The MYC proto-oncogene maps to 8q24.1 and encodes at least three major transcripts, two of them containing the transactivation domain (c-myc1 and c-myc2).^1–4 Depending on the context and on the isoform, this transcription factor can promote either cell proliferation or apoptosis.^1–4 Given the plethora of biological roles played by MYC gene products and the lack of reliable antibodies for immunohistochemical analysis, the analysis of the clinical relevance of c-myc overexpression in breast cancer has produced conflicting results.² On the other hand, MYC gene amplification has been extensively studied in breast cancer.^1,2 The prevalence of MYC amplifications ranges from 1.1% to 94.4% of cases, depending on the cohort of patients and the techniques used.^{1,2,5,6,7,8,9,10,11,12,13,14,15,16,17,18} In a recent meta-analysis, largely based on studies where MYC gene copy numbers were analysed by means of Southern blot, slot blot or PCR-based techniques,¹MYC gene amplification was shown to be associated with high histological grade, presence of lymph node metastasis, lack of progesterone receptor and poor survival.

In situ methods are currently considered the “gold standard” for gene copy number assessment. Fluorescent in situ hybridisation (FISH) has been applied to the study of MYC amplification in breast cancer,^{5–7,9–13,15–17,19,20} with frequencies ranging from 5.3%¹² to 86%.¹¹ With this method, the correlations between MYC amplification and clinicopathological parameters have been surprisingly inconsistent and, although MYC amplification has been shown to be associated with poor prognosis,¹ it is still unclear whether MYC amplification is an independent prognostic factor in invasive breast cancer.

Chromogenic in situ hybridisation (CISH) has been previously used to determine the prevalence of MYC amplification. In a cohort of 177 breast cancer patients, Rummukainen et al⁸ found that 15.25% of cases showed MYC amplification; this was correlated with aneuploidy, high histological grade, high S-phase fraction, lack of progesterone receptor and shorter overall survival.⁸ However, in that study, only a MYC gene specific probe was employed. Although a good agreement between MYC as defined by dual colour FISH and single probe CISH for MYC gene amplification has been described (unweighted κ = 0.68),⁸ the correlations with clinicopathological features and prognostic impact of MYC amplification defined by each technique seem to differ.⁸ Owing to (i) the very high frequency of 8q gains, in the form of additional isochromomes 8, in high grade breast cancer, and (ii) the fact that by single probe CISH analysis, cases defined as “MYC non-amplified but aneuploid” (4–5 copies/cell) also showed a poor prognosis when compared to cases with 1–3 copies of MYC gene,⁸ the use of a centromeric probe seems to be required to differentiate between polysomy of chromosome 8 and MYC gene amplification, and to reliably identify MYC low level amplifications (for example, five copies of MYC but only two copies of chromosome 8 centromere).

It has recently been shown that MYC gene amplifications are significantly more prevalent in tumours arising in BRCA1 mutation carriers.^15,16 Given the phenotypic and molecular similarities between these tumours and sporadic basal-like breast carcinomas,^21–26 we hypothesised that MYC amplification could also play a role in the biology of basal-like breast carcinomas.

In this study, utilising a centromeric probe for chromosome 8 (CEP8) and a gene specific probe for MYC, we investigated the correlation between MYC amplification and various clinicopathological parameters and outcome in a cohort of 245 women with invasive breast carcinoma treated with surgery followed by adjuvant anthracycline-based chemotherapy. A second aim was to define the prevalence of MYC amplification in “basal-like” breast carcinomas.

MATERIALS AND METHODS

Tissue microarrays

The tissue microarray contained replicate 0.6 mm cores of 245 invasive breast carcinomas (185 invasive ductal carcinomas, 27 invasive lobular carcinomas, 25 invasive mixed carcinomas and 8 invasive breast carcinomas of other special types). All patients were treated with therapeutic surgery (69 mastectomy and 155 wide local excision) and adjuvant anthracycline-based chemotherapy; those with oestrogen receptor (ER) positive tumours also received endocrine therapy. Follow-up was available for 244 patients, ranging from 0.5 to 125 months (median 67 months, mean 67 months). Full details of the characterisation of the tissue microarray and the cohort of patients are described elsewhere.^27,28 Tumours were graded according to the modified Bloom–Richardson scoring system²⁹ and size was categorised according to the TNM staging criteria. Details on the expression of ER, progesterone receptor (PR), HER2, epidermal growth factor receptor (EGFR), cytokeratin (Ck) 5/6, Ck 14 and Ck 17 are described elsewhere.^27,28 Tumours were classified into basal-like, luminal or HER2 groups according to the immunohistochemical panel proposed by Nielsen et al.³⁰ This study was approved by the Royal Marsden Hospital Ethics Committee.

