SEARCH

SEARCH BY CITATION

Keywords:

  • atherosclerosis;
  • biomarker;
  • CD93;
  • expression;
  • genetics;
  • myocardial infarction

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References
  10. Supporting Information

Abstract.  Mälarstig A, Silveira A, Wågsäter D, Öhrvik J, Bäcklund A, Samnegård A, Khademi M, Hellenius M-L, Leander K, Olsson T, Uhlén M, de Faire U, Eriksson P, Hamsten A (Karolinska University Hospital; Institute of Environmental Medicine, Karolinska Institutet, Stockholm; KTH-Royal Institute of Technology, Stockholm, Sweden). Plasma CD93 concentration is a potential novel biomarker for coronary artery disease. J Intern Med 2011; 270: 229–236.

Objectives.  A common nonsynonymous single nucleotide polymorphism (SNP) in the CD93 gene (rs3746731, Pro541Ser) has been associated with risk of coronary artery disease (CAD). CD93 is a transmembrane glycoprotein, which is detectable in soluble form in human plasma. We investigated whether the concentration of soluble CD93 in plasma is related to risk of myocardial infarction (MI) and CAD, using a case–control study of premature MI (n = 764) and a nested case–control analysis of a longitudinal cohort study of 60-year-old subjects (analysis comprising 844 of 4232 subjects enrolled at baseline). In addition, SNPs in the CD93 gene were studied in relation to plasma CD93 concentration and CD93 mRNA expression.

Methods and Results.  A sensitive and specific enzyme-linked immunosorbent assay was established for determination of the plasma CD93 concentration. Subjects were divided into three groups according to tertiles of the distribution of CD93 concentration. Lower odds ratios for risk of MI and incidence of CAD were observed in the middle CD93 tertile (142–173 μg L−1): odds ratio (95% confidence interval), 0.69 (0.49–0.97) and 0.61 (0.40–0.94), respectively. These associations were independent of traditional CAD risk factors. The minor allele of a SNP in the 3′ untranslated region of CD93 (rs2749812) was associated with increased plasma CD93 concentrations (P = 0.03) and increased CD93 mRNA expression levels (P = 0.02).

Conclusion.  The results of the present study suggest that the concentration of soluble CD93 in plasma is a potential novel biomarker for CAD, including MI.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References
  10. Supporting Information

Myocardial infarction (MI) is a clinical manifestation of preceding atherosclerosis, inflammation and thrombosis. The risk of MI and the presence of several risk factors for the condition (e.g. diabetes, hypertension, hyperlipidaemia, hypercoagulability and systemic inflammation) are influenced by both lifestyle and genetic susceptibility. Progress in molecular genetics has resulted in identification of novel susceptibility genes for MI, most of which have not been previously implicated in the aetiology of atherosclerosis or MI. The association between a single nucleotide polymorphism (SNP) in the CD93 gene and risk of MI was recently reported [1, 2].

CD93 is a transmembrane glycoprotein that together with human endosialin and thrombomodulin constitutes a small family of transmembrane proteins with similar domains [3]. The predicted molecular weight of CD93 is 68 kD, but in fact appears to be 126 kDa from SDS–PAGE, because of abundant glycosylation [4]. The unique domain architecture includes a C-type lectin-like domain, five EGF-like domains, a transmembrane domain and a short cytoplasmic tail [3, 4]. CD93 was originally designated as C1QRp, following the demonstration that CD93-binding antibodies could modulate the enhancement of phagocytosis when triggered by C1q [4, 5]. Although the relationship between CD93 and C1q remains elusive, more recent studies have demonstrated a role for CD93 in apoptosis, innate immunity and inflammation [6–9]. The function of CD93 has also been investigated in a murine CD93 knockout model [9]. Mice lacking CD93 were viable and had no severe developmental abnormalities, but phagocytosis of apoptotic cells was decreased in vivo in these animals. An interesting link between CD93 and inflammation has been reported by Bohlson et al. [10], who also identified the presence of a soluble form of CD93 in human plasma. These authors suggested that soluble CD93 is the result of proteolytic ectodomain cleavage (shedding) of membrane CD93 and demonstrated that treatment of human peripheral blood monocytes (PBMCs) with tumour necrosis factor α (TNF-α) and lipopolysaccharide induces shedding. Cleavage of CD93 seems to be executed by a metalloproteinase, as the general metalloproteinase inhibitor 1,10-phenanthroline inhibited phorbol ester-triggered shedding of CD93. By contrast, CD93 shedding was independent of the TNF-α-converting enzyme [10]. Membrane CD93 protein expression has been described in cells of myeloid origin, including bone marrow stem cells, endothelial cells and platelets, and is supported by studies of CD93 mRNA expression levels [11, 12].

