Phase II Trial of Karenitecin in Patients with Malignant Melanoma: Clinical and Translational Study

Patients and Methods

Enrollment criteria. A phase II trial of karenitecin (MCC 12824) was initiated at our institution on March 2002. Patient enrollment ended April 2003 once target enrollment was completed. The study design was approved by the Scientific Review Committee at Moffitt Cancer Center and the University of South Florida Institutional Review Board. Informed consent was obtained from all patients. Patients were required to have histologically or cytologically proven metastatic melanoma with measurable disease, and an ECOG performance status of 0 or 1. Patients could have had up to three prior chemotherapy regimens and central nervous system disease if the disease was stable or resected. Skin, mucosal, and uveal primary sites were eligible. Patients were excluded from this study if they were pregnant or nursing or had an active second malignancy. Patients were excluded if the following initial laboratory parameters were abnormal (bilirubin >1.5, creatinine <1.4 ANC <1,500, or platelets <100,000).

Statistical design and response assessment. This study was designed as an open-label, single-institution phase II trial to test the efficacy of karenitecin in melanoma. A Simon two-stage optimal trial design was followed. We planned to accrue 12 patients in the first stage of the trial and to close accrual if no response was seen (defined as a complete or partial response or as stable disease of ≥3 months duration). If one or more responses were seen, we would enroll up to 45 evaluable patients. We estimated that if this treatment had a 25% true response rate, we would have a 3% chance of stopping the study early and a 81% chance of considering it worthy of further study. At the end of stage 2, if seven or more responses of the 45 patients were seen, then the treatment would be considered worthy of further study. Bidimensional measurement (sum of the product of longest perpendicular diameters) was used to determine response/progression. A complete response was defined as resolution of all disease lesions with no new lesions and normalization of any elevated tumor markers confirmed at least 4 weeks after the initial observation. Partial response was defined as 50% reduction in bidimensional measurements confirmed at least 4 weeks after the first observed reduction. Minor response was defined as 25% to 49% reduction in bidimensional measurement. Stable disease was defined as neither shrinkage sufficient to qualify for partial response/complete response/minor response nor increase sufficient to qualify for progressive disease. Progressive disease was defined as increase in bidimensional measurement by 25%. Clinical benefit included patients with complete response, partial response, and stable disease for at least 3 months.

Treatment. All patients received 1 mg/m²/d karenitecin for 5 days consecutively. Treatment cycles were repeated every 21 days. Patients were premedicated with 8 mg ondansetron and 10 mg dexamethasone IV. The dexamethasone premedication was dropped in subsequent cycles if little or no nausea was observed. No colony stimulating factor support was utilized in this study; erythropoietin could be used for symptomatic anemia with a hemoglobin <10 g/dL and for any anemia with a hemoglobin <8 g/dL. Patients were evaluated for disease progression every two cycles. Treatment was continued only if there was stable disease (<25% increase in disease) or if there was disease reduction as evaluated by bidimensional measurements on computed tomography scan or on physical exam. Dose reductions for future cycles were based on hematologic or hepatic dysfunction; for febrile neutropenia, dose was reduced by 25% in succeeding cycles. If counts had not recovered to at least ANC ≥1,500 or platelets ≥100,000, treatment was held for 1 or 2 weeks until recovery of counts. If patients had two consecutive cycles of delay, the dose was reduced to 0.75 mg/m² and then to 0.5 mg/m².

