Phase I Study of Topoisomerase I Inhibitor Exatecan Mesylate (DX-8951F) Given As Weekly 24-Hour Infusions Three of Every Four Weeks

Vol. 7, 3963–3970, December 2001 Clinical Cancer Research 3963

Phase I Study of Topoisomerase I Inhibitor Exatecan Mesylate (DX-8951f) Given as Weekly 24-Hour Infusions Three of Every Four Weeks

Sunil Sharma,1 Nancy Kemeny, INTRODUCTION Gary K. Schwartz, David Kelsen, Eileen O’Reilly, CPT2 is an alkaloid extracted from the plant Camptotheca David Ilson, John Coyle, acuminata. The cytotoxic potencies of CPT and CPT analogues are related to their ability to inhibit the catalytic activity of the Robert L. De Jager, Martin P. Ducharme, nuclear enzyme Topo I, which is involved in gene transcription Sarah Kleban, Ellen Hollywood, and and DNA replication (1). Topo I is an enzyme that relaxes Leonard B. Saltz supercoiled DNA during replication and transcription. Topo I Division of Gastrointestinal Oncology, Memorial Sloan-Kettering binds covalently to double-stranded DNA and forms a break in Cancer Center, New York, New York 10021 [S. S., N. K., G. K. S., one strand; this intermediate is known as the cleavable complex. D. K., E. O., D. I., S. K., E. H., L. B. S.]; MDS Pharma Services, The intact strand is passed through the gap in the broken strand, Montreal, Canada H4R 2N6 [M. P. D.]; and Daiichi Pharmaceutical which is then resealed, and the enzyme dissociates from the Corporation, Montvale, New Jersey 07645 [J. C., R. L. D. J.] helix. CPTs bind to the Topo I-DNA cleavable complex and prevent resealing of the DNA (1). ABSTRACT DX-8951f (Fig. 1) is a synthetic analogue of CPT and was synthesized to increase solubility and enhance antitumor effi- Exatecan mesylate (DX-8951f) is a topoisomerase I inhib- cacy. The anhydrous free base form of the drug is referred to as itor that has increased solubility and antitumor activity com- DX-8951. In vitro studies demonstrated more potent activity of pared with other topoisomerase I inhibitors. The purpose of DX-8951f than SN-38 and topotecan against various types of this study was to establish a safe dose of DX-8951f given as a human tumor cell lines, including breast, lung, gastric, and weekly 24-h infusion 3 of every 4 weeks. DX-8951f was admin- colon cancer (2–4). In the human tumor cloning assay, DX- istered as a 24-h continuous infusion in escalating doses. Twen- 8951f showed dose-dependent inhibition of clonogenic cells ty-seven patients were treated with 81 courses of the drug. from head and neck, liver, non-small cell lung, breast, colon, Dose-limiting toxicities included neutropenia, thrombocytope- ovary, and prostate tumors (5, 6). DX-8951f also has a broad nia, and inability to administer all three doses in the first cycle. spectrum of activity in human tumor xenografts of breast, lung, In minimally pretreated patients, a dose of 0.8 mg/m2 was gastric, pancreatic, esophageal, and colon carcinomas implanted tolerable. In patients who were heavily pretreated, a slightly in nude mice (3, 4, 7–9). These include CPT-11-resistant non- lower dose, 0.53 mg/m2, was tolerated without any severe tox- small cell lung and pancreatic tumors. icities. Nonhematological toxicities were mild and consisted of Much like CPT-11 and topotecan, the active lactone form mild diarrhea, asthenia, mild nausea, and constipation. Phar- of DX-8951f exists in a pH-dependent equilibrium with the macokinetic parameters could be well described with a one- respective hydroxy species (10). In pharmacology studies in compartment model in most patients, although the application dogs and mice, the t1/2 values for lactone and total DX-8951 ϳ of the one-compartment model probably resulted in an under- were similar. The calculated AUC for lactone is 50% of the estimated elimination half-life. In conclusion, the recom- total drug AUC, and AUC is linearly related to dose. In metabolism studies carried out in vitro with human liver mended Phase II dose for DX-8951f administered as a weekly microsomes, major metabolites of DX-8951f were the 4- 24-h infusion on a 3-of-4 week schedule is 0.8 mg/m2 in mini- hydroxy-methyl form of DX-8951 (UM-1) and the 4-hydroxy- mally pretreated patients and 0.53 mg/m2 in patients who are lated form (UM-2; Ref. 10). There was good correlation be- heavily pretreated. tween the amounts of UM-1, UM-2, and UM-3 produced by the microsomes and the hydroxylating activity at the 6-␤ position of testosterone, which is an index of CYP3A activity. Production of UM-2 was also inhibited by the CYP1A-specific inhibitor ␣-napthoflavone. In addition, production of the metabolites was strongly inhibited by CYP3A-specific inhibitors such as keto- Received 2/14/01; revised 8/10/01; accepted 8/16/01. conazole, erythromycin, and troleandomycin. These results in- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 To whom requests for reprints should be addressed, at Division of Solid Tumor Oncology, Department of Medicine, Memorial Sloan- 2 The abbreviations used are: CPT, camptothecin; Topo I, topoisomerase Kettering Cancer Center, 1275 York Avenue, New York, NY 10021. I; AUC, area under the curve; ANC, absolute neutrophil count; MTD, Phone: (212) 639-8702; Fax: (212) 717-3320; E-mail: sharma1@ maximum tolerated dose; DLT, dose-limiting toxicity; HP, heavily mskcc.org. pretreated; MP, minimally pretreated; PK, pharmacokinetic.

