Herbert and coworkers 1 conducted a genome-wide association study (GWAS) on 694 subjects of the Framingham Heart Study with a family-based design to identify genetic loci (single nucleotide polymorphisms, SNPs) contributing to obesity using body mass index (BMI) as the primary phenotype. The authors reported strong statistical evidence that the SNP rs7566605 located near the 5' end of INSIG2 (Chr2:118,552,255) was associated with increased BMI 1 . In this initial study, association of rs7566605 with BMI was replicated in ethnically distinct populations, including European and African Americans. INSIG2 has been functionally linked to lipid metabolism, due to its role in endogenous cholesterol and fatty acid synthesis feedback inhibition 2 . It is an endoplasmic reticulum membrane bound protein that inhibits the proteolytic activation of Sterol Response Element Binding Proteins (SREPs) in response to cholesterol or insulin 2 . Thus, the biological function of the INSIG2 gene product is consistent with the initially observed association.

A series of follow-up studies sought to validate the association of INSIG2 rs7566605 with markers of obesity in humans, with several publications reporting inconsistent findings 3 4 5 6 7 8 9 10 11 12 13 .

There is ample evidence that excessive body weight is a risk factor for a number of diseases. In particular, the evidence that body fatness is a cause of postmenopausal breast cancer (BC) is convincing 14 . On the other hand, body fatness is inversely associated with BC risk in premenopausal women 14 .

Body fatness directly affects levels of many circulating hormones, such as insulin, insulin-like growth factors, and oestrogens, creating an environment that favors carcinogenesis and hinders apoptosis (WCRF/AICR, 2007). It also stimulates the body's inflammatory response, which may contribute to the initiation and progression of several cancers (WCRF/AICR, 2007).

We tested the SNP association with the body-mass index (BMI) and with BC risk in a large study of 3,973 women nested within the European Prospective Investigation into Cancer and Nutrition (EPIC).

We explored the association between a SNP in the INSIG2 gene promoter (rs7566605, Chr2:118,552,255) and BMI, as well as with BC risk. We included 1,269 incident BC cases from the EPIC cohort and 2,194 matched controls for the risk and 3,973 subjects for the BMI analysis. Table 1 summarizes the baseline characteristics of cases and controls. The genotype distribution was in Hardy-Weinberg equilibrium in controls, with a non-significant chi square value (data not shown). The frequencies and distribution of the genotypes and the odds ratios for the association with BMI and BC risk are described in Table 2.

We observed no statistically significant association between the SNP and the geometric means of BMI. We performed an additional analysis by categorizing subjects with BMI ≥ 30 as obese, 25 ≤BMI <30 as overweight and <25 as normal weight, and calculating odds ratios for obese or overweight versus normal weight with unconditional logistic regression, but we found no statistically significant association. We performed the analysis considering the cases and the controls separate or together and the results obtained did not change. We have also stratified for menopausal status at baseline and we did not find any difference in the genotype distribution. (data not shown).

We performed also an analysis of BC risk. We found no statistically significant associations with risk overall and stratified by age or menopausal status, with codominant model. Additional analyses performed by including cases of carcinoma in situ, or excluding cases diagnosed shortly after blood drawing, showed essentially the same results as using all cases and controls (data not shown).

In this report we explored the possible association between a SNP in the INSIG2 gene and BMI and with BC risk in a large population from the EPIC cohort.

In recent years the search for genetic factors predisposing to obesity has intensified. Several candidate and GWAS have been performed, and have led to the identification of a number of common common genetic variants related to obesity [ 6 17 18 19 , Walley, 2009 #167, 20 ].

The INSIG2 gene is a good candidate for being related to obesity because of its function in lipid metabolism, particularly in blocking the processing of SREBPs in response to cholesterol or insulin. A SNP lying approximately 10 kb upstream from INSIG2, and thus potentially affecting the transcriptional activity of the gene, rs7566605, has been found to be associated with BMI in a family-based GWAS 1 . Replication of this result has been attempted in multiple studies, with some studies confirming an association 3 4 6 7 8 9 10 11 12 13 21 and some others not finding it 3 4 6 7 8 9 10 11 12 13 21 . Recently a meta-analysis has been published on the association of this SNP in INSIG2 and obesity 22 The results of the study from Heid and collaborators suggests that the association is potentially valid but also suggests the possibility that heterogeneity could hinder possible findings if not taken into account properly.

Thus, evidence for or against the association of this SNP with obesity is still needed. We sought to provide it, at least for the Caucasian population, with a study nested within the EPIC cohort. EPIC is an ideal setting, since it has a large population size, accurate phenotypic information on BMI for all subjects, and the presence of both pre-menopausal and post-menopausal women. We tested the effect of rs7566605 on BMI and, at the same time, on BC risk.

We exhaustively analyzed the possible associations between the SNP, BMI and BC risk.. Moreover, subgroup analyses were performed based on menopausal status and age. We have found no statistical association between the polymorphism and the two selected end-points.

We had greater than 99% power to replicate the association of the SNP with BMI, and greater than 80% power to detect an association between rs7566605 and BC risk, with an odds ratio of at least 1.3 and a type 1 error of 0.05.

A possible limitation of this report could be the fact the majority of women and BC cases included in the study are post-menopausal. The power among pre-menopausal women was of 0.81 to detect an OR of 1.3 in a log additive model, hence we cannot exclude a minor effect of the variant allele on BC risk in pre-menopausal subjects. Moreover a possible explanation for a lack of association of this polymorphism with the two endpoints could be the role of gene-environment interactions.

In conclusion, we have studied association of the polymorphism rs7566605, located near the 5' end of the INSIG2 gene, with BMI and BC risk within the EPIC study and we can confidently exclude a major role of this polymorphism with respect to both end-points in this population.

The authors declare that they have no competing interests.

RK, FC conceived and designed the experiments, DC performed the experiments AH analyzed the data: DC, FC, AH, JDMK, OS, RK drafted the manuscript. All other authors contributed substantially to sample collection and manuscript editing.

All authors read and approved the final version of the manuscript.