Does chromosome 17 centromere copy number predict polysomy in breast cancer? A fluorescence in situ hybridization and microarray-based CGH analysis

Transcription

1 Journal of Pathology J Pathol 2009; 219: Published online in Wiley InterScience ( DOI: /path.2574 Original Paper Does chromosome 17 centromere copy number predict polysomy in breast cancer? A fluorescence in situ hybridization and microarray-based CGH analysis Caterina Marchiò, 1,2 Maryou B Lambros, 1 Patrizia Gugliotta, 2 Ludovica Verdun Di Cantogno, 2 Cristina Botta, 2 Barbara Pasini, 3 DavidSPTan, 1 Alan Mackay, 1 Kerry Fenwick, 1 Narinder Tamber, 1 Gianni Bussolati, 2 Alan Ashworth, 1 Jorge S Reis-Filho 1 * and Anna Sapino 2 * 1 The Breakthrough Breast Cancer Research Centre Institute of Cancer Research, London, SW3 6JB, UK 2 Department of Biomedical Sciences and Human Oncology, University of Turin, Turin, Italy 3 Department of Genetics, University of Turin, Turin, Italy *Correspondence to: Anna Sapino, Department of Biomedical Sciences and Human Oncology, University of Turin, Via Santena 7, Torino, Italy. anna.sapino@unito.it Jorge S Reis-Filho, The Breakthrough Breast Cancer Research Centre, Institute of Cancer Research, 237 Fulham Road, London, UK, SW3 6JB, UK. Jorge.Reis-Filho@icr.ac.uk No conflicts of interest were declared. Received: 1 April 2009 Revised: 30 April 2009 Accepted: 2 May 2009 Abstract Approximately 8% of breast cancers show increased copy numbers of chromosome 17 centromere (CEP17) by fluorescence in situ hybridization (FISH) (ie average CEP17 >3.0 per nucleus). Currently, this pattern is believed to represent polysomy of chromosome 17. HER2-amplified cancers have been shown to harbour complex patterns of genetic aberrations of chromosome 17, in particular involving its long arm. We hypothesized that aberrant copy numbers of CEP17 in FISH assays may not necessarily represent true chromosome 17 polysomy. Eighteen randomly selected CEP17 polysomic cases and a control group of ten CEP17 disomic cases, as defined by dual-colour FISH, were studied by microarray-based comparative genomic hybridization (acgh), which was performed on microdissected samples using a 32K tiling-path bacterial artificial chromosome microarray platform. Additional FISH probes were employed for SMS (17p11.2) and RARA (17q21.2) genes, as references for chromosome 17 copy number. Microarray-based comparative genomic hybridization revealed that 11 out of the 18 polysomic cases harboured gains of 17q with involvement of the centromere, one displayed 17q gain sparing the centromeric region, and only one could be defined as polysomic. The remaining five cases displayed of the centromeric region. Among these, one case, showing score 2+ by immunohistochemistry and 8.5 HER2 mean copy number, was classified as not amplified by HER2 /CEP17 ratio and as amplified by HER2 /SMS ratio. Our results suggest that true chromosome 17 polysomy is likely to be a rare event in breast cancer and that CEP17 copy number greater than 3.0 in FISH analysis is frequently related to gain or of the centromeric region. Larger studies investigating the genetic profiles of CEP17 polysomic cases are warranted. Copyright 2009 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. Keywords: array CGH; breast cancer; in situ hybridization; ; centromere; HER2 ; CEP17; Herceptin Introduction HER2 testing is currently performed either by immunohistochemical assessment of HER2 protein or by fluorescence in situ hybridization (FISH) analysis of the HER2 gene [1,2]. FISH can be performed with single- or dual-colour probes [1]. The use of a centromeric probe to correct the absolute HER2 gene copy number with the number of chromosomes 17 has been recommended, so as not to misinterpret chromosome 17 (Chr17) polysomy (ie increased numbers of chromosome 17) as HER2. According to the recently updated ASCO/CAP guidelines, patients with breast carcinomas displaying either a 3+ score by immunohistochemistry (IHC) or a HER2 /CEP17 ratio greater than 2.2 in dual-colour FISH or more than six HER2 gene copies in single-colour FISH and chromogenic in situ hybridization (CISH) are eligible for HER2-tailored therapies [2]. Equivocal cases using IHC, ie HER2-overexpressing carcinomas scored as 2+, should undergo FISH analysis to assess HER2 gene, and only patients with breast cancer harbouring HER2 copy number greater than 6 or HER2 /CEP17 ratios above 2.2 should be offered trastuzumab.

