CCND1 encodes for the cyclin D1 protein involved in G1/S cell cycle transition. In breast cancer the mechanism of CCND1 amplification, relationship between cyclin D1 protein expression and the key clinical markers estrogen receptor (ER) and HER2 requires elucidation. Tissue microarrays of primary invasive breast cancer from 93 women were evaluated for CCND1 amplification by fluorescent in-situ hybridization and cyclin D1 protein overexpression by immunohistochemistry. CCND1 amplification was identified in 27/93 (30%) cancers and 59/93 (63%) cancers had overexpression of cyclin D1. CCND1 amplification was significantly associated with cyclin D1 protein overexpression (p < 0.001; Fisher's exact test) and both CCND1 amplification and cyclin D1 protein expression with oestrogen receptor (ER) expression (p = 0.003 and p < 0.001; Fishers exact test). Neither CCND1 amplification nor cyclinD1 expression was associated with tumor size, pathological node status or HER2 amplification, but high CCND1 amplification (Copy Number Gain (CNG) ≥ 8) was associated with high tumor grade (p = 0.005; chi square 7.915, 2 df) and worse prognosis by Nottingham Prognostic Index (p = 0.001; 2 sample t-test). High CCND1 amplification (CNG ≥ 8) may identify a subset of patients with poor prognosis ER-positive breast cancers who should be considered for additional therapy.
The clinical benefits of molecular targeted therapy have recently become established in breast cancer with the advent of therapies targeting the HER2 gene.1, 2 Unlike haematological malignancies, technical challenges, the complex patterns of acquired genetic changes and high levels of chromosomal instability in human solid cancers have made further target identification and therapeutic development challenging.3
One promising gene with clinical potential is the CCND1 (PRAD1) gene located on chromosome 11q13 and amplified in 5–20% of primary breast cancers4, 5 but deleted in 5–9%.4, 5 A functional polymorphism of CCND1 may contribute to the development of breast cancer6, 7 and may also be associated with downregulation of HER2.7CCND1 is an estrogen-responsive gene with oncogenic potential via cell cycle regulation at the G1/S phase transition8, 9 and transcriptional potentiation of ER activity.10, 11
Gene amplification is an important cytogenetic manifestation of genetic instability. True gene amplification is mechanistically distinct from other forms of DNA copy number change and is associated with well recognized abnormal chromosome structures, double-minute chromosomes (DM) and homogeneously staining regions (HSR).12, 13 These selectable genetic elements are associated with acquired drug resistance in experimental systems14–16 and amplification of many different oncogenes in cancers.17–20 Most common epithelial cancers are characterized by a chromosomal instability (CIN) phenotype21 as opposed to a microsatellite instability phenotype (MIN) only seen in a minority of sporadic cancers but at high levels in some inherited cancers.22
The product of CCND1, cyclin D1, is overexpressed at the mRNA and protein level in over 50% of breast cancers, independent of gene amplification.5, 23 Cyclin D1 may thus provide a growth advantage to cancer cells,24 contribute toward resistance to endocrine therapy in ER positive cancers and has a potential prognostic and therapeutic role in breast cancer.
The aim of this study was to examine CCND1 amplification in primary breast cancer, compare this to cyclin D1 protein expression, estrogen receptor protein (ER) and HER2 amplification, predictive and prognostic markers in breast cancer.25, 26 and with the Nottingham Prognostic Index widely utilized as a surrogate for disease outcome.
