Association Studies of Calcium-Sensing Receptor (CaSR) Polymorphisms with Serum Concentrations of Glucose and Phosphate, and Vascular Calcification in Renal Transplant Recipients

Background Cardiovascular disease is the major cause of death in renal transplant recipients (RTRs) and linked to arterial calcification. The calcium-sensing receptor (CaSR), a G-protein coupled receptor, plays a pivotal role in extracellular calcium homeostasis and is expressed in the intimal and medial layers of the arterial wall. We investigated whether common CASR gene variants are predictors for aortic and coronary artery calcification or influence risk factors such as serum calcium, phosphate and glucose concentrations in RTRs. Methods Two hundred and eighty four RTRs were investigated for associations between three CASR promoter region single nucleotide polymorphisms (SNPs) (rs115759455, rs7652589, rs1501899), three non-synonymous CASR coding region SNPs (A986S, R990G, Q1011E), and aortic and coronary artery calcium mass scores, cardiovascular outcomes and calcification risk factors that included serum phosphate, calcium, total cholesterol and glucose concentrations. Results Multivariate analysis revealed that RTRs homozygous for the minor allele (SS) of the A986S SNP, when compared to those homozygous for the major allele (AA), had raised serum glucose concentrations (8.7±5.4 vs. 5.7±2.1 mmol/L, P<0.05). In addition, RTRs who were heterozygous (CT) at the rs115759455 SNP, when compared to those homozygous for the major allele (CC), had higher serum phosphate concentrations (1.1±0.3 vs. 1.0±0.2 mmol/L, P<0.05). CASR SNPs were not significant determinants for aortic or coronary artery calcification, and were not associated with cardiovascular outcomes or mortality in this RTR cohort. Conclusions Common CASR SNPs may be independent predictors of serum glucose and phosphate concentrations, but are not determinants of vascular calcification or cardiovascular outcomes.


Introduction
Cardiovascular disease is the major cause of premature death in renal transplant recipients (RTRs) [1]. Cardiovascular events and mortality in RTRs are strongly linked to the presence of substantial vascular calcification, which affects > 30% of transplanted patients [2]. Vascular calcification is an active disease process characterised by mineral deposition within the medial and intimal layers of the arterial wall [3,4]. Medial calcification is a consequence of dysregulated systemic mineral homeostasis and associated with the trans-differentiation of vascular smooth muscle cells (VSMCs) in the arterial media to osteochondrocytic cells that release matrix vesicles, which act as a nidus for mineralisation in the presence of elevated circulating calcium and/or phosphate concentrations. [4]. Medial calcification reduces the compliance of large arteries such as the thoracic aorta, thereby leading to hypertension and left ventricular dysfunction [5]. Intimal calcification develops within atherosclerotic plaques, and is the major form of mineral deposition within the coronary arteries [6]. Thus, the presence of intimal calcification is an indicator of advanced atherosclerosis and associated with myocardial infarction [7]. Major risk factors for atherosclerotic plaque development and intimal calcification include elevations in serum total cholesterol and glucose concentrations, together with increased systolic blood pressure [8].
Arterial calcification has been reported to have a substantial genetic component [9,10], and previous studies have demonstrated associations with common polymorphisms in genes encoding inhibitors of blood vessel mineralisation such as fetuin A and matrix Gla protein [11,12]. However, the contribution to vessel calcification from polymorphisms of the calcium-sensing receptor (CaSR), which is a G-protein coupled receptor (GPCR) that plays a pivotal role in systemic mineral homeostasis through its effects on parathyroid hormone (PTH) secretion and renal tubular calcium reabsorption [13], has not been investigated. Moreover, the CaSR is expressed and functionally active in the intimal and medial layers of large elastic arteries such as the aorta, and muscular arteries such as the coronary, tibial and internal mammary arteries [14][15][16][17], and abnormal functioning of the arterially expressed CaSR has been implicated in the trans-differentiation of VSMCs to mineralising cells, and development of vessel wall calcification [17]. We therefore hypothesised that CASR variants may be determinants for arterial medial and intimal calcification, and calcification risk factors that include serum calcium, glucose and phosphate concentrations, in high risk patient groups such as RTRs, and selected six single nucleotide polymorphisms (SNPs) (3 non-synonymous coding region SNPs and 3 promoter region SNPs) (Fig. 1), five of which have been previously associated with indices of mineral metabolism and/or cardiovascular disease [18][19][20][21][22]. We investigated a well characterised cohort of RTRs [2,23] for associations between these CASR SNPs and cardiovascular outcomes, mortality, coronary artery calcification (CAC) and aortic calcification (AoC), and vascular calcification risk factors, which included: systolic blood pressure, serum calcium, phosphate, total cholesterol and glucose concentrations.