Fig. 1 Chemical structure of DX-8951f (molec- ⅐ ⅐ ular formula, C24H22FN3O4 CH4O3S 2H2O. Mr 567.59).

dicate that DX-8951f is metabolized mainly by CYP3A en- experienced toxicity at a dose level, only one patient was treated zymes (10). at the next dose level. When moderate or worse toxicity was In toxicology studies in mice, rats, and dogs, bone marrow experienced at a dose level, at least three patients were treated suppression, gastrointestinal toxicity, and mutagenicity were at that dose level. If one patient of a group of three experienced principally observed. Hematological toxicity was dose limiting a DLT, the group size was increased to six patients. If two of a in all species, and the toxic dose low values in the most sensitive maximum of six patients developed DLT at the same dose, that species (dog) were 10 mg/m2 (0.5 mg/kg) for the single dose, dose was considered to be above the MTD. MTDs were evalu- 2 0.3 mg/m /day (0.015 mg/kg/day) for five daily doses, and 0.5 ated separately in HP and MP patients, according to pretreat- 2 mg/m (0.025 mg/kg) for the single 24-h continuous infusion ment status. HP patients were defined as those who had received schedule (10). In other in vivo studies, a cyclical dosing pattern more than six courses of alkylating agent-containing chemother- at lower doses of DX-8951f improved antitumor activity over apy (or more than four courses of carboplatin), radiation therapy single-dose administration. On the basis of these preclinical data, to Ͼ25% of hematopoietic reserves, or two or more courses of several schedules were chosen for intermittent administration. mitomycin C or a nitrosourea. In the MP group, a minimum of These included daily for 5 days every 3 weeks, 24-h infusions six patients were to be treated at the MTD. The MTD was weekly, and once a week for 3 of every 4 weeks (9, 11). defined as one dose level below the dose that induced DLTs in The clinical development of DX-8951f was based on its two of six patients. DLT was defined as grade 4 neutropenia that significantly higher potency in vitro and its relatively favorable was prolonged (Ͼ5 days) or accompanied by fever, grade 4 toxicity profile in preclinical studies. The present Phase I and thrombocytopenia, any grade 3 nonhematological toxicity (ex- pharmacological study had the following objectives: (a) char- cept vomiting, in which case the criterion was grade 4 with acterize the toxicities of DX-8951f administered as three weekly maximal support), inability to administer three weekly doses in 24-h infusions every 4 weeks; (b) determine the MTD and cycle 1, or Ͼ1 week delay in starting cycle 2. All toxicity recommended Phase II dose; (c) assess the pharmacological grading was performed according to the National Cancer Insti- behavior of DX-8951f on this schedule; and (d) seek prelimi- tute’s Common Toxicity Criteria. nary evidence of antitumor activity. DX-8951f was supplied by Daiichi Pharmaceuticals (Montvale, NJ) as 2 or 5 mg of anhydrous free base equivalent/ PATIENTS AND METHODS vial and was diluted in the vial with 0.9% NaCl USP to obtain Patient Selection. Patients with histologically confirmed a stock 0.5 mg/ml solution. The appropriate volume of stock solid tumors that had failed standard therapies were included in solution to yield the required dose was diluted in a polyvinyl the study. All patients were Ն18 years of age, signed a written chloride infusion bag with sterile 0.9% NaCl to a total volume informed consent, and had an Eastern Cooperative Oncology of 96 ml. This was administered by i.v. infusion over a 24-h Group performance status of 0–2. Other inclusion criteria in- period through a central venous catheter with a programmed cluded measurable or clinically evaluable disease, adequate peristaltic pump system. All treatment was done on an outpa- hematological (ANC Ն1.5 ϫ 103/mm3; platelet count Ն100 ϫ tient basis. After reconstitution with normal saline, DX-8951f is 103/mm3; hemoglobin Ն8.5 g/dl), renal (serum creatinine Յ2.0 stable for at least 24 h under ambient conditions of light and mg/dl), and hepatic function (serum bilirubin Յ1.5 mg/dl; as- temperature. partate aminotransferase Յ2.5 ϫ upper limit of normal or Pretreatment Evaluation and Follow-Up Studies. Յ5.0 ϫ upper limit of normal if patient had liver metastases). Baseline evaluation was carried out within 10 days prior to The study was approved by the Institutional Review Board of treatment initiation and included informed consent, a detailed Memorial Sloan-Kettering Cancer Center. medical history, comprehensive physical examination, 12-lead Drug Administration. The starting dose of DX-8951f, electrocardiogram, chest X-ray, complete blood count, differen- calculated as the human equivalent of one-third of the dog toxic tial count, serum chemistries, electrolytes, prothrombin time, dose low for a single 24-h infusion divided by 3, was 0.05 pregnancy test, and urinalysis. Patients were seen by a physician mg/m2/24 h. Dose levels explored were 0.05, 0.1, 0.15, 0.23, before each treatment course. Tumor measurements/serum tu- 0.35, 0.53, 0.8, 1, and 1.2 mg/m2/24 h. This escalation scheme mor markers were performed using appropriate tests at baseline is consistent with a modification of the Fibonacci sequence. One and after each course of therapy. patient was entered at the lowest dose level. When no patients A complete response was defined as the disappearance of