CD93 spans 6,985 base pairs (bp) and is positioned on chromosome 20 p11.21. It has two coding exons, separated by 285 bp of intronic sequence, and a long 3′ untranslated (UTR) region of approximately 4500 bp. A common nonsynonymous SNP (rs3746731, A>G, Pro541Ser) in exon 1 in CD93 was found to be associated with risk of MI in a multi-stage case–control study of high-risk individuals [1, 2]. The hazard ratio for coronary artery disease (CAD) associated with the rs3746731 GG genotype (Ser/Ser) was 1.26.

In the present study, we explored CD93 as a potential biomarker for atherothrombosis. Using a case–control study of premature MI and a population-based prospective study of initially healthy 60-year-old subjects, the primary aim was to investigate whether the concentration of soluble CD93 in plasma is related to risk of MI and other forms of CAD. For this purpose, a sensitive enzyme-linked immunosorbent assay (ELISA) for soluble CD93 in plasma was developed. In addition, SNP rs3746731 and other common SNPs in CD93 were evaluated as possible determinants of plasma CD93 concentration and CD93 mRNA production in human PBMCs.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References
  10. Supporting Information

ELISA for measurement of soluble CD93 in plasma

A sensitive and specific ELISA for determination of plasma CD93 concentration was developed using affinity-purified immunoglobulin (Ig)G antibodies produced in goats immunized with recombinant CD93 (R&D Systems, Minneapolis, MN, USA). As a result of the use of polyclonal antibodies, it is likely that equivalent sensitivities towards different CD93 isoforms are achieved. The intra-assay coefficient of variation was 5.6% (n = 5), and the inter-assay coefficient of variation was 6.5% for a total of 20 replicates. The specificity of the assay was checked by replacing the probing antibody used in the original assay (biotinylated goat IgG anti-human CD93 from R&D Systems) by a biotinylated monoclonal mouse IgG1 anti-human CD93 (Bio-Legend, San Diego, CA, USA). The ELISA with the replaced probing antibody produced similar results for the coagulation reference from Technoclone, Vienna, Austria, as well as for recombinant CD93 (obtained as a prerelease reagent from R&D Systems). Further details are provided in Data S1.

Study cohorts

The Stockholm Coronary Atherosclerosis Risk Factor (SCARF) study enrolled 377 survivors of a first MI before the age of 60, along with 387 healthy controls matched for age, sex and area of residence [13]. Whole blood for DNA extraction and citrated plasma samples were obtained 3 months after the index event. Coronary angiograms (n = 232) obtained routinely in two of the three participating hospitals were analysed by quantitative coronary angiography using the Medis QCA-CMS system, Leiden, The Netherlands [13].

The population-based study of 60-year-old men and women included a total of 4232 individuals (2039 men and 2193 women) recruited between 1997 and 1998 in the greater Stockholm area [14]. Participants underwent a physical examination, including laboratory tests, and completed a comprehensive questionnaire, including questions on medical history and demographic, socio-economic and lifestyle factors. They were then monitored regarding incident cardiovascular events [15]. After an average of 8 years of follow-up, a total of 211 cardiovascular events were recorded; 49 cases of stroke and 162 cases of CAD. All CAD cases were a first-time event and defined as death from coronary heart disease, acute MI (fatal and nonfatal), CAD-related death in hospital, unstable angina and events requiring procedures including percutaneous coronary intervention and coronary artery bypass grafting. Plasma CD93 concentration at baseline was measured in all subjects who experienced a cardiovascular event during follow-up (n = 211) and in three control subjects per case (n = 633), matched for gender.

Human PBMCs were obtained from venous blood from a group of 91 patients examined at the Department of Neurology at the Karolinska University Hospital (66 women and 25 men; mean age, 42.1 years). There were no signs of neuroinflammatory disease in any subjects, as determined by a variety of neurological and neuroradiological examinations and lumbar puncture as part of routine clinical examination. The following neurological diagnoses were observed in this group: unspecified sensory disturbance, idiopathic intracranial hypertension, unspecified headache, migraine, chronic idiopathic pain and chronic idiopathic fatigue.

All participants in the three studies gave informed consent, and the study protocols were approved by the ethics review board of the Karolinska Institutet.

