The Topoisomerase IIα Expression Correlates with Survival in Patients with Advanced Hodgkin’s Lymphoma

HD² has been presented as an example of a curable illness; however, ∼5% of patients suffer from progressive disease while undergoing therapy, and an additional 40% relapse in the advanced stages. Current progress is aimed at trying to optimize the treatment, individualizing it to the necessities of the patients, and, at the same time, inducing the least possible adverse effects. Efforts have been made to identify useful clinical prognostic factors, as in the International Prognostic Factor Project (1), and other biological factors such as β₂ microglobulin, and elevated levels of cytokine, for instance interleukin-10 (2), with controversial results or lack of confirmation in later studies.

The Topo family has been identified as the molecular target of many chemotherapeutic agents. Human cells are known to contain the following five topoisomerase family members: Topo I, Topo IIα and IIβ, Topo IIIα and IIIβ. Topo I, IIα, and IIβ are targets for several natural product-derived anticancer drugs. Topo I is targeted by camptothecins, whereas Topo II isozymes are the target for the epidophyllotoxins and DNA interacalators such as doxorubicin. Evaluation of Topo protein expression may be used to design rational combination therapies with Topo-targeting drugs. A number of studies have correlated the level of Topo II with a response to anti-Topo II drugs in cancer cell lines (3, 4, 5) Topo IIα has also been suggested as a cell proliferation marker (6), because Topo α expression increases during the late S phase and decreases at the end of the M phase, and anti-Topo α antibody labels cells in the S, G₂, and M phases of the cell cycle (7). The Ki67 protein is also demonstrable in all phases of the cell cycle except for G₀ (8).

The prognostic relevance of a high proliferation index, as measured with the Ki67 antibody, has been extensively shown in non-Hodgkin’s lymphomas (9); however, very few studies have been performed in HD (10), but it appears to be associated with shorter survival.

To our knowledge, only one study has been reported describing Topo IIα expression in HD cells (11), but this did not analyze the clinical evolution, response to treatment, or relationship with survival.

The aim of this study was to analyze a series of HD patients with a high probability of relapse, who were homogeneously treated and had a long-term follow-up, to determine whether the expression of Topo IIα and Ki67 in HD played a role as a prognostic factor or predictor of response to treatment.

This study was performed in patients with HD studied and treated in the Pathology and Medical Oncology Departments of the University Hospital “Puerta de Hierro.” Patients were eligible for this study if they fulfilled the following criteria: (a) patients with stage II, III, or IV, and two or more risk factors (presence of B symptoms, bulky disease, and ESR ≥30 mm/h); (b). They must have received primary treatment with Ch or Ch plus radiotherapy in bulky areas; (c) the initial diagnosis must have been made in a lymph node biopsy before treatment; (d) paraffin-embedded, formalin-fixed tissue blocks from the diagnosis lymph node must have been available for immunohistochemical studies; (e) a minimum follow-up of 2 years was required; and (f) HIV-infected patients were excluded. All of the cases were reviewed independently by three pathologists (S. R. C., C. C., C. S.). Controversial cases were excluded. The percentage of positivity in the immunohistochemical study was assessed independently by two pathologists (S. R. C., C. G.), and there were no significant differences between them in their evaluations. A total of 42 patients was accepted for this study.

The extension study and clinical classification followed the Rye criteria (1971), and those adopted in the Ann Arbor (12) and Cotswold consensus (13) revisions. Bulky disease was defined by the presence of adenopathies >10 cm or a mediastinal mass larger than one-third of the maximum diameter of the thorax. Ch included combinations of MOPP (14), which were used in our service until 1985, and, since then, have been alternated with ABVD (Ref. 15; doxorubicin 25 mg/m² i.v., bleomycin 10 mg/m² i.v., vinblastine 6 mg/m² i.v., and dacarbazine 375 mg/m², all administered on days 1 and 15; repeated 4 weekly) given in the doses and schedule described by Bonadonna et al. (16). Vincristine was capped at 2 mg. Ch was given in full doses if the granulocyte count exceeded 1 × 10⁹/liter and platelet count exceeded 100 × 10⁹/liter. If the granulocyte count was <1 × 10⁹/liter or the platelet count was <100 × 10⁹/liter, Ch was delayed by 1 week.

Complete response was defined as the absence of any clinical or radiological evidence of disease for at least 4 weeks after cessation of treatment. In cases of residual mass, a response was considered to be complete when the mass remained stable for at least 1 month after radiation therapy, or throughout two consecutive Ch cycles (17), and without clinical or analytical signs of active lymphoma. Incomplete response was any outcome that did not meet the above-mentioned criteria, although there was a partial response or stabilization.

The patients have been prospectively followed-up since their treatment; every 3 months during the first 2 years, 6 monthly between 2 and 5 years, and annually from the 5^th year. Salvage regimens included different combinations of chemotherapy and, in some cases, autologous stem cell transplantation.