2 HER2 assessment, chromosome 17 polysomy, and CEP17 copy number 17 Although clinical trials using dual- or singlecolour FISH or CISH for HER2 gene copy number assessment successfully identified patients that benefit from trastuzumab therapy, it has been suggested that systems assessing the HER2 /CEP17 ratio provide a more accurate evaluation of HER2 than single probe systems [3]. These claims are based on the fact that 8% of breast cancers show increased copy numbers of CEP17 by FISH (ie average CEP17 >3.0 per nucleus) and that these would represent Chr17 polysomy [2,4,5]. We and others [6,7] have demonstrated that in HER2-amplified cancers, Chr17 displays complex patterns of genetic aberrations, in particular involving its long arm. Furthermore, Isola et al [6] suggested that some cases considered polysomic for Chr17 may in fact harbour of the pericentromeric regions of CEP17. Troxell et al [8] recommended the use of probes for additional Chr17 loci (SMS and RARA mapping to 17p11.2 and 17q21.2, respectively) as surrogate Chr17 controls in cases with a complex CEP17 FISH pattern. Based on the observations described above, we hypothesized that increases in CEP17 copy number may not necessarily represent Chr17 polysomy (ie the occurrence in a nucleus of extra copies of chromosome 17) and that they could result from gains of centromeric/pericentromeric regions of Chr17, primarily on the long arm of this chromosome. To address this question, we analysed a series of FISH-defined Chr17 polysomic and disomic cases with tiling path microarray-based comparative genomic hybridization (acgh) and observed that increases in CEP17 copy number are only rarely caused by Chr17 polysomy. Materials and methods Case selection A series of 18 randomly selected Chr17 polysomic cases (six HER2-amplified) and a control group of ten Chr17 disomic cases (six HER2-amplified) were retrieved from the files of the Molinette Hospital, Turin. In our practice, all cases are subjected to dual-colour FISH analysis. Immunohistochemistry for HER2 had already been performed for diagnostic purposes. Immunohistochemistry slides were retrieved and HER2 scoring was reviewed according to the recently updated ASCO/CAP guidelines [2] by two of the authors. Out of the 12 cases harbouring a FISH ratio greater than 2.2, seven cases had a 3+ score and five a 2+ score (Supporting information, Supplementary Table 1). In the group of 16 cases with a FISH ratio less than 2.2, six were IHC-negative and ten were 2+ (Table 1). None of the Chr17 polysomic cases lacking HER2 gene (N = 12) displayed a 3+ score; however, eight of them (8/12, 66.7%) were equivocal (ie 2+ score) (Supporting information, Supplementary Table 1). The study was approved by the local Ethical Committee of the University of Turin. FISH analysis FISH was performed according to the manufacturers instructions with probes for HER2 (17q12), CEP17, SMS (17p11.2), and RARA (17q21.2) (Figure 1) (all from Vysis, Inc, Downers Grove, IL, USA). Briefly, slides were pretreated in sodium thiocyanate (30 min, 80 C) and then digested with pepsin (7 15 min, 37 C; Invitrogen, Carlsbad, CA, USA). They were then washed in 2 SSC, air-dried, covered with 10 µl of probe, co-denatured in HYBrite System (Vysis; 5min,72 CfortheHER2 /CEP17 dual-colour probe; 1min,85 CforSMS/RARA dual-colour probe) and hybridized overnight at 37 C. Slides were washed with post-hybridization buffer (73 C) and counterstained with 4,6 -diamidino-2-phenylindole. Six invasive areas on each slide were selected and automated acquisition was performed with the motorized Metafer Scanning System (Carl Zeiss MetaSystems GmbH) and AxioImager epifluorescence microscope (one focus plane for DAPI and 13 focus planes for green and red spots). Analysis of HER2 /CEP17 probes was performed automatically by Metafer through the PathVysion V2 software (FDA approved) and results were confirmed by one of the authors (AS). Both the HER2 /CEP17 ratio and the mean values of HER2 and CEP17 signals were considered independently in the result analysis. According to the ASCO guidelines [2], cases were scored as not amplified when harbouring a FISH ratio less than 1.8 or HER2 gene copy below 4.0; amplified if the FISH ratio was greater than 2.2 or HER2 gene copy was above 6.0; and equivocal if they displayed a FISH ratio between 1.8 and 2.2 or HER2 gene copy between 4.0 and 6.0. Chromosome 17 polysomy was defined as cases displaying a mean number of more than 3 copies of the CEP17 probe [2]. Analysis of the SMS /RARA probes was performed by counting red (SMS ) and green (RARA) spots on images taken by Metafer, and transferred into the ISIS software (at least 60 nuclei). SMS was used as an alternative to CEP17 to define the ratio; RARA was considered together with the other probes in order to describe Chr17 arrangement by FISH. All FISH analyses were performed with observers blinded to the results of immunohistochemistry and acgh analysis. Microdissection and DNA extraction Cases were microdissected to ensure greater than 90% purity of cancer cells. Microdissection was performed with a sterile needle under a stereo microscope (Olympus SZ61, Tokyo, Japan) from five consecutive 8-µmthick sections stained with nuclear fast red. DNA was extracted as previously described [9,10]. DNA concentration was measured with PicoGreen (Invitrogen,