Topoisomerase assay and fine-needle aspiration biopsy. Patients were consented to have biopsy only if they had metastases that can be biopsied with minimal risk. The biopsy was attempted on day 1 before the start of therapy and then on day 3 of the first cycle of treatment. Biopsy was done with a 23-gauge needle and two to four passes were done on each lesion by the study pathologist (Dr. B. Centeno). Aspirated cells were stained immediately in Diff-Quick and examined under a microscope to verify that a pure population of melanoma tumor cells was obtained. A sample was aspirated into 0.9% saline, serial cytospin slides were prepared, and frozen at −20°C until assayed. Quantitative immunflourescent staining for topoisomerase I, IIα, and IIβ was done as described previously (9). Briefly, slides were thawed and fixed at 20°C with 4% paraformaldehyde in PBS (Electron Microscopy Sciences, Fort Washington, PA) for 15 minutes. The cross-linking was stopped by rinsing slides in 1% glycine in PBS and cells were permeabilized for 24 hours by several changes of 0.25% Triton X-100 in 1% glycine-PBS. Rabbit polyclonal antibody 454 against topoisomerase IIα was combined with antihistone antibody diluted 1:100 with 0.1% NP40, 1% bovine serum albumin in PBS, and incubated for 1 hour at room temperature with the primary antibodies. The other combination included monoclonal anti–topoisomerase I and polyclonal anti–topoisomerase IIβ antibody. After three washes with PBS for 1 hour, the slides were incubated with 1:80 diluted goat anti-rabbit IgG-tetramethylrhodamine isothiocyanate labeled for the polyclonal primary antibodies and 1:125 diluted goat anti-mouse IgG-FITC for the monoclonal antibodies (Sigma, St. Louis, MO) in 0.1% NP40 and 1% bovine serum albumin in PBS for 35 minutes at room temperature. After three washes in PBS, the slides were dried and covered with coverslips in Vectashield mounting media of antifade/4′,6-diamidino-2-phenylindole (1:1; Vector Laboratories, Inc., Burlingame, CA). Immunofluorescence was observed with a Leitz Orthoplan 2 microscope and images were captured by a charged coupled device camera with Smart Capture program (Vysis, Downers Grove, IL). Quantification was done using the pixel count of the image obtained and was reported in arbitrary units.

Results and Analyses

Software package SAS version 8.2 was used in all of the following statistical analyses.

Patient characteristics. A total of 46 patients were enrolled on trial MCC 12824. One patient withdrew consent during his initial treatment. Another two patients were found to be ineligible for treatment due to violation of entry criteria (ECOG performance status >2). An additional two patients were not evaluable as they had stable disease after two cycles but decided not to continue on treatment. Therefore, a total of 41 patients were evaluable for response, 43 were evaluable for toxicity, and 45 patients were evaluable for survival. Patient characteristics are described in Table 1. The age, sex, and race distribution are close to our institutional and national norms. The majority of patients had visceral involvement and/or elevated lactate dehydrogenase; 72% of patients had stage M1C disease. There was a significant number of patients treated who had nonconjunctival ocular (6 of 41) and mucosal (2 of 41) primaries. Fifty-six percent of patients had previously received chemotherapy (mostly DTIC based, including biochemotherapy), and 26% had failed two prior chemotherapy regimens. Twenty-six patients had received radiation therapy, and 5 of 43 patients had brain metastases, which were “stable” before the start of therapy.

