Investigation of the Lith1 candidate genes ABCB11 and LXRA in human gallstone disease

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Genetic susceptibility in the causation of gallbladder diseases was recognized as early as 1937. A major gallstone susceptibility locus (Lith1) was identified in 1995 by quantitative trait locus mapping in mice. Two attractive positional and functional candidate genes in LXRA and ABCB11 are located in this interval. ABCB11 is associated with progressive familial cholestasis. This study was undertaken to investigate LXRA and ABCB11 as candidate genes for gallstone disease in humans. Eight hundred and ten patients who underwent cholecystectomy for symptomatic gallstone disease (median age of onset, 50 years) were compared with 718 sex-matched control individuals. Control individuals were sonographically free of gallstones. Haplotype tagging and all known coding single nucleotide polymorphisms (SNPs) were genotyped for ABCB11 (n = 29) and LXRA (n = 10). The investigated high-risk patient sample provides a power of greater than 80% for the detection of odds ratios down to 1.55. No evidence of association of the two genes in the single point tagging markers, coding variants or in the sliding window haplotype analysis was detected (all nominal single-point P values ≥ .08). In conclusion, in the investigated German sample, no evidence of association of ABCB11 and LXRA to gallstone susceptibility was detected. The gallstone trait is not allelic to progressive familial cholestasis at the ABCB11 locus. Systematic fine mapping of the Lith1 region is required to identify the causative genetic variants for gallstone in mice and humans. (HEPATOLOGY 2006;44:650–657.)

Cholelithiasis represents a frequent and economically relevant health problem in industrialized countries. A number of demographic and environmental risk factors, including age, female sex, and body mass index,1 have been identified. A genetic component in the susceptibility to gallstones has been recognized as early as 1937.2 The initial studies have used pathological section statistics2; more recent studies have turned to ultrasound surveys to investigate the familial clustering of cholelithiasis.3 Despite the heterogeneous study designs, there is both compelling evidence for familial clustering and an increased concordance of the trait in monocygotic twins as compared with dizygotic twins.4 Thus, the risk of gallstone disease has a strong genetic component.

Systematic, genome-wide investigation of gallstone traits in inbred mouse strains has yielded a susceptibility map of “Lith” loci.5–9 Several of these loci have been found in different mouse crosses, thus strengthening the case for robust susceptibility loci in the mouse.9 The human orthologs of these loci are thus natural positional candidates for gallstone susceptibility in humans. The Lith1 locus harbors two positional candidates that also are very interesting from a functional point of view: The bile salt export pump BSEP [HUGO gene symbol: ABCB11: ATP-binding cassette, sub-family B (MDR/TAP), member 11] and the liver X receptor alpha LXRA (HUGO gene symbol: NR1H3: nuclear receptor subfamily 1, group H, member 3).

The ABCB11 gene maps to the human chromosome 2q24, which corresponds to the mouse chromosome 2C2. It is a highly conserved member of the multi-drug resistance (MDR) gene family of adenosine triphosphate–binding cassette transporters. ABCB11 shares a high homology with MDR1. However, compared with MDR1, it does not show broad substrate specificity and mainly recognizes bile acids.10, 11 It is expressed exclusively at the canalicular membrane of hepatocytes and seems to be the predominant bile salt efflux system of hepatocytes.12 In addition, mutations in ABCB11 are associated with familial intrahepatic cholestasis.13–16 The liver X receptor alpha (LXRA) is located on human chromosome 11p11.2—its mouse homolog is also located in the mouse Lith1 locus at chromosome 2E1. LXR is a pleiotropic transcription factor that regulates a wide array of genes in cholesterol17 and bile transport and metabolism.18 Specifically, a number of ABC transporters are regulated by this gene, including ABCG119 and ABCD2.20 No large investigation of mutations in the ABCB11 or LXRA genes for human gallstone susceptibility has been published.

Despite the advanced stage of genetic investigation in the mouse system, a surprising lack of genetic studies in the human system have been performed. An association of pigment gallstones with a promoter variation in the UDP-glucuronosyltransferase 1A1 (UGT1A1) gene has been reported in patients with chronic hemolytic disorders21, 22 and cystic fibrosis.23 Investigations of the 7-alpha-hydroxylase and apolipoproteins A and B in 105 to 210 patients24–27 and apolipoprotein E in 37 to 169 patients28, 29 did not yield replicated association findings. A recent genome-wide linkage scan in Mexican Americans identified significant linkage (LOD 3.7) of gallstone susceptibility to chromosome 1p.30 This represents a genome-wide investigation of gallstone susceptibility in humans and is an excellent starting point for further systematic mapping efforts.

