DNase I digestion as described elsewhere (26). The protein-DNA complexes were fractionated on a 7% polyacrylamide gel, and the retarded bands were electroblotted onto a DEAE membrane (Schleicher & Schuell). Then the DNAs were eluted into a high salt buffer (1 M NaCl, 0.1 M EDTA, and 2 mM Tris, pH 8.0), followed by phenol/chloroform extraction and ethanol precipitation. Finally the precipitates, with equal amount of radioactivity loaded onto each lane, were electrophoresed through a 12% denaturing polyacrylamide gel (19:1).

RESULTS
The 6A Allele Has a Lower Promoter Activity Than the 5A Allele-Transient gene expression experiments were carried out to investigate whether the 5A/6A polymorphism has an effect on the strength of stromelysin-1 promoter. For these experiments, two reporter gene constructs were made, in which the stromelysin-1 promoter with either 5As or 6As at the polymorphic site was placed upstream of the reporter gene CAT. The resultant plasmids, 5A-CAT and 6A-CAT, were then separately transfected into cultured human fetal foreskin fibroblasts (HFFF2), together with a ␤-galactosidase transfection control. Levels of CAT expression driven by the allelic constructs was compared after standardization for ␤-galactosidase activity. The results of these experiments demonstrated that 5A-CAT transfectants expressed approximately 2-fold higher CAT activity than the cells transfected with 6A-CAT: 275 Ϯ 31 versus 162 Ϯ 23 (n ϭ 7, p ϭ 0.013). The stromelysin-1 promoter in the 5A-CAT and 6A-CAT constructs was derived from the stromelysin-1 gene cloned by Quinones et al. (16). However, a discrepancy with this promoter sequence was reported recently (27), showing that the XbaI-SacI fragment between Ϫ479 and Ϫ1303 is in the reverse orientation in the human genome. Therefore, two other constructs (5T-CAT and 6T-CAT) were made in which the Ϫ479 to Ϫ1303 fragment was inverted as compared to 5A-CAT and 6A-CAT, to restore it to the correct genomic orientation. In these constructs the polymorphic site (6T/5T) is located at Ϫ601. 5T-CAT and 6T-CAT were then transfected into HFFF2 cells, and the transiently expressed CAT activities were compared. Irrespective of the orientation of the Ϫ479 to Ϫ1303 fragment, very similar CAT activities were observed: 275 Ϯ 31 and 333 Ϯ 15 (5A and 5T, respectively), and 162 Ϯ 23 and 137 Ϯ 12 (6A and 6T, respectively). The activity of the 5T-CAT construct was again approximately 2-fold higher than that of the 6T-CAT construct: 333 Ϯ 15 versus 137 Ϯ 12 (n ϭ 10, p Ͻ 0.001).
Taken together, these results suggest that, in fibroblasts, the 5A/6A polymorphism may influence stromelysin-1 promoter activity in an allele-specific manner and that the effect is independent of the orientation and position of the regulatory sequence. Such allele-specific regulation was also observed in transient expression experiments on a rat vascular smooth muscle cell line (A10), in which the 5T-CAT transfectants expressed 1.5-fold more CAT activity than the 6T-CAT transfectants (125 Ϯ 8 versus 81 Ϯ 16; n ϭ 4, p Ͻ 0.05).
Nuclear Protein Binding to the Sequence at the Stromelysin-1 5A/6A Polymorphic Site-Having found that there is a significant difference in the stromelysin-1 promoter activity between the 6A and 5A alleles, EMSAs were carried out to investigate whether the sequence at the 5A/6A polymorphic site is a binding site for nuclear proteins. For these experiments, two 27/26mer oligonucleotide probes corresponding to the stromelysin-1 promoter sequence from Ϫ1189 to Ϫ1161 were synthesized, with either 6A or 5A in the polymorphic site. These probes were then reacted with crude nuclear extracts prepared from different cell types.
Representative autoradiograms of EMSAs using nuclear extracts of HFFF2 and primary human vascular smooth muscle cells are shown in Fig. 1 (A and B, respectively). The two autoradiograms show that there are two shifted bands (labeled 1 and 2) regardless whether the 5A or 6A probes are used. These bands may represent two different DNA-binding proteins or different forms of the same protein. It is also possible that complex 1 may be a dimer of complex 2. Although bands 1 and 2 are common to both alleles, their relative intensities differ with the two probes. Whereas there is no apparent difference in the signals of complex 2 between the two alleles, complex 1 is more readily detectable in the assays using the 6A probe as compared with those using the 5A probe, suggesting that there may be a preferential DNA-protein interaction with the 6 allele.This pattern of signal intensity was seen consistently with nuclear extracts prepared not only from fibroblasts and SMC, but also from HUVEC and Hep G2 cells (data not shown), suggesting that the nuclear protein(s) concerned are expressed in a wide variety of cell types.
