Tirapazamine Plus Carboplatin and Paclitaxel in Advanced Malignant Solid Tumors: a California Cancer Consortium Phase I and Molecular Correlative Study

4356 Vol. 9, 4356–4362, October 1, 2003 Clinical Cancer Research

Tirapazamine Plus Carboplatin and Paclitaxel in Advanced Malignant Solid Tumors: A California Cancer Consortium Phase I and Molecular Correlative Study

Primo N. Lara, Jr.,1 Paul Frankel, stable disease in at least one evaluation; 10 quickly pro- Philip C. Mack, Paul H. Gumerlock, Irina Galvin, gressed. One complete responder had normalization of vas- Cynthia L. Martel, Jeff Longmate, cular endothelial growth factor and plasminogen activator inhibitor-1 levels. James H. Doroshow, Heinz Josef Lenz, Conclusion: Dose levels 3 (carboplatin AUC of 6, 225 Derick H. M. Lau, and David R. Gandara mg/m2 paclitaxel, and 330 mg/m2 tirapazamine) and 2 (car- California Cancer Consortium at the University of California Davis boplatin AUC 6, 225 mg/m2 paclitaxel, and 260 mg/m2 tira- Cancer Center, Sacramento, California 95817 [P. N. L., P. C. M., pazamine) are the maximum tolerated doses for chemother- P. H. G., I. G., C. L. M., D. H. M. L., D. R. G.]; City of Hope apy naive and patients treated previously, respectively. Dose Comprehensive Cancer Center, Duarte, CA [P. F., J. L., J. H. D.]; and the University of Southern California Norris Comprehensive Cancer level 3 is the experimental arm of a Phase III Southwest Center, Los Angeles, CA [H. J. L.] Oncology Group trial (S0003) in advanced non-small cell lung cancer. Potential markers of tumor hypoxia may be useful correlates in studies of hypoxic cytotoxins and are ABSTRACT being prospectively investigated in S0003. Purpose: Tumor hypoxia confers chemotherapy resistance. Tirapazamine is a cytotoxin that selectively targets INTRODUCTION hypoxic cells and has supra-additive toxicity with platinums Hypoxia affects Յ50% of human tumor cells and nega- and taxanes in preclinical studies. We conducted a Phase I tively influences the response to both radiation and standard study of tirapazamine, carboplatin, and paclitaxel and as- chemotherapeutic agents (1–6). For example, cisplatin had little sessed potential plasma markers of hypoxia as surrogates or no cytotoxic activity against hypoxic tumor cells in a pre- for response. clinical model despite killing a large proportion of nonhypoxic Experimental Design: Forty-two patients with advanced cells (7). Therefore, therapies that potentiate cytotoxicity against solid tumors were treated at four dose levels; parallel dose hypoxic cancer cells should theoretically increase the efficacy of escalations were carried out in chemotherapy-naive and platinum-based chemotherapy (8, 9). previously treated subjects. Pre and post-therapy plasma Hypoxic conditions alter cellular homeostasis, e.g., malig- levels of the hypoxia-induced proteins plasminogen activator nant cells subjected to hypoxia are induced to transcribe a inhibitor-1 and vascular endothelial growth factor were discrete set of genes (10), including the genes coding for PAI-1 measured. and VEGF2 (11). Hypoxia may then exert selection pressure on Results: Three of four chemotherapy-naı¨ve patients de- tumor populations by promoting angiogenesis, p53 mutations, veloped dose-limiting toxicities at dose level 4 (grade 3 sto- and resistance to apoptosis (12). Furthermore, tumor hypoxia matitis/infection, grade 3 emesis, and grade 4 febrile neu- has been shown to increase the metastatic potential of cancer tropenia). Four of seven previously treated patients cells, leading to the conclusion that hypoxia influences tumor developed dose-limiting toxicities at dose level 3, including phenotype as well as clinical outcome (13). This notion is one death [grade 3 myalgia, grade 3 infection/grade 4 neu- supported by the observation that tumor hypoxia (as measured tropenia, grade 3 infection/grade 4 neutropenia, and grade 5 by Eppendorf oxygen electrodes) is correlated with an increased infection (death)/grade 4 neutropenia]. Of 38 patients assess- incidence of distant metastases and treatment failure in a variety able for response, 3 had a complete response, 1 a partial of tumors, including soft tissue sarcomas, cervical cancer, and response, 1 an unconfirmed partial response, and 23 had head and neck carcinomas (14–16). Tirapazamine, the prototype for a novel class of anticancer compounds called “hypoxic cytotoxins,” is a benzotriazine com- pound that exhibits substantial differential toxicity for hypoxic Received 3/4/03; revised 6/3/03; accepted 6/9/03. cells. At equivalent tirapazamine concentrations, hypoxic mu- The costs of publication of this article were defrayed in part by the rine and/or human tumor cells were found to be Յ200 times 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. This work was supported in part by CA62505, CA63265, Bristol Myers Squibb (P. N. L.), the Hope Foundation (P. N. L.), and the American Cancer Society CRTG 0019701CCE (P. N. L.). 2 The abbreviations used are: VEGF, vascular endothelial growth factor; 1 To whom requests for reprints should be addressed, at University of AUC, area under the curve; NCI, National Cancer Institute; DLT, California Davis Cancer Center, 4501 X Street, Sacramento, CA 95817. dose-limiting toxicity; NSCLC, non-small cell lung cancer; IGFBP, Phone:(916) 734-3771;Fax:(916) 734-7946;E-mail:primo.lara@ucdmc. insulin-like growth factor-binding protein-3; MTD, maximum tolerated ucdavis.edu. dose.

