BMC Medical Genetics BioMed Central

Background: Genes implicated in common complex disorders such as obesity, type 2 diabetes mellitus (T2DM) or cardiovascular diseases are not disease specific, since clinically related disorders also share genetic components. Cysteine protease Calpain 10 (CAPN10) has been associated with T2DM, hypertension, hypercholesterolemia, increased body mass index (BMI) and polycystic ovary syndrome (PCOS), a reproductive disorder of women in which isunlin resistance seems to play a pathogenic role. The calpain 5 gene (CAPN5) encodes a protein homologue of CAPN10. CAPN5 has been previously associated with PCOS by our group. In this new study, we have analysed the association of four CAPN5 gene variants(rs948976A>G, rs4945140G>A, rs2233546C>T and rs2233549G>A) with several cardiovascular risk factors related to metabolic syndrome in general population.


Background
Factors that increase the risk of cardiovascular disease (CVD) include obesity, dyslipidemia, glucose intolerance, type 2 diabetes mellitus (T2DM) and hypertension. When these factors cluster in an individual is called metabolic syndrome (MS), a complex disorder characterized by impaired glucose metabolism, dyslipidemia, hypertension and obesity [1].
Calpains are a class of cysteine proteases that requires calcium for its activation. They are implicated in a wide range of cellular functions, including apoptosis, intracellular signalling, proliferation and differentiation [2]. The human calpain family currently comprises 16 members and two subfamilies based on its functional domains [3]. The classic calpains are heterodimeric proteins that contain five sets of EF-hand calcium-binding structures in each subunit, similar to those in calmodulin. Calpain 10 (CAPN10), calpain 5 (CAPN5) and calpain 6 (CAPN6) proteins belong to the non-EF-hand subfamily that lacks these calcium-binding motifs; these calpains also differ from the remaining members of this subfamily in the presence of a T-domain homologous to C.elegans TRA-3, a protein implicated in the sexual determination of the hermaphrodite worm [4,5]. In fact, calpain 5 is the human ortologue of the TRA-3 nematode protein, a protease that exerts its action over a transmembrane protein called TRA-2 which binds and inhibits the male promoting FEM proteins (FEM-1, FEM-2 and FEM-3) [6,7]. Interestingly, the knockout mice of one of the three mammalian homologues of the C.elegans fem-1 gene, Fem1b, display abnormal glucose homeostasis, with abnormal glucose tolerance tests and defective glucose-stimulated insulin secretion [8]. In humans, the FEM1A gene maps to chromosome 19p13.3, a region linked to polycystic ovary syndrome (PCOS), a common endocrine disorder of women of reproductive age, characterized by chronic anovulation, infertility, hyperandrogenemia and frequently, insulin resistance resulting in an increased prevalence of obesity, CVD and T2DM, reason why PCOS is consider a phenotype closely related to metabolic syndrome. Maher y cols. [9] have recently reported a germline missense mutation in the FEM1A gene in a PCOS woman that was absent in 198 control chromosomes; the authors propose FEM1A as a candidate gene for PCOS.
The CAPN5 homologue CAPN10 was identified as a T2DM susceptibility locus by Horikawa et al. [10] and has been shown to be related to proinsulin processing, insulin secretion and insulin resistance [11,12]. Ehrman et al. [13] and our group [14] independently published association of CAPN10 gene with PCOS. A reanalysis of CAPN10 gene in a larger PCOS population, allowed us to identify specific haplotypes associated with hypercholesterolemia in PCOS patients [15]. Furthermore, CAPN10 has been associated with hypertension and elevated body mass index (BMI) by different groups [16,17].
In a previous work, we analyzed four CAPN5 gene variants (rs948976, rs4945140, rs2233546 and rs2233549) in 148 PCOS women [18]. We found that specific CAPN5 haplotypes were overrepresented in PCOS patients. In addition, we identified several CAPN5 alleles associated with phenotypic differences observed between PCOS patients, such as the presence of obesity, cardiovascular complications, and familial antecedents of obesity, hypertension and T2DM aggregation. These findings aimed us to investigate the relation of CAPN5 gene with cardiovascular risk factors in the general population. Silander et al. [19] have reported strong evidence of linkage within 11q14, the chromosomal region that contains CAPN5 gene in a large set of Finnish affected sibling pair families with T2DM.
Here we present the first population-based association analysis of CAPN5 gene in traits related to hypertension and other components of metabolic syndrome. Our results suggest that CAPN5 alleles could modulate diastolic blood pressure and cholesterol levels. In addition, a CAPN5 haplotype over-represented in obese individuals is also associated with the cluster of cardiovascular risk factors defined as metabolic syndrome.

