Functional Genetic Epidemiology Group

As genetic variants are effectively randomised at birth, genetic studies provide an unbiased methodology for identifying and characterising intermediate phenotypes (measurable traits associated with breast cancer risk) that influence risk. We combine genetic discovery with intermediate phenotype studies – including premenopausal hormone levels and the insulin-like growth factor (IGF) axis – to dissect the effects of individual genetic variants on pathways that influence breast cancer risk.

Given that breast cancer risk is strongly influenced by reproductive risk factors, we are prioritising genetic associations relevant to the protective effects of pregnancy and breast feeding; if we can understand the molecular basis of these protective effects, this may facilitate the development of novel risk-reduction and prevention strategies.

Identifying genetic variants associated with premenopausal hormone levels and breast cancer risk

We are using hormone levels, measured in urine samples, as a quantitative measurable trait that we can exploit to understand certain aspects of breast cancer predisposition and response to treatment. We have shown that a specific genetic variant (CYP3A7*1C) is associated with a 54% reduction in some oestrogen breakdown products in healthy premenopausal women. We have also demonstrated that this variant is associated with a worse prognosis for women with breast cancer, possibly via an effect on the way that these women respond to chemotherapy.

To extend these analyses, we are collecting serial urine samples from some of the Generations Study participants and using a variety of genetic techniques, including a genome-wide association study (GWAS) and targeted sequencing analysis, to identify new variants that are associated with hormone levels.

Defining the interaction targets of recently identified GWAS hits and prioritising regions for functional analyses

One of the main challenges in translating the findings of breast cancer GWAS is the fact that the vast majority of risk loci map to non-coding regions of the genome. We have developed a novel method, Capture Hi-C (CHi-C) to interrogate long-range physical interactions between (non-coding) regulatory elements and the genes that they regulate. We are using this technique to characterise more than 60 regions of the genome that have been shown to be associated with breast cancer risk.

We have been able to assign target genes to many of the risk loci; some of these are genes that have already been shown to be relevant to breast cancer, others are novel. We have also shortlisted individual variants at some of the risk loci; understanding how these variants influence breast cancer risk may – in the longer term – help us to develop new strategies for preventing breast cancer. In collaboration with the Breast Cancer Association Consortium (BCAC), we are carrying out more detailed analyses of two specific genomic regions – 2q35 and the 11p15.5.

Improving our understanding of pregnancy associated changes in BC risk

The focus of this project is the 11p15.5 risk locus. The association between this region of the genome and breast cancer risk is modified by whether or not a women has undergone pregnancy and childbirth. The risk allele at this locus also shows a parent of origin effect such that only the paternal allele is associated with risk.

We are carrying out experiments to try to understand which of the many genes that map to this locus are causally associated with risk; what types of cells these genes are expressed in; and whether their expression alters during pregnancy, breast feeding or involution (the process that occurs after a woman stops breast feeding).