Materials and Methods Patients
Patients were ascertained from the Brussels Renal Transplant Cohort [2,23], which was collected between February 3 rd 2004 and January 27 th 2005, and comprised >280 RTRs that had an isolated kidney graft functioning for >1 year. The study protocol was approved by The Ethics Committee of the Université catholique de Louvain (UCL) Medical School, Brussels and written informed consent was obtained from all patients [2].

Clinical parameters
At baseline, clinical parameters including a history of cardiovascular events (defined as myocardial infarction, cerebrovascular event or transient ischaemic attack, and lower limb necrosis or revascularisation) were recorded by a review of the medical charts. Blood samples were obtained at inclusion for biochemical analysis. Serum creatinine, total calcium, phosphate, glucose and total cholesterol concentrations were measured using a Synchron CX analyzer (Beckman Coulter). Serum intact PTH concentrations were measured by a two-site immunoradiometric method (Nichols Institute), and serum 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D concentrations were measured using a LIAISON analyzer (Diasorin Inc). At inclusion, calcification of the aorta and the main coronary arteries were assessed by multi-slice spiral CT scanning of the chest on a 16-slice scanner (Brillance 16; Philips Healthcare, http:// www.healthcare.philips.com), as described [2]. The calcium mass of the thoracic aorta and the 4 branches of the main coronary arteries were scored individually, as previously described [2]. Agatston scores and amount of hydroxyapatite (mg) of the coronary arteries and thoracic aorta were measured using a manufacturer algorithm (Heart Beat CS; Philips Healthcare) and by a single operator. Intra-operator variability was 3% and 8%, respectively, for CAC and AoC [2]. Substantial amounts of coronary artery and aortic calcification were defined as >100 mg and >600 mg, respectively, as previously reported [2]. The RTRs were followed up for a mean duration of 4.4 ± 0.3 years and underwent repeat measurement of aortic and coronary artery Agatston scores by spiral CT scanning, and assessment of cardiovascular event incidence and mortality, as described [23].

Statistical analysis
Analyses were performed using IBM SPSS version 20 software. Association analyses were conducted between CASR SNPs and baseline clinical and biochemical parameters. Radiological parameters measured at baseline and after a mean follow-up period of 4.4 ± 0.3 years were included in the analysis. Parameters showing a right skewed distribution were log transformed prior to parametric analyses. The power of the study was determined using the web-based program QUANTO v1.2 [28]. A univariate analysis of associations was performed using Pearson's cross product correlation and Chi-squared test for continuous and categorical variables, respectively [11]. All variables that showed associations at the P 0.2 statistical level, following Bonferroni correction, were entered into a stepwise multivariate linear regression model and an assessment of confounding variables was performed. Kaplan-Meier curves were used to analyse the impact of CASR SNPs on all-cause mortality and cardiovascular event free survival and compared by the Mantel (log-rank) test [29]. Results are presented as mean ± SD, or number of patients (N (%)), as appropriate. The results of the multivariate analysis are presented as regression coefficient (B) values ± 95% confidence interval (CI). A value of P <0.05 was considered significant for all analyses.

Patient characteristics
At baseline, the study cohort comprised 284 adult RTRs (168 males and 116 females) of European origin, with a mean ± SD age of 52.8 ± 12.6 years ( Table 1). The patients had a moderate reduction of kidney function (CKD stage II-IIIa), explained by the post-renal transplant status, and the mean concentrations of serum biochemical markers of mineral metabolism, glucose and total cholesterol were within normal limits (Table 1). Hyperparathyroidism (defined as serum PTH concentrations >6.5 pmol/L), hyperphosphataemia (defined as serum phosphate concentrations >1.50 mmol/L) and diabetes affected 25%, 9% and 15% of the study cohort, respectively (Table 1). Baseline assessment of arterial calcification by spiral CT scanning in 266 RTRs revealed substantial amounts of CAC (>100 mg hydroxyapatite) and AoC (>600 mg hydroxyapatite) in >20% and >35% of individuals, respectively (Table 1). Follow-up CAC and AoC spiral CT assessments, after a mean period of 4.4 ± 0.3 years, in 187 individuals revealed >50% of patients to have an increase in CAC or AoC, with 30% of patients having experienced 1 cardiovascular events, with an overall mortality rate for the cohort of 12%. Results are presented as mean ± SD or the number of patients with the % of the total of number of patients shown in parentheses. AgS, Agatston score.