CCND1: cyclin D1 gene; CNG: copy number gain; DM: double minute; ER: estrogen receptor; FISH: fluorescent in situ hybridization; HER2: human epidermal growth factor receptor type 2; HSR: homogeneously staining region; IHC: immunohistochemistry; NPI: Nottingham Prognostic Index; TMA: tissue microarray
Material and Methods
A cohort of 99 consecutive patients consenting to this biomarker study with primary, untreated, operable breast cancer was studied. Formalin-fixed, paraffin-embedded surgical specimens were retrieved from the pathology department, following local ethical committee approval. Clinical data from the 47 premenopausal and 52 postmenopausal women included pathology and Nottingham Prognostic Index (NPI).27
Tissue microarray (TMA)
Paraffin wax-embedded tissue samples from the 99 women were used to construct tissue microarrays (TMA). A whole tissue section from each specimen was stained with hematoxylin-eosin and examined for the presence of tumor. The optimum tumor areas were marked by a single pathologist (CAP) and four to six 0.6-mm diameter cores of cancer tissue from each specimen contributed to the TMA (Beecher Instruments, Silver Springs, MD).
Fluorescent In-Situ Hybridization (FISH)
CCND1 amplification was assessed by fluorescent in-situ hybridization (FISH) using LSI Cyclin D1 SpectrumOrange/CEP11 SpectrumGreen (Vysis, Downers Grove, IL) on TMA sections. The spectrum orange probe was directed against CCND1 (11q13) with CEP11 as the reference probe that maps to the centromeric region of chromosome 11. A readability score of 94% was achieved during the study; only 6 out of 99 cancers could not be assessed for gene amplification due to poor digestion, difficult morphology with overlapping cells and/or poor strength of fluorescent signal. The readability scores were substantially improved by probing at least 2 sections from each TMA, using different time periods for protein digestion (e.g., one section was digested for 20 min and the other for 25 min). The longer time period allowed adequate digestion of those samples that could not be assessed following a shorter period of digestion.
HER2 amplification was assessed by FISH using a locus specific probe (Qbiogene, MP Biomedicals Europe, France) on TMA sections, 4-μm thick. As with CCND1 FISH, TMA sections underwent protease digestion for 20 and 25 min to combat the effects of over and under digestion, before probe hybridization.
All HER2 and CCND1 slides were viewed using an Olympus BX51 epi-fluorescence microscope equipped with DAPI, SpectrumGreen™ and SpectrumOrange™ filter cubes. Images were captured using Applied Spectral Imaging acquisition software.
Two amplification criteria were employed: ratio and copy number gain. At least 10 nonoverlapping nuclei were scored for amplified tumors. For tumors with equivocal amplification, at least 30 nuclei were scored to achieve an average score. All the sections were scored by 2 independent observers (PGR, CAP). The ratio of orange (CCND1 or HER2) to green (CEP11 or CEP17) signals (PathVysion Her2 assay, Vysis, Downers Grove, IL) was calculated for all the cases studied. Ratio ≥ 2.2 and < 4.0 was counted as low amplification and ratio ≥ 4.0 as highly amplified. For copy number gain (CNG): CCND1 and HER2 (orange) signals were counted for each of the cases studied and CNG ≥ 8.0 qualified for high level of amplification, while CNG between 4.0 and 8.0 as low level and CNG < 4.0 as not amplified.
TMA sections 4-μm thick were dried, dewaxed and microwave treated in citrate buffer (pH 6.0). Staining was carried out in a Dako Autostainer Universal Staining System (DAKO; Ely, UK) using commercially available mouse monoclonal antibodies to study the expression of cyclin D1 (clone GM) and estrogen receptor α (ER) (clone 6F11) (Novocastra Laboratories, UK). The concentrations of primary antibodies used to achieve optimal staining were 1:40 for the antibody to cyclin D1 and 1:250 for anti-ERα. Negative controls (lacking primary antibody) were performed for all samples stained. All stained sections were scored for labeling index and intensity of staining and the immunohistochemistry score was obtained by multiplying the 2 indices using the Quickscore method.28 Tumors were labeled ER positive if the quickscore was greater than or equal to 4 (out of a maximum score of 18) as used for therapeutic decision-making in clinical practice. The immunohistochemistry scoring for ER on TMA sections was compared with the whole section receptor status from routine diagnostic specimens (also scored using the Quickscore method of Detre et al.) and greater than 95% concordance was found.
