|Home | About | Journals | Submit | Contact Us | Français|
Amplification or deletion of the topoisomerase II alpha (TOP2A) gene in breast cancers has been postulated to be more closely associated with responsiveness to anthracycline-containing chemotherapy than amplification of the human epidermal growth factor receptor type 2 (HER2) gene.
We studied 438 tumors from 710 premenopausal women with node-positive breast cancer who received cyclophosphamide, epirubicin, and 5-fluorouracil (CEF) or cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) as adjuvant chemotherapy in the randomized National Cancer Institute of Canada Clinical Trials Group Mammary 5 (MA.5) trial. TOP2A alterations and HER2 amplification were quantified by fluorescence in situ hybridization. The association of TOP2A and HER2 status with recurrence-free survival (RFS) and overall survival (OS) in the two treatment groups was analyzed using Kaplan–Meier curves, the log-rank test, and Cox proportional hazard models. All statistical tests were two-sided.
In patients whose tumors showed TOP2A alterations (either amplifications or deletions), treatment with CEF was statistically significantly superior to treatment with CMF in terms of RFS (adjusted hazard ratio [HR] = 0.35, 95% confidence interval [CI] = 0.17 to 0.73, P = .005) and OS (adjusted HR = 0.33, 95% CI = 0.15 to 0.75, P = .008). In patients without TOP2A amplification or deletion, the corresponding adjusted hazard ratios for RFS and OS were 0.90 (95% CI = 0.66 to 1.23, P = .49) and 1.09 (95% CI = 0.77 to 1.56, P = .62). Adjusted tests of interaction between treatment and TOP2A status were P = .09 for RFS and P = .02 for OS. Adjusted tests of interaction between treatment and HER2 status were P = .008 for RFS and P = .02 for OS.
TOP2A gene alterations (amplifications or deletions) are associated with an increase in responsiveness to anthracycline-containing chemotherapy regimens relative to non-anthracycline regimens that is similar to that seen in patients with HER2 amplification.
In Mammary 5 (MA.5) and other randomized breast cancer trials, women with amplification or overexpression of human epidermal growth factor receptor type 2 (HER2) were reported to benefit more from adjuvant therapies that included anthracyclines than from therapies without them. Because HER2 lies near the gene for topoisomerase II alpha (TOP2A), it was hypothesized that responsiveness to anthracyclines might be associated with TOP2A gene alterations.
Tissue microarrays were able to be constructed from 438 of 710 tumor blocks accrued in 1989–1993 from MA.5 patients. TOP2A and HER2 amplification were measured by fluorescence in situ hybridization on the arrays. The association of each with recurrence-free survival and with overall survival was statistically tested.
Patients whose tumors had alterations (either amplification or deletion) of TOP2A had statistically significantly longer overall survival and recurrence-free survival in response to chemotherapy with epirubicin than to chemotherapy without this anthracycline, whereas patients without TOP2A alterations showed no difference in responsiveness.
TOP2A alteration (amplification or deletion) is associated with a degree of responsiveness to anthracyclines that is similar to that observed in patients with HER2 amplification.
There is close but not complete concordance between TOP2A alteration and HER2 amplification. A larger number of patients will be needed to determine which measurement is most closely associated with responsiveness to anthracycline-containing regimens.
From the Editors
Our group and others have previously reported that women with breast cancers with human epidermal growth factor receptor type 2 (HER2) gene amplification or protein overexpression benefit more from adjuvant chemotherapies containing an anthracycline compared with therapies without an anthracycline (1–4). Three recent meta-analyses have confirmed this finding (2–4). It has been suggested, however, that HER2 gene amplification and protein overexpression may simply be markers for changes in topoisomerase II alpha, an enzyme that is integrally associated with the cytotoxic action of anthracyclines and whose gene, TOP2A, is close to HER2 on the long arm of chromosome 17. There have been several reports that in multivariate analyses, TOP2A gene amplification, or alteration (a combined category comprising amplification and deletion), was more closely associated than HER2 amplification or overexpression with the benefit of adjuvant anthracycline-containing therapy in primary breast cancer (5–7). Using samples from a randomized breast cancer trial, the National Cancer Institute of Canada Clinical Trials Group Mammary 5 (MA.5) trial, in which we previously had compared adjuvant cyclophosphamide, epirubicin, and 5-fluorouracil (CEF) with cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) and reported the association of HER2 amplification or overexpression with greater benefit from CEF (1), we further explored the association of TOP2A gene alterations with benefit from CEF compared with that from CMF. We also reexplored the association of HER2 amplification in the subset of MA.5 patients for whom TOP2A gene measurements were available.
