P h a s e I S t u d y o f a n O r a l F o r m u l a t i o n o f Z D 9 3 3 1 A d m i n i s t e r e d
`D a i l y f o r 2 8 D a y s
`
`By Michael B. Sawyer, Mark J. Ratain, Donna Bertucci, Robert P. Smith, Richard L. Schilsky, Nicholas J. Vogelzang, Keith Shulman,
`Edwin C. Douglass, and Gini F. Fleming
`
`Purpose: To define the maximum-tolerated dose and
`dose-limiting toxicities (DLTs) of an oral formulation of
`ZD9331, a novel thymidylate synthase inhibitor that is not a
`substrate for folylpolyglutamate synthase.
`Patients and Methods: Patients had Cancer and Leukemia
`Group B performance status < 2 and refractory solid tumors.
`Initially, patients received ZD9331 daily for 2 weeks, with the
`duration of treatment escalated to a maximum of 4 weeks,
`followed by a 2-week rest period. Once the maximum-toler-
`ated duration of treatment was determined, the dose of
`ZD9331 was increased until DLT occurred.
`Results: Fifty-five patients were enrolled at eight dose
`levels. The DLTs were thrombocytopenia and neutropenia. At
`3 mg/d, two of 19 patients developed DLT; one patient had
`
`grade 3 thrombocytopenia and grade 4 neutropenia, and the
`other patient had grade 3 thrombocytopenia only. Anemia
`was common, with a median hemoglobin nadir of 75% of
`baseline, before recovery or transfusion. The apparent oral
`clearance of ZD9931 was 11.6 ⴞ 6.3 mL/min. Dose-limiting
`myelosuppression was associated with both an increased
`24-hour ZD9931 concentration and blood urea nitrogen.
`Conclusion: The recommended phase II dose on this
`schedule is 3 mg/d for 4 weeks, followed by a 2-week rest
`period. ZD9331 seems to have a manageable toxicity pro-
`file, although it should be used with caution in patients with
`renal impairment.
`J Clin Oncol 21:1859-1865. © 2003 by American
`Society of Clinical Oncology.
`
`T HYMIDYLATE SYNTHASE catalyzes conversion of de-
`
`oxyuridine monophosphate to thymidylate and is the only
`de novo source of thymidylate for DNA synthesis.1 The central
`role of thymidylate synthase in nucleotide synthesis has led to
`the development of several folate analogs that inhibit thymidy-
`late synthase. CB3717, the first antifolate to specifically target
`thymidylate synthase, was abandoned because of its unpredictable
`nephrotoxicity,2 which was thought to be due to its low water
`solubility.3 Because CB3717 had significant antitumor activity, the
`Cancer Research Campaign Center for Cancer Therapeutics devel-
`oped antifolates with improved water solubility.
`This research led to the development of raltitrexed (Tomudex;
`ZD1694; AstraZeneca, Macclesfield, UK), which had improved
`water solubility and decreased nephrotoxicity in animal models.3
`Subsequent clinical studies demonstrated activity of ZD1694 in
`colorectal cancer and led to its approval in Australia, Canada,
`and Europe for treatment of metastatic colorectal cancer.4 In a
`later study in the United States, patients with metastatic colo-
`rectal cancer were randomly assigned to either fluorouracil and
`leucovorin or ZD1694; a slight survival advantage was found for
`fluorouracil and leucovorin.5 On the basis of this study, ZD1694
`was not developed further in the United States. ZD1694 required
`polyglutamation by folylpolyglutamate synthase to effectively
`inhibit thymidylate synthase.6 In vitro studies have shown that
`resistance to antifolates can occur by cancer cells decreasing
`accumulation of polyglutamated antifolates.7,8 In addition, vari-
`ability in polyglutamation may be a factor in the interpatient
`variability of antifolate pharmacokinetics.
`ZD9331 was developed to overcome decreased polyglutama-
`tion as a mechanism of resistance to antifolates and have more
`predictable pharmacokinetics and less interpatient variability
`than ZD1694. Preclinical studies demonstrated that ZD9331 was
`
`active against lymphoid and leukemia cell lines and small-cell
`lung, gastric, and colorectal cancer xenografts.9 ZD9331 was
`more active when given by a protracted schedule than an
`intermittent schedule. In vitro studies using L5178Y TK⫺/⫺
`mouse lymphoma cell lines demonstrated that exposure to 10
`␮mol/L of ZD9331 for 4 hours inhibited colony formation by
`80%, whereas exposure to 0.1 ␮mol/L of ZD9331 for 24 hours
`inhibited colony formation by 99.96%. In vivo studies also
`showed that ZD9331 had superior activity administered on a
`protracted schedule. Mice were implanted with L5178Y TK
`⫺/⫺ cells and treated with ZD9331; nine of 16 mice treated with
`100 mg/kg of ZD9331 given by continuous infusion for 7 days
`were cured, whereas none of the mice treated with 100 mg/kg of
`ZD9331 given by intraperitoneal injection were cured.10 Preclin-
`ical pharmacokinetic studies showed good oral bioavailability:
`30% to 60% in rats at 6 mg/m2 and 80% in dogs at 2 mg/m2.