CASR SNP frequencies in renal transplant recipients
Six CASR SNPs were selected for an assessment of genotype and allele frequencies ( Fig. 1 and Table 2). Three of these SNPs are coding region variants (A986S, R990G and Q1011E), and the other three SNPs are located in the promoter region (rs7652589, rs1501899 and rs115759455). The promoter region SNP, rs115759455, has not been previously reported and was detected in this patient cohort during the study. The allelic frequencies of these six CASR SNPs (Table 2) were not significantly different to those observed in large Caucasian population cohorts such as the NHLBI-ESP and 1000 Genomes Project cohorts [25,26]. Furthermore, the allelic frequencies of these six CASR SNPs did not deviate from the Hardy-Weinberg equilibrium ( Table 2). Linkage disequilibrium was observed between the rs1501899 and rs7652589 promoter region SNPs (r 2 = 0.79), but not between the other CASR variants (r 2 <0.015).
Association of CASR SNPs with serum glucose, indices of mineral metabolism, arterial calcification scores and clinical outcomes Investigations for associations between individual CASR variants and arterial calcification scores, systolic blood pressure, serum glucose, serum total cholesterol and serum markers of mineral metabolism (Table 3) revealed an association between CASR SNPs and serum metabolites, but not directly with cardiovascular disease. Thus, univariate analysis revealed a significant association between the A986S and serum glucose concentrations ( Table 3). The mean glucose concentration of patients that were homozygous for the major allele A986 allele (AA; N = 216) was 5.7 ± 2.1 mmol/L compared with 8.7 ± 5.4 mmol/L for patients that were homozygous for the S986 minor allele (SS; N = 6) (P <0.05). All patients were taking prednisolone as part of their immunosuppressant regimen. However, differences in the cumulative dosage of this glucocorticoid medication did not significantly affect serum glucose concentrations in the study cohort (S1 Table).
The CASR SNPs were not significantly associated with AoC or CAC scores at baseline, or with progression of arterial calcification, or the occurrence of cardiovascular events during the study follow-up period (Table 4). Univariate subgroup analyses that excluded patients with diabetes, hyperparathyroidism or hyperphosphataemia, as these factors are key determinants of  vascular calcification [30][31][32], also did not reveal any association between CASR SNPs and AoC or CAC scores (S2-S4 Tables). Furthermore, an analysis of CASR SNP genotypes in RTRs stratified according to the presence of low or high arterial calcification scores did not reveal any significant associations (Table 5). Using the Kaplan-Meier curve estimate, the effect of CASR variants on all-cause mortality was assessed and was shown to be absent for all the CASR SNPs (data not shown, P >0.05). An analysis of associations between multi-locus CASR haplotypes and serum glucose, indices of mineral metabolism and arterial calcification scores, was not performed due to the low prevalence of minor alleles in these haplotypes.  Following Bonferroni correction, all parameters that had associations at the P 0.2 statistical level were entered into a stepwise multivariate linear regression model, which corrected for potentially confounding influences ( Table 6). The homozygous S986 allele continued to remain significantly associated with serum glucose after correcting for the presence of diabetes, as well as correcting for age, gender, body mass index (BMI), renal function, glucocorticoid and immunosuppressant usage, and transplantation vintage ( Table 6). A significant association was also observed between the promoter region CASR SNP, rs115759455, and serum phosphate concentrations after correcting for confounding parameters such as calcium and vitamin D supplementation, serum creatinine and estimated glomerular filtration rate, serum calcium, 1.25-dihydroxyvitamin D and intact PTH concentrations. Thus, the mean serum phosphate concentration in patients homozygous for the rs115759455 major allele (CC) was 1.00 ± 0.23 mmol/L (N = 258) and 1.14 ± 0.31 mmol/L for patients harbouring the CT alleles (N = 23, P <0.05).