Table 1 Patient characteristics Table 2 Dose levels Characteristic Number Dose No. of No. of 2 patients courses No. of patients (MP/HP) 27 (22/5) (mg/m ) Sex (M/F) 15/12 0.05 1 1 Age (yr) 0.1 1 1 Median 64 0.15 1 9 Range 46–75 0.23 1 2 Serum albumin (g/dl) 0.35 1 4 Mean 3.8 0.53 5 22 Range 2.9–4.8 0.8 8 29 ECOG PSa 1.0 4 6 081.2 5 7 118 21 Previous treatment Chemotherapy 27 Radiation 11 by minimizing the Akaike information criterion test and the Tumor type residual variability. No difference in the quality of fitting was Colorectal 15 observed among all three models except for the Akaike infor- Esophagus 5 mation criterion test, which was lower for the one compartment. Pancreas 4 Liver 2 It was therefore the most appropriate one to use, and the pa- Gastric 1 rameters defined by this model were: volume of distribution 2 2 a ECOG PS, Eastern Cooperative Oncology Group performance (Vss; liters/m ) and clearance (CL; liters/h/m ). On average, 18 status. observations per patient were fitted simultaneously by the model. Individual PK parameter estimates were first derived with ADAPT-II (13) using maximum likelihood analysis. These estimates were then used as prior values for the population PK all disease, documented by measurements separated by at least analysis, which was performed using an iterative two-stage 4 weeks, and a partial response was defined as Ն50% reduction methodology (IT2S; Ref. 12). All concentrations were modeled ϭ 2 in the sum of bidimensional products of all measurable lesions using a weighting procedure of Wj 1/Sj , where the variance 2 2 ϭ documented by at least two measurements separated by at least Sj was calculated for each observation using the equation Sj ϩ 2 4 weeks. (a b*Y) , where a and b are the intercept and slope of each PK Sampling and Assay. Blood samples were collected variance model. The slope is the residual variability associated before the start of the first infusion (time 0) and at 0.25, 0.5, with each concentration, which includes the intraindividual var- 0.75, 1, 2, 6, 24, 24.25, 24.50, 24.75, 25, 26, 28, and 30 h after iability and the sum of all experimental errors, and the intercept the beginning of the first infusion. The complete urine output is related to the limit of detection of the analytical assay. Two was collected from time 0 to 24 h and from 24 to 48 h after the variance models were used, one for the urinary observations and beginning of the first infusion. Plasma and urine samples were the other for the plasma concentrations of DX-8951. Variance also collected at the same time before and after the beginning of parameter estimates were derived using maximum likelihood the third weekly infusion of DX-8951f in 17 patients. In the analysis (ADAPT II). These estimates were used as beginning remaining 10 patients, plasma and urine samples were collected priors and were updated iteratively during the population PK only after the first dose. analysis (IT2S) until stable values were found. Data from five Blood was collected in a heparinized tube and was centri- patients were not included in the population analysis because fuged at 3000 rpm for 15 min to separate the plasma. Plasma their concentration-time data on day 1 were not consistent with was stored at Ϫ20°C while awaiting dispatch. Urine samples those on day 15 (i.e., the values were judged to be anomalous). were collected in sterile wide-mouthed plastic bottles. After Concentration-time data for these patients were therefore fitted each bottle containing the urine was shaken, a measured quan- by allowing a difference between the day 1 and day 15 admin- tity of ϳ50 ml was drawn off and frozen at Ϫ20°C in a suitably istered doses. PK parameters were calculated in these patients labeled sample tube. with a Bayesian algorithm (MAP-B, ADAPT II) using the Total concentrations of DX-8951 (lactone plus hydroxy- results of the IT2S population analysis from the other patients. acid forms) were determined by high-performance liquid chro- matography. The plasma and urine concentrations were used to RESULTS determine the PK parameters by standard noncompartmental Twenty-seven patients were treated with 81 courses of and compartmental analysis (12). Three different compartmental DX-8951f. Patient characteristics are listed in Table 1. Eleven of PK models were compared in terms of their ability to describe 27 patients had previously received both chemotherapy and simultaneously the observed plasma concentrations and the ex- radiation, and all patients had been previously treated with creted urinary amounts of DX-8951. Discrimination between chemotherapy. The dose escalation scheme and number of pa- these candidate compartmental PK models was performed by tients treated at each dose level are listed in Table 2. The dose looking at pertinent graphics (e.g., fitted and observed concen- was increased stepwise for each dose escalation cohort, first by trations versus time, weighted residuals versus observed values), doubling the starting dose to dose level 2, and then by increasing by maximizing the coefficient of determination values (R2), and the dose by increments of 50% until hematological or nonhe-