Table 1 Dose escalation scheme pate. Patients who had previously received carboplatin and Dose Tirapazamine Carboplatin Paclitaxel paclitaxel concurrently were excluded. Patients who were preg- level (mg/m2) (AUC) (mg/m2) nant or who had unstable medical conditions (such as uncon- 1 260 6 200 trolled congestive heart failure) or other uncontrolled intercur- 2 260 6 225 rent illnesses precluding investigational therapy were also 3 330 6 225 excluded. Patients were required to use a medically acceptable 4 390 6 225 contraceptive throughout the treatment period and for 3 months after cessation of treatment with tirapazamine. This study was approved by the NCI-Cancer Therapy Evaluation Program and the institutional review boards of each California Cancer Con- more sensitive than aerobic cells to killing by tirapazamine (17). sortium site. All patients gave written informed consent to Preclinical studies have also tested the combination of tira- participate. pazamine with platinum-containing compounds in hypoxic tu- Baseline Evaluation. Before study entry, all patients un- mor models. Combinations of tirapazamine with either cisplatin derwent a complete history and physical examination. Baseline or carboplatin show enhanced antitumor activity in a number of imaging studies of all known sites of disease were obtained in vivo and in vitro models. In some models, the cytotoxicity of within 4 weeks of initial therapy. An audiometric evaluation was tirapazamine plus a platinum or a taxane is substantially greater required for all patients because of potential ototoxic effects of than the additive effects of both drugs. Studies in mice have tirapazamine. Laboratory studies included a complete blood demonstrated enhanced antitumor activity for the three-drug count with differential and platelet count, comprehensive met- regimen of tirapazamine, carboplatin, and paclitaxel versus car- abolic panel (which includes electrolytes, serum creatinine, total boplatin/paclitaxel in the MV-522 human lung carcinoma xe- bilirubin, and aspartate aminotransferase). Patient plasma was nograft model, forming the basis for this clinical study (18). collected before cycles 1 and 2 for molecular correlative studies The combination of carboplatin-paclitaxel has shown en- (see below). couraging activity in a number of different tumor types, includ- DLT. DLT was defined as any Grade 3 or 4 nonhema- ing NSCLC, and its toxicity profile has been well characterized. tological toxicity (including nausea/vomiting with maximal an- A recent Phase III trial (S9509) showed that paclitaxel-carbo- tiemetic prophylaxis), Grade 3/4 thrombocytopenia, and Grade platin has increased tolerability compared with vinorelbine- 4 neutropenia with fever (defined as a temperature of 38.3°C or cisplatin, prospectively defined as the ability to complete ther- 101°F) or a documented infection, occurring during the first apy and the percentage of patients taken off study attributable to course of treatment. Toxicity was graded according to the NCI toxicity (19). Furthermore, a recent Phase III trial in NSCLC of common toxicity criteria version 2.0. cisplatin with or without tirapazamine demonstrated improved MTD. The MTD was defined as the highest dose level at response and survival in patients treated with the combination which not Ͼ1 patient experienced a DLT, when Ն6 patients (20). Finally, a recently concluded Phase I trial demonstrated the were treated at that dose and were evaluable for toxicity. The feasibility of delivering a 3-h infusion of paclitaxel at 225 MTD was one dose level below the DLT level. Separate MTDs 2 2 2 mg/m with tirapazamine (390 mg/m ) and cisplatin (75 mg/m ) were determined for each stratified group (i.e., treated previ- every 3 weeks (21). In view of these preclinical and clinical ously and untreated previously). observations, we conducted a Phase I trial of tirapazamine plus Study Design and Treatment Plan. The dose escalation paclitaxel and carboplatin, administered on a single day, in schedule is summarized in Table 1. Study drugs were adminis- patients with advanced solid tumors, using therapeutic doses of tered once during each 3-week treatment cycle to each stratified paclitaxel and carboplatin and escalating doses of tirapazamine group according to the schedule outlined in Table 1. Each cycle over four dose levels (Table 1; Ref. 22). of treatment consisted of dosing on day 1 with follow-up through day 21. Tirapazamine was given i.v. before paclitaxel MATERIALS AND METHODS (as a 3-h infusion), which was followed by carboplatin. A Patient Selection. Patients with histologically or cyto- maximum of eight cycles was planned. Appropriate premedica- logically confirmed cancer not curable by standard means who tions were administered to prevent emesis, diarrhea, and hyper- had a life expectancy of Ͼ12 weeks and an adequate Zubrod sensitivity reactions. Treatment was continued in the absence of performance status (0, 1, or 2) were eligible. Patients were disease progression, provided that patients were clinically ben- stratified according to whether they had received previous efiting and tolerating the treatment. chemotherapy. Previous chemotherapy and/or radiotherapy Parallel dose escalations (three per cohort) were carried out must have been completed Ն4 weeks before study entry, and all in previously untreated patients and previously treated patients, treatment-related toxicities had to be completely resolved. Ad- but the dose level for the previously treated stratum was not equate hematological (WBC Ն 3,000/␮l, absolute neutrophil allowed to exceed that of the previously untreated stratum. Both count Ն 1,000/␮l, and platelet count Ն 100,000/␮l), renal strata were allowed to accrue to the same dose level simulta- (creatinine within institutional limits of normal and calculated neously. All patients treated at a particular dose level were creatinine clearance Ն 60 ml/min), and hepatic (serum biliru- observed for a minimum of 21 days after the start of the first bin Յ 1.5 mg/dl and aspartate aminotransferase within twice the course (i.e., the length of the treatment cycle) before accrual to institutional upper limit of normal) function were required. the next dose level was allowed. Because of possible tirapazamine-related ototoxicity, patients Patients received a minimum of one treatment cycle. Treat- with clinically severe hearing loss were not eligible to partici- ment continued in individual patients at the same dose level if