`These preclinical studies demonstrating increased activity
`with prolonged exposure to ZD9331 and good oral bioavailabil-
`
`From the Committee on Clinical Pharmacology, Department of Medicine,
`Cancer Research Center, and Section of Hematology/Oncology, University
`of Chicago; Committee on Clinical Pharmacology and Cancer Research
`Center, Chicago, IL; AstraZeneca Pharmaceuticals, Macclesfield, UK, and
`Wilmington, DE.
`Submitted January 30, 2002; accepted February 13, 2003.
`Supported in part by AstraZeneca and grant no. CA 14599 from the
`National Cancer Institute, National Institutes of Health.
`Michael B. Sawyer was the Gordon E. Richards Fellow of the Canadian
`Cancer Society.
`Address reprint requests to Mark J Ratain, MD, 5841 S Maryland Ave,
`MC2115, Chicago, IL 60637; email: mratain@medicine.bsd.uchicago.edu.
`© 2003 by American Society of Clinical Oncology.
`0732-183X/03/2109-1859/$20.00.
`
`Journal of Clinical Oncology, Vol 21, No 9 (May 1), 2003: pp 1859-1865
`DOI: 10.1200/JCO.2003.01.148
`
`Amerigen Exhibit 1103
`Amerigen v. Janssen IPR2016-00286
`Downloaded from ascopubs.org by Univ of Chicago Library on January 3, 2017 from 205.208.061.064
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`1859
`
`

`
`1860
`
`ity prompted this phase I study of oral ZD9331 with prolonged
`administration. Because the intravenous formulation of ZD9331
`had demonstrated that clearance was independent of body-
`surface area,11 this oral formulation study used fixed dosing
`rather than dosing that was based on body-surface area.
`
`PATIENTS AND METHODS
`
`Patient enrollment began in April 1998, and the study was closed in
`February 2000. Eligible patients had a solid tumor that was refractory to
`standard therapies or for which no standard therapy existed; Cancer and
`Leukemia Group B performance status ⱕ 2; age ⱖ 18 years; a life
`expectancy ⱖ 12 weeks; a baseline platelet count ⱖ 100 ⫻ 109/L; a total
`WBC count ⱖ 3.5 ⫻ 109/L or an absolute neutrophil count ⱖ 2 ⫻ 109/L;
`serum bilirubin concentration ⱕ 1.25 times the upper limit of normal; ALT
`and AST ⱕ 2.5 times the upper limit of normal in the absence of liver
`metastases and ⱕ 5 times the upper limit of normal if liver metastases
`were present; serum creatinine ⱕ 1.25 times the upper limit of normal; no
`severe or uncontrolled systemic disease, such as uncompensated respira-
`tory or cardiac conditions; no use of folate-containing vitamin supple-
`ments; and no chlorambucil, mitomycin, or nitrosoureas for more than 6
`months total duration.
`
`Treatment Plan
`
`Patients took ZD9331 as a single daily dose, between 8 and 10 AM (either
`30 minutes before food or 2 hours after). The initial patient was treated with
`ZD9331 0.5 mg/d given for 14 days, followed by a 14-day rest period.
`Because no toxicity was observed, the next patient was treated for 21 days,
`followed by a 2-week rest period. The treatment period was increased to 28
`days followed by a 2-week rest period for the next patient. The treatment
`period then remained at 28 days for the duration of the study. The dose of
`ZD9331 was doubled until the development of any grade of drug-related
`toxicity, with a minimum of one patient per dose level. Subsequent dose
`escalations followed a modified Fibonacci scheme starting at the 67%
`dose-escalation level, with a minimum of three patients enrolled per dose
`level.
`If any patient developed dose-limiting toxicity (DLT),
`three
`additional patients were enrolled at that dose level. At least 12 patients
`were to be enrolled at
`the recommended phase II dose, which was
`anticipated to be the dose at which less than one third of patients
`experienced DLT in the first cycle.
`DLT was defined as grade 4 neutropenia of any duration with fever, grade
`4 neutropenia without fever for at least 7 days, or grade 4 thrombocytopenia.
`Patients stopped treatment if they developed grade 2 neutropenia or throm-
`bocytopenia while still taking ZD9331, and this was considered a DLT.