3 18 CMarchiò et al Paisley, UK) and DNA quality was assessed using a multiplex PCR as previously described [9,11,12]. Microarray CGH The acgh platform used for this study was constructed at the Breakthrough Breast Cancer Research Centre and comprises BACs (Corning, NY, USA) [9,10]. This type of BAC array platform has been shown to be as robust as and to have comparable resolution to high-density oligonucleotide arrays [13 15]. Labelling, hybridization, and washes were carried out as previously described [10,11]. Slides were then scanned using an Axon 4000B scanner (Axon Instruments, Burlingame, CA, USA) and images were processed using Genepix Pro 5.1 image analysis software (Axon Instruments). Log 2 ratios were normalized for spatial and intensity-dependent biases using a two-dimensional loess regression followed by a BAC-dependent bias correction as previously described [16]. This left a final data set of clones with unambiguous mapping information, according to the March 2006 build (hg18) of the human genome ( Data were smoothed using a local polynomial adaptive weights smoothing (aws) procedure for regression problems with additive errors [9,10]. Thresholds for defining genomic gains and losses were obtained using data from female versus female and female versus male genomic DNA, as previously described [10,11]. A categorical analysis was applied to each clone on the array after classification as gain, loss, or no change according to their smoothed log 2 ratio values. Smoothed log 2 ratio values less than 0.12 were categorized as losses; those greater than 0.12 as gains; and those in between as unchanged. Amplifications were defined as smoothed log 2 ratio values above 0.45 [9,10]. Data processing and analysis were carried out in R ( and BioConductor 1.5 ( making extensive use of modified versions of the packages acgh, marray, and aws in particular. Cases were considered polysomic by acgh analysis when harbouring more than 50% of Chr17 BACs in both the short and the long arms with aws ratios above Amplification of CEP17 was defined as the centromeric/pericentromeric BACs of either the long or the short arms with aws ratios greater than Whole loss of a chromosomal arm was defined as more than 75% of clones with aws ratios less than 0.12 [17]. Results Comparison between acgh and FISH for the detection of polysomy We sought to investigate whether Chr17 polysomy as defined by FISH would correspond to gain of the whole Chr17. Out of the 18 cases diagnosed as polysomic by FISH, only case K22 displayed gains of both the short and the long arms of Chr17 (Table 1, Figures 1 and 2, and Supporting information, Supplementary Table 1) and was therefore classified as polysomic by acgh. Eleven cases harboured gains of 17q with involvement of the centromeric region (six with gains and five with of CEP17) (Figures 1 3 and Supporting information, Supplementary Table 1) and one case displayed 17q gain without affecting CEP17 (Table 1 and Supporting information, Supplementary Table 1). The remaining five cases (K11, K12, K3, K9, and K2) displayed only of the centromeric region (Tables 1 and 2 and Figure 1) without gains of the short or long arm. In the whole cohort, ten cases harboured (aws ratios >0.45) of two or more BACs mapping to the centromeric/pericentromeric region of Chr17 (Tables 1 and 2 and Figures 1 3), all of which were classified as polysomic by FISH (Table 1 and Supporting information, Supplementary Table 1). Likewise, all but two cases (K23 and K27) displaying CEP17 copy number gain were classified as polysomic by FISH; however, these two discordant cases displayed borderline ratios for gains by acgh analysis (Tables 1 and 2). One FISH-defined disomic case (K23, CEP17 copy number mean = 2.48, Supporting information, Supplementary Table 1) could be defined as polysomic by acgh (Table 1). Cases K24 and K25, considered disomic by FISH (CEP17 copy numbers of 2.30 and 1.55 per nucleus, respectively), displayed loss of CEP17 (aws ratios of 0.17) and loss of a substantial