Background

The pathogenesis of gallstone disease is complex and a variety of environmental predisposing factors such as obesity and rapid weight loss, nutrition, certain medications and number of pregnancies have been identified [1–6]. However, human and murine data suggest a strong genetic component for the risk of gallstone formation [7–18]. Very recently, Renner et al. have identified SLC10A2 (apical sodium-dependent bile acid transporter; protein name ASBT) as a novel susceptibility gene for cholelithiasis in humans [19]. SLC10A2 encodes the cholangiocyte bile salt transporter protein whose expression is reduced by a lithogenic diet in mice [20] and mediates intestinal bile acid absorption [21, 22]. ASBT is regulated by changes in gene expression in response to biliary bile salt concentration and inflammatory cytokines and is thought to enable cholangiocytes to sense biliary bile salts in order to activate intracellular signaling pathways [23] and to promote cholehepatic shunting of bile salts. Furthermore, ASBT expression in the cholangiocyte apical membrane is regulated by secretin [24]. It has been shown that in gallstone patients, ileal ASBT expression is diminished, and that this is associated with low cytosolic ileal lipid binding protein (ILBP) and basolateral organic solute bile acid exporter (OSTα-OSTβ) expression indicating impaired enterohepatic circulation of bile salts at least in a subset of patients with cholelithiasis [25, 26]. The regulation of ASBT expression also appears weight specific and so far, a diminished ASBT expression has only been shown in non-obese gallstone patients [25, 26].

The recent study by Renner et al. suggested that a single nucleotide polymorphism (SNP) rs9514089 mapping within the SLC10A2 locus is a genetic determinant of gallstone disease, expressing gender and weight specificity with higher risk observed in men and in normal-weight subjects [19]. Since the sample sizes of the two cohorts from Germany included in the recent study were rather small (N = 127 from Stuttgart and N = 368 from Aachen), we assessed the effects of rs9514089 on gallstone risk and related phenotypes of the metabolic syndrome in the self-contained population of Sorbs and performed a meta-analysis for effects of rs9514089 on susceptibility for cholelithiasis in the three independent cohorts (Stuttgart, Aachen and Sorbs).

Methods

Subjects

All subjects are part of a sample from an isolated population in Germany, the Sorbs [27, 28].

At present, about 1000 Sorbs are enrolled in the study. Sampling comprised unrelated subjects as well as families. Extensive phenotyping included standardised questionnaires for past medical history and family history, collection of anthropometric data (weight, height, waist-to-hip-ratio, body impedance analysis (BIA)) and a 75g-Glucose-tolerance-test (OGTT). Insulin was measured with the AutoDELFIA®Insulin assay (PerkinElmer Life and Analytical Sciences, Turku, Finland). Serum glucose was measured by the hexokinase method (Automated analyser Modular, Roche Diagnostics, Mannheim, Germany). Cases were defined as subjects with gallstones in the ultrasonic testing or those who underwent cholecystectomy due to symptomatic gallstones. Subjects with ultrasonic exclusion of gallstones were included as controls in this study. There were no other specific inclusion and exclusion criteria. A summary sample description is presented in Table 1. Currently, the entire family structure of all subjects is not yet known therefore, the estimated effect sizes might be biased by cryptic relatedness.

Table 1 Main characteristics of the Sorbs
Full size table

The study was approved by the ethics committee of the University of Leipzig and all subjects provided written informed consent before taking part in the study.

Genotyping of rs9514089

Genotyping of rs9514089 was performed using the TaqMan allelic discrimination assay (Assays-on-Demand (TM), SNP Genotyping Products; Applied Biosystems, Inc.) on an ABI PRISM 7500 sequence detector (Applied Biosystems Inc.) according to the manufacturer's protocol. The genotype distribution was consistent with Hardy-Weinberg equilibrium (minor allele frequency = 35%; p > 0.05). Genotyping success rate was >99%, and duplicate genotyping concordance was 100%.

Statistics

Standard descriptive and comparative statistics (ANOVA) were used to characterize and compare clinical parameters in cases and controls Genetic associations were assessed by linear or logistic regression using an additive model of inheritance unless stated otherwise, and adjusted for age, gender and BMI. All effect directions were standardized to the minor allele.

To obtain the combined effect of the three cohorts we performed a meta-analysis using the metan command in STATA based on the estimated effect sizes of each study and their 95% confidence intervals. The meta-analysis was performed in a fixed effect model by using the Mantel-Haenszel method.

All statistical analyses were performed using the SPSS 15.0.1 software package (SPSS, Inc.; Chicago, IL, USA) and STATA version 9.0, (StataCorp LP, Texas, USA).