Patients
This population based study included 606 non-related Caucasian men (n = 278, 46%) and women (n = 328, 54%) recruited by a simple random sampling approach from a cross-sectional population-based epidemiological survey in the province of Segovia in Central Spain (Castille) aimed at investigating the prevalence of anthropometric and physiological parameters related to obesity and other components of the metabolic syndrome [20]. Clinical characteristics of study subjects are summarized in table 1. Individuals with previous diagnosis of type 1 diabetes were excluded from the study.
All study subjects gave their written consent to participate in the study. The study protocol was approved by the Ethics Committee of the Hospital Clinico San Carlos of Madrid.

Phenotype measurements
We have analysed the existence of association between CAPN5 and twelve clinical parameters: body mass index (BMI), waist circumference (WC), systolic and diastolic blood pressures (SBP, DBP), fasting glucose, 2-hours glucose, fasting insulin, insulin resistance (IR-HOMA), total cholesterol, HDL-cholesterol, LDL-cholesterol and triglycerides (TGs). SBP and DBP were measured three times in the seated position after 10 minutes of rest to the nearest even digit by use of a random-zero sphygmomanometer.
After an overnight fast period, 20 ml of blood were obtained from an antecubital vein without compression. Plasma glucose was determined in duplicate by a glucoseoxidase method adapted to an Autoanalyzer (Hitachi 704, Boehringer Mannheim, Germany). Total cholesterol, triglycerides and high-density lipoprotein (HDL) cholesterol were determined by enzymatic methods using commercial kits (Boehringer Mannheim, Germany). Low-density lipoprotein (LDL) cholesterol was calculated by the Friedewald formula. Serum insulin was determined by RIA (Human Insulin Specific RIA kit, Linco Research Inc., St Louis MO, USA).
Oral glucose tolerance test (OGTT) using 75 g of glucose was performed according to the WHO recommendations. Two hours after glucose administration, blood samples were obtained for the determination of glucose levels and interpreted the results according to Genuth et al. [21]. The crude prevalences of T2DM, impaired fasting glucose (IFG) and impaired glucose tolerance (IGT) were respectively 8.7%, 13.6% and 15.0%.
The metabolic syndrome status was established according to ATPIII definition as it was recently modified: presence of at least three components between abdominal obesity (WC ≥ 102 cm men, 88 cm women), hypertriglyceridemia (TGs ≥ 150 mg/dl), hypertension (≥ 85/130 mmHg), HDL-c (< 40 mg/dl men, < 50 mg/dl women) and fasting glucose ≥ 100 mg/dl [23]. We have also classified subjects according to the recently International Diabetes Federation (IDF) worldwide definition of the metabolic syndrome for Europid populations [24]. In this last definition, central adiposity (defined as WC ≥ 94 cm for men and ≥ 80 cm for women) is a prerequisite risk factor for the diagnosis of the syndrome; two of the following factors are also necessary: raised TG level (>150 mg/dl) or specific treatment for this abnormality; reduced HDL-c (< 40 mg/dl males, < 50 mg/dl females) or specific treatment; raised blood pressure (SBP ≥ 130 or DBP ≥ 85) or treatment of previously diagnosed hypertension, raised fasting plasma glucose (FPG ≥ 100 mg/dl) or previously diagnosed T2DM. The crude prevalence of metabolic syndrome under ATPIII criteria is 12.6% and 19.8% under IDF definition.

