Effects of common haplotypes of the ileal sodium dependent bile acid transporter gene on the development of sporadic and familial colorectal cancer: A case control study

Colorectal cancer (CRC) is a growing burden on health services throughout Western countries. In contrast to familial CRC syndromes, little is known about the underlying molecular mechanisms in the majority of cases. However, the current paradigm is that CRC is due to the interaction of genetic and environmental factors [1]. The effect of each variable may vary considerably, ranging from monogenic familial cancer syndromes that are predominantly caused by genetic defects to sporadic CRCs, which are probably due to life-long exposure to environmental factors with only minor effects of genetic predisposition.

The genes underlying non-syndromic familial CRC are still unknown, or findings have not been reproducible among different studies. A common strategy in experimental genetics is to increase the power by identification of phenotypic extremes and employing association studies in these cohorts, since genetic effects can accumulate in these settings. We have recently successfully translated this strategy to the human situation [2, 3]. For the present study we chose to compare the effect of genetic variants in (i) familial CRC cases (after exclusion of Lynch's syndrome and familial adenomatous polyposis [FAP]) and (ii) sporadic CRC patients without a personal or familial history of CRC to a 'hyper-normal' control group with no signs of adenomas or cancer on screening colonoscopy and no personal or familial history of CRC.

Early clinical studies suggested that differences in bile acid metabolism and faecal bile acid composition between CRC patients and controls may contribute to cancer formation [4, 5]. In line with this observation, experimental data showed that bile acids can promote DNA adduct formation, induction of proliferation, inhibition of apoptosis, and promotion of tumor invasion [6]. Bile acid composition is influenced by environmental factors such as food intake as well as the bacterial flora in the gut [7, 8]. However, bile acid metabolism is also influenced by a number of transport proteins that mediate uptake, distribution or efflux of bile acids. Genetic variation in the genes encoding these transporters together with environmental factors might contribute to CRC development [9‐11]. In the intestine, primary bile acids are reabsorbed by active transport in the terminal ileum [12]. This ileal sodium dependent bile acid transporter (ISBT, synonym: apical sodium dependent bile acid transporter ASBT; gene code: SLC10A2) is located in the apical membrane of the enterocyte. Only a minor fraction of primary bile acids reaches the colon where the bacterial flora deconjugates the primary bile acids to secondary bile acids such as deoxycholic acid. Disruption of effective reabsorption diminishes the efficacy of the enterohepatic circulation of bile acids and increases the amount of secondary bile acids in the colon [13]. Of note, single nucleotide polymorphisms (SNPs) in the ileal bile acid transporter (SLC10A2) gene have been associated with the risk for development of sporadic colorectal adenoma, a precursor lesion for CRC [14]. It remains unclear whether genetic variation in this gene also contributes to the development of CRC.

A haplotype is defined as a specific combination of variant alleles that are inherited together on a single chromosome. In genetically complex diseases such as CRC, haplotypes might provide more information than the analysis of individual polymorphisms, and in some cases an association of a complex phenotype might even be missed if only single markers are analyzed instead of haplotypes [15]. Our aim now was to test whether common variants of the SLC10A2 gene are associated with sporadic and familial CRC, employing a haplotype-based association study.

A total number of 93 familial CRC, 150 sporadic CRC patients and 204 'hyper-normal' controls were included. Table 1 summarizes their clinical characteristics. 'Hyper-normal' controls were age-matched to sporadic CRC patients to reduce the possibility of inclusion of patients who might develop colorectal adenomas or CRC in later life. We found a slight over-representation of left sided tumors in familial CRC patients; in addition, these patients presented with more advanced tumor stages than sporadic CRC patients (Table 1).

Clinical and experimental data suggest a role of secondary bile acids as promoters of colorectal tumors [26, 27]. In contrast, water soluble bile acids may counteract, as they induce apoptosis and lower enterocyte proliferation rates in experimental tumor models and in vivo [28, 29]. Disturbance of the enterohepatic circulation, e.g. by impaired re-absorption of primary bile acids in the ileum due to genetic variants of the SLC10A2 gene, may lead to spillage of primary bile acids in the colon, where they are transformed into secondary bile acids by the endoluminal bacterial flora. In line with this concept, a recent study found an association of a specific mutation (c.507C>T;p.L169L) in the SLC10A2 gene and colorectal adenoma development [14]. However, no direct association was found with a nearby non-synonymous mutation (rs188096; c.511G>T;p.A171S), indicating that other variants in linkage disequilibrium to the associated locus may contribute to the effect. Although this study [14] found an association with colorectal adenomas, a CRC precursor lesion, the role of the observed association for the development of CRC remains unknown. The study of Wang et al. [14] also observed that a specific genotype combination of the silent mutation and the non-synonymous mutation is associated with colorectal adenoma development. Functional data for the two SNPs studied by Wang et al. are sparse. However, Oelkers et al. performed functional studies for c.511G>T;p.A171S in transfected COS cells but no effect on taurocholate uptake was noticed [14]. In our study, we reconstructed haplotypes of the locus covering the SNPs previously studied by Wang et al. and tested single SNPs as well as complete haplotypes in this block for association with familial and sporadic CRC [14].

By selecting tagging SNPs that covered a whole linkage disequilibrium block we should have been able to detect association with all common variants in linkage disequilibrium with the locus tested by Wang et al. [14]. Our study was sufficiently powered (see Methods and Figure 1), and we aimed to further increase the power of our analysis by testing phenotypic extremes (e.g. 'hyper-normal' controls and familial CRC patients), which is a common concept in experimental genetics [20, 30] and has been successfully applied by us in previous studies in our cohort of familial CRC patients [2, 3]. However, we could not detect such an association with any of the single markers or haplotypes. Hence, variation in the SLC10A2 gene appears to be associated with colorectal adenoma, but not with progression to colorectal carcinoma [14]. Our study is limited by a relatively small number of patients with familial CRC. In addition, our clinical information about the included cases is limited and information on life style such as preference of certain diets (e.g. high fiber vs. high fat/meat) and the presence or absence of cholecystectomy which may change bile acid turn over is totally lacking. Thus, we may have missed very small effects of SLC10A2 variants on CRC formation due to a type II error and we were unable to study genotype-phenotype interaction with environmental factors that may possibly influence bile salt metabolism.

In conclusion, development of sporadic and familial CRC is not associated with common haplotypes of the genomic region encompassing the SLC10A2 gene previously implicated in the development of colorectal adenomas. However, larger studies may be needed for the detection of very small effects of genetic variants in the SLC10A2 gene on CRC development.