Results

Associations of SLC10A2 variant rs9514089 with gallstones in the Sorbs

In our study, which includes 826 controls and 183 patients with gallstones, rs9514089 did not show any significant effect on gallstone prevalence, neither in the additive nor in the recessive or dominant mode of inheritance (all p > 0.05). In the subgroup of subjects with BMI ≤ 26 kg/m2 variant rs9514089 tended to be associated with gallstones (p = 0.05, OR = 0.57), whereas there was no effect in the group with BMI > 26 kg/m2 (p = 0.52) (Table 2). In the subgroup of females with BMI ≤ 26 kg/m2 the effects on gallstone risk reached nominal level of significance (p = 0.045, OR = 0.51, 188 controls, 102 cases, N(GG carriers) = 38, N(GA carriers) = 124, N(AA carriers) = 131), but would not sustain correction for multiple testing.

Table 2 Association of rs9514089 with gallstones in the Sorbs
Full size table

Association with extended phenotypes in the Sorbs

There was no significant association of rs9514089 with serum parameters of lipid metabolism (p = 0.17 for total cholesterol, p = 0.78 for HDL-cholesterol, p = 0.10 for LDL-cholesterol, p = 0.51 for triglycerides, p = 0.17 for APO-B) in the subgroup of Sorbs without lipid-lowering medication (Table 3).

Table 3 Association of rs9514089 with metabolic traits in the Sorbs
Full size table

Rs9514089 did not show any effects on BMI or fat mass, neither in the total cohort, nor in females or males separately. In a subgroup of subjects with available birth weight data (285 cases, 35 controls), the SNP tended to be associated with birth weight (p = 0.03). Gallstone carriers were characterized by increased age, BMI, fat mass, waist- and hip circumference (Table 1). In addition, women were affected more frequently than men.

Meta-analysis

In a combined analysis of the Sorbs and two previously published cohorts from Aachen and Stuttgart [19] there was no significant effect of the rs9514089 genotype on gallstone risk (Figure 1). Also the stratified analyses and assessment of different genetic models did not reveal any consistent significant effect on gallstone risk (Table 4). Furthermore, rs9514089 did not show any significant effect on serum cholesterol levels in the combined analysis of the three cohorts (p = 0.91, beta 0.001 [-0.014; 0.016] in the additive model) (Table 5).

Figure 1
figure 1

Meta-analysis of effects of rs9514089 on gallstone risk in three independent cohorts.

Full size image
Table 4 Meta-analysis of effects of rs9514089 on gallstone risk in the Sorbs, Aachen cohort and Stuttgart cohort
Full size table
Table 5 Meta-analysis of effects of rs9514089 on serum cholesterol and triglyceride levels in the Sorbs, Aachen cohort and Stuttgart cohort
Full size table

Discussion

Recently, SLC10A2 was suggested as a novel susceptibility gene for cholelithiasis in humans [19]. The data by Renner et al. indicated that the SNP effects were most pronounced when calculated in the recessive mode of inheritance, i.e. the risk of gallstone development was highest in subjects homozygous for the rs9514089 G allele. Therefore, we performed a meta-analysis including three independent cohorts (previously published cohorts from Stuttgart and Aachen and the cohort of Sorbs) and assuming a recessive mode of inheritance. Although there was no significant difference in minor allele frequencies between the three cohorts the risk allele was the G allele in the Stuttgart and Aachen cohorts but the A allele in the Sorbs. This effect is most likely driven by the small samples sizes in each genotype group in the cohorts. Furthermore, gene-environmental interactions based on factors such as eating behavior and nutrition composition as well as gene-gene interactions should be taken into account. In our study the meta-analysis of all three cohorts did not indicate any significant effect of the rs9514089 genotype on gallstone risk. The Sorbs and the Aachen cohort showed a trend of association in the subgroup of non-obese individuals indicating a possible relationship between the SNP effect and obesity as suggested by Renner et al. [19]. This would be in line with the reported weight specific regulation of the ASBT expression. However, so far, a diminished ASBT expression was confirmed only in non-obese gallstone patients [25, 26]. Also, since effect direction of the rs9514089 variant is different in the Sorbs the meta-analysis did not reveal any significant association in the combined analysis of all three cohorts.

In the Sorbs a nominal association of the SLC10A2 genotype and cholelithiasis was observed in the non-obese subgroup, even though the SNP effect could only be detected in women. This may be due to skewed gender distribution in favor of women and thereby greater statistical power for the female subset. The differences between the populations do not necessarily represent population specific effects of the gene variant. Consistent with the effect in the entire cohort the effect direction also in this subgroup was opposite in the Sorbs compared with the Aachen cohort. This could be attributed to population specific environmental factors or a different phenotype distribution in each population or it could be simply due to the lack of power based on the small samples sizes of the cohorts.

Furthermore, our meta-analysis did not confirm the reported effects of rs9514089 on serum total cholesterol [19]. However, the link between circulating lipid levels and development of gallstones has not been completely elucidated.

Conclusion

The effect of rs9514089 genotype on gallstone risk was not replicated in the Sorbs. Further analyses in larger cohorts are required to finally assess the role of genetic variants in SLC10A2 in human gallstone development and lipid metabolism