Chromosomal Instability Selects Gene Copy-Number Variants Encoding Core Regulators of Proliferation in Erþ Breast Cancer
Total Page：16
File Type：pdf, Size：1020Kb
Published OnlineFirst June 26, 2014; DOI: 10.1158/0008-5472.CAN-13-2664 Cancer Tumor and Stem Cell Biology Research Chromosomal Instability Selects Gene Copy-Number Variants Encoding Core Regulators of Proliferation in ERþ Breast Cancer David Endesfelder1,2, Rebecca A. Burrell3, Nnennaya Kanu4, Nicholas McGranahan3, Mike Howell7, Peter J. Parker5,6, Julian Downward8, Charles Swanton3,4, and Maik Kschischo1 Abstract Chromosomal instability (CIN) is associated with poor outcome in epithelial malignancies, including breast þ carcinomas. Evidence suggests that prognostic signatures in estrogen receptor–positive (ER ) breast cancer define tumors with CIN and high proliferative potential. Intriguingly, CIN induction in lower eukaryotic cells and human cells is context dependent, typically resulting in a proliferation disadvantage but conferring a fitness benefit under strong selection pressures. We hypothesized that CIN permits accelerated genomic evolution through the generation of diverse DNA copy-number events that may be selected during disease development. In support of this hypothesis, we found evidence for selection of gene amplification of core regulators of proliferation in CIN-associated cancer genomes. Stable DNA copy-number amplifications of the core regulators TPX2 and UBE2C were associated with expression of a gene module involved in proliferation. The module genes were þ enriched within prognostic signature gene sets for ER breast cancer, providing a logical connection between CIN and prognostic signature expression. Our results provide a framework to decipher the impact of intratumor heterogeneity on key cancer phenotypes, and they suggest that CIN provides a permissive landscape for selection of copy-number alterations that drive cancer proliferation. Cancer Res; 74(17); 1–11. Ó2014 AACR. Introduction viously experienced by the cell population (4). Highly aneuploid Induction of aneuploidy affects cellular fitness in various cancers are characterized by a large number of structural and organisms, including yeast (1), mice (2), and human cells (3). numerical DNA copy-number changes altering expression of Most aneuploid cells exhibit reduced proliferation rates, but most genes located in these regions. Aneuploidy is frequently fi studies in yeast have shown that aneuploidy can also be accompanied by chromosomal instability (CIN), de ned as an beneficial for the adaptation to stressful conditions not pre- increased rate of gains and losses of whole chromosomes or fractions of chromosomes (5). CIN generates intercellular chromosomal heterogeneity that may facilitate selection of 1Department of Mathematics and Technology, University of Applied genotypes tolerant to aneuploidy and strong selective pres- Sciences Koblenz, RheinAhrCampus, Remagen, Germany. 2Helmholtz sures in the tumor, and accelerate the proliferation of aneu- € fi Zentrum Munchen, Institute of Computational Biology, Scienti c Comput- ploid cancer cells (6, 7). The specific chromosomal changes and ing Research Unit, Neuherberg, Germany, Scientific Computing Research Unit, Neuherberg, Germany. 3Translational Cancer Therapeutics Labora- expression patterns that are implicated in this adaptive tory, Cancer Research UK London Research Institute, London, United response remain unclear. Kingdom. 4UCL Cancer Institute, London, United Kingdom. 5Protein Phos- phorylation Laboratory, Cancer Research UK London Research Institute, Several breast cancer gene expression signatures forecasting London, United Kingdom. 6Division of Cancer Studies, King's College clinical outcome and response to chemotherapy are strongly London, London, United Kingdom. 7High Throughput Screening Labora- associated with proliferation (8–11). It is still being debated tory, Cancer Research UK London Research Institute, London, United fl Kingdom. 8Signal Transduction Laboratory, Cancer Research UK London whether the prognostic value of these diverse gene lists re ects Research Institute, London, United Kingdom similar underlying biologic processes and pathways (12). Note: Supplementary data for this article are available at Cancer Research Intriguingly, there is increasing evidence that many breast Online (http://cancerres.aacrjournals.org/). cancer prognostic signatures are associated with CIN (13, 14). CIN is associated with inferior prognosis across multiple R. Burrell and N. Kanu contributed equally to this article. þ cancer types, including ER-positive (ER ) breast cancer (15). Corresponding Authors: Maik Kschischo, University of Applied Sciences Koblenz, RheinAhrCampus, Joseph-Rovan-Allee 2, Remagen 53424, Ger- The CIN70 signature (16) and a 12-gene genomic instability many. Phone: 