TOP2A Amplification in the Absence of That of HER-2/neu: Toward Individualization of Chemotherapeutic Practice in Breast Cancer

The relationship between human epidermal growth factor receptor 2/neu and topoisomerase IIα in breast cancer was investigated and their association with clinicopathological variables was examined.

Abstract

Primary objective.

To investigate the relationship between human epidermal growth factor receptor (HER)-2/neu and the gene encoding topoisomerase IIα (TOP2A) in breast cancer, while elucidating their association with clinicopathological variables.

Methods.

Real-time quantitative polymerase chain reaction (RQ-PCR) was performed on a 96-patient study group to assess gene amplification, and levels were determined using the comparative cycle threshold approach and Taqman assays. An immunohistochemistry (IHC) microarray (n = 76) was then employed to check for correlation between gene amplification and protein expression levels.

Results.

Amplification levels of TOP2A did not differ significantly according to HER-2/neu status by either RQ-PCR or IHC microarray. Of the HER-2/neu⁻ patients, 29.1% demonstrated levels of TOP2A above the third quartile, whereas 22.9% of the HER-2/neu⁺ patients had values in the first quartile (log TOP2A <0.62), thereby indicating low-level amplification. Of the 60 patients characterized as HER-2/neu⁻ using IHC and fluorescence in situ hybridization (FISH), 22.9% were classified as TOP2A⁺ on the IHC microarray. Of the 14 patients deemed HER-2/neu⁺ using IHC and FISH, meanwhile, the majority (n = 10) were classified as TOP2A⁺.

Conclusions.

Our results indicate that amplification of TOP2A in breast cancer is not confined to those who are concomitantly HER-2/neu⁺, and suggest that a significant proportion of HER-2/neu⁻ patients exhibit high levels of TOP2A.

Introduction

Although anthracycline-based chemotherapeutic regimens have been associated with superior outcomes when compared with standard cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) protocols since 1975 [1, 2], discordance remains over which factors can most accurately predict this treatment benefit. Although p53, BRCA1, human epidermal growth factor receptor (HER)-2/neu, and topoisomerase IIα (TOP2A) have all been investigated as potential biomarkers in assessing response to anthracyclines in breast cancer [3], it is the latter two, in particular, that have stimulated the greatest interest, with ongoing discordance as to their respective significance.

Retrospective analyses of two adjuvant studies—Cancer and Leukemia Group B 8869 and National Surgical Adjuvant Breast and Bowel Project (NSABP) protocol B-11—were the first to suggest that patients who overexpress the HER-2/neu oncogene may derive additional benefit from anthracycline-based regimens (doxorubicin) over their HER-2/neu⁻ counterparts [4–6]. Prospective trial results in the interim, including those of the National Cancer Institute of Canada's Mammary 5 Trial [7] and NSABP15 [8], have consistently agreed with this assessment, and a 2008 meta-analysis of eight of these trials, involving 5,354 women, demonstrated that the incremental benefit from anthracycline-based regimens over standard CMF is largely restricted to HER-2/neu⁺ patients [9]. The findings of the National Epirubicin Adjuvant Trials and Trial BR9601, published in 2006, conflict with these assessments, however [10]. This work compared CMF regimens alone with CMF regimens following epirubicin—patients with HER-2/neu nonamplified tumors showed significantly higher survival rates when treated with epirubicin–CMF than with CMF [11]. These findings were later supported by the work of the Danish Breast Cancer Cooperative Group Trial 89d, which randomly allocated patients to cyclophosphamide, epirubicin, and 5-fluorouracil or CMF. Those investigators also concluded that HER-2/neu was not a predictive marker for epirubicin sensitivity [12]. When one considers (a) the potentially toxic side effects of these drugs, particularly when given in conjunction with trastuzumab, and (b) that the vast majority of the currently used adjuvant regimens for almost all breast cancer patients worldwide, both HER-2/neu⁺ and HER-2/neu⁻, are anthracycline based [13], there is an urgent need to target these therapies to those subgroups of patients who will benefit most.