`Grade 3 or 4 nonhematologic toxicity that was not ameliorated by symp-
`tomatic directed therapy (with the exception of reversible elevations of AST
`or ALT) was also considered a DLT.
`Chemistry and hematologic measurements were repeated weekly during
`the first two cycles and within 3 days of starting the next cycle. Tumor
`measurements were repeated every two cycles.
`Patients remained on study until there was evidence of tumor progression,
`unacceptable toxicity, or until the patient or investigator felt continuing
`treatment with ZD9331 was not in the patient’s interest. Patients who
`experienced DLT but who were clinically benefiting from ZD9331 could
`continue to receive treatment at the investigator’s discretion at a reduced
`dose. The University of Chicago institutional review board approved the
`protocol and the consent form. Written informed consent was obtained from
`all patients.
`
`Pharmacokinetics
`
`Samples for pharmacokinetic analysis were collected on day 1 pretreat-
`ment and at 1, 2, 4, 6, 8, and 24 hours after starting treatment. Samples were
`also taken at predose and 1, 2, 4, and 6 hours postdose 1 week before the end
`of dosing, although these results are not presented here. The AstraZeneca
`Safety of Medicines Laboratory determined the plasma levels of ZD9331
`
`Table 1. Patient Characteristics
`
`SAWYER ET AL
`
`No. of Patients
`(n ⫽ 55)
`
`36:19
`
`Female/male
`Age, years
`Median
`Range
`CALGB performance status
`0
`1
`2
`Prior treatment
`Chemotherapy only
`Chemotherapy and radiation
`Chemotherapy regimens
`Median
`Range
`Cancer diagnosis
`Colorectal
`Ovarian
`Endometrial
`Esophagus
`Hepatocellular
`Others
`
`58
`33-79
`
`2
`1-8
`
`23
`29
`3
`
`36
`19
`
`31
`6
`2
`2
`2
`12
`
`Abbreviation: CALGB, Cancer and Leukemia Group B.
`
`(AstraZeneca) using a high-performance liquid chromatography–mass spec-
`trometry–mass spectrometry assay as previously described.8
`A noncompartmental approach was used to characterize the pharmacoki-
`netic parameters of ZD9331. The pharmacokinetic software used to analyze
`the plasma concentrations was WinNonlin (version 2.1; Pharsight Corp,
`Cary, NC). The area under the concentration-time curve (AUC) was
`calculated to the 24-hour sampling point (after the first dose) using the linear
`trapezoidal rule. The terminal half-life was estimated from the terminal part
`of the log concentration-time curve. The apparent oral clearance (Cl/F) was
`calculated as dose divided by AUC. Creatinine clearance was estimated
`using the formula of Cockcroft and Gault.12
`
`Statistics
`
`Stata 6 (version 6.0; Stata Corp, College Station, TX) was used to perform
`the statistical analysis. Pharmacokinetic parameters were examined for
`possible relationships with the development of myelosuppression using the
`Student’s t test. Exploratory analyses (using univariate and multivariate
`regression) were conducted to assess the possible relationship of various
`patient characteristics (eg, age, renal function) to ZD9331 clearance. In
`addition, an exploratory pharmacodynamic analysis was conducted after log
`transformation of the hematologic parameters. This analysis included an
`assessment of the relative contributions of variability in peak (maximum
`concentration; Cmax) and trough concentrations (24-hour minimum concen-
`tration; Cmin) and the AUC.
`
`RESULTS
`
`Demographics
`
`A total of 55 patients were enrolled, from April 1998 to
`February 2000. Patient characteristics are outlined in Table 1.
`Patients were assessable for toxicity if they were able to
`complete cycle 1 of treatment. Forty-five patients were assess-
`able for toxicity. One patient who developed suspected cellulitis
`during cycle 1 was withdrawn and therefore was not assessable
`for toxicity. Two patients died while on study, but the deaths
`were not considered to be related to ZD9331. One patient
`
`Downloaded from ascopubs.org by Univ of Chicago Library on January 3, 2017 from 205.208.061.064
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`
`PHASE I STUDY OF AN ORAL ZD9331 FORMULATION
`
`1861
`
`presented to her local hospital in the third week of treatment with
`acute shortness of breath; at presentation her WBC count was
`9.6 ⫻ 109/L, creatinine was 0.5 times the upper limit of normal,
`and blood urea nitrogen (BUN) was 9 mg/dL. Her respiratory
`status deteriorated, and she died from a presumed pulmonary
`embolus. At enrollment, she had a creatinine clearance of 81
`mL/min and BUN of 13.5 mg/dL. The other patient had a history
`of recurrent urinary tract infections and was taking nitrofurantoin
`to prevent recurrent infections. At enrollment, she had a calcu-
`lated creatinine clearance of 151 mL/min and a BUN of 9 mg/dL.