proportion of both Chr17 short and long arms. The genome-wide acgh profiles of case K24 demonstrated that this case had remarkably complex patterns of genomic aberrations and losses of Chr17 (Supporting information, Supplementary Figure 1) and case K25 displayed borderline deletion of CEP17 by FISH. Comparison between acgh and FISH for the detection of HER2 gene All but two cases displayed concordant results between dual-colour FISH and acgh (Tables 2 and 3 and Supporting information, Supplementary Table 1), with good agreement between the two techniques [unweighted Kappa coefficient = (95% confidence intervals of )]. All cases defined as HER2-amplified by FISH ratio greater than 2.2 displayed aws HER2 acgh ratios greater than 1.15 (high-level by acgh; Supporting information, Supplementary Table 1). However, two cases scored 2+ by IHC and identified as HER2-amplified by acgh would have been classified as not amplified according to the FISH ratio: (i) case K10 (Supporting information, Supplementary Table 1) displayed HER2 aws ratios of 1.31 () and HER2 average gene copy number as defined by FISH of 8.50 (); however, the HER2 /CEP17 FISH ratio was 1.41 (not amplified); (ii) case K31 (Supporting information, Supplementary Table 1) displayed HER2

4 HER2 assessment, chromosome 17 polysomy, and CEP17 copy number 19 Table 1. Chromosome 17 status assessment: comparison between HER2/CEP17 dual-colour FISH and microarray-based comparative genomic hybridization Short arm (17p) Long arm (17q) Case No CEP17 FISH results CEP17 FISH copy number mean (range) Gain %clones with aws ratio > 0.12 Loss %clones with aws ratio < 0.12 CEP17 aws ratio Gain %clones with aws ratio > 0.12 Loss %clones with aws ratio < 0.12 acgh description K23 Disomic 2.48 (1 4) Polysomic K24 Disomic 2.30 (2 4) CEP17 loss K25 Disomic 1.55 (1 2) CEP17 loss K26 Disomic 2.50 (2 4) Disomic K27 Disomic 2.48 (1 3) q gain K5 Disomic 1.88 (1 4) Disomic K19 Disomic 2.96 (2 4) p loss, 17q gain K20 Disomic 2.90 (2 4) Disomic K4 Disomic 2.61 (2 4) Disomic K7 Disomic 2.47 (1 3) q gain K11 Polysomic 5.20 (3 11) Not polysomic, CEP17 K12 Polysomic 3.30 (2 5) Not polysomic, CEP17 K13 Polysomic 5.10 (2 7) Not polysomic, 17q gain K21 Polysomic 3.20 (2 4) Not polysomic, 17q gain K22 Polysomic 3.50 (2 6) Polysomic K3 Polysomic 8.22 (2 12) Not polysomic, CEP17 K9 Polysomic 4.00 (3 5) Not polysomic, CEP17 K1 Polysomic 4.13 (2 10) Not polysomic, 17q gain, CEP17 K10 Polysomic 6.03 (3 9) Not polysomic, 17q gain, CEP17 K14 Polysomic 5.60 (2 6) Not polysomic, 17p loss, 17q gain K15 Polysomic 5.30 (2 8) Not polysomic, 17q gain K16 Polysomic 3.90 (2 5) Not polysomic, 17q gain K17 Polysomic 3.27 (2 5) Not polysomic, 17q gain K18 Polysomic 7.40 (3 10) Not polysomic, 17q gain, CEP17 K2 Polysomic 4.25 (3 7) Not polysomic, 17p loss, CEP17 K31 Polysomic 4.80 (3 12) Not polysomic, 17q gain, CEP17 K6 Polysomic 4.05 (2 6) Not polysomic, 17q gain K8 Polysomic 4.50 (3 6) Not polysomic, 17p loss, 17q gain, CEP17 acgh = microarray-based comparative genomic hybridization; aws = adaptive weights smoothing; CEP17 = chromosome 17 centromere; FISH = fluorescence in situ hybridization. Cases displaying CEP17 by acgh are shown in bold. aws ratios of 0.59 (), HER2 average gene copy number of 5.2 (equivocal), and a HER2 /CEP17 FISH ratio of 1.08 (not amplified). It should be noted that in these two cases, no evidence of true Chr17 polysomy was observed by acgh analysis (Table 1 and Figure 3). Four other cases scored as equivocal by IHC (K19, K13, K15, and K18) and classified as equivocal by the HER2 single probe did not display as defined by acgh; instead, these cases harboured gains of the HER2 region (0.12 < aws < 0.45) without true Chr17 polysomy (Supporting information, Supplementary Table 1). FISH analysis of SMS and RARA genes andcomparisonwithacghdata To investigate Chr17 rearrangement and polysomy by FISH further, we assessed SMS and RARA copy numbers (Supporting information, Supplementary Table 2). One (K10) of the two cases displaying

