Poster Presentation 39th Annual Lorne Genome Conference 2018

Invasive Lobular Breast Cancer: Using tumour genome-wide DNA methylation to further subtype and aid in the identification of susceptibility genes. (#254)

Medha Suman 1 , JiHoon Eric Joo 1 2 , Ee Ming Wong 1 2 , Tu Nguyen-Dumont 1 2 , Neil O’Callaghan 1 , Melissa Yow 1 , John Hopper 3 , Graham Giles 3 4 5 , Roger Milne 3 4 , Melissa C. Southey 1 2 4
  1. Department of Pathology, Genetic Epidemiology Laboratory, Victorian Comprehensive Cancer Centre, The University of Melbourne, Melbourne, Victoria, Australia
  2. Precision Medicine, School of Clinical Sciences at Monash Health, Monash University, Melbourne, Victoria, Australia
  3. Centre for Biostatistics and Epidemiology, School of Population and Global Health, The University of Melbourne, Melbourne, Victoria, Australia
  4. Cancer Epidemiology and Intelligence Division, Cancer Council Victoria, Melbourne, Victoria, Australia
  5. School of Public Health and Preventive Medicine, Monash University, Melbourne, Victoria, Australia

 

There is indirect evidence of an inherited component to risk of invasive lobular breast cancer (LBC). We and others have conducted a whole-genome sequencing project involving 120 population-based early-onset and clinic-based multiple-case LBC cases aimed at identifying susceptibility genes for LBC. The project has identified a vast amount of germline genetic variation but no strong candidate LBC susceptibility gene. The role of epigenetic alterations in the susceptibility and progression of cancers has been widely accepted. Here we have sought to identify methods to subtype LBC into groups of reduced heterogeneity using genome-wide DNA methylation patterns in order to interpret the whole genome sequences more precisely.

 

Formalin-fixed paraffin-embedded (FFPE) LBC-enriched DNA was prepared from 161 LBC samples using macrodissection and run on the Infinium HumanMethylation450K Beadchip (HM450K) array to generate genome-wide DNA methylation data. The raw methylation data was pre-processed and normalised using minfi Bioconductor package in R programming software. Unsupervised cluster analysis was used to identify subtypes of LBC based on DNA methylation involving 449005 probes. This dataset was also compared with similar data prepared from FFPE samples of all other breast cancer subtypes (n=341).

 

LBC samples were found to cluster into three main subtypes. We will present data that describes the analyses that have identified differentially methylated genomic regions between these three LBC subtypes. The identification of LBC subtypes using DNA methylation that increase the homogeneity of disease in each subtype could inform the analysis of the whole-genome sequencing project and lead to the identification of LBC susceptibility genes.