Quantification of CD93 mRNA expression in human PBMCs

The isolation of PBMCs, extraction of total RNA, synthesis of cDNA and extraction of genomic DNA have been described previously [16]. Real-time PCR was performed using the ABI 7900 HT instrument with Universal Master Mix (Applied Biosystems, Foster City, CA, USA) and predesigned primers and probe mix (Hs00362607_m1) from the same manufacturer. The probe hybridized to the exon 1–2 boundary in the CD93 gene.

Selection and genotyping of SNPs in the CD93 gene

A total of 13 SNPs located within 3000 bp of each flank of the CD93 gene were included in phase II of the HapMap project (Figure S1). Twelve of these SNPs were captured by our selection of five tag SNPs with a pairwise correlation coefficient of at least 0.9. The remaining SNP, rs7492, had a pairwise correlation coefficient with rs2749817 of 0.4, whereas the linkage disequilibrium (LD) coefficient (D′ coefficient) was 0.98. The selection of tag SNPs was performed using the Tagger algorithm in HaploView 4.0. (MIT/Harvard Broad Institute, Cambridge, MA, USA) SNPs rs2749817 and rs17682515 are located downstream of the CD93 3′ UTR region, SNP rs2749812 is located in the 3′ UTR region, SNP rs3746731 in exon 1 encodes a nonsynonymous P541S change, and SNP rs1884654 is located upstream of the CD93 promoter region (Figure S2). Genotypes were determined by allelic discrimination on real-time PCR with predesigned primers and probes from Applied Biosystems (design identification-numbers and call rates for the SNPs are provided in Table S1).

Statistical analysis

Continuous variables are presented as median (interquartile range). The Kruskal–Wallis or Mann–Whitney tests were used to compare median levels of continuous measurements between groups (e.g. across CD93 tertiles or between cases and controls). Association between plasma CD93 concentration and MI/CAD in SCARF in the cohort study of 60-year-old subjects was tested using a logistic regression model with and without adjustment for sex, smoking, body mass index (BMI) and concentrations of low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol and high-sensitivity C-reactive protein (CRP), all of which are established CAD risk factors and/or established biomarkers. A second model also included cystatin-C concentration as a covariate. Risk of MI/CAD is given as odds ratio (95% confidence interval) [OR (95% CI)]. Association between CD93 SNPs and plasma CD93 concentration was tested using analysis of covariance under both a general model and an additive genetic model, of which only the latter assumes an allelic dose-dependent effect. CD93 concentration was used as dependent variable in the genetic association analyses and therefore log-transformed to obtain a normal distribution. Calculation of LD was performed in HaploView 4.0. Statistical analyses were performed using spss 16.0 for Windows and plink 1.05 [17]. A two-tailed P-value <0.05 was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References
  10. Supporting Information

Baseline characteristics according to plasma CD93 concentration

Plasma CD93 concentration was first measured in the SCARF study (Figure S3). Subjects were divided into three groups according to tertiles of the distribution of the concentrations of soluble CD93 (Table 1). Median age, indicators of obesity, plasma lipoprotein concentrations and glucometabolic parameters were similar across CD93 tertiles. By contrast, cystatin-C and, to a lesser extent, plasma fibrinogen concentration and the number of smokers were increased in the middle and upper CD93 tertiles.

Table 1. Baseline characteristics of patients with MI and healthy controls in the SCARF study according to CD93 tertiles
 CD93 < 142 (μg L−1) (n = 256)CD93 142–173 (μg L−1) (n = 256)CD93 > 173 (μg L−1) (n = 256) P-value
  1. BMI, body mass index; LDL, low-density lipoprotein; HDL, high-density lipoprotein; SCARF, Stockholm Coronary Atherosclerosis Risk Factor.

  2. Values are number of subjects (%) or median (interquartile range). The P-values refer to tests of independence for categorical variables and to the Kruskal–Wallis test for continuous variables.