In this report, we have investigated two functional candidate genes within the mouse Lith1 locus for association with symptomatic gallstone disease in a large sample of cholecystectomy patients with a young age of onset for gallstone disease.

Abbreviations

MDR, multi-drug resistance; LXRA, liver X receptor alpha; SNP, single nucleotide polymorphism.

Materials and Methods

Patients and Phenotypes.

Patients who underwent cholecystectomy for cholecystolithiasis from 2001 to 2005 in the surgical departments of the hospitals at Kiel, Lüneburg, Rendsburg, Schleswig, Flensburg, Husum, Heide, and Eckernförde (all in Northern Germany) were contacted through mail by the respective hospitals and invited for participation in this study. For patients who did not respond, one written reminder was sent. Individuals who agreed to participate were contacted by the POPGEN (http://www.popgen.de) recruitment project31: They were interviewed by mail questionnaire, and venous EDTA blood sample was obtained either at the POPGEN offices or by the patients' general practitioners. For both cases and controls, the study was restricted to probands of German ethnicity; in other words, only individuals whose parents were born in Germany were included. The study protocols were approved by the institutional ethics committee and the public data protection agency. Written informed consent was obtained from all study participants. The first consecutive 810 participants were included in this study.

Gallstone-free control individuals were recruited from consecutive patients in the years 2003 to 2004 from the routine clinical ultrasound of the Department of General Internal Medicine Kiel. Patients with malignant disorders were excluded. Index patients were identified on the basis of the ultrasound reports, and n = 277 probands were recruited using the POPGEN infrastructure employing the protocols described previously. Additional controls, N = 441, were obtained from population-derived control individuals from the POPGEN project (patients identified on the basis of the inhabitants register31), who underwent an additional physical examination at the POPGEN facilities that included an abdominal ultrasound by a trained physician. No significant difference was seen in allele frequencies between the two control samples (P > .1).

DNA was prepared using the FlexiGene chemistry (Qiagen, Hilden, Germany) according to the manufacturer's protocols from all samples. An overview of the patient sample is given in Table 1.

Table 1. Overview of the Patient Sample
SampleNMedian Age (y)%Male
  1. NOTE. The median age of onset in the cases was 50.

Cases8105550%
Controls7186450%

Genotyping.

DNA samples were evaluated by gel electrophoresis and adjusted to 20 to 30 ng/μL DNA content using Picogreen fluorescent dye (Molecular Probes, Invitrogen, Carlsbad, CA). One microliter genomic DNA was amplified by the GenomiPhi (Amersham, Uppsala, Sweden) whole genome amplification kit and fragmented at 99°C for 3 minutes. One hundred nanograms of DNA were dried overnight in TwinTec hardshell 384well plates (Eppendorf, Hamburg, Germany) at room temperature. Genotyping was performed on these plates using the SNPlex chemistry (Applied Biosystems, Foster City, CA) on an automated platform with TECAN Freedom EVO and 384well TEMO liquid handling robots (TECAN, Männedorf, Switzerland). Genotypes were reviewed manually using the Genemapper 4.0 (Applied Biosystems) software. All process data were logged and administered through a database-driven LIMS system.32

Single Nucleotide Polymorphism Selection and Data Analysis.

Single nucleotide polymorphisms (SNPs) were selected from HAPMAP (www.hapmap.org) by the automated selection of haplotype tagging SNPs for Caucasians from the CEU dataset (settings: Mendel errors: 0, minor allele frequency: 0.01, HWE cutoff: 0.01). In addition, coding SNPs with a frequency > 0.01 were selected from dbSNP and the web resources from the “Pharmacogenetics of Membrane Transporters” project at http://pharmacogenetics.ucsf.edu/data.html.

The study was performed in case-control design: Sliding window haplotype analysis was performed using COCAPHASE through the UNPHASED suite of programs (http://www.rfcgr.mrc.ac.uk/∼fdudbrid/software/unphased/).33 Single point genotypic and allelic tests of association on sex-matched genotype results were performed using chi-squared statistics on contingency tables. For all tests, nominal P values are reported.