Table 3 Hematological toxicity (cycle 1) Dose level No. of WBC nadir ANC nadir Plt.a nadir Hgb nadir (mg/m2) patients (ϫ103/␮l) (ϫ103/␮l) (ϫ103/␮l) (g/dl) 0.05 1 6.4 4.4 202 9.9 0.1 1 8.8 4.7 332 12.4 0.15 1 7.5 5.1 167 12.2 0.23 1 6.6 5.2 253 10.9 0.35 1 2.3 1.4 43 9.8 0.53 5 5.3b (2.2–12.4) 3.3b (1.1–9.9) 215b (147–466) 11.8b (9.1–12.1) 0.8 8 3.35b (1.3–5.6) 2.05b (0.5–3.7) 132b (11–217) 9.3b (7.2–11.7) 1.0 4 2.05b (0.9–3.1) 0.95b (0.3–1.8) 164b (91–226) 8.05b (7.1–10.2) 1.2 5 1.7b (0.5–3.1) 0.6b (0.1–0.9) 99b (18–133) 9.1b (6.8–11.2) a Plt, platelet; Hgb, hemoglobin. b Median count.

Table 4 DLTs Grade 4 Grade 4 Dose level (mg/m2) No. of patients ANC Ͼ 5 days Pltsa Ͻ3 doses in cycle 1 Other 0.8 1/6 (MP) 1 0 0 0 2/2 (HP ) 0 0 1 1b (Grade 3 ANC Ͼ 5 days) 1.0 2/4 (MP) 0 0 1 1c (Grade 3 ANC Ͼ 5 days) 1.2 4/5 (MP) 2 0 2 0 (Grade 3 ANC Ͼ 5 days) a Plts, platelets. b One patient dose reduced after course 1 following grade 3 neutropenia and grade 3 thrombocytopenia lasting 6 and 4 days, respectively. c One patient dose reduced during course 1 following grade 2 neutropenia at study day 7. Second dose administration was delayed for 1 week for recovery of neutrophil count, extending the 4-week cycle to 5 weeks.

matological toxicity Նgrade 2 (moderate) was observed in at least one patient with the exception of certain grade 2 laboratory Table 5 Nonhematological toxicities tests already elevated to grade 1 at baseline (e.g., hemoglobin No. of courses and bilirubin). Thereafter, the dose was increased by increments Toxicity (Total courses ϭ 81) of 33%. Diarrhea 7 (9%) Hematological Toxicities. The most serious toxicities Grade 1 6 for DX-8951f were hematological (Table 3). The median plate- Grade 2 1 let and neutrophil nadir decreased with increasing doses (Table Constipation 20 (25%) 3), and four patients demonstrated dose-limiting hematological Grade 1 15 2 Grade 2 2 toxicity at a dose level of 1.2 mg/m (Table 4). None of the Grade 3 3 patients who experienced dose-limiting neutropenia had to be Nausea 20 (25%) hospitalized for sepsis. No platelet transfusions were required in Grade 1 14 patients experiencing dose-limiting thrombocytopenia. DLTs Grade 2 5 Grade 3 1 were seen at various dose levels for HP and MP patients. At Vomiting 11 (14%) 2 2 doses of 1.2 mg/m and 1.0 mg/m , four of five and two of four Grade 1 9 MP patients enrolled at each respective dose level experienced Grade 2 2 DLT. One of six MP patients experienced DLT at a lower dose Asthenia 22 (27%) of 0.8 mg/m2. Both of the HP patients receiving 0.8 mg/m2 Grade 1 8 2 Grade 2 13 experienced DLT. A slightly lower dose of 0.53 mg/m was Grade 3 1 tolerated without any severe toxicities or DLT in the three Neurosensory symptoms 21 (26%) patients treated at that dose level. Interpatient variability was Grade 1 21 observed at the various dose levels, but toxicity increased in grade at higher doses. Six patients required dose reductions because of severe dose-limiting hematological toxicity (two patients at 0.8 mg/m2 dose level (one HP, one MP), two patients Nonhematological Toxicities. DX-8951f was well toler- at 1.0 mg/m2 dose level (two MP), and two patients at 1.2 ated with only mild nonhematological toxicity (Table 5). Unlike mg/m2 dose level (two MP). Minor hematological toxicities irinotecan, diarrhea was observed in only 9% of courses, and it included grade 3 anemia (11%), grade 2 neutropenia (4%), and was mild. No patients required hospitalization for diarrhea or grade 3 neutropenia (7%). dehydration. Other toxicities included constipation (25%) and