`She developed urosepsis in the fifth week of treatment and
`presented to her local hospital. She progressed to septic shock
`and her advance directive precluded inotropic support. She died
`despite maximal support outside of the intensive care unit. The
`remaining patients not assessable for toxicity either withdrew
`consent or developed progressive disease, preventing completion
`of one cycle.
`
`Dose Escalation
`
`A total of 128 cycles were administered, and the median
`number of cycles completed by each patient was two (range, one
`to seven cycles). The treatment duration was escalated to 28 days
`(at a dose of 0.5 mg/d) without the occurrence of DLT. The dose
`of ZD9331 was then escalated from 0.5 to 5 mg/d. No DLTs
`occurred until the 5 mg/d dose level, at which two patients
`developed DLTs. At the prior dose level of 3 mg/d, no significant
`drug-related toxicity was observed, which led us to examine 4
`mg/d. The three initial patients evaluated at 4 mg/d experienced
`more severe toxicity than patients treated at 5 mg/d. A prelimi-
`nary analysis suggested that patients who experienced DLT at
`the 4 and 5 mg/d dose levels had decreased drug clearance and
`that this was correlated with an elevated BUN. Seven additional
`patients were enrolled at the 4-mg level to test this hypothesis;
`three patients had a high BUN, ⱖ 20 mg/dL (4.0H), and four
`patients had a low BUN,
`less than 20 mg/dL (4.0L). An
`additional 16 patients were subsequently enrolled at the recom-
`mended phase II dose of 3 mg/d to better define the toxicity and
`pharmacodynamics at this dosage level.
`
`Hematologic Toxicity
`
`The DLT of ZD9331 was myelosuppression, both thrombo-
`cytopenia and neutropenia (Table 2). Patients developed throm-
`bocytopenia at a median of 14 days after starting therapy (range,
`11 to 14 days). Platelet counts returned to baseline at a median
`of 10 days (range, 7 to 14 days). Neutropenia occurred at a
`median of 21 days (range, 11 to 21 days), with a median duration
`of 7 days before recovery (range, 7 to 11 days). At 3 mg/d, the
`recommended phase II dose, two of 19 patients developed DLTs;
`one patient had grade 3 thrombocytopenia and grade 4 neutro-
`penia, and the other patient had grade 3 thrombocytopenia only.
`Anemia was a common toxicity at doses ⱖ 3 mg/d (Fig 1), but
`was not dose-limiting. This was managed with erythropoietin
`and RBC transfusions. Midway through the study, one patient
`developed anemia associated with a low haptoglobin, elevated
`bilirubin and lactate dehydrogenase (LDH), and a negative
`Coombs test. After this index case, all subsequent patients had
`
`Table 2. Hematologic Toxicity by Dose Level (fully assessable patients only)
`
`Dose of
`ZD9331
`(mg/d)
`
`0.5
`0.5
`0.5
`1.0
`1.5
`3.0
`5.0
`
`4.0
`
`3.0
`
`4.0H*
`
`4.0L*
`
`Cycle Length
`(days)
`
`No. of
`Patients
`
`Grade 3 and 4 Adverse Events
`(no. of patients)
`
`Dose-Limiting
`Toxicity
`(no. of patients)
`
`14
`21
`28
`28
`28
`28
`28
`
`28
`
`28
`
`28
`
`28
`
`2
`2
`2
`3
`2
`3
`5
`
`3
`
`16
`
`3
`
`4
`
`—
`—
`—
`—
`—
`—
`Grade 3 leukopenia (1), grade
`3 thrombocytopenia (1),
`grade 3 neutropenia (1)
`Grade 3 thrombocytopenia
`(1), grade 3 neutropenia
`(1), grade 4 neutropenia (1)
`Grade 3 thrombocytopenia
`(2), grade 3 anemia (1),
`grade 4 neutropenia (1)
`Grade 3 thrombocytopenia
`(1), grade 2
`thrombocytopenia (1)
`Grade 4 thrombocytopenia
`(1), grade 2 nausea and
`vomiting (1)
`
`—
`—
`—
`—
`—
`—
`2
`
`2
`
`2
`
`2
`
`2
`
`*4.0H (high) blood urea nitrogen ⱖ 20 mg/dL, 4.0L (low) blood urea nitrogen ⬍
`20 mg/dL.
`
`haptoglobin levels measured along with routine biochemistry. At
`the 3-mg dose level, five of 16 patients had anemia with
`undetectable haptoglobin levels. As Fig 2 demonstrates, the
`decrease in hemoglobin was gradual over several weeks. There
`was no significant difference in nadir hemoglobin, bilirubin, or
`LDH between patients with or without low haptoglobin levels.