Cyclin D1 overexpression was defined by a score equal to or greater than 6 (out of 18) to ensure inclusion of cancers with greater than 50% labeling index for cyclin D1 and/or staining intensity greater than 1 (out of 3). All sections were analyzed for cyclin D1 nuclear staining independently by 2 observers (PGR, CAP). Greater than 95% concordance was seen between the findings of the 2 observers and the discordant results were reexamined to mutual consensus.
Statistical tests were performed using Minitab Release 14.1 (Minitab) and SPSS v13.0 (SPSS, Chicago, IL). For the purposes of statistical analysis, absolute levels of HER2 and CCND1 copy number average were used rather than the relative values provided by ratios (HER2/CEP17 and CCND1/CEP11), in keeping with the literature.
Cyclin D1 overexpression and CCND1 amplification were examined to assess their relation to tumor grade, tumor size, axillary lymph node status, menopausal status, HER2 amplification and NPI score.27
Fisher's exact test (2-tailed) was used to compare CCND1 gene amplification to protein overexpression, and amplification and overexpression with ER status. The χ2 test was used to compare the CCND1 amplification with intensity of cyclin D1 staining on IHC and to compare the difference in gene amplification between the amplified and nonamplified breast cancers for both CCND1 and HER2. The 2-sample t-test and Mann-Whitney U-test (parametric and nonparametric distributions, respectively) were used to compare average NPI values of amplified and nonamplified cancers. The null hypothesis was rejected at α level of 5% (p ≤ 0.05) for all analyses.
Evaluation of the 93 assessable cancers demonstrated 27 (29%) with CCND1 copy number gain (CNG) ≥ 4.0; 14/27 (52%) were highly amplified (CNG ≥ 8.0) (Fig. 1) with an average CNG of 12.9 while 13/27 (48%) exhibited low amplification (4.0 ≤ CNG < 8.0) (Fig. 2) with an average CNG of 5.4 (Table 1). No deletions of CCND1 were identified in this series. From the 93 cancers with CCND1 data, all were successfully examined for HER2 amplification and 18 demonstrated amplification on FISH with 16 highly amplified with an average CNG of 26 (Table 1). CCND1 and HER2 coamplification were not significantly associated.
Table 1. Patterns of amplification of CCND1 and HER2 in 93 primary breast cancers
Cyclin D1 immunohistochemistry
Cyclin D1 overexpression (Fig. 3) was observed in 63 of the 99 cancers (63.4%), and was comparable for premenopausal and postmenopausal women (62.7 and 64%, respectively). CCND1 amplification and cyclin D1 protein overexpression were significantly associated (p < 0.001, Fisher's exact test); cyclin D1 overexpression was observed in all 27 (100%) amplified cancers compared to cyclin D1 overexpression in 31/66 (47%) in the nonamplified group. CCND1 amplification also correlated significantly with the intensity of cyclin D1 staining on immunohistochemistry (p < 0.001, χ2 = 25.6, df = 1).
Clinical and pathological associations
CCND1 amplification was significantly associated with ER status (Table 2) (p = 0.003, Fisher's exact test); 26/27 amplified cancers were ER positive, with the odds ratio of 12.1 for CCND1 amplification in ER positive cancer compared to ER negative cancers. Cyclin D1 overexpression was identified in 57 of 76 (69.7%) ER positive cancers but only 6/23 (23.2%) ER-negative cancers (p < 0.001, Fisher's exact test).
Table 2. Clinical and pathological features of the patient cohort
CCND1 amplification was significantly associated with higher tumor grade (p = 0.005, df = 2, χ2 for trend = 7.915) with 70% Grade 3 but not with axillary lymph node status or tumor size. The mean NPI for the CCND1 amplified cancers was significantly higher at 5.55 (p = 0.001, 2-sample t-test) than for the nonamplified patients (4.57). The highly amplified CCND1 (mean NPI 5.64) had a significantly worse prognostic score (p = 0.005, 2-sample t-test) than the low amplification group (mean NPI 5.3). No significant difference was found between the mean NPI values of low-amplified and nonamplified patient groups.