Chromogenic in situ hybridisation

CISH for MYC and chromosome 8 centromere was performed on serial tissue microarray sections as previously described,^27,31 using the ready-to-use digoxigenin-labelled SPoT-Light C-MYC amplification probe (Zymed, San Francisco, California, USA) and biotin-labelled SPoT-Light Chromosome 8 Centromeric Probe (Zymed, San Francisco, California, USA). Heat pretreatment of deparaffinised sections consisted of incubation for 15 min at 98°C in CISH pretreatment buffer (SPOT-light tissue pretreatment kit, Zymed) and digested with pepsin for 6 min at room temperature according to the manufacturer’s instructions. An appropriate MYC gene-amplified breast tumour control was included in the slide run. CISH experiments were analysed by three of the authors (SMRP, SEP and JSR-F) on a multi-headed microscope. Only unequivocal signals were counted. Signals were evaluated at 400× and 630× magnification; 30 morphologically unequivocal neoplastic cells in each core were assessed for the presence of the gene and chromosome 8 centromere probe signals. Amplification was defined as a MYC:CEP8 ratio >2. The scoring was evaluated with observers blinded to the clinicopathological details and patients’ outcome.

Statistical analysis

The StatView V.5.0 software package (SAS Institute Inc., Cary, NC, USA) was used for all calculations. Correlations between categorical variables were performed using the χ² test, and Fisher’s exact test where appropriate. Correlations between continuous and categorical variables were performed with analysis of variance. Metastasis-free and overall survival was expressed as the number of months from diagnosis to the occurrence of an event (distant metastasis or disease-related death, respectively). Cumulative survival probabilities were calculated using the Kaplan–Meier method. Differences between survival rates were tested with the log-rank test. All tests were two-tailed, with a confidence interval of 95%.

Multivariate analysis was performed using the Cox multiple hazards model. A p value of 0.05 in the univariate survival analysis was adopted as the limit for inclusion in the multivariate model, and cases with missing values were excluded from this analysis.

RESULTS

Table 1 summarises he correlations between MYC amplification and clinicopathological features and immunohistochemical findings in 245 breast carcinomas. Briefly, 49 cores were either lost/fragmented in the CISH procedure, did not have invasive tumour or showed suboptimal signals for either MYC or CEP8. Of the 196 remaining tumours, 19 (9.7%) showed MYC:CEP8 ratios >2.0. Cases with MYC amplification were seen in the form of large clusters of MYC signals (fig 1) or multiple individual signals/nucleus (fig 1). MYC amplification was found only in grade 2 and grade 3 breast carcinomas (18 invasive ductal carcinomas and 1 mixed ductal-lobular carcinoma/pleomorphic lobular carcinoma³²). However, no significant correlation between MYC amplification and histological grade was found (only 19 grade 1 carcinomas were included in the series) and no correlation between MYC amplification and tumour size, presence of lymph node metastasis or lympho-vascular invasion was identified. No association between MYC amplification and ER, PR, HER2, EGFR, Ck 5/6, Ck 14 or Ck 17 and p53 expression was observed. MYC amplification showed a trend for a higher proliferation rate, as defined by MIB1 expression (p = 0.1084). No correlation between basal-phenotype, as defined by the immunohistochemical panel proposed by Nielsen et al,³⁰ and MYC amplification was found. In fact, only 4% of basal-like carcinomas showed MYC amplification, compared to 8.75% and 10.7% observed in luminal and HER2 tumours, respectively (p = 0.6134, NS).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