Median age (years)54 (49–57)53 (49–57)54 (49–57)0.77
Women (n, %)51 (20)36 (14)43 (17)0.21
Current smoker (n, %)49 (19)53 (21)57 (23)0.01
BMI (kg m−2)26.4 (24.3–28.5)26 (24.0–28.4)26 (24.3–28.7)0.69
Waist circumference (cm)97 (90–103)96 (90–102)96 (90–104)0.48
Glucose (mmol L−1)5.2 (4.7–5.7)5 (4.8–5.4)5.1 (4.7–5.6)0.16
Insulin (pmol L−1)42 (30–60)39 (29–58)41 (28–65)0.34
Diabetes (n, %)14 (6)12 (5)15 (6)0.83
LDL-cholesterol (mmol L−1)3.3 (2.7–4.2)3.3 (2.7–4.0)3.3 (2.7–3.9)0.79
HDL-cholesterol (mmol L−1)1.2 (1.0–1.4)1.2 (1.0–1.5)1.2 (1.0–1.5)0.87
Triglycerides (mmol L−1)1.4 (1.1–2.1)1.4 (0.9–1.9)1.4 (1.0–1.8)0.12
Statin treatment (n, %)41 (16)40 (16)48 (19)0.55
C-reactive protein (mg L−1)1.1 (0.6–2.5)1.1 (0.6–2.1)1.2 (0.6–2.9)0.19
Plasma fibrinogen (g L−1)3.6 (3.2–4.1)3.6 (3.1–4.1)3.8 (3.3–4.3)0.03
Cystatin-C (ng L−1)0.78 (0.69–0.88)0.83 (0.76–0.9)0.87 (0.8–0.97)<0.001
Estimated glomerular filtration rate (mL min−1)81 (72–90)78.4 (70–90)79.2 (70–91)0.64

Plasma CD93 concentration in relation to MI/CAD

Case–control study of premature MI.  The median plasma CD93 concentration was similar in patients and controls [median (interquartile range)] 159 (137–186) μg L−1 and 157 (132–189) μg L−1, respectively, Mann–Whitney P = 0.68). The association between CD93 tertile and MI risk was tested in a logistic regression model using the lower tertile as a reference. Plasma concentrations in the middle CD93 tertile showed a significant inverse association with risk of MI; OR (95% CI) 0.69 (0.49–0.97) (Table 2). The association was unaltered by adjustment for BMI, smoking and concentrations of LDL-cholesterol, HDL-cholesterol and plasma CRP, but was abolished by additional adjustment for cystatin-C concentration (Table 2).

Table 2. Association between the plasma concentration of soluble CD93 and risk of MI in the SCARF study (377 CAD cases and 387 controls)
CD93 tertileOR (95% CI) P-valueAdjusted ORa (95% CI) P-valueAdjusted ORb (95% CI) P-value
  1. OR, odds ratio; CI, confidence interval; BMI, body mass index; LDL, low-density lipoprotein; HDL, high-density lipoprotein; CRP, C-reactive protein; CAD, coronary artery disease; SCARF, Stockholm Coronary Atherosclerosis Risk Factor.

  2. aAdjusted for smoking, BMI and concentrations of LDL-cholesterol, HDL-cholesterol and CRP. badjusted for smoking, BMI, concentrations of LDL-cholesterol, HDL-cholesterol plasma CRP and cystatin-C.

<142 (μg L−1)Ref Ref Ref 
142–173 (μg L−1)0.69 (0.49–0.97)0.040.63 (0.44–0.93)0.030.76 (0.50–1.15)0.18
>173 (μg L−1)0.89 (0.63–1.31)0.510.78 (0.52–1.17)0.231.05 (0.67–1.63)0.83

Longitudinal study of incident CAD.  The association between plasma CD93 concentration and CAD was further evaluated in the cohort study of 60-year-old subjects. The median (interquartile range) of the plasma CD93 concentration among subjects who did not experience CAD during follow-up did not differ significantly from that of incident CAD cases [159 (137–184) μg L−1 and 154 (132–189) μg L−1, respectively, Mann–Whitney P = 0.24]. However, the middle CD93 tertile was significantly associated with a lower risk of incident CAD [OR (95% CI), 0.61 (0.40–0.94)], whereas no association was observed for the upper CD93 tertile (Table 3). Covariate adjustments, including for sex, smoking, BMI and concentrations of LDL-cholesterol, HDL-cholesterol and plasma CRP, did not significantly alter these associations (Table 3).

Table 3. Association between the plasma concentration of soluble CD93 and risk of incident CAD in the cohort study of 60-year-old subjects (160 CAD events among 844 subjects)
CD93 tertileOR (95% CI) P-valueAdjusted ORa (95% CI) P-value
  1. OR, odds ratio; CI, confidence interval; BMI, body mass index; LDL, low-density lipoprotein; HDL, high-density lipoprotein; CRP, C-reactive protein; CAD, coronary artery disease.

  2. aAdjusted for sex, smoking, BMI and concentrations of LDL-cholesterol, HDL-cholesterol and plasma CRP.

