Cancer Therapy: Preclinical Assessment of GS-9219 in a Pet Dog Model of Non-Hodgkin's Lymphoma

Purpose: To assess, in dogs with naturally occurring non-Hodgkin's lymphoma, pharmacokinetics, safety, and activity of GS-9219, a prodrug of the nucleotide analogue 9-(2-phosphonylmethoxyethyl) guanine (PMEG), which delivers PMEG and its phosphorylated metabolites to lymphoid cells with preferential cytotoxicity in cells with a high proliferation index such as lymphoid malignancies. Experimental Design: To generate proof-of-concept, a phase I/II trial was conducted in pet dogs (n = 38) with naturally occurring non-Hodgkin's lymphoma using different dose schedules of GS-9219. A subset of dogs was further evaluated with 3′-deoxy-3′F-fluorothymidine positron emission tomography/computed tomography imaging before and after treatment. Results: The prodrug had a short plasma half-life but yielded high and prolonged intracellular levels of the cytotoxic metabolite PMEG diphosphate in peripheral blood mononuclear cells in the absence of detectable plasma PMEG. Dose-limiting toxicities were generally manageable and reversible and included dermatopathy, neutropenia, and gastrointestinal signs. Antitumor responses were observed in 79% of dogs and occurred in previously untreated dogs and dogs with chemotherapy-refractory non-Hodgkin's lymphoma. The median remission durations observed compare favorably with other monotherapies in dogs with non-Hodgkin's lymphoma. High 3′-deoxy3′-F-fluorothymidine uptake noted in lymphoid tissues before treatment decreased significantly after treatment (P = 0.016). Conclusions: GS-9219 was generally well tolerated and showed significant activity against spontaneous non-Hodgkin's lymphoma as modeled in pet dogs and, as such, supports clinical evaluation in humans. Non-Hodgkin's lymphoma is the fifth leading cause of cancer deaths and the second fastest-growing form of cancer in the United States. In the latest compilation of Surveillance, Epidemiology and End Results reports, the incidence and death rate for non-Hodgkin's lymphoma continue to increase despite an overall decrease in overall cancer incidence (1). Therefore, there is still a major unmet medical need in non-Hodgkin's lymphoma patients for novel agents with improved efficacy compared with existing treatment modalities, especially in patients who have failed frontline therapy. The acyclic nucleotide phosphonate 9-(2-phosphonylmethoxyethyl) guanine (PMEG) forms an active phosphorylated metabolite, PMEG diphosphate, in cells and causes cytotoxicity in dividing cells due to potent inhibition of the nuclear DNA polymerases α, δ, and ε by causing DNA chain termination, resulting in inhibition of DNA synthesis (2). In rodent Authors' Affiliations: Department of Medical Sciences, School of Veterinary Medicine and the Paul P. Carbone Comprehensive Cancer Center, University of Wisconsin-Madison, Madison, Wisconsin; Animal Cancer Center, Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, Colorado; and Departments of Drug Safety Evaluation, Biology, Drug Metabolism and Chemistry, and Clinical Research (Oncology), Gilead Sciences, Foster City, California Received 11/28/08; revised 1/26/09; accepted 2/19/09; published OnlineFirst 5/5/09. Grant support: Gilead Sciences (D.M. Vail and D.H. Thamm). 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. Note: D.M. Vail and D.H. Thamm contributed equally to this work. Requests for reprints: David M. Vail, Center for Clinical Trials and Research, School of Veterinary Medicine, University of Wisconsin-Madison, 2015 Linden Drive, Madison, WI 53706. Phone: 608-890-1324; Fax: 608-265-8020; E-mail: [email protected]. F 2009 American Association for Cancer Research. doi:10.1158/1078-0432.CCR-08-3113 3503 Clin Cancer Res 2009;15(10) May 15, 2009 www.aacrjournals.org Cancer Research. on January 14, 2018. © 2009 American Association for clincancerres.aacrjournals.org Downloaded from models, PMEG has activity against leukemia and melanoma (3). However, the utility of PMEG as an anticancer agent is limited by its poor cellular permeability and nonspecific toxicity, especially for the kidney and gastrointestinal tract (3–5). GS-9219, a novel double prodrug of PMEG, was designed as a cytotoxic agent that preferentially targets lymphoid cells (6). In lymphocytes, GS-9219 is hydrolyzed intracellularly to 9-(2-phosphonylmethoxyethyl)-N-cyclopropyl-2,6-diaminopurine (cPrPMEDAP) and subsequently deaminated to PMEG. PMEG is then converted to its active phosphorylated form, PMEG diphosphate (6). The fact that rodents, unlike humans or dogs, have high plasma levels of carboxyesterase, which rapidly metabolizes GS-9219, precluded their use in preclinical models (7). Evaluation of GS-9219 in normal beagle dogs previously showed, at doses that were generally well tolerated, selective depletion of replicative lymphoid tissues while sparing other