Oral Presentation 40th Annual Lorne Genome Conference 2019

Non-coding RNAs underlie genetic predisposition to breast cancer (#42)

Mahdi Moradi Marjaneh 1 , Jonathan Beesley 1 , Tracy A O’Mara 1 , Pamela Mukhopadhyay 1 , Stephen Kazakoff 1 , Nehal Hussein 1 2 , Laura Fachal 3 , Nenad Bartonicek 4 , Kristine M Hillman 1 , Susanne Kaufmann 1 , Haran Sivakumaran 1 , Chanel E Smart 5 , Amy E McCart Reed 5 , Kaltin Ferguson 5 , Jodi Saunus 5 , Sunil R Lakhani 5 6 , Daniel R Barnes 7 , Antonis C Antoniou 7 , Marcel E Dinger 4 8 , Nicola Waddell 1 , Douglas F Easton 3 7 , Alison M Dunning 3 , Georgia Chenevix-Trench 1 , Stacey L Edwards 1 , Juliet D French 1
  1. Cancer Division, QIMR Berghofer Medical Research Institute, Brisbane, Australia
  2. Faculty of Medicine, The University of Queensland, Brisbane, Australia
  3. Department of Oncology, Centre for Cancer Genetic Epidemiology, University of Cambridge, Cambridge, UK
  4. Garvan Institute of Medical Research, Sydney, Australia
  5. UQ Centre for Clinical Research, Faculty of Medicine, The University of Queensland, Brisbane, Australia
  6. Pathology Queensland, The Royal Brisbane & Women's Hospital, Brisbane, Australia
  7. Centre for Cancer Genetic Epidemiology, Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK
  8. St Vincent’s Clinical School, UNSW Sydney, Sydney, Australia

To date, the focus of GWAS follow-up studies has been the impact of regulatory variants on the expression of protein-coding genes. However, stronger risk variants are frequently dismissed as causal because they do not fall in regions marked for enhancer or promoter activity. Fine-mapping of breast cancer GWAS regions has identified 195 high-confidence signals (conditional p-values<10-6) containing >5,000 credible causal variants (CCVs). The CCVs are predominantly noncoding and enriched in regulatory elements but the contribution of long non-coding RNAs (lncRNAs) is unknown. Using targeted RNA sequencing combined with de novo transcript assembly, we systematically annotated multi-exonic non-coding RNAs (mencRNAs) transcribed from 1.5Mb intervals surrounding breast cancer GWAS signals and assessed their contribution to risk. We captured >4000 mencRNAs (which were mostly novel) and showed that breast cancer CCVs were significantly enriched in the mencRNA exons, but not the promoters or introns (Pperm<10-3). Approximately 75% of the mencRNAs were detectable in TCGA breast samples and notably, using principal component analysis, they discriminated normal from tumour samples and distinguished cancer subtypes suggesting that these transcripts play an important role in breast cancer development. Using eQTL analysis we identified 16 mencRNAs whose expression was associated with CCVs in breast tumours (FDR<0.05). Approximately half of the mencRNAs contained the eSNP within the exons suggesting that the variants alter the stability of transcripts. We performed Capture Hi-C in 6 breast cell lines and provided evidence that these CCVs fall in distal enhancers that regulate mencRNAs through long-range chromatin interactions. In summary, we identified 874 mencRNAs as candidate breast cancer risk genes based on CCVs that fall in either 1) mencRNAs exons, 2) mencRNAs promoters or 3) regions that interact with the mencRNAs promoters through long-range chromatin interactions. These data indicate that modulation of lncRNAs by risk-associated variants is a predominant mechanism underlying breast cancer development.