<142 (μg L−1)Ref Ref 
142–173 (μg L−1)0.61 (0.40–0.94)0.020.62 (0.40–0.97)0.03
>173 (μg L−1)0.74 (0.49–1.12)0.150.78 (0.50–1.19)0.26

SNPs in the CD93 gene and plasma CD93 concentration

Genotype–phenotype association analyses were performed separately in controls and cases to minimize the confounding effects attributable to postinfarction sampling. SNP rs2749812 was weakly associated with soluble CD93 concentration in controls (Table 4). After adjustment for nongenetic covariates, the association between SNP rs2749812 and plasma CD93 concentration was reinforced, and statistical significance was attained for SNP rs3746731 (Table 4). A subsequent conditional analysis, controlling for the effect of SNP rs2749812, showed that the association between SNP rs3746731 and plasma CD93 concentration was because of its partial allelic association with SNP rs2749812. The pairwise D’ and R-squared correlation coefficients between SNPs rs2749812 and rs3746731, estimating the degree of linkage between the two SNPs, were 0.96 and 0.23, respectively, and HapMap genotype data indicate that three common haplotypes (combinations of allelic variants) exist in the Caucasian population: G–A (52%), G–G (30%) and A–G (18%). In patients, there was a weak association between SNP rs2749812 and plasma CD93 concentration in the general genetic model (two degrees of freedom), which was strengthened after taking into account the confounding effects of smoking and plasma cystatin-C concentration (Table 4). By contrast, SNPs rs17682515, rs2749817 and rs1884654 were not significantly associated with plasma CD93 concentration (data not shown).

Table 4. Association between SNPs rs2749812 and rs3746731 and CD93 plasma concentration and mRNA expression level in peripheral blood monocytes
 SCARF controls, plasma CD93 concentrationSCARF patients, plasma CD93 concentrationPBMC samples, CD93 mRNA
n Mean (SD) n Mean (SD) n Mean (SD)a
  1. PBMC, peripheral blood monocytes; SCARF, Stockholm Coronary Atherosclerosis Risk Factor; SNP, single nucleotide polymorphism.

  2. Values are displayed as mean (standard deviation). aLog-transformed arbitrary units relative to standard curve. bThe P-values were derived using a general linear model without or with adjustment for cystatin-C concentration and smoking.

rs2749812GG264163 (41)262168 (55)610.55 (0.47)
AG107170 (41)100177 (56)300.76 (0.46)
AA16176 (48)19141 (42)20.80 (0.25)
P-valueGeneral0.131 0.009 0.05
Additive0.047 0.621 0.04
Adjusted P-valuebGeneral0.024 0.004  
Additive0.006 0.279  
rs3746731AA118163 (42)99173 (62)290.68 (0.41)
AG187163 (39)191170 (52)420.62 (0.51)
GG84173 (47)92163 (53)220.59 (0.47)
P-valueGeneral0.098 0.409  
Additive0.057 0.268  
Adjusted P-valuebGeneral0.025 0.113  
Additive0.013 0.049 0.510

SNP rs2749812 is associated with CD93 gene expression level in human PBMCs

CD93 mRNA was found to be expressed to some degree in all PBMC samples. In separate analyses, SNP rs2749812 was associated with the CD93 mRNA expression level, whereas SNP rs3746731 was not (P = 0.04 and P = 0.51, respectively) (Table 4). PBMCs from subjects carrying one or two copies of the rs2749812 A allele had on average a higher CD93 mRNA expression level, which is the same direction of effect as with plasma CD93 concentration in the SCARF controls.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References
  10. Supporting Information

In the present study, we examined whether the soluble protein product of CD93, a recently identified susceptibility gene for CAD, can be used as a novel plasma biomarker for this condition. Nominally, significant associations were observed between plasma CD93 tertile and premature MI and incident CAD in two independent studies, with lower odds ratios being observed in the middle (142–173 μg L−1) as compared with the first tertile. It is important that this observation is replicated in a large prospective study to fully establish the relationship between plasma CD93 concentration and CAD.