Table 6 PK analysis of DX-8951f Data from five patients were not included in the population analysis because their concentration-time data on day 1 were not consistent with those on day 15 (i.e., the values were judged to be anomalous). Mean (CV, %) Median (Min–Max) Population Individual compartmental compartmental Parameter analysis (n ϭ 22) analysis (n ϭ 5) CL (liters/h/m2) 1.73 (59.3) 1.63 (57.1) 1.62 (0.24–4.96) 0.98 (0.95–2.92) FEa (%) 10.66 (63.8) 13.48 (61.0) 8.94 (0.87–30.29) 9.73 (4.86–22.14) 2 Vss (liters/m ) 12.72 (23.3) 13.74 (26.2) 12.84 (7.54–21.59) 13.29 (9.05–18.95)

t1/2 (h) 7.13 (75.5) 6.91 (42.4) 5.28 (2.89–25.68) 6.38 (3.94–10.84) a FE, fraction eliminated. Fig. 2 Relationship between the estimated clearance values (CL) fitted by compartmental analysis and the different administered doses (␮g/m2) of DX-8951f. nausea (25%) leading to grade 1 or 2 vomiting in a minority of courses (14%). Although asthenia was difficult to ascribe to DX-8951f, it was a frequent clinical toxicity (27%). It was almost always mild (grade 1 or 2), with only one patient experiencing grade 3 asthenia. Neurosensory symptoms were reported in 26% of courses and consisted mostly of grade 1 numbness of upper and lower extremities. Antitumor Activity. No objective responses were observed in this trial. Five patients exhibited stable disease. Three patients had a diagnosis of colorectal cancer and had stable disease for an average of six cycles. Interestingly, all of these patients had already received and had progressed on prior therapy with irinotecan. One patient with pancreatic cancer and another with hepatocellular carcinoma had stable disease for 9 and 22 cycles of therapy, respectively. PK Studies. Total plasma pharmacokinetics were ob- tained in all 27 patients and were analyzed using both noncompartmental and compartmental analyses. Compartmental PK results were more robust than noncompartmental analyses be- Fig. 3 Relationship between the estimated values of the volume of distribution (V ) fitted by compartmental analysis and the different cause 30 h of sampling in the plasma and 48 h of urinary data ss administered doses (␮g/m2) of DX-8951f. were simultaneously modeled. The results of compartmental analyses are presented in Table 6. The mean plasma elimination half-life was 7.13 h. Total-body clearance did not vary with dose (mean, 1.73 liters/h/m2), indicating linear pharmacokinetics within the administered dose range (Fig. 2). The percentage of labeling, nonlinear PK processes in these patients, concomitant the administered dose eliminated unchanged in the urine (FE) medications, and liver or renal dysfunction. It was found that the observed differences in the c were not consistent among was 10.66%. The Vss was modest and was independent of the max dose with a mean value of 12.7 liters/m2 (Fig. 3). Fig. 4 patients: sometimes it increased markedly on day 15 or de- demonstrates that 40% of the variability in plasma clearance of creased markedly. In the remaining 22 patients, the cmax be- the drug is explained by variation in the cmax. This is expected tween the two dosing days were consistent and indicative of because of the long duration of the infusion. PK parameters linear PK processes. The most likely hypothesis for the observed could be well described using a one-compartment model as differences in the five patients was that it was indicative of a demonstrated by the quality of fit in a representative patient “noisy” data set probably because of inaccuracies in the (Fig. 5). Five patients were excluded from the population PK amounts and timing of the doses administered. The PK data for analysis and were analyzed separately because of anomalies these patients were therefore analyzed separately, using a between the day 1 and day 15 data in these patients. Specifi- Bayesian control algorithm as described in the “Patients and cally, the cmax on day 15 was either much higher or much lower Methods” section. This allowed us to calculate the PK param- than that of day 1. Various reasons were retrospectively exam- eters for these patients and report them without biasing mark- ined, including time of sample collection, adequacy of sample edly the results of the population analysis itself.