tissues. Preliminary antineoplastic activity in dogs with naturally occurring non-Hodgkin's lymphoma has also been shown (6). Canine non-Hodgkin's lymphoma has proven to be a relevant model for preclinical evaluation of new therapeutics for both initial induction and rescue of drug-resistant relapse (8–14). Evaluation of novel therapeutic approaches in dogs with spontaneous cancer offers potential benefit to canine patients and a rapid assessment of therapeutic index. Because the tumors arise spontaneously in an immunologically intact host and have greater heterogeneity than passaged cell lines, it is not surprising that responses to standard chemotherapeutic agents in canine malignancies are similar to those of the corresponding tumors in man and that the preclinical results attained may be more predictive of activity in humans (8, 15, 16). Non-Hodgkin's lymphoma in dogs represents a relatively homogenous population with respect to histologic type as defined by the REAL/WHO or National Cancer InstituteWorking Formulation schema (85% are mediumto highgrade B-cell non-Hodgkin's lymphoma), with the majority being diffuse large B-cell lymphoma (17). As in people, non-Hodgkin's lymphoma in pet dogs is initially highly responsive to standard chemotherapy with ∼90% of dogs achieving a complete remission with multidrug (e.g., CHOP) chemotherapy; however, cures are rare in dogs as relapse of drug-resistant disease occurs in the majority of cases (85%), with median survival durations of ∼1 year following multidrug treatment protocols (16). To generate proof-of-concept, pharmacokinetic, safety, and activity data in lymphoid malignancies, a dose/schedule-finding and activity trial with GS9219 monotherapy was initiated in pet dogs with naturally occurring chemotherapy-naive or chemotherapy-refractory advanced-stage non-Hodgkin's lymphoma. Materials and Methods Subject population. Pet owners presenting to the School of Veterinary Medicine, University of Wisconsin-Madison, or the Veterinary Medical Center, College of Veterinary Medicine & Biomedical Sciences, Colorado Translational Relevance Canine non-Hodgkin's lymphoma bears many similarities to human non-Hodgkin's lymphoma and is an attractive model for preclinical studies. There remains a major unmet medical need in nonHodgkin's lymphoma patients for novel agents with improved efficacy compared with existing treatment modalities, especially in patients who have failed frontline therapy. We report here a proof-of-concept investigation of the novel nucleotide analogue prodrug GS-2919 through the inclusion of pet dogs with spontaneously arising non-Hodgkin's lymphoma. This study provided valuable information before initiating human investigations: adverse events were generally transient, manageable, and dose and schedule dependent; the active form of the drug persisted selectively in peripheral blood mononuclear cells, prompting further evaluation of regimens using more frequent administration, and a favorable therapeutic index was shown in both treatmentnaive and chemotherapy-refractory cases of nonHodgkin's lymphoma in pet dogs. Thus, GS-9219 is a promising development candidate and phase I investigations of GS-9219 in human patients with lymphoid malignancies are in progress. Table 1. GS-9219 scheduled treatment regimens evaluated Cohort GS-9219 dose (mg/kg)* Frequency Dogs treated Dose-limiting adverse events Best response observed Cohort 1-low 0.20 Daily for 5 d, every 21 d 7 1 5 CR/2 PR Cohort 1-high 0.29 Daily for 5 d, every 21 d 6 3 5 CR/1 PR Cohort 2-low 0.66 Once every 7 d 3 0 2 CR/0 PR Cohort 2-high 0.82 Once every 7 d 6 2 3 CR/2 PR Cohort 3-low 0.66 Once every 14 d 3 0 1 CR/0 PR Cohort 3-high 0.82 Once every 14 d 6 1 3 CR/2 PR Cohort 4-low 0.66 Once every 21 d 3 0 2 CR/0 PR Cohort 4-high 0.82 Once every 21 d 4 0 2 CR/0 PR *Based on GS-9219 free base. Each regimen was to be repeated for five total treatment cycles. One dog was diagnosed with a second malignancy after entry and a seventh dog was added to cohort. Cohort expanded to 6 dogs due to a single dose-limiting adverse event in first 3 dogs. One dog died of tumor-related complications before response evaluation and an additional dog was added to the cohort. 