5 20 CMarchiò et al Figure 1. Schematic representation of chromosome 17. The mapping positions of SMS (17p11.2), HER2 (17q12), and RARA (17q21.2) are illustrated. On the left-hand side, the genomic position corresponding to each cytoband is shown. CEP17 = chromosome 17 centromere; p = short chromosomal arm, q = long chromosomal arm. On the right-hand side, a schematic representation of chromosome 17 rearrangement as defined by acgh criteria (see the Materials and methods section and Table 1) is shown (grey = no change; green = gain; blue = ; losses are symbolized by the lack of a chromosomal portion ie white) high copy number of both HER2 and CEP17 benefited from correction of HER2 copies by SMS mean number of signals (Table 4 and Supporting information, Supplementary Tables 1 and 2). It should be noted, however, that the use of additional probes proved not to be sufficient to thoroughly define the rearrangement of Chr17 in all troublesome cases. For instance, case K31 displayed a FISH pattern which was highly suggestive for polysomy (similar mean number of signals for SMS, HER2, CEP17, and RARA). However, acgh analysis revealed a gain of the long arm of Chr17 involving the centromere and only a limited part of the small arm including the region where SMS maps to (Figure 3); hence the case was not defined as polysomic by acgh. The use of RARA copy numbers was of very limited help (Supporting information, Supplementary Table 2), given the frequent HER2 /RARA co. In our series, we found of this locus in 2/28 cases (K12 and K10) and equivocal RARA gene copy number in seven cases (K26, K27, K19, K21, K22, K15, and K31) (Table 4 and Supporting information, Supplementary Table 2). All cases displaying RARA as defined by singlecolour FISH displayed aws ratios greater than 0.45 by acgh analysis; an equivocal RARA status corresponded to copy number gains by acgh in all but two cases (K22 and K31), which were amplified by acgh (Supporting information, Supplementary Table 2). Only one sample, K31 classified as equivocal according to SMS copy number mean by FISH (4.01 copies per nucleus, Supporting information, Supplementary Table 2) harboured SMS as detected by acgh. Interestingly, seven out of 28 cases (25%) could be defined as hypodisomic (<1.5 signals per cell) [3,4] by single-colour FISH with SMS probes (Supporting information, Supplementary Table 2). Four of these cases displayed a loss affecting the SMS locus as defined by acgh (K5, K2, K6, and K8); two harboured borderline ratios for losses by acgh analysis (K3 and K14); and one displayed near-disomic acgh ratios (K12; Supporting information, Supplementary Table 2). It should be noted that HER2 /SMS ratio would have correctly assessed HER2 in K10, which was defined as not amplified/polysomic by HER2 /CEP17 ratio; on the other hand, it would have overestimated HER2 in K13, K14, K6, K18, and K8 (Supporting information, Supplementary Tables 1 and 2). Hence, our results demonstrate that the value of correcting HER2

6 HER2 assessment, chromosome 17 polysomy, and CEP17 copy number 21 Figure 2. acgh and FISH analysis of cases K11 (A C), K1 (D F), and K22 (G I). Representative chromosome plots and FISH images of FISH-defined polysomic cases. Chromosome 17 plots (A, D, G): log 2 ratios are plotted in grey and aws ratios in blue on the Y-axis against each clone according to genomic location on the X-axis; CEP17 is represented by a vertical dotted line. The genomic position of HER2 is highlighted by an orange box; the greenish box on the small arm highlights the SMS locus; and the greenish box on the long arm depicts the RARA locus. (B, C, E, F, H, I) Representative FISH images of the corresponding cases; insets highlight representative nuclei in each case. K11 harbours HER2 and of the centromeric region without gains of either the long or the short arm of chromosome 17 (A). FISH images demonstrate HER2 (red signals in B), a polysomic CEP17 copy number (green in B), and equivocal numbers of SMS (red in C) and RARA (green in C) signals. K1 displays of centromeric and proximal and distal pericentromeric BACs of chromosome 17 (A). Dual-colour FISH with probes for HER2 and CEP17 suggests polysomy of Chr17 (E): CEP17 (green in E) mean copy number of 4.13 and HER2 (red) copy numbers of SMS (red in F) and RARA (green in F) probes displayed a mean copy number of 2.6 and 2.35, respectively, and confirmed the acgh results, with only focal gains of Chr17. K22 was defined as polysomic by both acgh (G) and FISH (H, I) copy numbers using SMS copy numbers to determine HER2 gene status is limited. Discussion An accurate assessment of HER2 status and Chr17 polysomy is of paramount importance to identify patients eligible for trastuzumab therapy [18,19]. HER2 gene status assessment by dual-colour FISH is occasionally troublesome [2,6,8] and this is particularly the case in tumours displaying abnormal copy numbers of CEP17. Polysomy of Chr17, as defined by dual-colour FISH, is observed in approximately 8% of all breast cancer specimens [2,4,5], mostly among cases with four to six HER2 gene copies (the so-called equivocal range ) [4,5]. Polysomy is cytogenetically defined as the occurrence in a nucleus of extra copies of one or more individual chromosomes, and increases in copy number due to polysomy do not have the same biological impact as those caused by gene [20].