Immunohistochemical staining was performed on deparaffinized sections using the avidin-biotin complex immunoperoxidase technique with separate antibodies recognizing CD30 and the α isoform of Topo II. Sections of 2–3 μm were used on DAKO Chem Mate 75-μm slides. The slides were melted, dewaxed, and incubated in a 1:10 dilution of DAKO sodium citrate buffer in a pressure cooker that was heated until maximum pressure was reached. Immunohistochemical staining was performed with a HORIZON DAKO (MESIP program) automated immunohistochemical stainer in accordance with the manufacturer’s instructions. Topo IIα antibody was a purified mouse monoclonal antibody (NOVOCASTRA NCL, Clone 3FG), and was used at a dilution of 1:25. CD30 antibody (DAKO M0751, Clone VER-H2) was used at a dilution of 1:25. Detection of bound antibodies was accomplished using a peroxidase detection kit (DAKO K5001) with diaminobenzidine as the chromogen. RS cells were identified by positive staining with CD30 in all of the cases. At least 50 RS or RS-variant cells were studied, depending on the tissue size and number of RS or RS-variant cells present. The percentage of cells expressing cytoplasmic CD30 and nuclear Topo IIα staining simultaneously was determined for each case.

OS was calculated from the date of diagnosis to the date of the last follow-up or death from any cause. FFS was defined as the interval from the start of treatment to death, disease progression during treatment, or relapse. The actuarial survival analysis was performed according to the method described by Kaplan and Meier (18), and differences between curves were evaluated with the log-rank test (19). The effect of patient and tumor characteristics on FFS and OS were evaluated in a univariate analysis using the Cox proportional hazards model (20) calculating the RR of each variable and CIs. The Cox model was also implemented in a multivariate analysis using stepwise selection. Two-tailed Ps of <0.05 were considered to be significant. Statistical analyses were performed using the SPSS version 9.0 software package.

Forty-two patients were selected for the study, 27 (64%) were male, the mean age was 29, (range, 11–61). For stages, 30 (71%) patients were III+ IV, and all were treated with chemotherapy; 12 (28%) additionally received radiotherapy in bulky disease. Table 1 summarizes the main initial characteristics of the patients. Univariate analysis of pretreatment variables for prognostic significance included age, sex, histology, number of areas involved, mediastinal involvement, presence of B-symptoms, ESR, LDH, serum albumin, and hemoglobin. Treatment types were also included in the analysis.

Positive nuclear staining for Topo IIα in the RS or RS-variants was seen in 90% (38 of 42) of HD, 88% (37 of 42) for Ki67, and 74% (31 of 42) for CD30. The mean Topo IIα was 30%, with a range of 1–90%.

Coexpression of Ki67 and Topo IIα was seen in 79% (33 of 42), and coexpression of CD 30 and Topo IIα in 72% (30 of 42). The mean Ki67 was 20%, with a range of 1–50%.

No significant difference in percentage of neoplastic cells expressing Topo IIα, Ki67, and CD 30 was found among histologic subtypes.

The median follow-up was 193 months (range, 24–256 months), and no patient was lost at follow-up. The OS was 94% (95% CI, 88–100) at 2 years, 79% (95% CI, 67–91) at 5 years, and 66% (95% CI, 46–86) at 10 years. The FFS was 72% (95% CI, 60–84) at 2 years, 65% (95% CI, 52–70) at 5 years, and 43% (95% CI, 25–61) at 10 years.

The results of the univariate Cox test are shown in Table 2. The male gender had an RR of 2.81 (95% CI, 1.01–8.07; P = 0.04), high LDH at diagnosis (RR, 2.87; P = 0.02), and Topo IIα, when using a cutoff of 30%, had an RR of 2.91 (95% CI, 1.23–6.89; P = 0.014), and were significantly associated with more relapses and shorter FFS. Ki67 ≤20 (RR, 2.21; P = 0.068) showed a tendency toward significance. A greater tendency to relapse also existed when MOPP Ch was used compared with alternant MOPP/ABVD regimens (P = 0.14). There were no statistically significant differences when the chemotherapy type was correlated with the degree of Topo IIα expression.

In the multivariate analysis for FFS, only Topo IIα <30% was statistically associated with shorter FFS, with RR of 3 (95% CI, 1.26–7.15; P = 0.013; Figs. 1 and 2). Male gender showed a tendency to shorter FFS, but it did not reach statistical significance (RR, 2.50; P = 0.09).

Only Topo IIα, when using a cutoff of 30% (RR, 2.83; P = 0.05), was significantly associated with shorter OS. High LDH at diagnosis (RR, 2.60; P = 0.08) showed a tendency toward significance.