The MA.5 study included 710 premenopausal women with axillary lymph node–positive breast cancer (T1–T3a, N1–N2, M0) who had completed primary breast cancer surgery no more than 10 weeks before random assignment (8,9,11). Patients were accrued between 1989 and 1993 at 35 centers in Canada. The MA.5 protocol was approved by the institutional review board at each participating center and registered as NCI-V90-0027 on cancer.gov. Written informed consent was obtained from each woman before random assignment.
The adjuvant CEF regimen consisted of six cycles of epirubicin (Pharmorubicin; Pfizer, New York, NY) 60 mg/m2 and 5-fluorouracil (5-FU) (Efudex; Valeant Pharm, Aliso Viejo, CA) 500 mg/m2, both delivered intravenously on days 1 and 8, and oral cyclophosphamide (Cytoxan; Bristol Myers Squibb, New York, NY) 75 mg/m2 daily on days 1 through 14. During this regimen, patients received antibiotic prophylaxis with trimethoprim–sulfamethoxazole (Septra; Glaxo, Philadelphia, PA; Merck, Whitehouse Station, NJ) 400 mg orally twice daily or ciprofloxacin (Cipro; Bayer, Berlin, Germany) 500 mg orally twice daily. The CMF regimen consisted of six cycles of methotrexate (Wyeth, formerly Lederle, Madison, NJ) 40 mg/m2 and 5-FU 600 mg/m2, both delivered intravenously on days 1 and 8, and oral cyclophosphamide 100 mg/m2 daily on days 1 through 14.
Representative formalin-fixed, paraffin-embedded tumor blocks from the primary surgical specimen were retrospectively requested for each woman enrolled in the study, as previously described (1). Usable tissue microarrays (TMAs) were able to be constructed for 438 (61%) of 710 study patients. Preparation of the TMAs involved marking representative areas of tumor on hematoxylin and eosin–stained sections to guide the coring of the corresponding blocks. Sector maps were designed using Microsoft Excel spreadsheets to identify the location of each tumor on the array blocks. Each tumor had two (0.6-mm) core biopsies sampled with an array machine from Beecher Instruments (Sun Prairie, WI). Four-micrometer sections were cut from each tissue array block containing 0.6-mm cores from each tumor and placed on slides for subsequent fluorescence in situ hybridization (FISH) studies.
TOP2A amplification was measured by FISH. The probes used were the Locus Specific Identifier (LSI) TOP2A SpectrumOrange/Centromere Enumeration Probe (CEP17) and the SpectrumGreen DNA Probe kit (Vysis-Abbott, Downers Grove, IL). Slides were prepared according to the manufacturer's instructions for paraffin sections. Slides were analyzed using a Leica DMBRX epifluorescence microscope (Bannockburn, IL) equipped with filters for the separate detection of 4′,6-diamidino-2-phenylindole, spectrum green, and spectrum orange, and with a triple bandpass filter for simultaneous detection of the three signals. A minimum of 60 nuclei was counted per case. To determine TOP2A amplification or deletion, the LSI TOP2A probe (orange signals) and CEP17 probe (green signals) were recorded for each cell, and the average number of orange and green signals per cell was calculated. Images were captured by a charge-coupled device camera using software from Applied Imaging (Santa Clara, CA). A tumor was considered to have amplified TOP2A if the TOP2A:CEP17 ratio was 2:1 or greater, to have deleted TOP2A if the ratio was 0.8:1 or less, and to have normal TOP2A if the ratio was between 0.8:1 and 2:1. Cut points were decided in advance of the study based on previous work (6,10). “Altered” TOP2A was a combined category that included both tumors with deleted and amplified TOP2A.