7 22 CMarchiò et al Figure 3. acgh and FISH analysis of cases K10 (A C) and K31 (D F). Representative chromosome plots of cases K10 (A) and K31 (D). Log 2 ratios are plotted in grey and aws ratios in blue on the Y-axis against each clone according to genomic location on the X-axis; CEP17 is represented by a vertical dotted line. The genomic position of HER2 is highlighted by an orange box; the greenish box on the small arm highlights the SMS locus and the greenish box on the long arm depicts the RARA locus. Representative FISH images of the corresponding cases for HER2/CEP17 (B, E), SMS (red), and RARA (green) (C, F). Insets highlight representative nuclei in each case. Both cases lack concurrent gain of both the short and the long arm, ruling out chromosome 17 polysomy; on the other hand, they both display gain of 17q and CEP17 (aws ratios = 1.31 and 0.59, respectively). FISH images of case K10 show high HER2 and CEP17 copy number (mean = 8.5 and 6.03, respectively); SMS was considered normal (red signals in C) and RARA showed (green signals in C). FISH images of case K31 illustrate a pattern highly suggestive of chromosome 17 polysomy: an average of 5.12 for HER2 (red signals in E), 4.8 for CEP17 (green signals in E), 4.01 of SMS signals (red signals in F), and 4.07 of RARA signals (green signals in F) FISH assessment is currently performed in routine histological sections and pathologists have employed probes for chromosome centromeres as surrogates of cytogenetic analysis to detect polysomy. However, there is currently no accepted definition of true polysomy when routine histological sections are employed and different groups have employed different criteria to define it [2]. There are data to suggest that Chr17 polysomic breast cancers have clinicopathological features similar to those of HER2-negative disease; however, 60 75% of polysomic breast cancers devoid of HER2 display equivocal immunohistochemical results (2+) [19]. Here we have demonstrated that the presence of abnormal CEP17 copy numbers by FISH is the result of CEP17 polysomy in a minority of cases (5.5%). It has been previously hypothesized that abnormal CEP17 signals could stem from of the centromeric region [6,8]. Coupling FISH and acgh analysis, we have provided the first direct evidence that additional copies of CEP17 by FISH are frequently caused by CEP17 copy gains or and that these phenomena are more prevalent than true Chr17 polysomy: out of the 18 polysomic cases as defined by FISH, seven cases (38.9%) displayed CEP17 copy number gains and ten (55.5%) CEP17. The reliability of the results on CEP17 by acgh is proven by the high concordance between FISH and acgh analysis in the evaluation of HER2 status. These findings are further supported by preliminary data presented in abstract form, where true CEP17 polysomy was also suggested to be an uncommon phenomenon [21]. Our results have important implications for a correct definition of HER2 gene status and therefore for the selection of patients eligible for trastuzumab treatment in IHC equivocal cases. In fact, case K10, which was equivocal by IHC analysis (2+), displayed a dual-colour FISH ratio of 1.41 (not amplified), and this patient would not be eligible for trastuzumab therapy. Microarray CGH analysis confirmed the presence of of the HER2 gene and of the centromeric region. Interestingly, the mean copy numbers of HER2 and CEP17 were 8.5 and 6, respectively, demonstrating that in contrast to previous claims that all cases should be corrected by CEP17 copy numbers [3], there are important exceptions to this rule. Moreover, dual-colour FISH assessment of case K31 proved to be troublesome, as it was diagnosed as not