49-2642932330; Fax: 49-2642932399; E-mail: signature (13), both derived from their associations with [email protected]; and Charles Swanton, Cancer Research UK, London Research Institute Lincoln's Inn Fields Laboratories 44 Lin- aneuploidy and CIN, have prognostic value in many cancer coln's Inn Fields London WC2A 3LY. Phone: 44-0-20-7269-3463; Fax: 44- types and also correlate with proliferation markers. However, 0-20-7269-3094; E-mail: [email protected] little is known about possible genotypes crucial for the CIN doi: 10.1158/0008-5472.CAN-13-2664 phenotype or the relationships between CIN and proliferation Ó2014 American Association for Cancer Research. in vivo. Indeed, it has been proposed that CIN is a consequence, www.aacrjournals.org OF1 Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 2014 American Association for Cancer Research. Published OnlineFirst June 26, 2014; DOI: 10.1158/0008-5472.CAN-13-2664 Endesfelder et al. rather than a cause of the selective pressure for proliferative database (22). Genes inducing cell death (P < 0.05) in at least þ drive in tumors. four cell lines were considered as proliferation genes in ER In this study, we sought to better characterize the relation- breast cancer. ships between CIN and proliferation through an analysis of the effects of copy-number alterations associated with CIN and the Microarray expression analysis after UBE2C and TPX2 downstream consequences of such alterations upon prognos- silencing in T47D cells tic signature gene expression. T47D cells were maintained in 5% CO2 at 37 C, in RPMI-1640 medium supplemented with 10% FBS, glutamine, and penicil- Materials and Methods lin/streptomycin. All siRNA (Dharmacon, Thermo Scientific) SNP and expression data processing experiments were performed at 40 nmol/L final concentrations Microarray expression data and SNP array based copy- by reverse transfection with Dharmafect2 reagent according to þ number data of 264 patients with ER breast cancer with the manufacturer's instructions. Transfections were per- pathologist estimates of more than 60% tumor cell content formed using TPX2 [L-010571-00-0005 and M-010571-00- or more than 60% tumor nuclei were selected from The 0005 and UBE2C (L-004693-00-0005 and M-004693-03-0005] Cancer Genome Atlas (TCGA). Agilent 244K Custom Gene siRNA pools. Non-targeting control siRNA and scrambled Expression G4502A-07 two-color microarrays were normal- control 2 were used. At 72 hours after transfection, RNA was ized (print-tip-group loess normalization, Bioconductor extracted and knockdowns validated by quantitative PCR to be package marray; ref. 17). Probes with missing values in more at least 85% to 90%. RNA was extracted and samples were than 15% of the samples were excluded and duplicated hybridized to HG_U133 Plus 2.0 arrays. Expression calls were probes were averaged. Gene mappings based on NCBI build generated by the MAS5.0 algorithm (R-package simpleaffy; 36.3 were downloaded from the TCGA data portal (18). For ref. 23). Gene expression values were subtracted from the gene probes mapped to the same ENTREZ Gene ID, the probe expression values after silencing TPX2 and UBE2C, respectively, with the highest Pearson correlation coefficient of gene from the siRNA control experiment. This difference was mul- expression with wGII score was chosen. For the remaining tiplied by the sign of the Spearman correlation of the gene genes, missing gene expression data were imputed with expression with TPX2 or UBE2C in the tumor samples. Negative nearest neighbor (k ¼ 10) averaging implemented in the values indicate that silencing of TPX2 or UBE2C results in Rpackageimpute (19). Genes with standard deviation lower downregulation of genes positively correlated and in upregula- than 0.25 were removed. The expression values of all remain- tion of genes negatively correlated with TPX2 or UBE2C ing genes were standardized. expression in tumors. Deviations from zero were tested with SNP 6.0 samples were normalized with the Affymetrix the Wilcoxon signed-rank test. Genotyping Console using standard settings. Samples not passing the quality check criteria were excluded. Integer copy Weighted genome integrity index numbers were estimated by the GAP algorithm (20). The ploidy of a tumor sample was determined as the weighted median integer copy number, with weights equal to the lengths RNA interference screening data of the copy-number segments. For each sample and each of the Whole-genome RNA interference (RNAi) screening data (21) 22 autosomal chromosomes, the percentage of gained and lost analyzing cell numbers after gene silencing were available for genomic material was calculated relative to the ploidy of the five cell lines: PC9 (lung), RCC4 (kidney), HCT116 (colon), MCF- sample. The use of percentages eliminates the bias induced by 10A (breast), and HT1080 (fibrosarcoma).