LB59
An integrative analysis of genome copy number, whole genome expression and cancer gene methylation status in breast cancer cell lines of basal-like phenotype
Elodie Noel1, Anita Grigoriadis1, Peijun Wu1, Pierfrancesco Marra1, Alan Mackay2, Rachael Natrajan2, Jessica Taylor2, Chris Lord2, Alan Ashworth2, Jorge S Reis-Filho2, Andrew Tutt1
1Breakthrough Breast Cancer Research Unit, Guys Hospital, London, UK; 2Breakthrough Breast Cancer Research Centre, The Institute of Cancer Research, London, UK
Background
Basal-like breast cancer (BLBC) is usually associated with the absence of
expression of ooestrogen, progesterone and ERBB2 receptors and displays an
aggressive phenotype with few therapeutic options. Target validation studies
for breast cancer subtypes have largely employed established breast cancer cell
lines (BCCLs) which have proven to be effective tools for drug discovery. BLBC
has a heterogenous clinical and histopathological phenotype and differing
morphological and functional phenotypes are observed among BCCLs of the basal
type. This prompted us to investigate the genetic, epi-genetic and somatic
footprints of these BCCLs to identify similarities and differences with
available profiles obtained from tumours. We aimed to identify those cell lines
which better represent an in vitro model for the study of BLBC based
upon similarities in genetic, epigenetic and transcriptional pattern.
Method
We performed array comparative genomic hybridisation on a 32K tiling path
BAC array, alongside SNP genotyping on SNP 370CNV, Illumina HumanWG-6v2.0
arrays gene expression and methylation profiling using the Golden Gate Cancer
Panel on 26 BCCL predominantly of basal-like phenotypes. Genomic
aberrations were validated by FISH on an in-house cell line tissue microarray
and selected genes of interest were verified by quantitative real-time RT-PCR.
Results
Clustering, centroid correlations and PAM analysis of expression profiles
determined HCC1187, HCC1569, HCC1143, MDA-MB-468 and HCC38 as the better
basal-like BCCL models. Different gene signatures (proliferation signature,
immune response, 70 gene signature) derived from primary breast cancers
provided further annotation of BCCLs. Patterns of genomic instability indicated
a predominantly sawtooth pattern for basal-like BCCL. Integration of gene
expression with methylation analysis suggests epigenetic influence on EPHA2,
SNRPN and HOXB2 gene expression in basal-like BCCLs.
Conclusion
By integrating datasets from BCCL and performing a detailed annotation
based on breast tumour profiles, we have built an in-depth resource to assess
the suitability of breast cancer cell lines as experimental models for studying
triple negative and basal-like breast cancer.