Treatment responses. In this trial, we had a total of 41 patients evaluable for response and 14 responses (complete response + partial response + minor response + stable disease). Estimated response rate is 0.341, with a SE of 0.074. Details of response of patients to this treatment are shown in Tables 2 and 3. One patient experienced a complete response to treatment. He had received 16 cycles of treatment with regression of skin lesions after the first four cycles and developed stable necrotic appearing lesions in the pancreas on computed tomography scan. These lesions showed no uptake on positron emission tomography scan and were resected. Pathology from resected lesions showed intrapancreatic and peripancreatic lymph nodes completely replaced with necrotic tumor (Fig. 1). Interestingly, this patient had previously progressed on concurrent biochemotherapy (cisplatin, vinblastine, dacarbazine, interleukin 2, and IFN-I) and then on an investigational trial with Epothilone B (Epo 908, Novartis, East Hanover, NJ). A significant number of patients (32%) or 13 of 41 experienced minor responses or disease stabilization (Tables 2-4). The longest to date is a patient with ocular melanoma with bone, brain, liver, and soft tissue metastatases had stable disease lasting for 12 cycles (36 weeks). He had stable brain metastasis on karenitecin for 36 weeks without any other treatment during this entire period. His treatments were held as he needed surgery for an unrelated condition and he progressed soon afterward. Several other patients had relatively long periods of stable disease. Patients with ocular melanoma particularly appeared to benefit from karenitecin; three of five patients with ocular melanoma had stable disease. Of these, one patient had previous chemoembolization of liver lesions with progression of disease before treatment and remained stable for 18 weeks on karenitecin; one patient had previous radiation therapy and remained stable for 36 weeks (see above); and one patient who had no prior therapy was diagnosed with concurrent ocular and liver disease and remained stable for 18 weeks on therapy.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Patient SAB 4 was treated with karenitecin for 16 cycles (48 weeks). He had regression of subcutaneous disease with stable intra-abdominal disease by computed tomography scan. He had a positron emission tomography scan showing no glucose uptake (data not shown). He underwent resection of two intrapancreatic and peripancreatic lesions. Left, a low-power view of an intrapancreatic lymph node replaced with metastatic malignant melanoma that showed nearly 100% necrosis. Right, central region at high power showing melanotic cellular debris and ghost-like remnants of tumor cells.

Survival data analysis. Forty-five patients were evaluable for survival. Progression-free survival was computed as the time between beginning treatment and disease progression (Fig. 2). Overall survival time was defined as the time beginning at treatment until death, last known follow-up, or off treatment (Fig. 3). Kaplan-Meier curves were used to describe progression-free survival and overall survival. The median progression-free survival time was 7.9 weeks and corresponding 95% confidence limit was (7.0, 9.1) weeks. The Kaplan-Meier progression-free survival curve for all evaluable patients is shown in Fig. 2.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

The Kaplan-Meier progression-free survival curve for all evaluable patients treated with karenitecin.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

The Kaplan-Meier overall survival curve for all evaluable patients treated with karenitecin.

For overall survival, the median was 33.4 weeks and corresponding 95% confidence interval was (21.0, 38.7) weeks. The percentage of the patients that are still living is 26.67%. The Kaplan-Meier overall survival curve for all evaluable patients is shown in Fig. 3.

Toxicity. Karenitecin was well tolerated in this trial. Adverse effects seen on therapy are shown in Table 4. The predominant toxicity was hematologic. Anemia was the most frequent toxicity (79% of patients experiencing some degree of anemia) with 19% of patients experiencing grade 3 anemia. Neutropenia and thrombocytopenia were also relatively frequent (44% each). We did not use growth factor support during this trial and in most patients neutropenia resolved with dose delay. Thirty-two percent of patients experienced either grade 3 or 4 neutropenia. Neither neutropenia nor thrombocytopenia seemed to be cumulative. One patient died while on treatment with an intracranial hemorrhage that occurred during the platelet nadir (this patient was also taking aspirin). While gastrointestinal symptoms had been anticipated, only mild abdominal pain and nausea were seen with any frequency. The overall incidence of diarrhea (only grade 1 and 2 were seen) was 16%. Local infusional site discomfort was a notable complaint in 18% of patients and a sterile phlebitis developed in three patients. Fatigue (usually mild and transient) was also frequently noted.

Assessment of tumor specimen topoisomerase levels by immunofluorescence. The fine needle aspiration biopsy sample provided an abundant (>100,000 cells per sample) and nearly homogenous (>95% pure) population of melanoma cells (Fig. 4A) for immunostaining. Thirteen patients had pretreatment and posttreatment biopsies; a further seven patients had only pretreatment biopsies. Abundant expression of topoisomerase I was seen in patients who underwent biopsy (Figs. 4 and 5). We also profiled topoisomerase IIα and IIβ expression and location, which have been reported to increase following topoisomerase I inhibitor treatment (10). We found no increase on day 3 of treatment compared with pretreatment levels; we do report high levels of these enzymes, especially topoisomerase IIβ in most melanoma cells (Fig. 5). No relationship was seen between pretreatment topoisomerase I expression and response to therapy or to localization with karenitecin treatment.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Topoisomerase I, IIα, and IIβ expression in melanoma cells. Fine-needle aspiration was done immediately before treatment and on day 3 of treatment with karenitecin. Aspirated cells were stained to verify that a pure population of melanoma tumor cells was obtained (A). Immunoflourescent staining for topoisomerase I, IIα, and IIβ, respectively, was done (B). Western blotting was also done (bottom). Controls are HL-60 cells at days 5, 2.5, and 1 after differentiation, and melanoma cells are shown at day 1 pretreatment and day 3 of treatment (d3).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Cumulative topoisomerase I, IIα, and IIβ expressions in nuclei of melanoma cells. Bars, SE.