for MP patients, whereas neutropenia and thrombocytopenia were dose limiting in HP patients (23–27). Neutropenia is also the principal toxicity described with other Topo I inhibitors (irinotecan and topotecan). The incidence of grade 3 or 4 neutropenia with irinotecan and topotecan is 12–20% and 30–80% of treatment cycles (22). In other trials of DX-8951f, two cases of acute pancreatitis not predicted by preclinical toxicology were also observed (28). The PK parameters for most patients indicated high inter- and intrapatient variability, which is not unlike that observed with irinotecan (29–31). The reasons for this variation are not well understood at this time and may be related to pharmaco- genetic variations in DX-8951f metabolism. For most patients, the pharmacokinetics of the drug appeared linear as evidenced by a lack of relationship between the administered dose and

either clearance or Vss (Figs. 2 and 3). The PK analysis was well Fig. 4 Relationship between the estimated clearance values (CL) fitted described using a one-compartment model, although the appli- by compartmental analysis and the different observed c values max cation of this model probably resulted in an underestimated (␮g/L) of DX-8951 after a 24-h i.v. infusion (day 1). elimination half-life. This is not attributable to an inadequacy of the model in explaining the data, but simply to the fact that only 30 h of sampling were modeled in each subject, preventing a DISCUSSION robust characterization of the true elimination half-life. This is in contrast to rich-sampling studies performed with this drug, Topo I is an important nuclear enzyme that is required for where two- or three-compartment PK models were necessary to relaxation of torsionally strained duplex DNA (14). Several correctly characterize plasma concentrations and excreted uri- studies have indicated that tumor cells may have a higher nary amounts (28). In the present study, the one-compartment amount of Topo I compared with surrounding normal tissues. These include studies with colon, ovarian, and esophageal can- model explained the data just as well as a two- or three- cers (15–18) and in non-Hodgkin’s and other lymphomas (19). compartment PK model. This is probably attributable to mask- Topo I is a target for CPT (20), which appears to stabilize ing of the distribution phase by the long duration of the infusion adducts containing the Topo I bound to the cleavable complex and to underestimation of the “true” terminal elimination phase (21). In recent years, several analogues of CPT, including iri- because of the limited number of plasma samples after the end notecan (CPT-11) and topotecan, have been approved for human of infusion. Nevertheless, the results of this PK analysis are very use against a variety of tumors in the United States. similar to the ones that have been reported in studies using more DX-8951f is a synthetic CPT with increased solubility and extensive sampling and a two- or three-compartment PK model. 2 antitumor efficacy in vitro. In preclinical studies, the principal The CL was 1.73 liters/h/m versus the previously reported 2 2 DLT of this drug was myelosuppression on all tested schedules. value of 1.63 liters/h/m , the Vss was 12.72 liters/m versus 2 We performed a Phase I trial in humans with advanced cancer to 17.65 liters/m , and the terminal elimination half-life was 7.13 h determine the MTD of DX-8951f administered as a weekly 24-h versus 12.3 h (27). Some patients exhibited a PK profile that infusion 3 of every 4 weeks. appeared to be different between the first and third dose admin- DX-8951f was generally well tolerated. The DLT of neu- istered; these patients were analyzed separately using a tropenia, thrombocytopenia, or inadequate doses (less than Bayesian algorithm. As mentioned before, these differences are three) in cycle 1 was seen at 1.0 mg/m2 in MP patients and at 0.8 probably the result of noisy data rather than a true inconsistency mg/m2 in HP patients. The MTD was defined as 0.8 mg/m2 for in the model. This is because although nonlinearity can exist MP and 0.53 mg/m2 for HP patients. Neutropenia and throm- with drugs in clearance processes, it is very rare to observe bocytopenia were dose related, although no patients required nonlinearity in the volumes of distribution. hospitalizations for neutropenic fevers. In alternative schedules, DX-8951f has been given daily for Nonhematological toxicity was mild with diarrhea ob- 5 days, weekly for 3 of every 4 weeks, every 3 weeks, and as a served in 9% of courses. No patients required hospitalization for continuous infusion over 5–21 days (23–27). The daily for 5 diarrhea or dehydration. This is in contrast to treatment with days every 3 weeks schedule has been chosen for Phase II irinotecan, where overall incidence of delayed diarrhea was 87% studies. This decision was based on preclinical models showing (in ϳ60% of the cycles administered; Ref. 22). Other nonhe- higher antitumor activity with cyclical dosing at lower doses matological toxicities included constipation (19%) and nausea compared with single-dose administration, antitumor activity (25%) leading to grade 1 or 2 vomiting in a minority of courses observed with the daily for 5 days and weekly (30-min infusion) (14%). Although asthenia was difficult to ascribe to DX-8951f, regimens, and less nausea and vomiting observed with the daily it was a frequent clinical toxicity (27%). for 5 days regimen than with the weekly (30-min infusion) and Neutropenia was the DLT of DX-8951f on all of the single-dose (30-min infusion) schedules. There are no immedi- alternative schedules (daily for 5 days, weekly for 3 of every 4 ate plans to develop the present schedule for Phase II studies weeks, every 3 weeks, and continuous infusion over 5–21 days) because of the inconvenience of administration of DX-8951f as

Fig. 5 Fitted (solid line) and observed (F) plasma concentrations (top) and excreted urinary amounts (bottom) of DX-8951 after a 24-h i.v. infusion in a representative patient.