Discussion
Our study has revealed that the A986S CASR SNP is a predictor of serum glucose concentrations independently of BMI and the presence of diabetes mellitus. Thus, patients who were homozygous for the S986 minor allele had elevations in serum glucose, and this highlights a potential role for the CaSR as a regulator of systemic glucose homeostasis. Indeed, this would be consistent with the reported expression of the CaSR in pancreatic islet beta-cells and that CaSR activation by extracellular calcium or calcimimetic drugs in isolated islets can stimulate beta-cell activity and insulin secretion [33][34][35]. In addition, the CaSR is expressed and functionally active in adipocytes [36], and may potentially regulate the peripheral actions of insulin, as highlighted by the finding of an association with the A986S CASR polymorphism and insulin resistance in patients with the polycystic ovarian syndrome [37]. Moreover, altered CaSR function in diabetic patients may contribute to the development of atherosclerosis and increased cardiovascular risk [38]. A significant association was also observed with the rs115759455 5'UTR CASR SNP and increased serum phosphate concentrations. Common genetic variants have been previously linked to serum phosphate concentrations, including a SNP (rs17265703) located on chromosome 3q21.1, which was shown to be in strong linkage disequilibrium with the A986S CASR variant [39]. Our study revealed the rs115759455 minor CASR allele (T) to be an independent predictor of serum phosphate concentrations, as heterozygosity (TC), when compared to homozygosity (CC) for the major allele, was associated with significantly increased serum phosphate concentrations; such effects of homozygosity (TT) for the minor allele could not be established as only two RTRs were homozygous. Circulating phosphate concentrations are regulated by the actions of PTH and fibroblast growth factor-23 (FGF-23) on phosphate reabsorption by the proximal renal tubule, and by 1,25-dihydroxyvitamin D mediated intestinal phosphate reabsorption [40]. Our findings indicate that the CaSR may regulate phosphate homeostasis independently of its effects on circulating PTH and 1,25-dihydroxyvitamin D concentrations. Moreover, studies of mice with the combined ablation of Casr and Pth alleles indicate that such effects of the CaSR on phosphate homeostasis are also not mediated by FGF-23 [41]. Indeed, the CaSR is likely to have a direct role in regulating circulating phosphate concentrations, as highlighted by micro-perfusion studies of isolated proximal renal tubules, which have revealed CaSR activation in this nephron segment to promote renal phosphate reabsorption [42]. Our findings provide further support for the CaSR being an independent regulator of phosphate metabolism.
However, CASR SNPs were not significantly associated with the development and progression of aortic medial calcification or coronary arterial intimal calcification, or cardiovascular outcomes in RTRs. These findings contrast with a report of the A986S CASR SNP as an independent predictor of coronary artery disease and cardiovascular mortality [22]. In addition, a mouse model, Nuf, with an activating Casr mutation, has been reported to have mineralisation within the aorta, elastic and muscular arteries [43]. The differences in these studies may be partly explained by differences in the cohort characteristics e.g. patient age, ethnicity, medication history, and underlying pathologies of the different patient groups (renal transplant versus non-renal disease); as well as by species differences (man versus mouse). However, these differences may also reflect limitations in our study of the cohort of RTRs, which include the small sample size and low prevalence of some alleles. Moreover, these findings may be affected by survival bias, which favours patients with less severe cardiovascular disease and calcification. Our study of 284 RTRs had a power of 99% and 89% to detect a locus that contributed 10% or 5%, respectively, of the genetic variance, assuming a type 1 error of 0.01 and marker frequency of 0.2. This indicates that our study was sufficiently powered to detect effects that would explain up to 5% of the variance, but not 1% of variance [28]. Nevertheless, the findings of this study indicate that the six common CASR polymorphisms are unlikely to play a major role in the development or progression of aortic or coronary artery calcification in patients with renal transplants.
In conclusion, our investigation of associations between common CASR variants, arterial calcification and other cardiovascular risk factors in a cohort of renal transplant recipients indicates that these CASR variants are unlikely to represent major determinants for calcification of either the aorta or coronary arteries. However, the CaSR was demonstrated to be an independent predictor of serum glucose and phosphate concentrations, thereby highlighting potential new metabolic roles for this GPCR.
Supporting Information S1 Dataset. Clinical, biochemical, radiological and genotype data for Brussels Renal Transplant Cohort. (XLS) S1