In multivariate analysis for OS only Topo IIα <30% was statistically associated with shorter survival, with RR = 2.82 (95% CI, 1.05–7.53; P = 0.038; Fig. 3).

In spite of the success obtained in the treatment of HD, there are still no clear and unanimously accepted risk factor identifiers. One of the biggest efforts to clarify this issue was made by the German Hodgkin’s Lymphoma Study Group, and three big prognostic groups were identified. Early stages encompass former clinical stages I and II without the presence of any known risk factor, such as large mediastinal tumor mass, extranodal involvement, high ESR, or three or more lymph node areas being involved. If any of these risk factors is present, patients are categorized in the intermediate stage group. Only patients in stage IIB with a large mediastinal tumor and/or extranodal involvement are included within the advanced stage group in addition to all of the patients with stages III or IV.

The group of patients in our study was homogeneously treated; all of them had a long follow-up period, and they presented a significant risk of relapse. The clinical characteristics, as well as the treatments administered and survival data, are similar to other large series with advanced HD (21, 22) treated with standard chemotherapy. Therefore, we believe that no significant biases existed to prevent valid conclusions being reached.

Knowledge of the molecular bases related to the progression of HD is very scarce. The role of p53, Ki67, and bcl2 have been studied, and, contrary to non-HD lymphomas (23, 24, 25) where they have been clearly related with shorter survival, studies on the various cellular cycle regulatory factors in HD are contradictory (26, 27, 28). In our study, the most important prognostic factor for disease-free survival as well as global survival was Topo IIα expression inferior to 30%, with an RR of 2.83 (95% CI, 1.05–7.53; P = 0.038), being much greater than any other factor. No study has been published to date that relates clinical evolution and Topo IIα expression in HD patients. One study has been published recently in non-HD lymphomas where it seems that Topo IIα determination could be useful in identifying a greater tendency to relapse (29).

Topo IIα expression in HD is very high, occurring in >90% of the tumoral cells, greater even that the expression of Ki67 (88%) or CD 30 (74%). No differences were found between the different HD cellular subtypes in function of the different degrees of Topo IIα expression. Topo IIα expression has been studied in low-grade astrocytomas with 4% of immunopositivity, or in multiform glioblastomas with 13.8%, both clear examples of chemoresistant tumors (30), whereas it is detectable in other tumors that are curable with chemotherapy, such as seminoma (31), or it is more frequently identified in the more chemosensitive type of lung tumors such as the small cell variety than in less chemosensitive non-small cell types (32). Topo IIα is also a target for several chemotherapeutic agents such as adriamicin or etoposide. Hypothetically the level of Topo IIα would contribute with additional information to select groups with different response to the chemotherapy. In our experience, the level of Topo IIα is more useful to predict relapse than other molecular parameters such as Ki67, or than any of the clinical parameters studied in patients with advanced HD treated with standard chemotherapy.

More relapses were found in patients treated with MOPP than in the group treated with ABVD, with a tendency toward significance but not reaching statistical value. If we admit that a high Topo IIα expression exists in HD, and that moreover it is a proliferation marker, as is Ki67, it could represent a role as a response predictor, because of its function as a target for various chemotherapeutic agents, and could explain the well-known greater relapse and lesser cure rate expected in patients only treated with MOPP against combinations that have Topo IIα inhibitors (33). It is difficult to reach valid conclusions because of the small number of patients exclusively treated with MOPP in our study. We believe that the determination of Topo IIα in advanced HD may open new prospects as possible predictor of response to treatment in patients with HD, selecting patients with more sensitivity to combinations with Topo IIα inhibitors. Combination chemotherapies such as bleomyein, etoposide, adriamycin, cyclophosphamide, vincristine, procarbazine, prednisone. have demonstrated high activity in high risk patients (34).

In our study, Ki67 presented a tendency toward statistical significance in disease-free survival and global survival in the univariate analysis. The cutoff point in our study was 20%, which was the median Ki67 expression, and was similar to the study by Morente et al. (10). This is lower than the found as a cutoff point to determine bad prognosis in high-grade non-HD lymphomas and close to the level found in other intermediate-grade non-HD lymphomas (35). The loss of significance in the multivariate analysis could be because of the selection of patients, so whereas in the study by Morente et al. (10) the factors of bad prognosis associated with Ki67 were advanced stages and B symptoms, in our study most of the patients belonged to this category.

In conclusion, Topo IIα expression seems to be useful in HD to identify patients with a high risk of relapse and worse OS. Level of Topo IIα presents a higher significance than other clinical, molecular, or analytic parameters studied and may have a potential utility in the design of specifically directed treatments. Because of its high expression, easy determination, and reproducibility, randomized clinical trials should include studies of Topo IIα to determine its true value as a predictor of response to chemotherapy.

We thank Martin Hadley-Adams for assistance with the English language.