Discussion

Current therapy for metastatic melanoma remains highly unsatisfactory (1, 11, 12). Most chemotherapy regimens contain the alkylating agent dacarbazine. This drug and its congener, temozolomide, produce objective response rates in patients with melanoma in 10% to 20% of patients. It is currently not known if dacarbazine prolongs survival in patients with melanoma. Combination chemotherapy in the form of the “Dartmouth regimen” has higher reported response rates, but in a prospective randomized trial showed no advantage in survival over dacarbazine alone (13). More recently, IFN-α and interleukin-2 have been incorporated into biochemotherapy regimens that also showed very high response rates (45-60%) in single-institution phase II trials. However, in a prospective randomized clinical trial, biochemotherapy failed to show any survival benefit over chemotherapy alone (4). Therefore, it seems that varying combinations of currently used drugs will have only limited utility in this disease.

The topoisomerase I inhibitors have shown a wide spectrum of activity in human malignancies. Whereas initial studies of sodium camptothecin, which was formulated in sodium hydroxide that lead to hydrolysis of the E-ring lactone, had limited clinical utility due to its severe side effect profile (especially hemorrhagic cystitis) and absence of antitumor activity (probably because the drug was in the carboxylate form), other topoisomerase I inhibitors, such as topotecan and irinotecan, have proved useful in the clinic. Two other trials of the topoisomerase I inhibitor topotecan have been reported to our knowledge (14, 15). In the first study, 28 patients with melanoma were treated, 27 were evaluable, and 1 patient with M1a disease had a partial response. No statistics were reported on stable disease. In the second study, presented in abstract form only (15), 17 patients were treated in this study and no activity was seen. The major side effect seen was myelosuppression (severe grade 3-4 in 70% of this small sample). Topoisomerase II inhibitors have also been explored in melanoma. Epirubicin was tested in several phase II trials and not found to be effective (16–18). In our preliminary analysis, melanoma cells expressed high levels of topoisomerase I. Furthermore, we found that karenitecin, a novel camptothecin analogue, showed activity against melanoma cell lines in vitro. This molecule has a novel silyl substitution in the camptothecin B-ring, which enhances the drug's lipophilicity, increases its lactone stability, and the agent cannot undergo glucuronidation, which, in turn, reduces intrapatient variability, and the agent has the lowest observed susceptibility to breast cancer resistance protein–mediated drug resistance (19, 20). Although not clinically proven, based on its liphophilic nature, karenitecin seems to have some degree of central nervous system penetration, which would be a major advantage in melanoma with its high proportion of central nervous system involvement.

We found in the current trial that karenitecin has appreciable activity in melanoma. Even in the poorest prognosis patients (visceral disease with prior chemotherapy treatment), it was possible to have prolonged disease stabilization. Indeed, the only patient who experienced a complete response on this trial had 16 cycles of therapy after having progressed on biochemotherapy and on an investigational trial of Epothilone B. Having said that, the major treatment effect we observed in this trial was disease stabilization rather than tumor shrinkage. Some of the disease stabilization was accompanied by a decrease in lactate dehydrogenase for the duration of treatment (the longest being 9 months). Again, this effect was observed both in patients with progression on previous regimens and on previously untreated patients. Another interesting aspect of this study was the inclusion of patients with nonconjunctival ocular melanoma and mucosal melanoma. In patients with ocular melanoma 60% (three of five patients) had disease stabilization. Whereas this sample is too small to draw any conclusions, they do deserve to be included in future trials of novel agents in melanoma and in light of their biological differences with cutaneous melanoma may actually have better responses to treatment.