Results

A power estimation in the utilized sample of 810 cases and 718 controls was performed for single-point allelic effects over an odds ratio of one to two at a nominal significance level of 0.0534 for (minor) allele frequencies of 0.1, 0.2, and 0.5 of a potential susceptibility marker. As shown in Fig. 1, the power for the detection of odds ratios greater than 1.55 is approximately 80% or greater under all models. For more frequent susceptibility factors, odds ratios of 1.4 also would be detectable at this power level. No marker showed departure from Hardy-Weinberg equilibrium (P > .1), indicating robust genotyping performance in this experiment.

Figure 1.

Graphical representation of a power estimation in the sample size at a significance level of P < .05 for a two-sided test over an odds ratio range of 1 to 2. The graph was generated using PS-power34 and shows the test power as a function of the odds ratio (x-axis).

For both candidate genes, haplotype-tagging SNPs were generated using the HAPMAP genotype data (http://www.hapmap.org) for the candidate gene regions.35 Twenty-nine SNP markers in the ABCB11 gene were selected. These 29 markers provide good coverage as determined by the tagging functionality36 in HAPLOVIEW37 (settings: see Materials and Methods). Figure 2 shows the distribution of markers across the gene and the regional haplotype structure as generated from the HAPMAP Caucasian genotypes (category CEU) by HAPLOVIEW. As shown in Fig. 2, these markers tag all major haplotype blocks of this gene. Additionally, all coding SNPs that were polymorphic in whites with an allele frequency greater than 0.01 were included in the genotyping—rs2287622 is a common coding SNP that leads to the exchange of alanine to valin at position 444 in the protein. The allele frequency reported in HAPMAP (40.1%) corresponded very closely to the frequencies observed in this experiment with a minor allele frequency of 40% in cases and 43% in control individuals. The relatively rare coding SNP rs11568364 was also included in the panel and showed no difference in allele frequency between cases and controls (0.015 vs. 0.012). The remaining annotated coding SNPs (e.g., rs1521808, rs11568364, rs2287617, rs11568357) in the databases and on the dedicated transporter websites (http://pharmacogenetics.ucsf.edu/data.html) were not polymorphic in whites. Their database annotation as polymorphisms is based on population differences between Asians, Africans, and whites. The SNP panel was genotyped in 810 cases and 718 controls (Table 1). Single-point allelic and genotypic association tests were performed, yielding nominal P values ranging from 0.08 to 0.95. To improve power, that is, to detect association to variants on the haplotypes not directly tagged by one of the SNPs in the experiments, a sliding window haplotype analysis using window sizes of two to five markers was performed. Here, a nominal significance level of .04 was obtained for the window size 5 for the haplotype spanning from rs2287622 to rs13416802. None of the neighboring haplotypes showed evidence of association. The frequent coding SNP rs2287622 was not associated with gallstone risk in the single-point allelic and genotypic analysis. Table 2 lists all results for this gene.

Figure 2.

Overview of the physical and genetic structure of the ABCB11 gene region: The gene is annotated on the genomic minus strand. The physical position of the SNPs investigated and a schematic chart of the gene structure are shown in the top panel. The coordinates refer to the genome assembly build 34. The lower panel gives an overview of the linkage disequilibrium structure of the locus (D′) as generated by Haploview37 from the Caucasian HAPMAP data.

Table 2. Results of the Genetic Association Analyses of the Tagging SNPs in the ABCB11 Gene
dbSNP idPositionMAFcaseMAFcontORRecORdomPallelicPgenoHAP2HAP3HAP4HAP5
  • *

    No odds ratio calculated because of low allele frequency.

  • #

    P value reported for the first marker in the haplotype window.