Measurements of HER2 amplification were those reported in our previous study, carried out on 639 (90%) available whole sections using the PathVysion HER2 DNA Probe kit (Vysis-Abbott) (1). The methodology was identical to that described above for the analysis of TOP2A amplification. To determine HER2 amplification, the LSI HER2 probe (orange signals) and CEP17 probe (green signals) were recorded for each cell, and the average number of orange and green signals per cell was calculated. The ratio of HER2 signals per cell to CEP17 signals per cell was then calculated. An HER2/CEP17 ratio greater than or equal to 2:1 was considered positive for HER2 gene amplification. An HER2/CEP17 ratio less than 2:1 was considered negative for HER2 gene amplification.
Primary outcomes for this study were recurrence-free survival (RFS) and overall survival (OS). RFS was defined as time from random assignment to recurrence including local breast chest wall and regional and/or distant recurrence. OS was based on death from any cause (8). For the analysis of TOP2A gene alterations, all patients were divided into three groups (amplified, deleted, and normal) as described above. RFS and OS for the groups defined by TOP2A gene status or treatment were plotted using Kaplan–Meier curves and compared by the log-rank test. The Cox proportional hazard model with single covariates was used to obtain the hazard ratios (HRs) and associated 95% confidence intervals (CIs) for the groups compared. The Cox model with treatment, TOP2A status, and their interaction as covariates was used to assess the interaction between treatment and TOP2A status. In this and prior analysis of MA.5, estrogen receptor (ER) data (see Table 1) were those collected from local clinical tests done on all patients by the ligand-binding assay. The primary multivariable analyses in this article were performed using the Cox model, and adjustments were made for age (≥50, <50 years), number of positive nodes (≤3, ≥4), ER protein levels from the clinical database of the MA.5 trial (ie, local ER measurements from each center [≥10, <10 fmol protein/mg tissue]), surgical type (total vs partial mastectomy), tumor size (T1, T2, T3), tumor grade [1, 2, 3, using the Nottingham grading system (12), performed by central review on whole sections], and HER2 gene amplification as measured by FISH on whole sections (amplified or not amplified and read centrally). The assumptions of proportionality for the Cox model were examined based on Schoenfeld residuals.
Among the patients with TOP2A measurements from FISH on TMAs, statistical analyses similar to those described above were carried out to explore the role of HER2 gene amplification as a prognostic factor and the interaction between treatment and HER2 gene amplification to duplicate and compare our previous analyses from this trial (1) in this smaller subset. The analyses reported used HER2 amplification as measured by FISH on whole sections (1). Sensitivity analyses were also performed adjusting for HER2 protein expression (as measured by immunohistochemistry [IHC] from whole sections and TMAs) and for ER protein (results from centrally reviewed IHC from TMAs) in multivariable analyses including all other variables listed above. These adjustments did not change the results appreciably, and results of the sensitivity analyses are not reported here. All statistical tests were two-sided.
Among the 438 (62%) of the 710 patients from the study for whom we had usable TMAs, 54 (12.0%) had tumors with amplification of TOP2A, 26 (6%) had deletion of TOP2A, and 358 (82%) had the normal TOP2A, as defined by preset cut points (see Subjects and Methods). Patient characteristics are summarized in Table 1, which shows that TOP2A alterations were more common in tumors with HER2 amplification and in larger tumors. This subset of 438 patients was similar to the original set of 639 (90%) MA.5 patients that we previously studied (1) in terms of all the characteristics described in Table 1 and in terms of outcomes (data not shown).
