Genetic Determinants of Breast Cancer

*Douglas F. Easton, University of Cambridge 

Keywords:

Common cancers, including breast cancer, exhibit familial aggregation, consistent with substantial variation in inherited susceptibility. Over the past 25 years, the underlying genetic basis for this susceptibility has been increasingly understood. Three important classes of genetic loci have been identified: • "High-penetrance” genes, principally BRCA1 and BRCA2. Mutations in these genes confer a high lifetime risk of breast, ovarian and other cancers. However, mutations in these genes are rare and explain less than 20% of the familial risk of breast cancer. • "Itermediate-penetrance” genes currently include ATM, CHEK2, BRIP1 and PALB2. Mutations in these genes confer a 2-3 fold increased risk of breast cancer. • Comon “low-penetrance” alleles, principally identified through genome-wide association studies (GWAS). More than 20 such loci have been identified to date. For the most significant such loci, in FGFR2 and TOX3, the susceptibility allele confers relative risks of 1.2 to 1.3 fold.

Altogether, the known breast cancer loci probably explain 25-30% of the familial aggregation of breast cancer, implying that many additional loci remain to be found. It is unlikely that there any further important high-penetrance loci, but many more lower penetrance loci should be identifiable, through GWAS, large-scale replication studies and resequencing. The mutations in BRCA1 and BRCA2 (and, so far, those in the intermediate risk genes) involve inactivating coding alterations. In contrast, many if not all the low-penetrance loci involve non-coding regions (including “gene deserts”) and are presumably regulatory. Most of these regions have not previously been related to carcinogenesis and point to novel mechanisms. Fine-scale mapping of the FGFR2 locus has shown that the likely causal variant regulates FGFR2 expression through differential transcription factor binding. Similar mechanisms may operate for the other loci, but the causal variants at the other loci are not yet known and, in most cases, even the target genes have not been unambiguously identified. While the risks associated with the common low-penetrance alleles are modest, these risks combine multiplicatively with each other, and this raises the possibility that genetic profiles based on combinations of alleles may become important in risk prediction. Based on the currently estimated relative risks, the 5% of the population of at highest risk have a risk that is approximately twice the population average, while the 5% of the population at lowest risk are at about half the risk of the general population. These differences are unlikely to be sufficient to change prevention strategies by themselves, but may be more important when combined with other risk factors such as family history and breast density. Many of the breast cancer susceptibility loci show strong subtype specificity. Specifically, many of the loci (such as FGFR2) predispose predominantly to ER-positive breast cancer, while some (such as TERT and ESR1) predispose predominantly to ER-negative breast cancer. These pathological differences may have implications both for risk prediction and prevention. There is also strong evidence that most of the common susceptibility alleles modify the breast cancer risks in BRCA2 mutation carriers, to a similar relative extent as in the general population, while fewer (notably those that predispose to ER-negative disease in the general population) appear to modify the risks in BRCA1 carriers. These differences mirror the differences in pathology between tumours in BRCA1 and BRCA2 carriers. Thus, SNP profiling may significantly improve risk prediction and genetic counselling in BRCA1 and BRCA2 carriers.