br Our study improves on previous
Our study improves on previous studies in that we examined the interaction between a number of biologically plausible poly-morphisms in different genes that are involved in a wide range of cancer-related processes, not just SNPs in a single gene or in genes acting on a specific pathway. The observed interactions have not been previously reported. Our results also confirm the notion that SNP-SNP interactions could facilitate the understanding of the additional missing heritable components of BC risk, possibly through interplay in several pathways contributing to BC etiology. We think that our interaction results are relevant in BC screening because genotyping a small number of SNPs (using the widely available PCR techniques) and testing for interaction is less cumbersome and easier to interpret than genome-scale methods that disregard such interactions.
We acknowledge the following limitations of our study. First, selection bias was inevitable; 56% of our patients had a family history of BC, which might have enriched the study with a genetic
component. This result is higher than the 17% to 22% frequency reported in large caseecontrol studies,44,69 but it is consistent with
the 58% frequency in another Egyptian study.70 However, we controlled for this by adjusting all results with family history. Also, the majority of our patients had nonmetastatic disease (72%), so association of haplotypes with bone metastasis could be biased. Second, although studied SNPs demonstrated weak single-locus effects on BC susceptibility in our study, results were inconsistent among different populations23-29 and thus might not be Olfat G. Shaker, Mahmoud A. Senousy
reproducible. Third, we used a statistically oriented approach for N,N-Dimethylsphingosine detection; with adjustments for multiple comparisons, the BsmI association did not meet the conservative Bonferroni P value and thus might be due to chance. Fourth, we limited our analyses to Egyptian women, which may limit the generalizability of our findings; however, population homogeneity is a study strength by reducing genetic variability. Fifth, our findings are based on a small number of case subjects, and their replication in a larger study is unclear and so should be interpreted with caution. Future large independent studies are warranted to further validate our results. SNPeenvironment interactions should be also evaluated to comprehensively account for BC risk. Nevertheless, we believe that our findings reflect important, sound biology to clinical medicine rather than simply statistical findings.
In summary, our results suggested that the stronger combined effect of SNPs in the RANKL, OPG, CHI3L1, and VDR genes via geneegene interaction may help to predict BC risk and its prog-nosis. Our data have potential implications in genetic counseling, BC screening, and BC prognosis.
Clinical Practice Points
Most of the heritability of BC is still unexplained. Geneegene interactions (epistasis) are assumed to contribute to overall BC risk. Genetic variants of RANKL, OPG, CHI3L1, and VDR play an overall role in breast carcinogenesis; however, their epistatic effects on BC susceptibility remain unknown.
We provided the first evidence of an association of multiple SNP-SNP interactions and TCTCTG haplotype of 6 SNPs in RANKL, OPG, CHI3L1, and VDR with a genetic predisposition of BC. We also demonstrated the association of certain haplo-types with BC prognosis.
The TCCCTG haplotype was inversely correlated with positive ER/PR. The TCTCTA, TCTCTG, and TGTCTG haplotypes were correlated with bone metastasis.
Bioinformatic analysis of protein interactions revealed that the 4 genes interact most significantly with STAT5B. Our results suggested that a stronger combined effect of SNPs in the RANKL, OPG, CHI3L1, and VDR genes via geneegene inter-action may help predict BC risk and prognosis.
Our work explored the complex biology of BC via geneegene interaction as well as proteinegene interaction that may shed more light into unexplained heritability of BC. Our work also identified new genetic markers for BC screening and prognosis.
Our findings reflect important, sound biology relevant to clinical medicine, rather than simply statistical findings, and have po-tential implications in genetic counseling, BC screening, and BC prognosis.
The authors have stated that they have no conflict of interest.
2. Ibrahim AS, Khaled HM, Mikhail NN, Baraka H, Kamel H. Cancer incidence in Egypt: results of the National Population-Based Cancer Registry Program. J Cancer Epidemiol 2014; 2014:437971.
3. Tung N, Lin NU, Kidd J, et al. Frequency of germline mutations in 25 cancer susceptibility genes in a sequential series of patients with breast cancer. J Clin Oncol 2016; 34:1460-8.