Anthracycline-based therapies have been postulated to exert their cytostatic and cytotoxic actions through a number of different mechanisms, including inhibition of DNA synthesis in the tumor cell, free radical formation with consequent induction of DNA damage, or lipid peroxidation, interference with DNA unwinding or DNA strand separation and helicase activity, and the initiation of DNA damage via the inhibition of topoisomerase II [14]. First identified by Wang in 1971 [15], interest in TOP2A (the gene that encodes the protein topoisomerase IIα) increased dramatically in 1999 following the finding by Järvinen et al. [16–18] that it is frequently coamplified with HER-2/neu at 17q12–21. Those authors postulated that perhaps it was the subgroup of patients with coamplification of HER-2/neu and TOP2A that benefited most from anthracyclines. Although subsequent work, including the results of the Breast Cancer International Research Group 006 Trial [19], the Danish Breast Cancer Cooperative Group 89-D trial [12], and the Scandinavian Breast Group Trial 9401 [20], has all returned discordant results concerning the relationship between TOP2A amplification and response to anthracyclines, a 2009 meta-analysis of four large phase III trials suggested that TOP2A-amplified patients gain at least a modest additional benefit from anthracycline-based over nonanthracycline-based treatment regimens [21]. Identification of those who derive this benefit could obviate the need for anthracyclines in many patients, thereby avoiding the aforementioned side effects of these drugs, and, in particular, the greater incidence of these side effects with the combination of anthracyclines with trastuzumab.

Studies have returned discordant results, however, both concerning the relationship between gene amplification of TOP2A, as assessed using fluorescence in situ hybridization (FISH), and the resultant expression of the topoisomerase II protein, and furthermore, the relationship between the amplification levels of TOP2A and those of HER-2/neu. Much of this work has concluded that amplification of TOP2A is restricted to those patients who are HER-2/neu coamplified, leading a few commentators to conclude that HER-2/neu⁻ patients should not receive anthracyclines as part of their treatment [13]. These findings are complicated by the complex relationship that exists among mRNA, gene amplification, and protein expression levels of TOP2A, however, with the former being particularly poorly understood. mRNA levels of this target have taken on a new significance in light of recent work that has demonstrated their association with metastasis-free survival and response to adjuvant anthracylines in breast cancer patients [22], and further assessment of the relationship between these levels and those of HER-2/neu is warranted, particularly given the previous concentration of work on DNA levels, which do not always translate to gene or protein expression because of transcriptional or translational control mechanisms [22].

The primary aim of this work was to investigate the relationship between mRNA levels of HER-2/neu and TOP2A, and their association with clinicopathological variables using real-time quantitative polymerase chain reaction (RQ-PCR). A secondary aim was to further define the relationship between protein expression levels of HER-2/neu and TOP2A using a tissue microarray (TMA).

Materials and Methods

Study Populations

The study group for RQ-PCR consisted of a 96-patient cohort; the mean patient age at diagnosis of breast cancer was 57.6 years (standard deviation [SD], 13.3; range, 28–92 years), and the median follow-up period was 49 months (range, 1–224 months) (Table 1). Patient samples were selected from the Biobank at the Department of Surgery, National University of Ireland Galway; the primary selection criterion was HER-2/neu status, such that the resulting study group would have approximately equal proportions of HER-2/neu⁺ and HER-2/neu⁻ patient samples. HER-2/neu status had been previously determined by scoring membranous staining according to the HercepTest™ (Dako, Glostrup, Denmark) protocol. Similarly, routine clinical determination of estrogen receptor (ER) and progesterone receptor (PR) status had been performed at Galway University Hospital using the Allred scoring method [23]. Patients were classified according to tumor characteristics as follows: luminal A, ER⁺ and/or PR⁺ and not HER-2/neu⁺; luminal B, ER⁺ and/or PR⁺ and HER-2/neu⁺; HER-2/neu overexpressing, HER-2/neu⁺, ER⁻, PR⁻; and basal-like or triple negative (ER⁻, PR⁻, and HER-2/neu⁻). According to this classification, 35 (36.5%) patients in the study group had luminal A breast cancer, 31 (32.3%) had luminal B breast cancer, 17 (17.7%) were HER-2/neu overexpressing, and 13 (13.5%) had basal-like breast cancer. Fifty-seven patients (59.4%) were ER⁺ and 55 patients (57.3%) were PR⁺; the PR status for one patient was unknown. Forty-eight patients (50.0%) were HER-2/neu⁺, and 48 patients (50.0%) were HER-2/neu⁻.

Table 1.

Study groups

Abbreviations: ER, estrogen receptor; HER-2, human epidermal growth factor receptor 2; NPI, Nottingham Prognostic Index; PR, progesterone receptor; RQ-PCR, real-time quantitative polymerase chain reaction.