This study has, for the first time, examined in detail CCND1 copy number on TMAs of primary breast cancer in relation to cyclin D1 protein expression, ER, HER2 amplification, clinical and pathological parameters and prognostic index (NPI) as a surrogate for clinical outcome.
FISH for CCND1 amplification worked well with TMA sections.29CCND1 amplification, identified in 30% of cancers, was strongly associated with ER expression in keeping with previous studies, but was also associated with a higher NPI, an established surrogate for the poor prognostic link demonstrated in previous studies.
Few studies have looked at the prognostic significance of cyclin D1 overexpression and CCND1 gene amplification in the same group of patients, although cyclin D1 overexpression alone or cyclin D1 overexpression and CCND1 amplification have been associated with tamoxifen resistance in some but not all30, 31 studies of ER positive breast cancer. It may be that alternatively spliced cyclinD1b transcripts may overcome endocrine therapy induced cell cycle arrest.32 Cyclin D1 can potentiate ER activity, even in the absence of estrogen, which may give it carcinogenic or promoter potential.10, 11
To examine the importance of CCND1 amplification further, and in the absence of established criteria for CCND1 gene amplification, low level amplification was defined as an average copy number ≥4 and <8, and high level amplification as ≥8.0 for both CCND1 and HER2. This identified, for the first time, that high-level amplification of CCND1 was particularly linked to worse prognosis using the NPI. This difference between high and low CCND1 copy number may reflect an underlying mechanistic difference of amplification between the patient groups. Low-level amplification would be consistent with CCND1 copy number gain attributable to chromosome 11 polysomy which rarely occurs in breast cancer as confirmed by karyotyping and CGH, or other structural chromosomal rearrangement such as partial duplication or unbalanced translocation involving 11q13. By contrast, high-level amplification (CNG ≥8.0) is likely to reflect “true” gene amplification as reported for other oncogenes such as HER2 and N-MYC.19 Our understanding of gene amplification associated with the cytogenetic manifestations of DM and HSR remains incomplete. It is widely accepted that DM and HSR are hallmarks of advanced disease and are intimately associated with a chromosomal instability “mutator” phenotype (CIN). As such they may represent selectable genetic elements that can amplify many different oncogenes and confer drug resistance in experimental systems. It is therefore perhaps not surprising that high CCND1 amplification is associated with poor prognosis in breast cancer.
Using FISH data for both CCND1 and HER2 we confirmed the lack of association between HER2 and CCND1 coamplification. The absolute level of copy number gain was lower for CCND1 (average CNG of 12.9) than for HER2 (average CNG of 26), which may reflect underlying differences in the structure and biological nature of the respective amplicons. Whether the pattern of amplification is HSR or DM in type, “true” gene amplification reflects significant chromosome instability associated with a well-recognised CIN phenotype. Unfortunately, as much classical cancer cytogenetic literature predates our current molecular understanding of cancer, the underlying genetic targets of these early gene amplification studies remain obscure. Renewed research interest in the nature and extent of coamplified genes within both the HER2 and CCND1 amplicons has identified a number of other contiguous candidate genes. While such additional FISH and array-CGH investigations were beyond the scope of present study; this clearly suggests a productive area for future research.
In terms of clinical utility, high level CCND1 amplification (CNG ≥ 8) may be useful to identify a subset of patients with ER positive breast cancer who have a predicted poorer prognosis. Since the response of CCND1 amplified cancers to endocrine agents is controversial, until CDK4 inhibitors are implemented in the clinic, CCND1 amplification may be an indication for additional chemotherapeutic treatment in women with ER positive breast cancer.
The authors thank the Tayside Tissue Bank, Dundee (http://www.tissuebank.dundee.ac.uk) for its valuable support and cooperation. The authors also thank Ms. Kathleen Rooney, Mr. George Thomson and Ms. Nicola Foster for technical support and Dr. Simon Ogston for statistical advice.