3504 Clin Cancer Res 2009;15(10) May 15, 2009 www.aacrjournals.org Cancer Therapy: Preclinical Cancer Research. on January 14, 2018. © 2009 American Association for clincancerres.aacrjournals.org Downloaded from State University, were offered entry into this study for treatment of their dogs with GS-9219 under compliance with the Animal Care and Use Committees of the University of Wisconsin-Madison and Colorado State University and the Clinical Review Board of the Colorado State University Veterinary Medical Center. All owners signed informed consent documents, including a statement to allow necropsy in the event of their animal's death, whenever possible. Dogs were evaluated by complete physical examination, routine clinical biochemistry with electrolytes, hematology and urinalysis, thoracic radiographs, bone marrow aspiration cytology, and diagnostic lymph node biopsy inclusive of histology and immunohistochemistry to confirm non-Hodgkin's lymphoma and determine immunophenotype. Concurrent antineoplastic therapy was not allowed. Previous cytotoxic chemotherapy was allowable, with at least a 3-week washout from the most recent treatment. GS-9219 monotherapy protocols. GS-9219 was administered in all protocols by a 30 min intravenous infusion in 5% dextrose for injection, USP (2 mL/kg body weight). Four different schedules were evaluated, each with two dosing levels for a total of eight treatment cohorts (Table 1). Initial starting dose was chosen based on results from previous standard toxicology studies in beagle dogs (6). Several dosing schedules were initiated to evaluate potential treatment regimens. A standard 3 + 3 phase I cohort design was employed where 3 dogs were initiated in each cohort and observed for dose-limiting toxicities (DLT). A DLT was defined as persistent grade ≥3 for all adverse events, excluding neutropenia where a grade 4 adverse event was deemed doselimiting. If noDLTs were observed in the first cohort of animals, a second cohort was treated on the same regimen at the higher dose. If a DLT was observed in 1 dog, the cohort was expanded to 6 dogs. If two or more DLTs were noted in any cohort, no further dose escalations were done. Pharmacokinetic analysis of plasma and peripheral blood mononuclear cell levels of GS-9219 and its metabolites. A subset of dogs was evaluated for pharmacokinetics. Blood for plasma levels was collected at various time points into sodium fluoride potassium oxalate Vacutainer tubes (BD Biosciences) and centrifuged at 4°C to separate plasma. Plasma was prepared by protein precipitation by adding acetonitrile and formic acid to a final concentration of 60% and 0.2%, respectively, in the presence of internal standards (a methylated version of PMEG and its corresponding phosphonamidate prodrug). Samples were dried in a MiVac Duo Concentrator (Genevac) and reconstituted in sterile water. Blood for intracellular peripheral blood mononuclear cell (PBMC) levels was also collected 24 h post-dose into sodium citrate CPT Vacutainer tubes (BD Biosciences) for PBMC isolation. Following centrifugation through a thixotropic polyester gel and a Ficoll-Hypaque solution, isolated PBMCs were resuspended in normal saline. The cell suspension was centrifuged to pellet cells and isolated cells were resuspended in 70% methanol lysis buffer. Lysed samples were dried and resuspended at 1 million cells/10 μL calf intestinal phosphatase buffer with internal standards and treated with 1 unit calf intestinal phosphatase at 37°C for 2 h (buffer and enzyme obtained from Sigma-Aldrich). Reactions were stopped by adjusting to 60% acetonitrile followed by drying and resuspension at 1 million cells/10 μL in water. All liquid chromatography-tandem mass spectrometry analyses used a HTS PAL autosampler (Leap Technologies) with cooled stacks, a Shimadzu LC-20AD ternary pump system (Shimadzu Scientific Instruments), and a Sciex API-4000 mass spectrometer (Applied Biosystems) operating in multiple-reaction monitoring and positive ionization modes. Liquid chromatography separation for the analysis of GS-9219, cPrPMEDAP, and PMEG was done using a Synergi Max RP-80 4 μm 150 × 2.0 mm column (Phenomenex) and a multistage linear gradient from 4% to 91% acetonitrile in 0.2% formic acid at a flow rate of 250 μL/min. Standard curves and quality-control samples were used to ensure appropriate accuracy and precision. Noncompartmental pharmacokinetic analysis was done using WinNonLin version 5.0.1 (Pharsight). Safety evaluation. Adverse events were graded according to the Veterinary Cooperative Oncology Group-Common Terminology Criteria for Adverse Events version 1.0 (18) modified from the National Cancer Institute-Common Terminology Criteria for Adverse Events used in humans. Toxicity evaluations were based on physical examination, client questionnaire, CBC, biochemistry profile, and urinalysis and were done pretreatment and 1 week post-treatment for all five cycles in the once every 7 days, once every 14 days, and once every 21 days cohorts. For dogs in cohort 1 (five consecutive daily treatments out of every 21 days), toxicity evaluations were done pretreatment and day 9 of every cycle. In dogs experiencing prolonged grade 3 adverse events, dose reductions or prolongation of the treatment interval was instituted. In the event of death or euthanasia, necropsy evaluation was done whenever possible. Activity evaluation. Although assessment of lymph node size, computed tomography (CT), positron emission tomography (PET)/CT, and flow cytometric analysis is routine for determining standard response criteria