The 5-year RFS and OS for patients with TOP2A amplification, deletion, and normal TOP2A gene status are given in Table 2. There was no statistically significant difference between the three groups in terms of either RFS (P = .19) or OS (P = .11). The difference in both RFS and OS between those with amplified and normal TOP2A status was apparently, but not statistically significantly, smaller than that between those with deleted and normal TOP2A status (P value of interaction test was .2 for RFS and .5 for OS). In the primary multivariable analysis model adjusted for age, node status, tumor grade, ER protein, surgical procedure, tumor size, and HER2 amplification, TOP2A status among these three groups was not independently associated with either RFS or OS (P = .13 and .32, respectively).
When the patients with either amplification or deletion of TOP2A were combined into one group, “altered” TOP2A status, there was no statistically significant difference between the altered and the normal TOP2A groups in terms of RFS (P = .23), but a marginally significant difference was observed in OS (P = .05; Table 2). After adjustment for age, node status, tumor grade, ER, surgical procedure, tumor size, and HER2 status, however, altered TOP2A status was no longer statistically significantly associated with either RFS (adjusted HR = 0.94, 95% CI = 0.63 to 1.39, P = .75) or OS (adjusted HR = 1.01, 95% CI = 0.66 to 1.56, P = .95).
Table 3 shows the 5-year RFS and OS and the hazard ratios for the effects of CEF compared with CMF, adjusted for the potential predictive factors described above, for patients with amplified, deleted, or normal TOP2A status. The adjusted P values for the interaction among the three categories of TOP2A and treatment were .22 for RFS and .07 for OS.
For patients in the combined category, altered TOP2A, the 5-year RFS and OS with CEF and CMF treatment are shown in Figures 1 and and2,2, and the adjusted hazard ratio for the treatment effect of CEF compared with CMF were calculated (Table 3). The hazard ratio for interaction between treatment and TOP2A status (altered vs normal) after adjustment for all primary covariates was 0.53 for RFS (95% CI = 0.26 to 1.10, P = .09) and 0.38 for OS (95% CI = 0.17 to 0.85, P = .02).
From 438 patients for whom there were TOP2A FISH measurements, the 430 patients for whom there were also HER2 FISH results from our previous study (1) were included in analyses of the association of HER2 amplification and relative effects of CEF or CMF. Among them, 116 (27%) had tumors with amplified HER2 and 314 (73%) had tumors without HER2 amplification. The 5-year RFS and OS for patients grouped by HER2 status are shown in Table 2. After adjustment for age, node status, tumor grade, ER, surgical procedure, tumor size, and TOP2A alterations, HER2 amplification was not associated with either RFS (HR = 1.20, 95% CI = 0.85 to 1.71, P = .30) or OS (HR = 1.32, 95% CI = 0.90 to 1.93, P = .16).
Adjusted hazard ratios for the effects of CEF compared with CMF were calculated (Table 3). The P value for the adjusted test for interaction between HER2 status and treatment was .008 for RFS (HR = 0.40, 95% CI = 0.21 to 0.79) and .02 for OS (HR = 0.44, 95% CI = 0.22 to 0.90).
The concordance between the presence of TOP2A alterations and HER2 amplification on whole sections is shown in Table 4. The adjusted hazard ratio for the treatment effect of CEF compared with CMF is given in Table 5 for each combination of TOP2A alteration and HER2 amplification status. In an exploratory analysis, we examined whether a combination of altered TOP2A and HER2 amplification would be associated with outcomes or interact with treatment. Those patients with both altered TOP2A and amplified HER2 were compared with the remaining patients. There were few patients with both alteration in TOP2A and amplification in HER2 (23 in CEF group and 27 in CMF group). Coalteration or amplification was not independently associated with either RFS (P = .61) or OS (P = .11). However, there was a trend toward statistically significant interaction between treatment and RFS (HR for patients treated with CEF compared with those treated with CMF = 0.39, 95% CI = 0.15 to 1.06, P = .07) and OS (HR = 0.40, 95% CI = 0.15 to 1.06, P = .06).
TOP2A gene alterations (deletions or amplifications) were not independently associated with either RFS or OS in our multivariable analyses. However, they were associated with the relative benefits of CEF and CMF in terms of both RFS and OS. Tests for interaction were of borderline statistical significance for RFS (HR = 0.53, P = .09) and statistically significant for OS (HR = 0.38, P = .02).