DNA extraction
DNA extraction from 5 ml of frozen peripheral blood was performed in a MagNa Pure LC Instrument (Roche Diagnostics), using MagNa Pure LC DNA Isolation kits (Roche Diagnostics) according to the manufacturer's instructions. Aliquots of DNA of 5 ng/ul were obtained to carry out PCR reactions.

Genotyping
We have analysed two polymorphisms in intron 1 (rs948976 A>G and rs4945140 G>A) and two polymorphisms located in the, intron 3 of CAPN5 gene (rs223546 C>T and rs223549 G>A) ( Figure 1). These last two SNPs are located in the coding region of the OMP gene, an intronic gene integrated in the non-coding region of CAPN5 gene, so the aminoacids encoded by the OMP gene are not present in the CAPN5 protein. These SNPs were selected because they had been associated with PCOS, an insulin resistance related phenotype, in a previ- ous work by our group [18]. The genotyping of the selected SNPs was performed according to described procedures [18].

Statistical analysis
To analyze differences in genotype distribution, chisquared association studies for metabolic syndrome, T2DM, IFG and IGT and test for deviation from Hardy-Weinberg equilibrium, we have used tests adapted from Sasieni [25] at the online resource available at the Institute for Human Genetics, Munich, Germany [26] or SPSS software (Ver. 11.0.0., LEAD Technologies, Inc).
For genotype quantitative analysis, we have performed an analysis of variance using the GLM procedure included in SPSS software (Ver. 11.0.0., LEAD Technologies, Inc). Normality of the dependent variables was assessed by Kolmogorov-Smirnov test and, when necessary, we applied mathematical transformations (natural logarithm or inverse of the square root) prior to the analysis. Each analysis was adjusted for age and sex. The adjusted percentage of the phenotypic variance explained by each genotype was estimated by subtracting the adjusted r2 value for a model that includes the genotype from the r2 for a model that excludes it. The Levene's test of equality of error variances was also performed for each analysis. In addition, all values from individuals under antihypertensive, antidiabetic or hypocholesterolemic medication were eliminated of the QTL analysis.
Linkage disequilibrium (LD) and haplotype analysis was performed using THESIAS software based on the SEM algorithm [27,28]. This method allows to estimate haplotype frequencies and haplotype effects by comparison to a reference (the intercept) taken here as the most frequent one. For quantitative analysis, haplotypes effects are Scheme of CAPN5 gene organization and LD pattern of CAPN5 with neighboring genes (region spanning 140 kb aproximately) Figure 1 Scheme of CAPN5 gene organization and LD pattern of CAPN5 with neighboring genes (region spanning 140 kb aproximately). Analysed polymorphism are indicated by triangles. expressed as increases/decreases of the phenotypic mean with respect to the intercept's one; for qualitative analysis, results are expressed as Odds ratio (OR). All the quantitative studies were corrected by age and sex. In a first approximation, all haplotypes identified were computed but, in order to minimize loss of power, haplotypes with a frequency lower than 3% were excluded from further analysis.
We have also performed a regression-based haplotype association test using Whap software [29,30] to confirm results obtained with the SEM approach. This software also allows executing a Monte-Carlo test to obtain the empirical p values. All values were standardized before the analysis according to software instructions.

Allele frequencies and genotype distribution
Polymorphism status and allele frequencies were determined in all 606 individuals. The observed allelic frequencies in our population were 0.74 for A allele in rs948976 A>G, 0.57 for G allele in rs4945140 G>A, 0.94 for C allele in rs223546 C>T and 0.80 for G allele in rs223549G>A. The rs223546 is the less informative one, with only 0.06 frequency for the T allele and an observed heterozygosity of 0.11. All genotype frequencies fit the Hardy-Weinberg equilibrium (p ≥ 0.129). Genotype distribution and Pearson's goodness-of-fit chi-square test for deviation from Hardy-Weinberg equilibrium are shown in table 2.