for human patients with non-Hodgkin's lymphoma (19, 20), availability of advanced imaging limits similar assessments in veterinary patients. Clinical response in veterinary cancer patients with non-Hodgkin's lymphoma is generally based on summation of longest diameters of affected peripheral nodes, as measured with calipers, followed by the application of Response Evaluation Criteria in Solid Tumors of solid tumors for the pretreatment and post-treatment measurements (21). In the study reported here, response was assessed using a combination of lymph node size reduction (Response Evaluation Criteria in Solid Tumors), radiographic assessment of thoracic lesions, and cytologic assessment of previously affected nodes. Additionally, antitumor activity was further evaluated in a subset of 9 dogs by 3′-deoxy-3′-F-fluorothymidine (FLT) PET/CT before and after treatment. Antitumor activity was evaluated 7 days following the initial treatment with GS-9219, at the day of each subsequent treatment cycle, and monthly following completion of all GS-9219 treatments. A complete response (CR) was defined when post-treatment nodal size was deemed within normal limits (minimum of 75% reduction in the sums of the longest diameters) and was confirmed with cytologic assessment of a previously affected lymph node. A partial response (PR) was defined as a 30% decrease in the sums of the longest diameters of measurable involved nodes, and progressive disease was defined as a 20% Table 2. Subject demographics Characteristics Value (%) Age (y) Mean 7.5 Median 8 Range 2-14 Gender Male intact 1 Male neutered 22 Female intact 3 Female neutered 12 Tumor stage* III 12 IV 13 V 13 Tumor immunophenotype B cell 29 T cell 9 Prior Treatment Naive 17 Prior chemotherapy 21 CHOP combination 15 Doxorubicin monotherapy 4 CCNU/L-asparaginase/prednisone 1 Chlorambucil/prednisone 1 Abbreviations: CHOP, cyclophosphamide, doxorubicin, vincristine, prednisone (16); CCNU (lomustine), L-asparaginase, prednisone rescue protocol (16). *WHO staging for tumors in domestic animals (22). 3505 Clin Cancer Res 2009;15(10) May 15, 2009 www.aacrjournals.org Assessment of GS-9219 in Pet Dogs with Lymphoma Cancer Research. on January 14, 2018. © 2009 American Association for clincancerres.aacrjournals.org Downloaded from increase in the sums of the longest diameters of measurable involved nodes or newly arising lesions. Dogs that experienced progressive disease before completion of all GS-9219 cycles were removed from study and allowed to seek additional antitumor therapy at the discretion of the owner. Dogs that experienced relapse of their non-Hodgkin's lymphoma after completion of treatment with GS-9219 were offered either retreatment with GS-9219 monotherapy or other additional antitumor therapy at the discretion of the owner. Median first remission duration (FRD) was defined as the time from documented attainment of remission until relapse (in the case of complete responders) or the documentation of progression (in those dogs achieving only a partial remission). FLT-PET/CT. Nine dogs were evaluated by FLT-PET/CT before and after treatment based on scheduling and availability of scanner access. The FLT-PET/CT imaging studies were done on a clinical GE Discovery LS PET/CT scanner at University of Wisconsin hospitals and clinics. FLT was obtained from the cyclotron and radiopharmaceutical laboratory at the Department of Medical Physics at the University of WisconsinMadison. Approximately 6 mCi (220 MBq) of FLT activity was administered per scan. The whole-body FLT-PET/CT image acquisition with 10 min scan per bed position was initiated after 60 ± 10 min postinjection. All dogs were initially scanned in the preceding 24 h before their initial GS-9219 treatment and scheduled for subsequent scans 5 days (±1 day) following initial dose and 3 weeks following completion of five cycles of treatment (n = 4 dogs) or just before the fourth cycle of chemotherapy and 3 weeks following completion of five cycles of treatment (n = 5 dogs). For each study, the mean maximum body mass standardized uptake values for FLT were calculated. Amira (Mercury Computer Systems) was used to identify regions of interest and perform data analysis. Statistical analysis. Median FRD time for dogs in this study was determined from survival curves generated using the KaplanMeier method, which accounts for dogs that were alive, lost to follow-up, died from unrelated causes, or had not relapsed at the time of analysis. The log-rank test was used to determine differences in survival between groups analyzed. To compare response rates between groups, Fisher's exact test (two-tailed, 90% confidence interval) was done. A two-tailed paired nonparametric t test (Wilcoxon) was used to compare FLT-PET/CT-derived standardized uptake values before and after GS-9219 treatment. For all statistical analyses, P < 0.05 was considered significant.