 MYC gene copy numbers in breast cancer: grade 1 invasive ductal carcinoma (A) with 1–3 chromosome 8 centromere (CEP8, B) and MYC (C) gene signals. Metaplastic spindle cell breast carcinoma (D) with chromosome 8 polysomy (E) and an average of 4.5 copies of MYC (F); the MYC:CEP8 ratio was 1.59. Grade 2 invasive ductal carcinoma (G) harbouring “low level” MYC amplification: note the presence of 1–3 copies of CEP8 (H) but increased copy numbers of MYC (I); the MYC:CEP8 ratio was 2.24. Grade 3 invasive ductal carcinoma (J) with 1–3 copies of chromosome 8 centromere (K) and clusters of MYC gene signals (L). Inset: MYC gene signal clusters. Mixed ductal-lobular carcinoma (M) with 1–3 CEP8 copies (N) and MYC large signal clusters (O). (Original magnification ×200—A, D, G, J and M; ×630—B, C, E, F, H, I, K, L, N and O; ×1000—inset L.)

When data on the replicate cores of each tumour were treated independently, the agreement for MYC amplification, considering the categories not amplified and amplified was good, (unweighted κ = 0.7916 (0.6152–0.968)). MYC amplification was found in both cores in 11 cases and in one of the cores in 5 cases. In 3 cases, one of the cores showed amplification, whereas the other was lost during hybridisation. No statistically significant differences in tumour size, histological grade, presence of vascular invasion or lymph node metastasis, ER, PR, HER2, EGFR, Ck 14, Ck 5/6 and Ck 17 expression, and MIB1 proliferation rates were observed between cases homogeneously (ie, both cores showing MYC amplification) or heterogeneously (ie, one core showing MYC amplification and the other with normal copy numbers) amplified for MYC gene (data not shown). However, given the limited sample size of the cohort of MYC amplified cases, type 2 or β errors cannot be excluded in the present study. Interestingly, all cases with heterogeneously amplified MYC (n = 5) lacked p53 nuclear expression, whereas 6 of 11 cases with homogeneous MYC amplifications showed p53 positivity (p<0.100, χ² test).

If MYC amplifications were defined based only on the MYC gene probe copy numbers as described by Rummukainen et al,⁸ 12 of 19 amplified cases would have been classified as amplified, whereas 1 non-amplified case by MYC:CEP8 ratios would have been classified as amplified (unweighted κ = 0.7286 (0.55–0.9072). MYC amplification as defined by single probe CISH showed a borderline statistical correlation with larger tumour size and higher proliferation rate (p<0.1000, χ²). No basal-like carcinomas showed MYC amplification (0/26), whereas 7% of both luminal and HER2 tumours were MYC amplified (data not shown).

Survival analysis

In this cohort, MYC amplification, tumour size, presence of lympho-vascular invasion, lymph node metastasis at time of diagnosis, ER, PR, Ck 14, Ck 5/6, Ck 17, p53 and proliferation index as defined by MIB1 were statistically significant prognostic factors for metastasis-free survival on univariate analysis (table 2, fig 2A). On multivariate Cox hazard analysis, including all parameters that showed an association with metastasis-free survival on univariate analysis, presence of MYC amplification and lymph node metastasis were shown to be independent prognostic factors (table 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

 Univariate analysis of the prognostic impact of MYC gene amplification on metastasis-free (A) and overall survival (B).

Univariate survival analysis revealed MYC amplification, lymph node metastasis at time of diagnosis, ER, PR, Ck 14, Ck 5/6, Ck 17, p53, and proliferation index as defined by MIB1 as prognostic factors for overall survival (table 2, fig 2B). On multivariate Cox hazard analysis, including all parameters that showed an association with overall survival on univariate analysis, MYC amplification, the presence of lymph node metastasis and p53 expression were shown to be independent prognostic factors for OS (table 4).

MYC amplification defined by single probe CISH was not significantly correlated with metastasis-free or overall survival (p>0.05, data not shown).

DISCUSSION

Chromogenic in situ hybridisation has proven to be a useful technique to determine gene copy numbers and gene amplification on formalin-fixed, paraffin-embedded tissue sections.^8,27,31 Unlike FISH, CISH allows a direct comparison between morphological features of neoplastic cells and the presence of gene amplification.^8,27,31 Furthermore, CISH analysis is relatively quick; in the present study, the whole analysis of two probes (ie, CEP8 and MYC) in 245 replicate cores took less than a week, the neoplastic cells were easily recognisable and only 20% of the cases were not interpretable for one or the other probe.