DNase I Footprinting of the Polymorphic Site-Based on the observations from the EMSAs that there are nuclear protein(s) that bind to the promoter sequence at the 5A/6A polymorphic site, DNase I footprinting was carried out to determine the actual sequences to which these nuclear protein(s) bind. As the two DNA elements bound by nuclear factors 1 and 2 may overlap in the same region (both factors bind to the 26/27-mer oligonucleotide probes in EMSAs), conventional DNase I footprinting techniques would not be suitable for this study. Therefore, a modified footprinting approach was used. Two oligonucleotide probes corresponding to the sequences of the 5A or 6A alleles were 5Ј-end-labeled on either the top or bottom strands. The probes were incubated with nuclear extract and then subjected to partial DNase I digestion. The nucleoprotein-DNA complexes were then separated by polyacrylamide gel electrophoresis and electroblotted to a DEAE membrane. The shifted bands 1 and 2, as well as the bands of the free probes, were then isolated from the membrane and electrophoresed in a DNA sequencing gel. Comparing the DNA sequence ladders from bands 1 and 2 with those from the free probes revealed the regions protected by the nuclear protein(s). In contrast to the conventional footprinting methods, where a long stretch of DNA is employed as a probe, showing the entire protein binding region as a gap in the DNA ladder, the probes used in this study were 36/37-mer oligonucleotides. Therefore, only one end of the protein binding sequences could be deduced from one direction. However, by combining the footprints from both directions (top and bottom strands; see Fig. 2), the entire extent of the protein binding elements was determined. These are TGATGGGGGGGAAAAA(A)CCATGTCTT for complex 1 and GGGAAAAA(A)CCATGTCTT for complex 2, respectively (Fig. 2).
A Random DNA Sequence at the 5A/6A Polymorphic Site Abolishes Nuclear Protein Binding-As mentioned above, cells transfected with constructs containing the 6A allele expressed less CAT than transfectants of constructs containing the 5A allele, and nucleoprotein factor 1 binds preferentially to the 6A probe, suggesting that this nucleoprotein factor may be a transcriptional repressor. Experiments were carried out to test whether replacement of the 5A/6A cassette would abolish the binding of the proteins and diminish their repressive effect on stromelysin-1 promoter activity. The SIGSCAN computer program was used to search for random DNA sequences to replace the poly(A) cassette, which would not introduce new cis-elements into the promoter. The results showed that replacing the AAAAA(A)CC sequence at the 5A/6A polymorphic site with ATACTTAG would not create any site for binding of known transcription factors.
EMSA with the mutant probe showed that the replacement of the 5A/6A cassette completely abolished the binding of both nuclear factors 1 and 2 to the probe (Fig. 3). Transient trans- fection experiments were then carried out to test whether the abolition of protein binding would increase the strength of the stromelysin-1 promoter. The M-CAT construct was introduced into HFFF2 and A10 cells, and CAT activities in these cells were compared with those in the transfectants of 6T-CAT and 5T-CAT. The results of these experiments showed that in HFFF2, the CAT activity was 2-fold higher in the M-CAT transfectants than the cells transfected with 6T CAT (M-CAT 275 Ϯ 13 versus 6T-CAT 137 Ϯ 12; n ϭ 10, p Ͻ 0.001), but there was no difference between the M-CAT and 5T-CAT transfectants. In A10 cells, however, M-CAT expressed 2.7-fold more CAT activity than 6T-CAT (M-CAT 220 Ϯ 12 versus 6T-CAT 8 Ϯ 16; n ϭ 4, p Ͻ 0.001), and 1.8-fold more CAT activity than 5T-CAT (M-CAT 220 Ϯ 12 versus 5T-CAT 125 Ϯ 4; n ϭ 4, p Ͻ 0.001).

DISCUSSION
The MMPs play important roles in connective tissue remodeling during tissue repair, cell migration, angiogenesis, tissue morphogenesis, and growth (2,3). The normal operation of such physiological processes requires a tightly controlled balance between the MMPs and specific tissue inhibitors of metalloproteinases (TIMPs). Disruption of the balance, however, could lead to various pathological states. For instance, unregulated expression of MMPs is partially the cause of the accelerated breakdown of extracellular matrix in arthritic disease, tumor invasion, and metastasis (1, 28 -30), while on the other hand, inadequate production of these enzymes is associated with the excessive accumulation of connective tissue in systemic sclero-sis (31). Clear evidence supporting the importance of maintaining the balance between active enzyme and inhibitor is provided by the recent discovery of mutations in the TIMP-3 gene, which cause abnormal connective tissue remodeling in the eye (32,33).