a 24-h infusion and lack of superiority to the daily for 5 days tured in vitro and xenografted into nude mice. Jpn. J. Cancer Res., 88: schedule. 760–769, 1997. In conclusion, the recommended Phase II doses of DX- 10. Daiichi Pharmaceuticals. DX-8951f clinical investigator’s bro- 8951f given as weekly 24-h infusions 3 of every 4 weeks is 0.8 chure. Fort Lee, NJ: Daiichi, 1998. mg/m2 for minimally and 0.53 mg/m2 for heavily pretreated 11. Kumazawa, E., Jimbo, T., Ochi, Y., and Tohgo, A. Potent and broad antitumor effects of DX-8951f, a water-soluble camptothecin-deriva- patients. tive, against various human tumors xenografted into nude mice. Cancer Chemother. Pharmacol., 42: 210–220, 1998. REFERENCES 12. Gibaldi, M., and Perrier, D. Pharmacokinetics, Ed. 2, New York: 1. Slichenmyer, W. J., Rowinsky, E. K., Donehower, R. C., and Kauf- Marcel Dekker Inc., 1982. mann, S. H. The current status of camptothecin analogues as antitumor 13. D’Argenio, D. Z., and Schumitzky, A. A program package for agents. J. Natl. Cancer Inst., 85: 271–291, 1993. simulation and parameter estimation in pharmacokinetic systems. Com- 2. Ikeda, K., Terashima, M., Yegashi, Y., Kuboi, M., Sasaki, N., put. Programs Biomed., 9: 115–134, 1979. Kawamura, H., Takiyama, I., Maesawa, C., Takagane, A., Abe, K., et al. 14. Gellert, M. DNA topoisomerases. Annu. Rev. Biochem., 50: 879– Antitumor activities of camptothecin analogues against human esopha- 910, 1981. geal cancer. Proc. Am. Assoc. Cancer Res., 36: 453, 1995. 15. Giovanella, B. C., Stehlin, J. S., Wall, M. E., Wani, M. C., Nicholas, 3. Mitsui, I., Ohsuki, S., Hirota, Y., Sugimori, M., Terasawa, H., and A. W., Liu, L. F., Silber, R., and Potmesil, M. DNA topoisomerase Sato, K. High potency of a camptothecin derivative DX-8951 might be I-targeted chemotherapy of human colon cancer in xenografts. Science attributed to good membrane permeability. Proc. Am. Assoc. Cancer (Wash. DC), 246: 1046–1048, 1989. Res., 35: 455, 1994. 16. Kim, R., Ohi, Y., Inoue, H., and Toge, T. Expression and relation- 4. Mitsui, I., Kumazawa, E., Hirota, Y., Aonuma, M., Sugimori, M., ship between topoisomerase I and II ␣ genes in tumor and normal tissues Ohsuki, S., Uoto, K., Ejima, A., Terasawa, H., and Sato, K. A new in esophageal, gastric and colon cancers. Anticancer Res, 19: 5393– water-soluble camptothecin derivative, DX-8951f, exhibits potent anti- 5398, 1999. tumor activity against human tumors in vitro and in vivo. Jpn. J. Cancer 17. McLeod, H. L., Douglas, F., Oates, M., Symonds, R. P., Prakash, Res., 86: 776–782, 1995. D., Van Der Zee, A. G., Kaye, S. B., Brown, R., and Keith, W. N. 5. Eckhardt, S. G., Degen, D., Tohgo, A., and Von Hoff, D. Activity of Topoisomerase I and II activity in human breast, cervix, lung and colon DX-8951f, a water soluble camptothecin derivative, on primary human cancer. Int. J. Cancer, 59: 607–611, 1994. tumor colony-forming units. Proc. 9th NCI-EORTC Symp., A447, 1996. 18. Van Der Zee, A. G., Hollema, H., de Jong, S., Boonstra, H., Gouw, 6. Weitman, S., Degen, M., and Von Hoff, D. Antitumor activity of A., Willemse, P. H., Zijlstra, J. G., and De Vries, E. G. P-Glycoprotein topoisomerase I inhibitors against a pediatric solid tumor panel. Proc. expression and DNA topoisomerase I and II activity in benign tumors Am. Assoc. Cancer Res., 37: 435, 1996. of the ovary and in malignant tumors of the ovary, before and 7. Kumazawa, E., Ochi, Y., Nakayama, Y., Iwahana, M., Aonuma, M., after platinum/cyclophosphamide chemotherapy. Cancer Res., 51: Tanaka, N., and Tohgo, A. Antitumor effect of DX-8951f, a new 5915–5920, 1991. camptothecin derivative in various murine models. Proc. Am. Assoc. 19. Gosky, D. M., Belfi, C. A., Dordal, M. S., Berger, N. A., and Cancer Res., 36: 440, 1995. Berger, S. J. Topoisomerase I and II levels in human lymphoma cell 8. Takiyama, I., Terashima, M., Ikeda, K., Kawamura, H., Sasaki, N., lines SU-4 and SU-4R. Proc. Am. Assoc. Cancer Res., 38: 19, 1997. Hayakawa, Y., Ishida, K., and Saito, K. Remarkable synergistic inter- 20. Liu, L. F. DNA topoisomerase poisons as antitumor drugs. Annu. action between camptothecin analogs and cisplatin against human Rev. Biochem., 58: 351–375, 1989. esophageal cancer cell lines. Proc. Am. Assoc. Cancer Res., 38: 15, 21. Hsiang, Y. H., Hertzberg, R., Hecht, S., and Liu, L. F. Camptoth- 1997. ecin induces protein-linked DNA breaks via mammalian DNA topoi- 9. Takiguchi, S., Kumazawa, E., Shimazoe, T., Tohgo, A., and Kono, somerase I. J. Biol. Chem., 260: 14873–14878, 1985. A. Antitumor effect of DX-8951, a novel camptothecin analog, on 22. Rothenberg, M. L. Topoisomerase I inhibitors: review and update. human pancreatic tumor cells and their CPT-11 resistant variants cul- Ann. Oncol., 8: 837–855, 1997.