`
`Nonhematologic Toxicity
`
`The most frequent toxicities encountered at the 3-mg dose
`level are listed in Table 3. The most common toxicity was
`fatigue, which occurred in almost all patients. Nausea was a
`frequent side effect but was not dose-limiting. Four patients
`developed an erythematous maculopapular rash, which preceded
`the development of myelosuppression. Six patients developed
`elevated AST and ALT while receiving therapy, but the ALT and
`AST levels returned to baseline during the rest period and did not
`exceed grade 2 toxicity.
`
`Tumor Response
`
`Patients were evaluated after the first two cycles and every
`two cycles thereafter. There were no objective minor or major
`responses observed, but 14 patients had prolonged stabilization
`of disease. The median duration of stable disease was 12 weeks
`(range, 12 to 48 weeks). At the 3-mg dose level, five of 19
`patients achieved stable disease, including two patients who had
`prolonged stable disease (36 and 48 weeks). The majority of
`patients with stable disease had colorectal cancer (n ⫽ 10), but
`two patients with ovarian cancer also had disease stabilization.
`
`Pharmacokinetics
`
`None of the patients with DLT reached the day 21 sampling
`time. Therefore, to avoid biasing the pharmacokinetic analysis,
`
`Downloaded from ascopubs.org by Univ of Chicago Library on January 3, 2017 from 205.208.061.064
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`
`1862
`
`SAWYER ET AL
`
`Fig 1. Relationship between dose
`and development of anemia. Symbols
`differentiate dose levels.
`
`the pharmacokinetic analysis is limited to the day 1 pharmaco-
`kinetic samples that are available for all patients. ZD9331
`concentrations were measured in 51 patients on day 1 (Table 4).
`A sample concentration-time profile of a patient at the 3-mg dose
`level is shown in Fig 3. Peak concentrations occurred within the
`first 2 hours, followed by a rapid decline. Some patients who had
`decreased Cl/F reached a plateau in the terminal part of the
`concentration-time curve. There was a rebound in concentra-
`tions seen at 6 to 8 hours, which may represent enterohepatic
`circulation. The mean terminal half-life at 3 mg/d was 25 ⫾
`
`22 hours (range, 9 to 109 hours). The mean half-life is likely
`an underestimate of the true half-life; only 24-hour data were
`available on all patients because patients with an estimated
`long half-life came off study early, before the late pharma-
`cokinetic sample times.
`As the dose of ZD9331 increased from 3 to 5 mg (a 67%
`increase in dose), the mean AUC increased from 2,839 to
`3,431 ng/mL 䡠 h (a 21% increase), and the mean peak
`concentrations were 206 ng/mL and 298 ng/mL, respectively
`(a 45% increase).
`
`Time course of the develop-
`Fig 2.
`ment of anemia. Symbols differentiate
`time points.
`
`Downloaded from ascopubs.org by Univ of Chicago Library on January 3, 2017 from 205.208.061.064
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`
`PHASE I STUDY OF AN ORAL ZD9331 FORMULATION
`
`Table 3. Common Toxicities at the 3 mg/d Dose Level
`(n ⴝ 19)
`
`Fatigue
`Nausea
`Anorexia
`Constipation
`Alopecia
`Diarrhea
`Vomiting
`Increased AST
`Dyspnea
`Increased ALT
`Ankle edema
`Rash
`Mucositis
`Skin hyperpigmentation
`Fever
`Increased bilirubin
`
`1
`
`8
`14
`10
`10
`6
`6
`5
`5
`5
`3
`4
`4
`3
`2
`1
`
`2
`
`9
`1
`5
`
`2
`1
`1
`1
`
`2
`
`1
`
`Grade
`
`3
`
`1
`1
`1
`
`4
`
`1
`
`1
`
`1
`
`The Cl/F, AUC, Cmax, and Cmin for each dose level are listed
`in Table 4. At doses of ⱖ 3 mg/d, there was no relationship
`between dose and the pharmacokinetic parameter AUC or Cmin.
`Thus data for the 3-, 4-, and 5-mg dose levels were combined for
`the purpose of analysis of the determinants of pharmacokinetic
`variability. At doses ⱖ 3 mg/d, there was no significant corre-
`lation between AUC and creatinine clearance (r ⫽ 0.14) but
`BUN was correlated with AUC (r ⫽ 0.42; P ⫽ .007). At doses ⱖ
`3 mg/d, Cmin was significantly related to BUN (r ⫽ 0.49; P ⬍
`.001) but not to creatinine clearance (r ⫽ 0.20).
