Oncogene (2010) 29, 2013–2023 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 $32.00 www.nature.com/onc ORIGINAL ARTICLE Integrative molecular profiling of triple negative breast cancers identifies amplicon drivers and potential therapeutic targets N Turner1, MB Lambros1, HM Horlings2,3, A Pearson1, R Sharpe1, R Natrajan1, FC Geyer1, M van Kouwenhove2, B Kreike4, A Mackay1, A Ashworth1, MJ van de Vijver2,3 and JS Reis-Filho1 1The Breakthrough Breast Cancer Research Centre, Institute of Cancer Research, London, UK; 2Department of Pathology, Academic Medical Center, Meibergdreef, Amsterdam, The Netherlands; 3Division of Experimental Therapy, The Netherlands Cancer Institute, Plesmanlaan, Amsterdam, The Netherlands and 4Division of Radiation Oncology, The Netherlands Cancer Institute, Plesmanlaan, Amsterdam, The Netherlands Triple negative breast cancers (TNBCs) have a relatively Introduction poor prognosis and cannot be effectively treated with current targeted therapies. We searched for genes that The identification of distinct subgroups of breast cancer have the potential to be therapeutic targets by identifying has led to the development of therapeutic strategies that genes consistently overexpressed when amplified. Fifty-six exploit the underlying biology of the subtype, with TNBCs were subjected to high-resolution microarray- hormonal therapies and HER2-targeting agents for based comparative genomic hybridization (aCGH), of hormone receptor-positive and HER2-positive breast which 24 were subjected to genome-wide gene expression cancers, respectively (Reis-Filho and Tutt, 2008; Schnei- analysis. TNBCs were genetically heterogeneous; no der et al., 2008). Tumours lacking expression of individual focal amplification was present at high freq- hormone receptors (oestrogen receptor (ER) and pro- uency, although 78.6% of TNBCs harboured at least one gesterone receptor (PR)) and HER2 (triple negative focal amplification. Integration of aCGH and expression breast cancers, TNBCs), on the other hand, pose a data revealed 40 genes significantly overexpressed when significant clinical challenge due to a poor under- amplified, including the known oncogenes and potential standing of the genetic alternations that underlie the therapeutic targets, FGFR2 (10q26.3), BUB3 (10q26.3), development of TNBC (Reis-Filho and Tutt, 2008; RAB20 (13q34), PKN1 (19p13.12) and NOTCH3 Schneider et al., 2008) and this is reflected in a lack of (19p13.12). We identified two TNBC cell lines with subtype-specific targeted therapies. FGFR2 amplification, which both had constitutive activa- TNBCs comprise a heterogeneous group of breast tion of FGFR2. Amplified cell lines were highly sensitive cancers and account for 10–15% (Agrawal et al., 2007; to FGFR inhibitor PD173074, and to RNAi silencing of Carey et al., 2007; Dent et al., 2007; Reis-Filho and FGFR2. Treatment with PD173074 induced apoptosis Tutt, 2008; Schneider et al., 2008) of all invasive breast resulting partly from inhibition of PI3K-AKT signalling. cancers. Histologically, the majority of TNBCs are Independent validation using publicly available aCGH grade III invasive ductal carcinomas of no special type, data sets revealed FGFR2 gene was amplified in 4% although the majority of medullary, metaplastic and (6/165) of TNBC, but not in other subtypes (0/214, adenoid cystic carcinomas also display a triple negative P ¼ 0.0065). Our analysis demonstrates that TNBCs are phenotype (Reis-Filho and Tutt, 2008). These tumours heterogeneous tumours with amplifications of FGFR2 in a are more prevalent in young women (o50 years) and in subgroup of tumours. women of African and Hispanic descent. TNBCs have a Oncogene (2010) 29, 2013–2023; doi:10.1038/onc.2009.489; poor prognosis characterized by early relapse (Tischko- published online 18 January 2010 witz and Foulkes, 2006; Dent et al., 2009), potentially reflecting the high proliferative rate of TNBCs. Simi- Keywords: triple negative breast cancer; microarrays; larly, women with TNBCs have a significantly shorter gene expression; comparative genomic hybridization; survival following recurrence when compared with those FGFR2 with non-triple negative cancers (Harris et al., 2006; Dent et al., 2007). Therefore, the identification of novel therapeutic targets for TNBCs is important if the outcome of patients with these tumours is to be improved. On the basis of the concept of oncogene addiction, we Correspondence: Dr N Turner, or Dr JS Reis-Filho, The Break- and others have demonstrated that genes that are through Breast Cancer Research Centre, Institute of Cancer Research, consistently overexpressed when amplified may be 237 Fulham Road, London, SW3 6JB, UK. E-mails: [email protected] or jorge.reis-fi[email protected] selectively required for the survival