A TMA containing cores from a series of 82 consecutive patient samples was constructed for protein immunohistochemical (IHC) analysis; of these, 76 were usable (Table 1). The mean patient age at diagnosis was 55.4 years (SD, 11.2; range, 31–78 years). Forty-four (57.9%) of these patients were classified as luminal A, eight (10.5%) were classified as luminal B, eight (10.5%) were classified as HER-2/neu overexpressing, and 12 (15.8%) were classified as having basal-like breast cancer. Four patients (5.3%) had not been categorized by subtype because of unknown ER, PR, or HER-2/neu status. Fourteen patients were confirmed as HER-2/neu⁺ using IHC and FISH; 60 were confirmed as HER-2/neu⁻ and the HER-2/neu status of two patients was unknown.

RQ-PCR

Breast tumor tissue was obtained at the time of surgical resection and immediately snap frozen. This tissue was later homogenized and the total RNA was extracted. The RNA concentration and purity were determined using a Nanodrop spectrophotometer (Nanodrop Technologies, Wilmington, DE). Aliquots of RNA equivalent to 1 μg were reverse transcribed using SuperScript™ III (Invitrogen, Carlsbad, CA). RQ-PCR reactions were then carried out in final volumes of 10 μl using a 7900HT sequence detection system (Applied Biosystems, Carlsbad, CA), using Taqman probes, optical 96-well FastPlates, and sequence detection system software (Applied Biosystems). RQ-PCR reaction volumes were: cDNA, 1 μl; mastermix, 5 μl; Taqman probe mix, 0.5 μl; and nuclease-free water, 3.5 μl.

cDNA, synthesized from commercially available breast cancer cell line RNA, was included on each 96-well plate as an interassay control. All reactions were performed in triplicate, and the threshold SD for intra- and interassay replicates was 0.3.

The relative quantity of mRNA expression was calculated using the comparative cycle threshold (ΔΔCt) method and Qbase software (Biogazelle, Ghent University, Belgium). The geometric mean of the cycle threshold value of the endogenous control genes PPIA and MRPL19 was used to normalize the data, and the lowest expressed sample was used as a calibrator. Relative quantities were logarithmized (log 10) and the distribution of the dataset was assessed using the Kolmogorov–Smirnov test in SPSS version 15.0 (SPSS, Inc., Chicago, IL). The logarithmized values did not vary significantly from a normal distribution. Statistical tests employed to assess the data included the Pearson correlation coefficient, Student's t-test, and one-way analysis of variance for between-groups analysis. Kaplan–Meier curves and the log-rank test were employed to investigate significant relationships between gene amplification and disease-free survival (DFS) and overall survival (OS) rates. The former were performed by comparing survival outcomes between those patient samples with gene amplification levels less than and greater than the mean amplification level of the gene under study across the entire patient cohort.

TMA

IHC staining for HER-2/neu and TOP2A was optimized on 10 patient samples for each target. A representative donor tumor block was selected from each case in the study group above (Table 1). Hematoxylin and eosin (H&E)-stained sections were employed to identify the area of tumor from which cores were retrieved. A wax block was constructed into which the donor cores were inserted. A tissue microarrayer (Beecher Instruments, Madison, WI) was then used to take tissue cores (6-mm diameter) from a single donor block for each case, and these were arrayed into the recipient paraffin block according to our predefined map and coordinates. From the initial 82 cases, only 76 were represented on the array because for some cases the blocks were too thin or the tumor area was either too small or could not be accurately marked on the donor block.

Sections from the TMA were cut at 3.5–4 μm and placed on ChemMate™ Capillary Gap microscope slides (Dako). Immunostaining was performed using a Ventana Discovery instrument (Ventana Medical Systems, Illkirch Cedex, France) according to the manufacturer's protocols. Slides were dried overnight at 40°C, dewaxed, and washed thoroughly. A standard 3′ diaminobenzidine peroxidase-conjugated streptavidin-biotin method was used to detect the reaction. Tumors and tissues with known staining patterns were used as immunostaining controls and normal tissues served as nontumor controls.