`
`Pharmacodynamics
`
`In phase I studies, the large number of dose levels studied can
`confound analysis of determinants of pharmacodynamic vari-
`ability.13 Toxicity is usually correlated with dose administered
`and therefore dose level. The relationship between dose and
`AUC will lead to a false-positive association between AUC and
`toxicity unless the confounding issue of dose is removed. At the
`3-, 4-, and 5-mg dose levels, there was no correlation between
`dose and AUC. The pharmacodynamic analysis was therefore
`confined to the 3-, 4-, and 5-mg dose levels to eliminate potential
`confounding by dose level.
`Demographic characteristics, pharmacokinetic parameters,
`and organ function were examined for possible relationships
`
`1863
`
`Fig 3.
`
`Typical concentration profile of ZD9331.
`
`with thrombocytopenia and anemia. In the univariate analysis,
`significant predictors of log platelet nadir were RBC folate, dose,
`Cmin, half-life, AUC, clearance, bilirubin, BUN, and log Cmin
`(Table 5). Serum folate, Cmax, performance status, albumin
`creatinine clearance, BUN/creatinine ratio, age, and log baseline
`platelet count were not associated with log platelet nadir. In the
`multivariate analysis, only log Cmin and BUN were correlated
`with log platelet nadir (r2 ⫽ 0.44; P ⬍ .001).
`In univariate analysis, significant predictors of log hemoglobin
`nadir were baseline hemoglobin, dose, Cmin, bilirubin, albumin,
`creatinine clearance, and log Cmin. Serum folate, Cmax clearance,
`AUC, performance status, BUN, and BUN/creatinine ratio did
`not reach statistical significance. In multivariate analysis, signif-
`icant predictors of log nadir hemoglobin were log baseline
`hemoglobin, albumin, and log Cmin (r2 ⫽ 0.61; P ⬍ .001).
`AUC (during the first 24 hours), Cmin, creatinine clearance,
`and BUN were examined as predictors for dose-limiting myelo-
`suppression (Table 6). BUN levels were significantly higher in
`patients who developed dose-limiting myelosuppression com-
`pared with patients who did not (Student’s t test, P ⫽ .04). In
`addition, Cmin was higher in patients who developed dose-
`limiting myelosuppression than in patients who did not (Stu-
`dent’s t test, P ⫽ .03).
`
`DISCUSSION
`
`The recommended phase II dose for ZD9331 is 3 mg/d for 4
`weeks followed by a 2-week rest period. Some patients were able
`to tolerate ZD9331 doses of 4 and 5 mg/d without toxicity.
`Because preliminary pharmacokinetic analysis suggested that
`
`Table 4. Pharmacokinetic Parameters
`
`Dose Level
`(mg/d)
`
`0.5
`1.0
`1.5
`3.0
`4.0
`5.0
`
`No. of Patients
`
`Cl/F (mL/min)
`
`AUC (ng/mL 䡠 h)
`
`Cmin (ng/mL)
`
`Cmax (ng/mL)
`
`6
`3
`1
`23
`12
`6
`
`9.9 ⫾ 4.8
`22.1 ⫾ 9.5
`13.8
`12.1 ⫾ 8.8
`10.0 ⫾ 4.5
`17.9 ⫾ 8.2
`
`423 ⫾ 125
`414 ⫾ 138
`1203
`2,839 ⫾ 857
`3,441 ⫾ 1241
`3,431 ⫾ 1045
`
`13 ⫾ 6
`12 ⫾ 4
`27
`79 ⫾ 35
`110 ⫾ 50
`80 ⫾ 50
`
`32 ⫾ 13
`34 ⫾ 14
`96
`206 ⫾ 66
`215 ⫾ 58
`298 ⫾ 70
`
`Abbreviations: Cl/F, apparent oral clearance; AUC, area under the curve; Cmin, minimum concentration; Cmax, maximum
`concentration.
`
`Downloaded from ascopubs.org by Univ of Chicago Library on January 3, 2017 from 205.208.061.064
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`
`1864
`
`SAWYER ET AL
`
`Table 5. Univariate Analysis
`
`Correlation
`
`P
`
`Log platelet nadir variable
`RBC folate
`Dose
`Cmin
`Half-life
`AUC
`Clearance
`Bilirubin
`BUN
`Log Cmin
`Log hemoglobin nadir variable
`RBC folate
`Log baseline hemoglobin
`Dose
`Cmin
`Bilirubin
`Albumin
`Creatinine clearance
`Log Cmin
`
`⫺0.30
`⫺0.43
`⫺0.60
`⫺0.36
`⫺0.48
`0.38
`⫺0.30
`⫺0.48
`⫺0.50
`
`⫺0.32
`0.59
`⫺0.32
`⫺0.38
`⫺0.31
`0.59
`0.36
`⫺0.44
`
`.04
`.004
`.0001
`.02
`.001
`.01
`.02
`.001
`.0004
`
`.04
`.0001
`.03
`.01
`.04
`.0001
`.02
`.003
`
`Abbreviations: Cmin, minimum concentration; AUC,
`area under the curve; BUN, blood urea nitrogen.
