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.

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 glucose-oxidase 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%.

Insulin resistance was estimated by the homeostasis model assessment (HOMA-IR) method according to the formula: Insulin (mU/l) × Glucose (mmol/l)/22.5 22.

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 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.

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 previous work by our group 18. The genotyping of the selected SNPs was performed according to described procedures 18.

To analyze differences in genotype distribution, chi-squared 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 2728. 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 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 2930 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.

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.

In this analysis we have found significant differences in BMI values associated with the two markers of the first cluster (table 3). For the rs948976 A>G polymorphism, wild type homozygotes have the highest BMI values (means: AA, 27.4 ± 3.7; AG, 26.6 ± 4.1; GG, 26.8 ± 3.5 kg/m2; p = 0.035, r2 = 1.5%), whereas the presence of the polymorphic A allele at the rs4945140 G>A locus is associated with increased BMI values, especially in the homozygous state (means: GG, 26.5 ± 3.7; GA, 27.2 ± 3.8; AA, 27.8 ± 4.2 kg/m2; p = 0.041, r2 = 2.0%). The rs4945140 G>A marker is also associated with DBP values, being again the polymorphic A allele which confers increased risk (means: GG, 75.3 ± 8.7; GA, 77.7 ± 7.9; AA, 77.9 ± 7.8 mmHg; p = 0.015, r2 = 2.0%). Finally, HDL-c levels are associated with rs223549G>A genotypes; raised HDL-c values are related to the A allele (means: GG, 58.7 ± 17.2; GA, 64.0 ± 21.4; AA, 65.9 ± 18.8 mg/dl; p = 0.025, r2 = 1.7%). We have not find any other significant associations at genotype levels.

The four polymorphisms analysed are in two LD blocks: rs948976 A>G and rs4945140 G>A in intron 1 (D' = -1, r² = 0.26) and rs223546 C>T and rs223549G>A in intron 3 (D' = -1, r² = 0.02), separated from each other by a recombination hot-spot (D' between blocks = 0, r2 = 7.8 × 10^-5). With this LD pattern, nine different haplotypes are possible with frequencies ranging from 1.5% to 28.5% (table 4). The leading results of haplotype analysis are shown in table 5; the complete study is available online [see Additional files 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]. The highest global haplotypic effect was found for DBP (χ²_5d.f = 14.96 with, p = 0.010) followed by total cholesterol (χ²_5d.f = 11.43, p = 0.043). There is also a trend toward being associated for BMI, but it did not reach the statistical significance level (χ²_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 individuals) and non-obese subjects, being this haplotype over-represented in obese population (OR = 1.63 [1.05–2.54], p = 0.029). The haplotype analysis of BMI values in the two clusters independently (intron 1 from one side and intron 3 from another) showed that only the first cluster is informative for BMI (χ²_2d.f 6.07, p = 0.048) according to the genotype analysis. We have not found evidence of association of CAPN5 alleles with abdominal obesity estimated by waist circumference [see Additional file 2].

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 (χ²_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].

Considering metabolic syndrome under de ATPIII definition, we have found association with the obesity related AA-CA haplotype (p = 0.029, OR = 1.80 [1.06–3.06], n = 90), although the global haplotypic effect is not significant (p = 0.166). For the metabolic syndrome diagnosed following the IDF criteria, the global haplotypic effect was significant (χ²_5d.f = 11.74, p = 0.039), but only a suggestive protective effect for the GG-CG haplotype could be determined (OR = 0.63 [0.37–1.08], p = 0.092).