To the best of our knowledge, this is the first study to investigate whether plasma CD93 concentration can be used as a biomarker for CAD. A potential link between CD93 and CAD was first suggested by the results of a genetic association study [2]. However, CD93 may also be regarded as a biological candidate gene for atherosclerosis and CAD development, given the reported functions of CD93 in the immune system [18]. An interesting finding of a study in CD93 knockout mice was that CD93 participates in the regulation of phagocytosis of apoptotic cells in vivo [9]. This process is essential for tissue macrophages and believed to be of major importance in the progression of atherosclerotic plaques [19, 20]. In addition, in vitro studies have shown expression of CD93 in macrophages and endothelial cells [10], both of which are cell types known to participate in early lesion development and atherosclerosis progression. It could be speculated that the relationship between plasma CD93 concentration and CAD risk is accounted for by decreased phagocytosis of apoptotic cells at low CD93 levels and a more pronounced inflammatory and/or proteolytic activity at high CD93 levels, the latter being reflected by the strong association with the cystatin-C concentration.

To what extent the soluble form of CD93, which was measured in the present study, reflects activity of membrane-bound CD93 is currently unknown. Bohlson et al. [10] showed that soluble CD93 may be the product of ectodomain cleavage of membrane CD93 and that the cleavage is mediated by a matrix metalloproteinase (MMP). It is unknown whether soluble CD93 reflects active CD93 on the cell membrane. MMPs have been extensively studied in both humans and mice and are now known to be key factors in vascular remodelling [21]. Another family of proteases, which is also involved in extracellular matrix remodelling, is the cysteine protease family of cathepsins [22]. Cystatin-C is an important extracellular inhibitor of the activity of cathepsins, and low concentrations of this protease inhibitor have been observed in smooth muscle cells in atherosclerotic plaques [23]. In addition, we have reported that functional sequence variants in the CST3 gene, encoding cystatin-C, are associated with coronary stenosis [24]. In the present study, we found that the plasma CD93 concentration correlated strongly with the plasma cystatin-C concentration in both healthy controls and patients with MI and that the association between plasma CD93 concentration and MI was not independent of the cystatin-C concentration (Table 2). One may speculate that the close relationship between CD93 and cystatin-C implies a role for CD93 in plaque remodelling or that soluble CD93 reflects general cathepsin and metalloproteinase activity.

We identified a SNP in the 3′ UTR region of the CD93 gene (rs2749812), the rare A allele of which was associated with an increased plasma concentration of soluble CD93 and a higher expression level of CD93 mRNA in PBMCs from healthy donors. Data from the HapMap project show that the G allele of SNP rs3746731 is present in the haplotype that also includes the rare rs2749812 A allele and may suggest that the association between CAD and SNP rs3746731 is attributable to the strong LD with SNP rs2749812. It is also possible that the association between CAD and SNP rs3746731 is related to both the functional consequences of the proline to serine amino acid shift and the effects on CD93 mRNA expression level that are mediated by SNP rs2749812.

There are several mechanisms by which genetic variants in the 3′ UTR region of a gene may alter the mRNA expression level. Most examples involve alterations of polyadenylation sites, with an impact on the stability of the mRNA, but examples of binding sites for micro RNAs that have been altered by SNPs have also been reported [25, 26]. Future genotype–phenotype association studies targeting CD93 variants should include the rs2749812 SNP, as the results of our study suggest that this SNP affects CD93 gene transcription.

The clinical usefulness of plasma CD93 determination in risk assessment for CAD requires further substantiation considering that the association with CAD risk was moderate and that the statistical power of the two cohorts examined in the present study was limited. These restrictions notwithstanding, the results of the present study suggest that plasma CD93 is a potential novel biomarker for CAD.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References
  10. Supporting Information

We would like to thank Karin Husman and Fariba Foorogh for excellent technical assistance. The study was funded by the the Leducq Network of Excellence in Atherothrombosis, the Foundation for Strategic Research, the European Commission (LSHM-CT- 2007- 037273), the Swedish Heart-Lung Foundation, the Swedish Research Council (8691, 09533, 7429), the Knut and Alice Wallenberg Foundation, the Torsten and Ragnar Söderberg Foundation, the Strategic Cardiovascular Programme of Karolinska Institutet, the Stockholm County Council (560183), the Magnus Bergvall Foundation and the Nanna Svartz Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References
  10. Supporting Information

Data S1. Methods.

Figure S1. Linkage disequilibrium between CD93 SNPs, as estimated in the HapMap CEU population.

Figure S2. Location and identity of CD93 SNPs examined in the present study.

Figure S3. Distribution of plasma CD93 concentration in the SCARF study.

Table S1. Location and genotyping assay identification numbers for the SNPs genotyped in the present study.

FilenameFormatSizeDescription
JOIM_2364_sm_MethodsTabS1-FigS1-S3.doc95KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.