In this study, we attempted to replicate as closely as possible the analytical approach used in our previous publication on HER2 overexpression and amplification in the MA.5 trial (1). We had tissue available from 639 (90%) of the 710 patients in our previous analyses (1), but in the current study, we were able to obtain enough tissue to make usable TMAs for only 438 (61%) of 710 patients. Furthermore, we were able to conduct multivariable analyses on only 395 (56%) patients using the HER2 data based on whole sections and ER from the MA.5 clinical database. Tumor grade was not available for our previous analysis and therefore was not used in multivariable analysis at that time (1).
In spite of these differences, our current HER2 results are similar to those previously published (1–4,13). Although HER2 amplification does not appear to be statistically significantly associated with either RFS or OS in current analyses, after adjusting for two additional factors, TOP2A status and tumor grade, we continue to show in multivariable analysis that HER2 amplification retains a statistically significant association with treatment for both RFS (HR = 0.40, P = .008) and OS (HR = 0.44, P = .02), even when corrected for the presence of TOP2A amplification or deletion. TOP2A alterations, however, also have a borderline or statistically significant association with treatment when corrected for HER2 status (RFS: HR = 0.53, P = .09, and OS: HR = 0.38, P = .02).
The obvious question is which of these factors is most closely associated with treatment benefit from CEF? We found a trend toward statistically significant interaction between treatment and the simultaneous presence of TOP2A alteration and HER2 amplification for both RFS and OS, but the small numbers of patients with alteration in both genes made the interpretation of these analyses difficult. One might conclude from the hazard ratios shown in Table 5 that TOP2A alteration and HER2 amplification appear to be associated to a similar degree with clinical responsiveness to anthracycline-containing chemotherapy regimens. However, a larger sample size would be needed to definitely answer this question. We hope to obtain further data on this matter not only from the results of individual studies but also from an ongoing meta-analysis of individual patient data in five similar trials that compared anthracycline with non-anthracycline regimens in which the association of treatment outcomes with the presence of either TOP2A alterations or HER2 amplification is being assessed.
Some investigators have suggested that patients whose tumors have TOP2A amplification or deletion will always also have HER2 amplification (14). In work by Jarvinen et al., detailed analysis of metaphase FISH, and in some cases fiber FISH estimations of HER2 and TOP2A alterations in 136 primary breast tumors, suggested that HER2 amplification is the initial chromosomal event, followed by TOP2A amplification or deletion (14,5). We found close but not complete concordance between TOP2A gene alterations and HER2 amplification (P < .001; Tables 4 and and5).5). Others have reported cases in which TOP2A is amplified whereas HER2 is not (15,16). Undoubtedly, technical measurement issues will continue to complicate these types of analyses.
In summary, we have shown that both TOP2A alteration and HER2 amplification may be associated with a larger benefit from CEF than from CMF. Patients whose tumors do not have TOP2A alterations or HER2 amplification appear to receive virtually no benefit from CEF as compared with CMF. These data suggest that patients whose tumors do not exhibit TOP2A gene alterations or lack amplification of HER2 could potentially be treated with a less toxic regimen, such as CMF, whereas patients whose tumors show TOP2A alterations or HER2 amplification should receive dose-intensive anthracycline-containing regimens such as CEF. Our data suggest that measurements of TOP2A alteration and HER2 amplification appear to have similar value in guiding the selection of anthracycline-containing regimens.
This study was supported by the National Cancer Institute of Canada, the Canadian Cancer Society, and the Canadian Breast Cancer Research Alliance, and by unrestricted grants from Sanofi-Aventis for some of our early tissue banking efforts and from Adria Canada (subsequently Pharmacia, now Pfizer) for the Mammary 5 trial. Discounted-price kits for the HER2 and TOP2A gene analyses were provided by Vysis-Abbott.
These funding sources had no influence on the design of the study, data collection, interpretation of data, and writing or decision to submit the manuscript.