  • NOTE. The minor allele frequencies (MAF) for cases and controls are reported. ORdom and ORrec refer to the odds ratios under the recessive or dominant modes of inheritance, respectively. P values are reported for allelic (pallelic) and genotypic association (pgeno). The columns HAP2 to HAP5 refer to a sliding window haplotype analysis using COCAPHASE.

rs5636943′ Flanking region0.360.351.071.030.680.920.380.540.720.68
rs4783333′ Flanking region0.470.480.930.930.610.800.850.890.920.74
rs7572878Intron0.0040.004n/a*1.040.950.950.540.970.890.82
rs6709087Intron0.250.231.121.130.280.500.760.920.760.77
rs4148797Intron0.30.311.040.950.880.740.950.600.710.27
rs3770582Intron0.390.410.940.820.720.340.600.590.51
rs4148790Intron0.390.420.80.830.080.190.290.190.270.14
rs17267869Intron0.160.151.741.090.270.290.570.580.240.29
rs3770585Intron0.440.431.071.080.590.790.640.120.290.39
rs11568364Coding [met→val]0.0150.012n/a*1.240.500.490.150.260.230.17
rs2287622Coding [ala→val]0.400.430.790.840.100.250.310.290.120.04
rs2058996Intron0.470.470.960.940.730.830.490.450.780.74
rs3770589Intron0.470.461.11.120.480.600.370.470.420.62
rs3770594Intron0.450.441.031.060.790.820.530.420.780.93
rs13416802Intron0.140.121.31.20.130.310.240.520.830.62
rs7563233Synonymous SNP0.020.02n/a*1.380.250.240.570.870.740.90
rs2287618Intron0.320.330.890.990.620.690.860.580.910.71
rs6433102Intron0.250.270.820.950.420.600.120.290.290.26
rs10196426Intron0.050.041.361.390.080.220.390.250.220.23
rs10209995Intron0.030.03n/a*1.140.500.570.510.470.500.51
rs3770601Intron0.10.081.831.180.180.360.450.500.420.36
rs4148772Untranslated0.030.02n/a*1.170.430.540.410.350.410.41
rs3755161Untranslated0.030.02n/a*1.180.410.520.350.410.410.32
rs11892966Untranslated0.050.041.351.260.210.450.330.340.200.24
rs7602171Untranslated0.330.350.830.990.440.320.260.140.230.10
rs13430236Untranslated0.490.481.041.140.710.310.220.250.40#
rs42338235′ Flanking region0.060.070.860.770.090.220.110.26##
rs21610375′ Flanking region0.450.460.910.990.560.630.84###
rs104901355′ Flanking region0.450.431.061.190.460.16####

Ten SNPs were selected for LXRA using the same approach as described for ABCB11. No published or database coding SNPs exist for this gene. Therefore, a pure haplotype tagging approach was chosen. Single-point analyses for genotypic and allelic association yielded nominal P values in the range of 0.12 to 0.94. The sliding window-haplotype analyses resulted in nominal significance levels of 0.14 to 0.87. The SNP overview (Fig. 3) and all results of the association analysis are presented in Table 3 for LXRA.

Figure 3.

Overview of the physical and genetic structure of the LXRA gene region: This gene is transcribed from the genomic plus strand. The panels are otherwise organized as in Fig. 2.

Table 3. Results of the Genetic Association Analyses of the Tagging SNPs in the LXRA Gene
dbSNP idPositionMAFcaseMAFcontORRecORdomPallelicPgenoHAP2HAP3HAP4HAP5
  • NOTE. Column headings are identical to the ABCB11 gene overview.

  • #

    P value reported for the first marker in the haplotype window.

rs73955815′ Flanking region0.350.341.140.990.700.570.870.770.660.66
rs29578735′ Flanking region0.240.231.3710.610.390.500.610.520.70
rs119885′ Flanking region0.370.390.860.860.220.380.370.430.600.76
rs110391435′ Flanking region0.050.050.880.920.620.880.660.680.500.54
rs37586685′ Flanking region0.160.171.000.960.780.940.730.420.590.60
rs10838681Intron0.300.291.141.060.490.780.340.440.480.17
rs11039149Intron0.270.290.90.820.120.130.320.410.14#
rs7120118Intron0.350.341.151.010.620.640.360.23##
rs105013203′ Flanking region0.260.280.870.840.140.240.40###
rs22911193′ Flanking region0.270.261.211.030.590.69####

Discussion

The investigation of gallstone susceptibility in mice has yielded multiple genomic risk regions for this disorder in this animal model. “Lith1” is longest known among these and has been identified in independent mouse strains.5, 9 The investigation of human homologs of the most striking candidate genes in these intervals represents an intuitive approach to potential risk variants for the common gallstone trait in humans.9, 38 A special strong case can be made for ABCB11, because mutations in this gene are associated with familial intrahepatic cholestasis.13–16