8 HER2 assessment, chromosome 17 polysomy, and CEP17 copy number 23 Table 2. Comparison between FISH and acgh data in the assessment of HER2 and CEP17 status FISH HER2/CEP17 ratio CEP17 copy number Not amplified (N = 16) Amplified (N = 12) (N = 10) > 3 (range ; mean 4.8) (N = 18) acgh HER2 Not amplified 14 (87.5%) 0 4 (40%) 12 (66.7%) Amplified 2 (12.5%) 12 (100%) 6 (60%) 6 (33.3%) CEP17 Normal 4 (25%) 3 (25%) 6 (60%) 1 (5.6%) Gained/Amplified 12 (75%) 7 (58.3%) 2 (20%) 17 (94.4%) Lost 0 2 (16.7%) 2 (20%) 0 acgh = microarray-based comparative genomic hybridization; CEP17 = chromosome 17 centromere; FISH = fluorescence in situ hybridization. Table 3. Concordance between single- and dual-colour FISH and acgh in the assessment of HER2 HER2/CEP17 FISH ratio HER2 FISH copy number > 6 Not amplified (N = 16) Amplified (N = 12) Not amplified (N = 15) Amplified (N = 13) HER2 acgh Not amplified 14 (87.5%) 0 14 (93.3%) 0 Amplified 2 (12.5%) 12 (100%) 1 (6.7%) 13 (100%) acgh = microarray-based comparative genomic hybridization; CEP17 = chromosome 17 centromere; FISH = fluorescence in situ hybridization. Table 4. Comparison between HER2/SMS FISH ratio and acgh in the assessment of HER2 and RARA status in the whole cohort as assessed by FISH mean copy number FISH HER2/SMS ratio RARA copy number Not amplified (N = 9) Amplified (N = 19) < 4(N = 19) 4 6 (N = 7) > 6(N = 2) acgh HER2 Not amplified 9 (100%) 6 (31.6%) 12 (63.2%) 2 (28.6%) 0 Amplified 0 13 (68.4%) 7 (36.8%) 5 (71.4%) 2 (100%) CEP17 Lost 0 2 (10.5%) 2 (10.5%) 0 0 Normal 4 (44.4%) 3 (15.8%) 4 (21.1%) 3 (42.9%) 0 Gained/Amplified 5 (55.6%) 14 (73.7%) 13 (68.4%) 4 (57.1%) 2 (100%) acgh = microarray-based comparative genomic hybridization; CEP17 = chromosome 17 centromere; FISH = fluorescence in situ hybridization. amplified by FISH ratio (HER2 /CEP17 ratios <2.2), borderline by HER2 mean copy number, and equivocal by IHC. Array CGH analysis demonstrated that this tumour harboured a gain of the long arm of Chr17 with of the centromeric region. Probes for genes mapping to the short (SMS ) and the long (RARA) arms of Chr17 [8] have been proposed as alternative methods to define accurately HER2 gene status (Figure 1). Although these additional probes may be of assistance in some cases, it is likely that in a substantial proportion of cases the information provided by this analysis would not be sufficient to address this diagnostic dilemma, as demonstrated in the present study. This is mainly the result of the frequent loss of the short arm of Chr17 encompassing the SMS locus in breast cancer [22] and of RARA, which maps to the smallest region of of the TOP2A amplicon [7,23] and is found in up to 57% of HER2-amplified breast cancers [24]. In our study, almost perfect agreement between the single HER2-FISH probe and acgh was observed, and all cases defined as amplified by FISH were amplified by acgh. Further re-analysis of trastuzumabsensitive tumours considered HER2-not amplified in central analysis with dual-colour FISH [25] using other methods, such as acgh, may help to clarify the reasons for the response to trastuzumab despite the apparent lack of HER2. Taken together, our results confirm that Chr17 usually displays very complex rearrangements and demonstrate that true Chr17 polysomy is not the common denominator of increased numbers of CEP17 in dual-colour FISH analysis. We have also demonstrated that abnormal CEP17 copy numbers may actually stem from high-level gains/ of CEP17, regardless of copy number gains of the short and long arms. These findings demonstrate that in some cases, correction with CEP17 probes may provide misleading HER2 gene status assessment results and that Chr17 polysomy cannot be accurately defined by FISH with CEP17 probes.