Quantitative genotype analysis
In this analysis we have found significant differences in BMI values associated with the two markers of the first cluster (
With this LD pattern, nine different haplotypes are possible with frequencies ranging from 1.5% to 28.5% ( There is also a trend toward being associated for BMI, but it did not reach the statistical significance level (χ 2 5d.f = 10.8, p = 0.056). No significant global associations have been found for any of the other traits analysed (p ≥ 0.081).
According to the global haplotypic analysis, none of the CAPN5 haplotypes are consistently associated with BMI values and only the AA-CA haplotype shows a suggestive association (p = 0.086), being this result significant under de regression-based model (empirical p-value after 5.000 permutations = 0.039). We determined the frequency of the AA-CA haplotype in obese (BMI ≥ 30 kg/m2, 148 indi- With respect to blood pressure, haplotypes AG-CA (p = 0.0005) and GG-CG (p = 0.006) are related to a significant decrease of DBP values (-14.2% and -6.1%, respectively). The, independent analysis of the two clusters showed that the second one is not significant whereas the global measure of association of cluster 1 is of a similar magnitude that in the full-length analysis (χ 2 5d.f = 7.84, p = 0.019) Haplotype GG-CA is associated with higher total (p = 0.001, +29.6%) and LDL-cholesterol (p = 0.008, +36.1%). In contrast, haplotype GG-CG is related to a reduction in both values (p = 0.029, -8.0% and p = 0.044, -10.3%). None of the regions are independently associated with the trait (global p ≥ 0.090), suggesting that both are necessary to modulate the phenotype and explaining the absence of association in the genotype analysis. No associations have been found for HDL-c or TG levels in the haplotypic analysis [see Additional files 10 and 12].
We have not found evidence of CAPN5 being related to glucose or insulin levels [see Additional files 5,6,7]. Although the AG-CA haplotype was significantly associated with fasting glucose and 2 h-glucose levels (p ≤ 0.034), the global haplotypic effect was not significant (p ≥ 0.107). In this way, T2DM, IFG or IGT do not seem to be related to CAPN5 alleles, although the number of affected individuals is small (49, 77 and 74 individuals respectively) [see Additional files 15,16,17].