`
`elevated BUN predisposed patients to develop ZD9331 toxicity,
`we therefore re-examined the 4 mg/d dose level, enrolling
`patients on the basis of BUN levels. However, even in the group
`with low BUN, we could not safely administer ZD9331 at 4
`mg/d and could not recommend 4 mg/d as the initial phase II
`dose. Given that only two of 19 patients at 3 mg/d had DLT,
`patients who tolerate ZD9331 3 mg/d might be considered for
`escalation to a dose of 4 mg/d.
`The DLT of ZD9331 on this oral schedule was neutropenia
`and thrombocytopenia; this contrasts with other studies in which
`cutaneous toxicity and myelosuppression were found to be the
`DLTs.11,14 In the study by de Jonge et al,14 dose-limiting
`cutaneous toxicity occurred at 10 mg twice a day given for 5
`days. In the study by Goh et al,11 dose-limiting skin toxicity
`occurred at 12 mg/m2 and 25 mg/d given daily for 5 days. In
`this study,
`the daily administration of ZD9331 was more
`prolonged at 4 weeks, and the highest daily dose administered
`was 5 mg/d. The total dose of ZD9331 administered per cycle
`in the three studies was not substantially different (100 v 125
`v 84 mg/cycle). ZD9331 was administered four times more
`rapidly in the two schedules associated with skin toxicity than
`in this study. This indicates that skin toxicities may be related
`
`Table 6. Parameters Examined for Relationship to Dose-Limiting
`Myelosuppression (fully assessable patients only)
`
`Parameter
`
`AUC
`Cmin
`Creatinine clearance
`BUN
`Half-life
`
`Patients Without
`Myelosuppression
`
`2,975 ⫾ 794
`77.6 ⫾ 32.6
`94 ⫾ 34
`14 ⫾ 4
`20.8 ⫾ 10.9
`
`Patients With
`Myelosuppression
`
`3,877 ⫾ 1,249
`128 ⫾ 57.1
`86 ⫾ 30
`22 ⫾ 9
`39.2 ⫾ 10.2
`
`P
`
`.07
`.03
`NS
`.04
`.12
`
`Abbreviations: AUC, area under the curve; Cmin, minimum concentration; BUN,
`blood urea nitrogen; NS, not significant.
`
`to peak concentrations and the rate at which thymidylate
`synthase was inhibited.
`Anemia was a common toxicity at or above the 3-mg dose
`level, but was not dose-limiting and was easily managed with
`erythropoietin and RBC transfusions. Midway through the study,
`one patient developed anemia associated with a low haptoglobin
`level. The decrease in hemoglobin was gradual over several
`weeks, which is more likely secondary to decreased production
`than hemolysis.
`Three studies have implicated methotrexate (in combina-
`tion with other agents) with elevations of LDH or anemia
`associated with a low haptoglobin level. Klimo and Connors15
`first reported elevation of LDH levels in the majority of
`patients treated with a regimen consisting of methotrexate,
`doxorubicin, cyclophosphamide, vincristine, prednisone, and
`bleomycin. The patients with elevated LDH were in complete
`or partial remission and had no elevations in other hepatic
`enzymes. In another study of methotrexate, doxorubicin,
`cyclophosphamide, vincristine, prednisone, and bleomycin,
`McAdam et al16 noted an increase in LDH during treatment
`that returned to baseline before the next cycle. In a study of
`cyclophosphamide, doxorubicin, methotrexate, bleomycin,
`vincristine, etoposide,
`ifosfamide, and prednisolone, Ma-
`ruyama et al17 observed an elevation of LDH isoenzymes 1
`and 2. In addition, six patients had undetectable haptoglobin
`levels. We speculate that anemia with elevation of LDH and
`suppressed haptoglobin may be an unappreciated side effect
`of the antifolate class of chemotherapeutic agents as a result
`of ineffective erythropoiesis.
`Given orally, ZD9331 showed a less than dose-proportional
`increase of AUC at doses ⱖ 3 mg/d. Potential explanations
`are solubility-limited absorption (discounted because of the
`known physiochemical properties of the compound), satura-
`ble gastrointestinal absorption, nonlinear protein-binding, or
`saturable renal tubular reabsorption. Although interpretation
`is difficult because of the variability of the data and low
`numbers of subjects at lower doses, the increase of Cmax with
`dose could suggest that saturable gastrointestinal absorption
`is not the entire explanation. Because preclinical studies did
`not demonstrate saturable protein-binding, it is possible that
`the nonlinearity may be secondary to saturation of
`the
`extensive renal reabsorption of this folate analog.