Treatment with karenitecin was generally well tolerated. The major complaint voiced by patients was the inconvenience of having treatment 5 of 21 days resulting in a disruption of lifestyles. The major side effect observed in this study was noncumulative reversible hematologic toxicity. Twelve of 38 patients had grade 3 or 4 neutropenia. Because no growth factor support was allowed, dose delays were the major consequence of this toxicity. Anemia was also frequent with 87% of patients having some anemia and 23% having moderate or severe (grade 3 or 4) anemia. Thrombocytopenia occurred in 50% of patients treated and six patients had moderate to severe thrombocytopenia. The one death on the study that was considered related to treatment occurred in a patient who had had severe thrombocytopenia in a previous cycle and had a bleed related to a brain metastasis.

There is very limited prospective human data on topoisomerase expression. One of the major aims of this study was to profile topoisomerase expression in melanoma before and during treatment. In preclinical studies, topoisomerase IIα levels have been shown to increase with topoisomerase I inhibitor treatment (10). We found no significant change of topoisomerase IIα with karenitecin treatment. The discrepancy may relate to the tumor type studied or due to a difference between the model studied by Whitacre et al. and human subjects studied in our trial. Previous studies have suggested that topoisomerase expression may correlate to sensitivity to topoisomerase poisons (21–23). Other studies have not shown a correlation between topoisomerase expression and outcome (24, 25). One caveat with these studies is that determination of topoisomerase expression in paraffin-embedded blocks may allow translocation of topoisomerase and, therefore, make it hard to evaluate expression, this may be reduced by rapid fixation of live cells, such as those used in our study (9). Another concern with topoisomerase expression has been evaluation of nuclear topoisomerase as opposed to total cellular topoisomerase (which is the active enzyme; refs. 26–29). Therefore, we quantified topoisomerase expression in nuclei taking precautions to reduce processing time. Because topoisomerase treatment could reduce levels of topoisomerase, we sampled tumors both before and on day 3 of treatment (30, 31). Additionally, we measured topoisomerase isoforms that are not targets of karenitecin to evaluate if any compensatory change in expression may occur during treatment 10. We found abundant expression of topoisomerase in the majority of tumors examined (Fig. 4). There was especially high expression of topoisomerase I, suggesting that this may be an interesting target for therapy in melanoma. Most of the topoisomerase expression was nuclear, rather than cytoplasmic, suggesting its utility. Topoisomerase I expression did increase (although this was not significant) with treatment (Figs. 4 and 5). Additionally, high level expression of topoisomerase IIβ was also seen, suggesting that this may be an interesting target for future trials. We did not find evidence of nuclear to cytoplasmic translocation of topoisomerase isoforms, which has been reported as a potential mechanism of resistance to topoisomerase poisons. Because only easily accessible tumors were biopsied and most patients with clinical benefit tended not to have tumors that can easily be biopsied, only three patients with clinical benefit had topoisomerase expression profiling. Hence, no conclusions could be drawn regarding topoisomerase expression and/or localization and response. Another caveat to these analyses is that functional analysis of topoisomerase function and/or cleavable complexes was not done, which could further help to determine the effectiveness of these drugs. However, it is clear that fine-needle aspiration biopsy represents a easy means to access and evaluate a population of a given patient's tumor cells, which can be useful for biochemical and histologic analyses. We are currently developing an assay to determine the frequency of cleavable complexes in fine-needle aspiration specimens that may be useful in future trials.

Based on these results, we believe the topoisomerase inhibitor, karenitecin, has moderate activity in the current schedule in melanoma. We are currently examining this drug in oral form using a prolonged dosing in combination with another currently used antimelanoma agent to determine if higher response rates may be generated.