Discussion
In this population-based survey, we have isolated CAPN5 as a candidate gene for cardiovascular risk factors such as BMI, diastolic blood pressure and cholesterol. In addition, the haplotype related to obesity (AA-CA) is more frequent in individuals with metabolic syndrome (ATPIII criteria), supporting the role of obesity as an important underlying risk factor for cardiovascular diseases [31,32]. Two groups of haplotypes can be drawn regarding cardiovascular disease risk factors: protective haplotypes (GG-CG and AG-CA) and risk haplotypes (AA-CA and GG-CA).  In a previous study, we analysed these four CAPN5 gene variants in PCOS women, an anovulatory disorder of women in which insulin resistance plays a central role [33]. In that work, the GG-CA haplotype, associated with cholesterol levels in this work, was over-represented in PCOS patients. In addition, we associated several CAPN5 haplotypes with phenotypic differences observed between PCOS patients, such as the presence of obesity (AA-CA) and familial antecedents of obesity (AA-CA and GG-CG), hypertension (AA-CA and GG-CG) and hypercholesterolemia (GG-CA). [18]. Presence of familial antecedents of hypertension in PCOS women is a frequent finding; the higher frequency of this haplotype among women with familial antecedents of hypertension suggested a hypertensive phenotype associated to this allele. However, in this work, a measure of the effect of this haplotype in a population-based analysis has revealed a blood pressure lowering effect of the GG-CG haplotype. It is possible that smaller sample size in PCOS study, the effect of PCOS phenotype or PCOS related alleles have been acting as confounding factors. Anyway, the identification of CAPN5 gene haplotypes affecting cardiovascular risk factors also in general population reinforce CAPN5 as a candidate gene for metabolic syndrome related phenotypes. It is unlikely that the OMP gene itself be causative for the described associations in our population, given that haplotype association analysis at OMP region (intron 3) was not significant for any of the traits analysed, being necessary the CAPN5 intron 1 region to modulate the phenotype.
Little is known about CAPN5 function in humans. It is expressed in almost all human tissues, including testis, ovary, liver, heart, colon, kidney and brain [5] [6]. Its homologue CAPN10 has been related to insulin secretion and action in both functional and genetic analyses [34]. Some authors have also pointed out a role for CAPN10 in adipocyte biology. Paul et al. [35] described how CAPN10 facilitates GLUT4 vesicle translocation during insulinstimulated glucose uptake in adipocytes. Another report by Hoffstedt et al. [36] describes association of CAPN10 gene with the lipolytic sensitivity of β3-adrenoreceptors in subcutaneous fat cells in overweight subjects. Calpains have been also associated with adipocyte differentiation [37]. Regarding hypertension, calpains have been involved in mitochondrial mechanisms of oxidative stress that leads to vascular dysfunction which causes systemic hypertension and compensatory cardiac stiffness and diastolic heart failure [38].
The CAPN5 ortologhe in C.elegans, TRA-3, mediates the proteolitic cleavage of TRA-2A, a transmembrane protein that binds and inhibits the activity of the male-promoting FEM proteins (FEM-1, FEM-2 and FEM-3) [7]. The proteolitic action of TRA-3 over TRA-2 protein; produce a soluble fragment that is delivered to the nucleus where it enhances the transcriptional activity of the zinc finger transcription factor TRA-1 which promotes female development. Because mechanisms of sexual differentiation are not conserved between nematodes and human, it is possible that this pathway has more general functions than in C.elegans. In fact, the finding of abnormal glucose homeostasis in the knockout mice for Fem1b gene, the identification of a germline missense mutation in FEM1A gene in a PCOS patient and the observation of FEM1A being downregulated in human Rhabdomyosarcoma cell lines, support this hypothesis [8,9,39]. In humans, CAPN5 is localized to the cytoplasma as well as to the nucleus [40]. It is possible that, in a similar way of its ortologue TRA-3, CAPN5 could act over nuclear proteins that affect DNA transcription such as transcription factors or related proteins. Calpains are known to cleave several transcription factors: c-Fos, c-Jun, Oct-1, p53 or C/EBP transcription factor family among others, are calpain substrates [41][42][43]. Calpains can also affect DNA transcription in an indirect way. For example, activation of NF-kappa B requires the proteolitic degradation of its inhibitor I kappa B by calpains [44].
It is also possible that the identified associations are not related to CAPN5 gene itself, but other upstream gene in linkage disequilibrium with the polymorphic variants analysed. However, HapMap shows that LD block that includes the 5' region of CAPN5 is interrupted 29 kb upstream CAPN5 5'UTR. This region only includes the gene that codifies B3Gn-T6, a member of the N-acetylglucosaminyl transferase protein family related to muscular dystrophies, neoplastic disorders and altered glucose homeostasis, although B3Gn-T6 has never been analysed in humans [45]. The adjacent 5' LD block comprises 193 kb and includes the PHCA gene that codifies an alkaline phytoceramidase, an integral membrane protein related to protein biosynthesis and lipid membrane ceramide metabolism [46]. Beyond this region, is not probably to detect association with our marker selection. The absence of association in the independent cluster analysis of BMI, cholesterol levels and diastolic blood pressure with the markers of the second LD block that contains the OMP gene, makes improbable that the associations described here can be due to this intragenic gene. To elucidate this question, further analysis is being performed.

Conclusion
Results obtained from our analysis in PCOS women and this population-based study suggest that CAPN5 is a gene related to metabolic syndrome and related phenotypes such as obesity, hypertension and hypercholesterolemia, at least in the Spanish population. These results are in accordance with the evidence of linkage of T2DM within the chromosomal region that contains CAPN5 gene reported by Silander et al. [19]. However, a confirmation of our findings in independent populations is necessary to clearly establish the role of CAPN5 in metabolic syndrome.