`The dominant role of reabsorption in the renal handling of
`ZD9331 may explain the relationship between Cmin and BUN of
`ZD9331 and the poor correlation between Cmin and calculated
`creatinine clearance; the latter is an estimation of the amount of
`blood that is filtered at the glomerulus and enters the proximal
`tubule. Creatinine clearance is a useful surrogate for predicting
`the clearance of drugs that are predominantly filtered, such as
`carboplatin, but may not be well correlated with the ability of
`the proximal tubule to reabsorb solutes and electrolytes. BUN
`is a better surrogate for tubular function because urea is
`reabsorbed. The parallels between urea reabsorption and
`ZD9331 reabsorption may be the basis for the apparent
`relationship between Cmin and BUN. Thus ZD9331, like other
`
`Downloaded from ascopubs.org by Univ of Chicago Library on January 3, 2017 from 205.208.061.064
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`
`PHASE I STUDY OF AN ORAL ZD9331 FORMULATION
`
`1865
`
`renally excreted drugs, should be used with caution in patients
`with renal insufficiency or in combination with drugs that
`affect proximal tubular reabsorption.
`Plasma 2⬘-deoxyuridine (dUrd) is the substrate of thymidylate
`synthase that is elevated when thymidylate synthase is inhibited.
`Jackman et al18 and Mitchell et al19 have shown that plasma
`levels of dUrd are correlated with the degree of thymidylate
`synthase inhibition. We had planned to measure dUrd levels in
`
`this study as a measure of thymidylate synthase inhibition;
`unfortunately, plasma interference prevented analysis of dUrd in
`samples obtained.
`In this study, we observed disease stabilization of colorectal
`and ovarian cancers at a dose of 3 mg/d given for 28 days
`followed by a 2-week rest period. The results from this study
`indicate that phase II studies of ZD9331 are warranted in
`metastatic colorectal and ovarian cancer.
`
`REFERENCES
`
`1. Hillman RS: Hematopoietic agents: Growth factors, minerals and
`vitamins, in Hardman JG, Limbird LE, Molinoff PB, et al (eds): Goodman
`and Gilman’s Pharmacological Basis of Therapeutics (ed 9). New York, NY,
`McGraw-Hill, 1996, pp 1311-1340
`2. Vest S, Bork E, Hansen HH: A phase I evaluation of N10-propargyl-
`5,8-dideazafolic acid. Eur J Cancer Clin Oncol 24:201-204, 1988
`3. Jodrell DI, Newell DR, Morgan SE, et al: The renal effects of
`N10-propargyl-5,8-dideazafolic acid (CB3717) and a non-nephrotoxic ana-
`logue ICI D1694, in mice. Br J Cancer 64:833-838, 1991
`4. Clarke SJ, Hanwell J, de Boer M, et al: Phase I trial of ZD1694, a new
`folate-based thymidylate synthase inhibitor, in patients with solid tumors.
`J Clin Oncol 14:1495-1503, 1996
`5. Blackledge G: New developments in cancer treatment with the novel
`thymidylate synthase inhibitor raltitrexed (Tomudex). Br J Cancer 77:29-37,
`1998 (suppl)
`6. Jackman AL, Boyle FT, Harrap KR: Tomudex (ZD1694): From
`concept to care, a programme in rational drug discovery. Invest New Drugs
`14:305-316, 1996
`7. Pizzorno G, Mini E, Coronnello M, et al: Impaired polyglutamylation
`of methotrexate as a cause of resistance in CCRF-CEM cells after short-term,
`high-dose treatment with this drug. Cancer Res 48:2149-2155, 1988
`8. Rhee MS, Wang Y, Nair MG, et al: Acquisition of resistance to
`antifolates caused by enhanced gamma-glutamyl hydrolase activity. Cancer
`Res 53:2227-2230, 1993 (suppl)
`9. Jackman AL, Kimbell R, Brown M, et al: The antitumour activity of
`ZD9331, a non-polyglutamatable quinazoline thymidylate synthase inhibitor.
`Adv Exp Med Biol 370:185-188, 1994
`10. Jackman AL, Kimbell R, Aherne GW, et al: Cellular pharmacology
`and in vivo activity of a new anticancer agent, ZD9331: A water-soluble,
`
`nonpolyglutamatable, quinazoline-based inhibitor of thymidylate synthase.
`Clin Cancer Res 3:911-921, 1997
`11. Goh BC, Ratain MJ, Bertucci D, et al: Phase I study of ZD9331 on
`short daily intravenous bolus infusion for 5 days every 3 weeks with fixed