of cancer cells Received 18 August 2009; revised 30 November 2009; accepted 7 harbouring their amplification, and can be exploited as December 2009; published online 18 January 2010 potential therapeutic targets (Reis-Filho et al., 2006; Integrative molecular profiling of TNBC N Turner et al 2014 Bernard-Pierrot et al., 2008; Natrajan et al., 2009a). et al., 2009b). This analysis revealed that the expression Earlier studies have examined TNBCs with expression of 4972 out of 24 565 genes (20.2%) correlated with copy profiling (Kreike et al., 2007; Bertucci et al., 2008) and number changes (Pearson’s correlation adjusted microarray-based comparative genomic hybridization Po0.05, Supplementary Table 6). This analysis suggests (aCGH) (Han et al., 2008; Andre et al., 2009). These that a large proportion of the genes expressed in TNBCs studies have found TNBCs to be heterogeneous, with are at least in part regulated by copy number changes. complex genomic profiles and infrequent amplifications. Within this list, we determined the genes whose To identify amplicon drivers and genes that have the expression was significantly upregulated or downregu- potential to be therapeutic targets in TNBCs, we lated in the presence of copy number gain or loss, integrated aCGH and gene expression data from a large respectively (Mann–Whitney U adjusted Po0.05, Sup- series of TNBCs. Our aims were to characterize the plementary Table 7). We identified 324 genes whose genomic and transcriptomic profiles of TNBCs and mRNA expression levels were significantly higher in identify and validate genes that are recurrently amplified tumours harbouring DNA copy number gains and 39 and consistently overexpressed when amplified in genes whose mRNA expression levels were significantly TNBCs. We found more frequent high level, focal lower in tumours displaying DNA copy number loss. amplifications than described earlier (Chin et al., 2006), The lower number of genes downregulated when lost and identified potential therapeutic targets that are may stem from the slightly lower sensitivity of array consistently overexpressed when amplified. We vali- CGH to detect loss of a single copy of a genomic region dated the fibroblast growth factor receptor 2 (FGFR2) in highly aneuploid tumours (Ng et al., 2006), coupled gene as one of these targets and provided functional with the limitations of expression arrays to accurately data to suggest that this tyrosine kinase receptor may be determine the expression of genes expressed at low a novel therapeutic target in a subset of TNBCs levels. harbouring FGFR2 gene amplification. Ingenuity pathway analysis (IPA) of genes expressed at higher levels when gained revealed that 5 networks and 27 canonical pathways were significantly enriched (Supplementary Table 8), and only one network/ Results canonical pathway from genes expressed at lower levels when lost (Supplementary Table 9). Importantly, the TNBCs display complex genomic profiles canonical pathways of growth factors that have been We established the genomic array CGH profiles of 56 shown to have a function in breast cancer development TNBCs with an array CGH platform with an earlier and progression, or to be potential therapeutic targets, validated effective resolution of 50 kb. This revealed a were significantly enriched with genes upregulated when high level of genetic instability, with gain or loss gained (Table 1). Copy number gains of the tyrosine affecting a median of 44.4% of the genome (range kinase receptors were rare, and the growth factor 9.3–76.7%). Using the genomic pattern classification pathway enrichment reflected frequent copy number proposed by Hicks et al. (2006), 15 of the TNBCs (27%) gains and overexpression of key adapter molecules and were classified as ‘simplex’ and 41 (73%) were con- downstream signal transduction kinases (Table 1). The sidered to have ‘complex’ genomic patterns, of which 28 p53 signalling pathway was enriched for genes that are (50%) were ‘sawtooth’ and 13 (23%) were ‘firestorm’ upregulated when gained, including survivin (BIRC5). (Hicks et al., 2006). Given that the majority of TNBCs displayed ‘sawtooth’ patterns, it is not surprising that multiple regions of recurrent gains and losses were Recurrent amplified regions and potential amplicon identified (Figure 1a; Supplementary Table 4). In drivers agreement with earlier studies (Han et al., 2008; Andre We next analysed the data to identify the genes that were et al., 2009), loss of 1p36-p34, 5q11-q35, 8p23.3-p12 and significantly overexpressed when amplified in TNBCs, 17p13-q21 and gain of 3q22-q26, 6p25-p21, 7q32-q36, compared with non-amplified cancers (Figure 2). This 8p11-q24, 10p15.3-p11.21 and 12p13-p11 were found in analysis revealed only 40 genes (Mann–Whitney U-test 430% of TNBCs