Two observers evaluated the IHC for the 76 cases (R.W.G. and S.J.M.). Each was blinded to the other's results and to the clinicopathological data. Both observers reviewed discrepant scores (eight of 76) together and agreed on a final score. Membranous staining was scored for HER-2/neu according to the HercepTest™ as follows: 0, no staining or faint staining in <10% cells; 1, faint staining in >10% cells; 2, weak to moderate staining in >10% cells; and 3, strong staining in >10% cells. In scoring TOP2A, for each case, the number of cancer cells was first counted at 20× magnification using a light microscope on the H&E-stained slide. If the cells were not clear at this magnification, each punch was divided into four quadrants (40×) and these were then added to provide an overall cell count. The positive cells were then counted again at either 20× magnification or 40× magnification on the antibody-treated slide; the TOP2A index was calculated as the percentage of positive cells over total cells—a cutoff of 10% was used in this work.

Results

RQ-PCR

Gene amplification levels of HER-2/neu correlated moderately with those of TOP2A (r = 0.337; p < .001). Levels were significantly higher in those patients who had been classified clinically as HER-2/neu⁺ using IHC and FISH (p < .001)). In contrast, amplification levels did not differ significantly between patients according to either their ER (p = .481) or PR (p = .935) status. Post hoc analysis using a Bonferroni adjustment demonstrated that levels of HER-2/neu were significantly higher in the HER-2/neu overexpressing and luminal B subgroups than in both the luminal A (p < .001) and basal-like (p < .001) subgroups.

Analysis of the relationship between amplification levels of HER-2/neu and DFS (p = .651) and OS (p = .505) using Kaplan-Meier curves failed to demonstrate significant relationships. Similarly, no differences were noted when we looked at amplification levels in relation to DFS (p = .915) and OS (p = .495) up to the 60-month time point.

Amplification levels of TOP2A did not differ significantly between patients who had been classified clinically as HER-2/neu⁺ and those who had been classified as HER-2/neu⁻ (p = .475) (Fig. 1). Investigation of this finding revealed that 14 of the 48 (29.1%) HER-2/neu⁻ patients demonstrated levels of TOP2A above the third quartile (log TOP2A >1.33) (Fig. 2). Similarly, 11 of the 48 (22.9%) HER-2/neu⁺ patients returned values in the first quartile (log TOP2A <0.62), thereby indicating low-level amplification.

Figure 1.

Relationship between amplification levels of TOP2A by RQ-PCR and HER-2/neu status by IHC and FISH. Amplification levels of TOP2A did not differ significantly between patients who had been classified clinically as HER-2/neu⁺ and those who had been classified as HER-2/neu⁻.

Abbreviations: FISH, fluorescence in situ hybridization; HER-2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; RQ-PCR, real-time quantitative polymerase chain reaction; TOP2A, topoisomerase IIα.

Figure 2.

Levels of TOP2A relative to HER-2/neu status. Fourteen of 48 (29.1%) HER-2/neu⁻ patients demonstrated levels of TOP2A above the third quartile. Similarly, 11 of 48 (22.9%) HER-2/neu⁺ patients returned values in the first quartile, thereby indicating low-level amplification.

Abbreviations: HER-2, human epidermal growth factor receptor 2; RQ-PCR, real-time quantitative polymerase chain reaction; TOP2A, topoisomerase IIα.

Amplification levels of TOP2A did not differ significantly between those patients who had been classified as ER⁺ and those classified as ER⁻ (p = 0.125), or between PR⁺ and PR⁻ patients (p = .688). In addition, no differences were found in amplification levels of TOP2A across the intrinsic subtypes (p = .536) and no differences were noted for TOP2A in relation to DFS or OS, or other clinicopathological variables.

TMA

Significant correlation was noted between HER-2/neu status as assessed clinically using IHC and FISH and that as determined on scoring the TMA (r = 0.614; p = .000). Of the 60 patients characterized as HER-2/neu⁻ using IHC and FISH, 14 (22.9%) were classified as TOP2A⁺ on the TMA (Fig. 3a). Of the 14 patients deemed HER-2/neu⁺ using IHC and FISH, meanwhile, 10 were classified as TOP2A⁺. Similar results were returned when comparing HER-2/neu status by TMA and TOP2A status by TMA—of the 54 patients deemed HER-2/neu 0 or 1+, 14 (25.9%) were scored as TOP2A⁺ (Fig. 3B); of the eight patients deemed HER-2/neu 3+, seven were scored as TOP2A⁺.

Figure 3.

Relationship between TOP2A as assessed on the TMA and HER-2/neu status as assessed using IHC and FISH (A) and on the TMA (B).

Abbreviations: FISH, fluorescence in situ hybridization; HER-2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; TMA, tissue microarray; TOP2A, topoisomerase IIα.