9 24 CMarchiò et al Acknowledgements This study was funded by Breakthrough Breast Cancer, Compagnia San Paolo di Torino, Cassa di Risparmio di Torino, AIRC , Ricerca Finalizzata Regione Piemonte 2008, and Regione Piemonte-CIPE CM was funded in part by Breakthrough Breast Cancer. We also acknowledge NHS funding to the NIHR Biomedical Research Centre. References 1. Lambros MB, Natrajan R, Reis-Filho JS. Chromogenic and fluorescent in situ hybridization in breast cancer. Hum Pathol 2007;38: Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 2007;25: Dal Lago L, Durbecq V, Desmedt C, Salgado R, Verjat T, Lespagnard L, et al. Correction for chromosome-17 is critical for the determination of true Her-2/neu gene status in breast cancer. Mol Cancer Ther 2006;5: Ma Y, Lespagnard L, Durbecq V, Paesmans M, Desmedt C, Gomez-Galdon M, et al. Polysomy 17 in HER-2/neu status elaboration in breast cancer: effect on daily practice. Clin Cancer Res 2005;11: Reddy JC, Reimann JD, Anderson SM, Klein PM. Concordance between central and local laboratory HER2 testing from a community-based clinical study. Clin Breast Cancer 2006;7: Isola J, Tanner M, Forsyth A, Cooke TG, Watters AD, Bartlett JM. Interlaboratory comparison of HER-2 oncogene as detected by chromogenic and fluorescence in situ hybridization. Clin Cancer Res 2004;10: Arriola E, Marchio C, Tan DS, Drury SC, Lambros MB, Natrajan R, et al. Genomic analysis of the HER2/TOP2A amplicon in breast cancer and breast cancer cell lines. Lab Invest 2008;88: Troxell ML, Bangs CD, Lawce HJ, Galperin IB, Baiyee D, West RB, et al. Evaluation of Her-2/neu status in carcinomas with amplified chromosome 17 centromere locus. Am J Clin Pathol 2006;126: Marchio C, Iravani M, Natrajan R, Lambros MB, Savage K, Tamber N, et al. Genomic and immunophenotypical characterization of pure micropapillary carcinomas of the breast. J Pathol 2008;215: Marchio C, Natrajan R, Shiu KK, Lambros MB, Rodriguez- Pinilla SM, Tan DS, et al. The genomic profile of HER2- amplified breast cancers: the influence of ER status. J Pathol 2008;216: Arriola E, Lambros MB, Jones C, Dexter T, Mackay A, Tan DS, et al. Evaluation of Phi29-based whole-genome for microarray-based comparative genomic hybridisation. Lab Invest 2007;87: van Beers EH, Joosse SA, Ligtenberg MJ, Fles R, Hogervorst FB, Verhoef S, et al. A multiplex PCR predictor for acgh success of FFPE samples. Br J Cancer 2006;94: Coe BP, Ylstra B, Carvalho B, Meijer GA, Macaulay C, Lam WL. Resolving the resolution of array CGH. Genomics 2007;89: Gunnarsson R, Staaf J, Jansson M, Ottesen AM, Goransson H, Liljedahl U, et al. Screening for copy-number alterations and loss of heterozygosity in chronic lymphocytic leukemia a comparative study of four differently designed, high resolution microarray platforms. Genes Chromosomes Cancer 2008;47: Tan DS, Lambros MB, Natrajan R, Reis-Filho JS. Getting it right: designing microarray (and not microawry ) comparative genomic hybridization studies for cancer research. Lab Invest 2007;87: Huang J, Gusnanto A, O Sullivan K, Staaf J, Borg A, Pawitan Y. Robust smooth segmentation approach for array CGH data analysis. Bioinformatics 2007;23: Natrajan R, Lambros MB, Geyer F, Marchiò C, Tan DSP, Vatcheva R, et al. Loss of 16q in high grade breast cancer is associated with oestrogen receptor status: evidence for progression in tumours with a luminal phenotype? Genes, Chromosomes Cancer 2009;48: Rosenberg CL. Polysomy 17 and HER-2 : true, true, and unrelated. J Clin Oncol 2008;26: Vanden Bempt I, Van Loo P, Drijkoningen M, Neven P, Smeets A, Christiaens MR, et al. Polysomy 17 in breast cancer: clinicopathologic significance and impact on HER-2 testing. J Clin Oncol 2008;26: Albertson DG. Gene in cancer. Trends Genet 2006;22: Yeh I-T, Martin MA, Robetorye RS, Gunn S. Polysomy 17 by FISH for HER2 in breast cancer is a misnomer: results of chromosome 17 analysis by array comparative genomic hybridization. Cancer Res 2008;69:207S 207S. 22. Natrajan R, Lambros MB, Rodriguez-Pinilla SM, Moreno-Bueno G, Tan DS, Marchio C, et al. Tiling path genomic profiling of grade 3 invasive ductal breast cancers. Clin Cancer Res 2009;15: Jarvinen TA, Tanner M, Barlund M, Borg A, Isola J. Characterization of topoisomerase II alpha gene and deletion in breast cancer. Genes Chromosomes Cancer 1999;26: Mano MS, Rosa DD, De Azambuja E, Ismael GF, Durbecq V. The 17q12 q21 amplicon: Her2 and topoisomerase-iialpha and their importance to the biology of solid tumours. Cancer Treat Rev 2007;33: Paik S, Kim C, Wolmark N. HER2 status and benefit from adjuvant trastuzumab in breast cancer. N Engl J Med 2008;358: SUPPORTING INFORMATION ON THE INTERNET The following supporting information may be found in the online version of this article. Table S1: Summary of HER2 immunohistochemical expression results, FISH and array-cgh data for HER2 and CEP17 in the whole cohort. Table S2: Comparison between FISH and array-cgh data on SMS and RARA genes and HER2 status as assessed by HER2/SMS ratio. Figure S1: Representative chromosomes plot of case K24. Supplementary Data - Adaptive Weight Smoothed Data.