In the present study, MYC amplification was shown to be a poor prognostic factor for distant metastasis and overall survival, in agreement with previous studies.^{1,2,8,13,18,33–35} However, we also show that the prognostic impact of MYC amplification is independent of size, histological grade, lympho-vascular invasion, lymph node status, and expression of ER, PR, basal markers, p53 and proliferation rate. Although this study is retrospective, the differences in survival rates are not confounded by the therapeutic regimens, given that all patients were treated with surgery and adjuvant anthracycline-based chemotherapy. Alternatively, the independent prognostic impact of MYC amplification observed in this cohort of patients could be explained by a reduced sensitivity of MYC amplified cases to anthracycline-based chemotherapy. Although this possibility could not be completely ruled out without a non-treated control arm, a direct mechanistic association between MYC amplification and reduced sensitivity to anthracyclines has never been described in vivo.

The distribution of MYC amplification within subclones of breast carcinomas is reported to be remarkably heterogeneous.¹⁴ In our analysis, of 16 cases harbouring MYC amplification where two cores of each tumour rendered results for both CEP8 and MYC, MYC was homogeneously amplified (ie, both cores showed amplification) in 68.75% and heterogeneously amplified in 31.25% of cases (ie, one core with MYC amplification and another with normal copy numbers). If only one core per tumour was used in the present study, 3 of 19 cases with MYC amplification would have been classified as “non-amplified”. These results suggest that tissue microarray analysis of MYC amplification need more than one core per tumour.

In the present study, single probe CISH analysis was slightly less sensitive that CISH with two probes (MYC and CEP8). As expected, all but one case with large signal clusters and/or >5 signals/cell using one probe also displayed MYC:CEP8 ratios >2. However, the use of a centromeric probe allowed for the identification of cases with MYC low level amplification (MYC:CEP8 ratios >2, but ⩽5 MYC gene signals on average). This is not surprising given that the presence of >5 copies of a given chromosome in primary breast cancer and breast cancer cell lines is an exceedingly rare biological phenomenon³⁶ (The Cancer Genome Anatomy Project, http://cgap.nci.nih.gov/Chromosomes). The agreement between MYC amplification as defined by CISH using one or two probes was similar to that described for CISH with one probe and dual colour FISH.⁸

We, and others, have recently shown that sporadic basal-like tumours share several morphological, immunophenotypic, epigenetic and genetic characteristics with tumours arising in BRCA1 mutation carriers.^22–26 Although MYC amplification has been reported to be remarkably frequent in tumours arising in BRCA1 mutation carriers, our results suggest that this is not one of the underlying genetic events driving the biology of sporadic basal-like carcinomas, as MYC amplification was seen in only 1 of 25 cases (4%). This finding further corroborates the results of Adler et al,³⁷ who showed that MYC amplification is one of the drivers of the activated serum signature, but is not related to the basal-like phenotype. In fact, the basal-like signature does not seem to be activated by MYC.³⁷ However, some special types of basal-like breast carcinomas, namely medullary carcinomas¹² and metaplastic breast cancers (Reis-Filho et al, unpublished observations), seem to have frequent MYC amplifications. Further studies analysing the prevalence of MYC copy number gains in a larger cohort of basal-like breast carcinomas, including the special types, are warranted.

In the present study, MYC amplification was an independent predictor of distant metastasis. Interestingly, the activated serum/“wound response” gene expression signature,³⁸ which has been shown to be induced by MYC amplification,^37,39 is strongly associated with increased risk of distant metastasis. Taken together, these results suggest that MYC amplification may play an important biological role in the later stages of tumour progression,¹³ in particular in the activation of a transcriptional programme that promotes the development of breast cancer metastasis.^38–40

In conclusion, this study shows that although MYC amplification is heterogeneous in breast carcinomas, it can be reliably and rapidly assessed on tissue microarrays by means of CISH using a combination of a gene specific probe for MYC and a centromeric probe for chromosome 8. However, more than one core of each tumour should be included. Further studies analysing the concordance of results on MYC amplification obtained with CISH performed on whole tissue sections and tissue microarrays are warranted. MYC amplification has proven to be an independent prognostic factor for metastasis-free and overall survival. Prospective studies assessing the additional prognostic information provided by MYC amplification analysis in addition to that offered by traditional clinicopathological and immunohistochemical parameters are required to further confirm these findings.