Discussion

As mentioned in the introduction to this paper, the relationship among TOP2A mRNA, gene amplification, and protein expression levels is complex. Whereas multiple FISH-based analyses examining the relationship between TOP2A gene amplification and HER-2/neu gene amplification have concluded that TOP2A gene amplification is extremely rare in the absence of HER-2/neu gene amplification [11, 12, 24], this current work has demonstrated conflicting results when measuring both TOP2A mRNA and protein expression levels.

Until recently, major caveats to these results would have been the knowledge that TOP2A protein expression does not correlate with gene amplification as measured using FISH [25, 26] and the lack of any work that examined the relationship between TOP2A mRNA levels and protein expression. Fortunately, Brase et al. [22] recently demonstrated that, whereas amplification of TOP2A by FISH did not correlate with protein expression as measured by IHC (p = .283), in contrast, levels of TOP2A by RQ-PCR correlated strongly with those by IHC (p = .001). The authors then proceeded to investigate the prognostic significance of TOP2A in a cohort of 782 node-negative breast cancer patients; amplification levels of TOP2A, as measured using RQ-PCR, were associated with metastasis-free survival in both univariate and multivariate analyses. Finally, the authors investigated the relationship between response to anthracyclines and levels of TOP2A mRNA in a cohort of 80 patients treated with neoadjuvant chemotherapy (epirubicin and cyclophosphamide); patients with complete remission had significantly higher levels of TOP2A than those with no change (p = .002) and those who demonstrated a partial response (p = .002). These findings led the authors to surmise that: (a) gene amplification (by FISH) alone is not sufficient to study a possible prognostic or predictive effect of TOP2A, and (b) TOP2A (by RQ-PCR) RNA analysis may be sufficient, because it correlates well with protein expression Taken together, our results show that TOP2A RNA is a powerful prognostic marker in breast cancer [22].

This validation of RQ-PCR as a technique in the assessment of TOP2A holds particular relevance for our own results, because it strongly suggests that the large cohort of previously unrecognized HER-2/neu⁻, TOP2A amplified breast cancer patients identified by us will respond to anthracycline-based therapies; perhaps even more importantly, it demonstrates that this relatively simple technique can be used to identify the even larger cohort of HER-2/neu⁻, TOP2A nonamplified patients who will not benefit from anthracyclines but currently must still suffer from the associated side effects, most notably cardiotoxicity. Furthermore, it may also go some way to explaining the conflicting results of studies that have examined the role of HER-2/neu in predicting response to anthracycline-based therapies; although the majority have demonstrated the significance of HER-2/neu as a predictive marker, a few have returned paradoxical findings [11, 12]. This may perhaps now be explained by the significant proportion of HER-2/neu⁻ patients who are TOP2A amplified and should therefore respond to therapy.

Neither amplification of HER-2/neu nor that of TOP2A correlated with DFS or OS in this present work. Although these findings conflict with previously published evidence, it should be noted that the median follow-up duration in our study group was just 49 months, and it is thus difficult to extrapolate on the significance, or otherwise, of our findings.

Conclusion

In conclusion, our results suggest that a significant cohort of patients with breast cancer amplify or overexpress TOP2A in the absence of that of HER-2/neu. These findings, which have employed the gold standard of RQ-PCR in the assessment of gene amplification, demonstrate the need for additional work to clarify the relationship between HER-2/neu and TOP2A, and we suggest that, until this is achieved, a consensus will not be reached on the potentially important association between these biomarkers and response to anthracycline-based treatment regimens.

Author Contributions

Conception/Design: Ronan W. Glynn, Nicola Miller, Sarah Mahon, Catherine Curran, Grace Callagy, Michael J. Kerin

Provision of study material or patients: Ronan W. Glynn, Nicola Miller, Sarah Mahon, Catherine Curran, Grace Callagy, Michael J. Kerin

Collection and/or assembly of data: Ronan W. Glynn, Nicola Miller, Sarah Mahon, Catherine Curran, Grace Callagy, Michael J. Kerin

Data analysis and interpretation: Ronan W. Glynn, Nicola Miller, Sarah Mahon, Catherine Curran, Grace Callagy, Michael J. Kerin

Manuscript writing: Ronan W. Glynn, Nicola Miller, Sarah Mahon, Catherine Curran, Michael J. Kerin

Final approval of manuscript: Ronan W. Glynn, Nicola Miller, Sarah Mahon, Catherine Curran, Grace Callagy, Michael J. Kerin