Topoisomerase IIα in Wilms’ tumour: gene alterations and immunoexpression
- 1Department of Pathology, University of Chicago, Chicago, Illinois, USA
- 2Department of Hematology and Oncology, University of Chicago
- 3Department of Health Studies, University of Chicago
- 4Department of Pediatrics, University of Chicago
- 5Van Andel Research Institute, Grand Rapids, Michigan, USA
- 6Department of Pathology, Northwestern Memorial Hospital, Northwestern University, Feinberg School of Medicine, Chicago
- Correspondence to: X J Yang Department of Pathology, Feinberg 7-334, Northwestern Memorial Hospital, Northwestern University, Feinberg School of Medicine, 251 E Huron Street, Chicago, IL 60611,USA;
- Accepted 24 November 2005
- Published Online First 23 March 2006
Background: Topoisomerase IIα (topoIIα) is an essential enzyme gene in regulating DNA structure and cell proliferation and is encoded by the TOP2A. Using cDNA microarray analysis, TOP2A has been reported to be one of the top genes overexpressed in Wilms’ tumour.
Aim: To evaluate the role of TopoIIα in Wilms’ tumorigenesis and its prognostic value.
Methods:TOP2A gene copy numbers were determined using the fluorescence in situ hybridisation technique, and protein expression levels of TopoIIα by immunostaining in 39 samples of primary and 18 samples of metastatic Wilms’ tumour.
Results:TOP2A gene amplification was detected only in anaplastic Wilms’ tumours, and none of the Wilms’ tumours showed deletion of the TOP2A gene. TopoIIα protein overexpression was detected in 97% of Wilms’ tumours, and correlated strongly with proliferation, as measured by Ki-67 (r = 0.85). The high TopoIIα expression was associated with the presence of vascular invasion, prominent apoptosis, metastases and adverse clinical outcomes (p<0.05).
Conclusions: Our findings suggest that TopoIIα overexpression in Wilms’ tumours is caused by a change at the transcription level, except for anaplastic Wilms’ tumours, in which gene amplification was present. High levels of TopoIIα protein are correlated with tumour aggressiveness. The assessment of TopoIIα expression in Wilms’ tumour may have prognostic value.
- FISH, fluorescence in situ hybridisation
- IHC, immunohistochemistry
- PFS, progression-free survival
- TMA, tissue microarray
- topoIIα, topoisomerase IIα
Wilms’ tumour, also known as nephroblastoma, is one of the most common solid malignant tumours in children, with an incidence of 1/8000.1–3 Wilms’ tumour is histologically characterised by a triphasic pattern of blastemal, epithelial and stromal components, which have different proliferating potential.1 Although, with effective treatment, patients with Wilms’ tumour have nearly 90% long-term survival for localised disease and >70% for metastatic disease,2,4 there is still controversy regarding the best therapeutic strategy and potential molecular targets for antitumour drugs, which can further improve prognosis, reduce treatment-related toxicity and prevent late complications.5–8
In our previous cDNA microarray study, we reported that TOP2A had the highest expression change (80.8-fold increase) compared with non-cancerous kidney.9 Topoisomerase II α (TopoIIα) is an essential nuclear enzyme that regulates DNA topological structure, chromosome condensation and segregation, DNA recombination, transcription and cell cycle progression. Amplification of TopoIIα is predominant in actively proliferating cells and has been reported in non-neoplastic kidney lesions and several malignancies, including breast, colon and pancreatic cancer, and meningiomas.10–14
TopoIIα inhibitors, antibiotics actinomycin D and anthracycline doxorubicin, along with vincristine, are used as the main chemotherapeutic agents for the treatment of Wilms’ tumour.6,7 It was suggested that TopoIIα gene copy numbers and enzyme expression may correlate with the efficacy of anthracycline treatment in breast cancer.10,15–17 However, the mechanisms and molecular pathways of chemotherapy-inhibiting effects are not well understood.
We evaluated TOP2A copy numbers and protein expression levels of TopoIIα in primary and metastatic Wilms’ tumours, and examined their correlation with proliferation indices and various clinicopathological characteristics.
MATERIALS AND METHODS
Fifty seven cases of Wilms’ tumours (39 primary and 18 metastatic) were retrieved from the files of the Department of Pathology, University of Chicago, Chicago, Illinois, USA, with a protocol approved by the institutional review board. Data on histopathological patterns, presence of anaplasia, vascular invasion, capsule or ink penetration were obtained from pathology reports. Clinical data from 39 patients with primary tumours were obtained from the patients’ charts.