The study reported here has used more than 800 patients who have been operated on for symptomatic gallstone disease and thus represents the largest case-control study in this trait. The population median of affection in North Germany for gallstone is approximately 651; that is, approximately 50% of patients develop the disease up to this age. To further improve power, only patients with an age of onset younger than 65 years were included in this study, yielding a median age of onset of 50 in cases. In a further attempt to maximize the power of this investigation, only gallstone-negative (as determined by abdominal ultrasound investigation) control individuals with a moderately higher median age (64 years) were used. Many polygenic disorders are characterized by a strong correlation between the age of onset of affected relatives, as has been documented, for instance, in breast cancer39 and Alzheimer disease.40 The genetic influence on the development of such disorders may be reflected by the age at which individuals develop the disorder.41 The gallstone hazard ratio therefore may be higher in these younger patients than the background familial risk of 2 to 4 reported for gallstones.2, 3, 42–46 Thus, the investigated sample represents a “high-risk” group of patients that should provide an improved experimental power as compared with a general population sample. The sample size used in this study has a power of greater than 80% for the detection of effects with allelic odds ratios of 1.4 to 1.55 as shown in the formal power calculation for a significance level of .05 (Fig. 1). This range of odds ratios corresponds to findings in other complex disorders that could be replicated later in independent populations such as CARD15 (allelic OR approximately 2.5)47–49 and the 5q31 risk haplotype (allelic OR approximately 1.6) in Crohn disease.50, 51 In an additional effort to maximize power, nominal P values were reported for all tests, including the haplotype associations.

The international HAPMAP project (www.hapmap.org) has generated a wealth of genotype and marker information that significantly aids in the design of candidate gene studies.35 For this candidate gene study, a primary haplotype tagging approach was chosen; that is, the genetic variation at both loci was captured with a set of carefully selected variants. This tagging approach is able to detect signals from undisclosed regulatory or functional elements in a genetic region.36, 52, 53 It thus potentially offers even advantages over a direct mutation detection of the coding region of a gene only, because disease susceptibility also may be conferred by other variations, for instance, in splice sites or intronic enhancers.54, 55 Tagging SNPs were selected from the public HAPMAP (www.hapmap.org) SNP resources.35 The genotype and allele frequencies for the SNPs investigated here were not significantly different from the ones for the Caucasian HAPMAP individuals (P > .11; allele frequencies in Tables 2 and 3), thus supporting the selection of variants from this resource. The ABCB11 locus has been subjected to extensive specific mutation detection efforts—the two Caucasian SNPs with a population frequency greater than 0.01 were also incorporated into this study.

None of the single-point nominal P values in any of the tests passed the P < .05 threshold. No correction for multiple testing was employed to maximize the power of the analyses.

In summary, this association study is negative for both genes. Thus, in the investigated German sample, these two loci do not contribute as major risk factors for the common gallstone trait within the power considerations quoted. This result has some important implications: First, it shows that the common gallstone trait in humans is not allelic to the familial cholestasis syndromes at the ABCB11 locus, which is also supported by the very different clinical presentation of the two entities. Second, the mouse Lith1 locus clearly spans a large amount of genomic sequence in the mouse and, consequently, also in the human homolog. Thus, further fine mapping of the mouse loci and identification of the causative variants in the animal system is needed to successfully transfer these discoveries to the human disease. In fact, potentially, association mapping of the respective regions in the human may offer advantages because of the lower level of disequilibrium in human populations and the abundant genomic resources in humans. A combination of these approaches may ultimately be able to untangle the genetic basis of gallstones.

Acknowledgements

The cooperation of all patients, their families, and physicians who participated in this study is gratefully acknowledged. Especially, the help of Birgit Timm and Huberta von Eberstein (POPGEN Biobank project), and of the heads of the surgical departments Ilka Vogel (Städtisches Krankenhaus Kiel), Hermann Dittrich (Rendsburg), Jürgen Belz (Husum), Rainer Quäschling (Eckernförde), Hodjat Shekarriz (Schleswig), Volker Mendel (Flensburg), Werner Neugebauer (Flensburg), Friedrich Kallinowski (Heide), Anton Schafmayer (Lüneburg) Jiri Klima (Niebüll), Marco Sailer (Hamburg-Bergedorf) is greatfully acknowledged.

Ancillary