In this study, we investigated the possible function of the stromelysin-1 promoter 5A/6A polymorphism in regulating the expression of this metalloproteinase gene. In transient expression experiments, cultured fibroblasts and vascular smooth muscle cells transfected with the constructs containing 6As expressed a roughly 2-fold lower amount of reporter gene product as compared with the transfectants of the constructs containing 5As. DNA-protein interaction assays showed the binding of one or more nuclear protein(s) to the sequence surrounding the polymorphic site. One of the nucleoprotein factors bound preferentially to the 6A allele (the allele associated with lower promoter strength), suggesting that it may be a transcriptional repressor. Replacing the 5A/6A cassette with a random DNA sequence abolished the binding of the nucleoprotein(s) and diminished the repressive effects. In vascular smooth muscle cells, the construct with a random DNA sequence replacing the 5A/6A cassette (M-CAT) expressed more CAT activity than the constructs containing the 5A or 6A allelic promoters (5T-CAT and 6T-CAT, respectively), suggesting the effective removal of the repressive effect. In fibroblasts, however, although CAT activity was higher in the M-CAT transfectants than the cells transfected with 6T-CAT, no difference was observed in CAT activity between the M-CAT and 5T-CAT It has been demonstrated that variation in the promoters of a number of genes can affect the expression of gene products directly. For example, the 4G/5G polymorphism in the plasminogen activator inhibitor-1 (PAI-1) promoter regulates the transcription of the PAI-1 gene in an allele-specific manner, and is associated with inter-individual differences in plasma PAI-1 levels (26,34,35). The promoter activity of the 4G allele is 2-fold higher than that of the 5G allele in transiently transfected HepG2 cells (26). Similar findings have also been reported on the genes encoding apolipoprotein AI and ␤-fibrinogen (36 -38). The results in the present study suggest that analogous mechanisms may be involved in the regulation of stromelysin-1 gene expression.
In a study of 72 patients with coronary heart disease, we reported previously that the 6A allele of the stromelysin 5A/6A polymorphism was associated with a more rapid progression of coronary stenosis due to atherosclerosis (23). In the present study, the results suggest a lower promoter activity with the 6A allele. Thus, reduced stromelysin-1 expression seems to be associated with more rapid progression of atherosclerosis. But what are the mechanisms?
One of the characteristics of atherosclerosis is the alteration in the content of extracellular matrix in the arterial wall (39 -41). Thus, it is suggested that, during atherogenesis, there is continuous connective tissue remodeling, which involves synthesis and degradation of the extracellular matrix proteins including interstitial collagens, proteoglycans, and elastin (42,43). Recently, stromelysin-1 and several other MMPs have been implicated in this remodeling process (18 -22, 44). A number of studies have demonstrated an increase in MMP expres-sion in certain lesional areas such as the plaque shoulders, and it is suggested that MMP activity may contribute to the weakening of atherosclerotic lesions and consequently to plaque rupture (18 -22). However, in many atherosclerotic lesions, there is a net increase in the content of extracellular matrix, suggesting that the overall matrix turnover may have favored deposition rather than degradation. This is in agreement with the observations by Tyagi et al. (44), who recently studied the MMP levels in relation to the extracellular matrix contents in normal and atherosclerotic arteries, and showed that weightfor-weight atherosclerotic vessels contained more collagen and proteoglycans but lower collagenolytic activity than normal vascular tissues.
Based on the results in this study, we speculate that, compared with other genotypes, individuals homozygous for the 6A allele could have lower stromelysin-1 levels in their arterial walls because of reduced transcription of the gene. We have previously reported augmented expression of stromelysin-1 in advanced human atherosclerotic plaques (18). These data, supported by subsequent work by others (19 -22), demonstrate that there is a qualitative difference in expression of MMPs in atherosclerotic tissues compared with normal vessels, where active remodeling is not taking place. The majority of people in Western populations will develop atherosclerosis to some extent, and these enzymes will have a role in the matrix remodeling events associated with atherogenesis. The genetic studies identify inter-individual differences in the regulation of expression of these genes that may result in a quantitative difference between individuals in the level of enzyme activity, and which may therefore have an impact on the extent to which the matrix is degraded. It is emphasized that these changes are likely to be small, consistent with the contribution expected from a single gene operating in a complex multifactorial disorder. Our data support the concept that such differences may exist, potentially affecting the amount of enzyme in the vessel wall. In the model presented here, reduced levels of stromelysin-1 in the 6A homozygotes compared with other genotypes might affect the balance between synthesis and degradation during matrix turnover to favor increased deposition of matrix, leading to more rapid chronic growth of the atherosclerotic plaque, consistent with our previous findings (23).
In conclusion, the data presented here suggest that the 5A/6A polymorphism may play a functional role in regulating the expression of the stromelysin-1 gene. The genetic variant appears to influence stromelysin-1 expression in an allele-specific manner, a mechanism that could underlie its association with the progression of atherosclerosis in patients with coronary heart disease, as reported previously (23). As this is a common variant, with over a quarter of the population having the 6A6A genotype (23), its effect on stromelysin-1 expression with a 2-fold difference between the two alleles is likely to be biologically important. Although this common variant has been studied in relation to atherosclerosis, it is possible that it may also be relevant in the progression of other common chronic connective tissue disorders.