Paraffin-wax tissue blocks with primary, metastatic and normal adjacent kidneys were subjected to tissue microarray (TMA). A minimum of two tissue cylinders with a diameter of 1.5 mm were arrayed into a recipient block with a manual tissue microarrayer (Beecher Instruments, Sun Prairie, Wisconsin, USA). The recipient TMA block and 21 conventional blocks were cut into 4-μm-thick sections for analysis by immunohistochemistry (IHC) and fluorescence in situ hybridisation (FISH).
Fluorescence in situ hybridisation
To evaluate the TOP2A gene copy number, we carried out FISH using an LSI TOP2A probe, with two other probes HER-2 and CEP17 as controls in multicolour probe mixture (Vysis/Abbott, Downers Grove, Illinois, USA). LSITOP2A is about 160 kb, a unique sequence probe direct labelled with SpectrumOrange that hybridises to the 17q21–22 region containing TOP2A. The HER-2 probe contains about 190-kb DNA sequences specific for the HER-2(HER-2/neu) gene on 17q11.2–q12 and direct labelled with SpectrumGreen. The CEP17 probe direct labelled with SpectrumAqua contains α-satellite DNA that hybridises to the centromeric region of chromosome 17 (17p11.1–q11.1). The CEP17 probe was used to distinguish true gene amplification or deletion, and as an internal control for chromosome 17 aneusomy. FISH assay was conducted on TMA samples and conventional sections, according to protocols described previously.18 In each sample, an average of 82 (range 30–202) well-defined nuclei were scored.16,18 Tumours with TOP2A:CEP17 and HER-2:CEP17 ratios of >2 were considered to be amplified, and tumours with ratios of 1.5:2.0 as indicating the gain of gene copies. TOP2A was considered to be deleted if the ratio was >0.8.10
Antigen retrieval was carried out by heating sections in TRIS–EDTA (pH 9) for 15 min in a microwave oven. Sections were incubated for 1 h at room temperature with mouse monoclonal antibodies against TopoIIα (clone 3F6; 1:50; Novocastra, Newcastle, UK) and Ki-67 proliferation marker (clone Ki-S5; 1:50; Dako, Carpinteria, California, USA). This step was followed by 30 min of incubation with goat anti-mouse immunoglobulin G conjugated to a horseradish peroxidase-labelled polymer (EnVision, Dako, Hamborg, Germany). Slides were then developed for 5 min with 3-3′-diaminobenzidine chromogen and counterstained with haematoxylin. For detection of HER-2 protein expression, we used the clinically approved HercepTest kit (Dako) according to the manufacturer’s recommendations. Negative controls were prepared by substituting primary antibody with non-immune mouse immunoglobulins.
Evaluation of immunostaining
Nuclear staining of TopoIIα and Ki-67 was quantified by using the Chromavision Automated Cellular Imaging System (Clarient Inc, Aliso Viejo, California, USA). The nuclear staining intensity was measured on the basis of three related colour parameters: hue, luminosity and saturation.19 The Automated Cellular Imaging System software was programmed, by setting colour-specific thresholds, to determine brown (positively stained) and blue (negative) nuclei in the outlined areas of interest, and to calculate the ratio (%) of positively stained nuclei to all nuclei (nuclear index). For each section, at least 10 representative regions were analysed and roughly 50 000 cells were counted.
The association between the continuous expression levels of TopoIIα and Ki-67 in the primary tumours (n = 39) and the clinicopathological variables was assessed using the t test. Expression levels were dichotomised into low-expression and high-expression groups on the basis of the median expression levels (20% and 25% for TopoIIα and Ki-67, respectively).20 Their relationship with clinicopathological variables was assessed using Fisher’s exact test. Differences in expression levels between primary, metastatic and normal tissues were assessed with a mixed-effects analysis of variance model. The mixed-effects model was used to account for the within-subject correlation in patients for whom several tissue samples were available. Progression-free survival (PFS), defined as time from surgery to developing metastases, to relapse or death, was estimated by the Kaplan–Meier method. Differences in survival between groups were assessed using the log rank test. Results were considered significant only if p⩽0.05. Data were analysed with the Stata software (College Station, Texas, USA).
Our study included samples of 39 primary and 18 metastatic tumours. Primary tumours were from 39 patients, with one tumour per patient. Of these 39 patients, 8 presented with metastases; however, only 3 metastatic tumour tissue samples from 2 of the 8 patients were available for analysis. The remaining 11 metastatic tissue samples included in the analysis were collected from patients who developed metastases later, and 4 more tissue samples were from 3 patients whose primary tumours were not available for study.