23. Bates, N. P., Boven, E., Dobbs, N., Varcoe, S., Ruijter, R., administered by 24-hour continuous infusion to patients with advanced O’Grady, J., Pinedo, H. M., and Talbot, D. C. Phase I and pharmaco- solid tumors. Proc. Am. Soc. Clin. Oncol., 18: A682, 1999. kinetic study of DX-8951f, a novel topoisomerase inhibitor. Proc. Am. 28. De Jager, R., Cheverton, P., Tamanoi, K., Coyle, J., Ducharme, M., Soc. Clin. Oncol., 18: A684, 1999. Sakamoto, N., Satomi, M., and Suzuki, M. DX-8951f: summary of 24. Geyer, C., Hammond, L. A., Johnson, T., Smetzer, L., Coyle, J., Phase I clinical trials. Ann. NY Acad. Sci., 922: 260–273, 2000. Drengler, R., Von Hoff, D., De Jager, R., and Rowinsky, E. Dose- 29. Chabot, G. G., Abigerges, D., Catimel, G., Culine, S., de Forni, M., schedule optimization of the hexacyclic camptothecin (CPT) analog DX-8951f: a Phase I and pharmacokinetic study with escalation of both Extra, J. M., Mahjoubi, M., Herait, P., Armand, J. P., and Bugat, R. treatment duration and dose. Proc. Am. Soc. Clin. Oncol., 18: A813, Population pharmacokinetics and pharmacodynamics of irinotecan 1999. (CPT-11) and active metabolite SN-38 during Phase I trials. Ann. Oncol, 6: 141–151, 1995. 25. Kamiya, Y., Yamamoto, N., Tamura, T., Shimada, Y., and Oguma, T. Phase I and pharmacokinetic (PK) study of DX-8951f, a novel 30. Rothenberg, M. L., Kuhn, J. G., Burris, H. A. III, Nelson, J., camptothecin analog, given as 30 minute infusion daily for 5 days. Proc. Eckardt, J. R., Tristan-Morales, M., Hilsenbeck, S. G., Weiss, G. R., Am. Soc. Clin. Oncol., 18: A832, 1999. Smith, L. S., and Rodriguez, G. I. Phase I and pharmacokinetic trial of 26. Minami, H., Sasaki, Y., Onozawa, Y., Fujii, H., Ohtsu, T., Igarashi, weekly CPT-11. J. Clin. Oncol., 11: 2194–2204, 1993. T., Itoh, K., Tamanoi, K., and Oguma, T. Phase I and clinical pharma- 31. Rothenberg, M. L., Eckardt, J. R., Kuhn, J. G., Burris, H. A. III, cology of DX-8951f, a new camptothecin derivative, infused over 30 Nelson, J., Hilsenbeck, S. G., Rodriguez, G. I., Thurman, A. M., Smith, minutes every 3 weeks. Proc. Am. Soc. Clin. Oncol., 18: A685, 1999. L. S., Eckhardt, S. G., Weiss, G. R., Elfring, G. L., Rinaldi, D. A., 27. Royce, M., Lassere, Y., Hoff, P., Brito, R., Ravandi, F., Medgysey, Schaaf, L. J., and Von Hoff, D. D. Phase II trial of irinotecan in patients D., Zukowski, T., Coyle, J., De Jager, R., and Pazdur, R. DX-8951f: with progressive or rapidly recurrent colorectal cancer. J. Clin. Oncol., Phase I and pharmacokinetic study of a novel camptothecin analogue 14: 1128–1135, 1996.

Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2001 American Association for Cancer Research. Phase I Study of Topoisomerase I Inhibitor Exatecan Mesylate (DX-8951f) Given as Weekly 24-Hour Infusions Three of Every Four Weeks

Sunil Sharma, Nancy Kemeny, Gary K. Schwartz, et al.

Clin Cancer Res 2001;7:3963-3970.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/7/12/3963

Cited articles This article cites 28 articles, 5 of which you can access for free at: http://clincancerres.aacrjournals.org/content/7/12/3963.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/7/12/3963.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/7/12/3963. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.