In all, 24 of the 39 patients with primary tumours were African-American and 15 were Caucasian; mean age was 3.4 (standard deviation (SD) 2.36; range 0.5–9) years. The mean primary tumour size was 11.2 (3.2; 0.5–18) cm. The primary cases of Wilms’ tumour were classified as stage I (n = 15), stage II (n = 9), stage III (n = 11) and stage IV (n = 4). In all, 11 of 20 (55%) male and 4 of 19 female patients (21%) had stage III or stage IV disease. All patients received surgical treatment in combination with pre-surgical (26%) or post-surgical (100%) chemotherapy with either actinomycin D or anthracycline oxorubicin, or both, in combination with vincristine. The median follow-up for all patients was 69.05 (range 3–206) months. Eight patients presented with metastases at the time of diagnosis and 11 patients developed metastases or relapsed 3–76 months later (mean 29.45, median 12 months). Of the 39 patients, 5 (13%) died from Wilms’ tumour, including two of the three patients with unfavourable histology.
We found a significant association between PFS, sex and stage: patients who were female (p = 0.02) and had a low tumour stage (I and II; p = 0.05) had significantly better prognosis. Worse outcome in male patients may be attributed to uneven stage distribution among male and female patients in our cohort, as confirmed by Fisher’s exact test (p = 0.048). We found a weak association between longer PFS in patients without lymph node metastases (p = 0.09) and renal vein invasion (p = 0.10), and in patients treated with TopoIIα inhibitors before surgery (p = 0.08).
TOP2A gene copies by FISH
Inherent probe hybridisation efficiency was found in 93.75% of Wilms’ tumour samples and normal kidney tissues. TOP2A amplification or gain was detected in 4 (12%) samples of Wilms’ tumour with anaplasia: 2 primary and 2 late metastases from the same patients, which also had a simultaneous gain of HER-2 gene (fig 1). TOP2A was not deleted in all studied tumours. We found a strong positive correlation between the ratios of TOP2A:CEP17 and HER-2:CEP17 in samples of Wilms’ tumour (r = 0.923) and in adjacent normal kidney tissues (r = 0.862). The average percentage of aneusomic (monosomic and polysomic) cells was higher in primary Wilms’ tumour than in adjacent normal kidney tissue (p = 0.003), and higher in metastases than in primary Wilms’ tumour (p = 0.032). We found that a higher percentage of chromosome 17 alterations (above the median value of 45% for monosomy and 48% for aneusomy) was weakly associated with the presence of metastases at initial presentation (p = 0.063 v p = 0.09) and tumour progression (p = 0.053 v p = 0.10).
Positive nuclear TopoIIα and Ki-67 staining was observed in all 57 samples, whereas staining in adjacent normal kidney tissues was virtually negative (<0.4% positive cells). TopoIIα and Ki-67 expression levels were found to be the highest in metastatic or recurrent tumours (22.85% and 31.81%, respectively), high in primary tumours (17.36% and 20.94%) and low in adjacent normal kidney (0.38% and 0.26%). Expression was also low in the two stromal Wilms’ tumour (2.9% and 4.85%) samples. Mixed-effects analysis of variation models showed that expression levels of TopoIIα and Ki-67 were significantly higher in primary and metastatic tissues than in adjacent normal kidney (p<0.0001 for each) and higher in metastatic than in primary tumours (p = 0.05 and p = 0.02 for TopoIIα and Ki-67, respectively). Figure 2 shows representative images of TopoIIα expression in primary and metastatic Wilms’ tumours.
We found significant differences between primary tumours expressing high and low levels of TopoIIα with respect to the sex of the patients (p = 0.048), presence of vascular invasion (p = 0.02), prominent apoptosis (p = 0.02) and metastases (p = 0.04). TopoIIα levels were higher in patients with stages III and IV, anaplastic tumours and no pre-operative chemotherapy; however, the differences were not significant. The association of Ki-67 expression levels with clinical and pathological features was similar. The group of patients with low TopoIIα and Ki-67 expression levels tended to have longer PFS (p = 0.09; fig 3).
We found a strong positive correlation between TopoIIα and Ki-67 expression levels (fig 4) in 50 000 cells/case, by automated image analysis (r = 0.85).
Comparison of FISH and IHC
TopoIIα nuclear indices were non-significantly higher (27.75% (11.78%)) than the median value of 20% in all four tumour stages with TOP2A amplification or gain. We found a weak correlation between TOP2A:CEP17 ratios and TopoIIα expression levels in Wilms’ tumour (r = 0.292).
HER-2 amplification or gain was low, with a mean HER-2 copy number ranging from 1.5 to 3.71 per cell (HER-2:CEP17 maximum ratio of 2.23). Immunohistochemistry for the HER-2 protein was essentially negative in all samples of Wilms’ tumour studied.
By applying FISH, immunohistochemistry, tissue microarray and automated imaging analysis, we showed that TOP2A is amplified in only four samples with anaplastic Wilms’ tumour, whereas TopoIIα is overexpressed in most samples. This is apparently caused by mechanisms other than gene amplification. Higher TopoIIα expression levels strongly correlate with tumour proliferation and progression features, such as vascular invasion, local and distant metastases and marked apoptosis (p⩽0.05). There was no evidence of TopoIIα differential expression in predominantly epithelial, blastemal and mixed histological subtypes reported in a semiquantitative study by Granzen and coauthors.21 However, we found that two stromal variants of Wilms’ tumour showed a sixfold lower TopoIIα expression than other variants (p = 0.0031), and that normal epithelium from adjacent kidney tissues was essentially TopoIIα negative.
The results of TopoIIα immunohistochemistry strongly correlate with those of our previous study of mRNA levels by cDNA microarrays and real-time polymerase chain reaction9 in tissues with Wilms’ tumour, which showed marked overexpression (80.8-fold) compared with non-cancerous adjacent tumour kidney tissue. By automated immunostaining, the mean TopoIIα protein overexpression in primary tumours was 45-fold that in non-cancerous kidney tissues (17.36% (SD 11.61%) v 0.39% (0.058%); p<0.0001). Our finding of a higher expression of TopoIIα protein in high-stage Wilms’ tumour is consistent with our previously reported data, an increase of mRNA levels in high-stage Wilms’ tumour.9
Over the past 25 years, long-term survival rates for children with Wilms’ tumour approached 90% for all patients; however, the prognosis for patients who relapse is not as good, with only 30–50% survival expected after retrieval therapy.2,7 Therefore, it is important to uncover the mechanisms of drug resistance and its reliable detection to further improve prognosis for children.22 Two main mechanisms have been suggested to explain resistance to TopoIIα-inhibiting drugs: deletion and mutation of the TOP2A gene ,23,24 whereas TOP2A amplification may cause increased sensitivity to chemotherapy.17,25
In the cases analysed, we found no TOP2A deletion, whereas TOP2A amplification or gain was detected in only a small subset of four tumours. All tumours with TOP2A amplification had unfavourable anaplastic histology and showed simultaneous HER-2 amplification or gain. Absence of TOP2A deletions may explain the overall high sensitivity of patients with Wilms’ tumour to anthracyclines and actinomycin D. However, individual drug sensitivity may vary depending on TopoIIα protein expression levels.
Our findings suggest that measuring the TopoIIα protein expression may have some value in predicting sensitivity to chemotherapy in patients with Wilms’ tumour. For instance, patients who underwent pre-operative chemotherapy tended to have lower levels of TopoIIα expression and lower proliferation indices (13.26% and 17.54%) than untreated patients (18.99% and 22.02%). Interestingly, a cohort of patients who were treated before the operation and who did not show any sign of tumour progression had the lowest nuclear indices for TopoIIα and Ki-67 (7.67% and 11.06%) in tumours that were removed regardless of stage. Our findings coincide with the previous observation of reduced TopoIIα expression in Wilms’ tumours after chemotherapy.26 Our results also indirectly support in vitro data25,27 that the sensitivity of breast cancer cells to TopoIIα inhibitors is dependent on the expression levels of TopoIIα rather than on the TOP2A amplification status.
In contrast with the strong correlation of TopoIIα mRNA and protein levels, we found no correlation between TOP2A status using FISH and TopoIIα using IHC (r = 0.29). Mueller et al28 obtained similar results in breast tumours, which showed no correlation between TOP2A copy number and TopoIIα expression levels. These findings suggest that TopoIIα expression is highly regulated at the level of transcription and translation.
Ki-67 protein expression in Wilms’ tumour tissue was about 25% higher than that of TopoIIα expression, but nuclear indices of TopoIIα and Ki-67 strongly correlated with each other (r = 0.85), and were higher in tissues with metastatic, relapsed and anaplastic Wilms’ tumour. High TopoIIα expression seems to be a general feature of rapidly proliferating cells, with strong, cell cycle-dependent connection and accumulation of the TopoIIα active form of the enzyme during S, G2 and M phases of replicating chromatin.29
In summary, the overexpression of TopoIIα in all samples of Wilms’ tumour and detection of TOP2A amplification in only 12% of those suggest that the abnormality responsible for raised expression of TopoIIα mRNA and protein is mainly at the transcription level. To our knowledge, this is the first study to correlate clinical and prognostic features with TOP2A status and TopoIIα expression levels in tissues with Wilms’ tumour. Our results suggest the potential clinical utility of TopoIIα immunostaining in providing guidance for designing treatment for patients with Wilms’ tumour.
We thank Dr Olofunmilayo Olopade for reviewing this manuscript and the Doris Duke Foundation for supporting FISH studies by TG. We also thank Can Gong for her excellent technical assistance with immunostaining and TMA.
Published Online First 23 March 2006
Competing interests: None declared.
Presented in abstract form at the 95th Annual Meeting of the United States and Canadian Academy of Pathology, San Antonio, March 2005.