Table 2 Patient characteristics Table 3 Most common hematologic and nonhematologic toxicities (Grade 3 or 4): Cycle 1 Patients enrolled (number) 42 Untreated previously 26 Toxicity Grade 3 Grade 4 Total Treated previously 16 Neutrophils/granulocytes 12 23 35 Median age, years (range) 61 (41–80) Total WBC 17 7 24 Gender Lymphopenia 13 0 13 Male (%) 20 (48%) Anemia requiring transfusions 10 0 10 Performance status Thrombocytopenia 9 0 9 PS 0 17 Fatigue 8 0 8 PS 1 14 Nausea/vomiting 6 0 6 PS 2 11 Sensory neuropathy 6 0 6 Tumor types Infection 4 1 5 NSCLC 22 Myalgia 3 0 3 Unknown primary 9 Anorexia 3 0 3 Other 11 Febrile neutropenia 3 0 3 Bladder cancer (urothelial) 3 Small cell lung cancer 2 Esophageal, gastric, parotid, 1 each uterine, colon, and thyroid cancers No. previous chemo regimensa 2 (median) on the E-max microplate reader with Softmax software (Molec- a For patients in the previously treated stratum only. ular Devices, Sunnyvale, CA). All plasma samples were analyzed in duplicate in two independent ELISA experiments. Reported values are an average of all data points.

no DLT was observed. Although a maximum of eight treatment RESULTS cycles was planned, patients who were clinically benefiting Patient Demographics. Forty-two eligible patients with from treatment were allowed to continue therapy in the absence a median age of 61 years (range 41–80) were accrued between of unacceptable toxicity. There was no intra-patient dose esca- March 2000 and June 2002. Patient characteristics are summa- lation allowed. rized in Table 2. Twenty-six patients were chemotherapy naı¨ve, Before each treatment, a complete blood count and serum whereas 16 patients had been treated with at least one previous chemistries, including creatinine and liver function tests, were chemotherapy regimen. Patients treated previously had received obtained. Intercurrent toxicities were also assessed every 3 a median of two previous regimens before protocol therapy. weeks. Doses of carboplatin and paclitaxel were reduced by There were 20 males and 22 females. Most patients had a good 25% on subsequent cycles for neutropenic fever or a nadir performance status, with the majority having PS 0/1 (74%). The platelet count Ͻ25,000 during the previous cycle. Any nonhe- most common tumor types enrolled were NSCLC (22) and matological toxicity more than or equal to grade 3 required a carcinoma of unknown primary (9). Three patients had bladder reduction to the next lower dose level. Treatment was discon- cancer; 2 had small cell lung cancer; and 1 each had esophageal, tinued for evidence of disease progression, development of parotid, gastric, endometrial, colon, and thyroid cancer. unacceptable toxicities, or at the request of the patient or treating Dose Escalation and Toxicities. At dose level 4 of the physician. previouly untreated stratum, 3 of 4 simultaneously enrolled Patient Assessments. Patients were followed by serial patients developed DLTs: grade 3 emesis, grade 3 infection, and physical examinations, laboratory tests, and radiographic eval- febrile neutropenia with grade 3 infection. One patient in dose uations where appropriate. Patients completing at least one level 4 also developed complete hearing loss, which resolved treatment course were assessable for toxicity. Response was within 4 days. Subsequently, dose level 3 was identified as the assessed after every two cycles. Standard RECIST criteria for MTD for previoulsy untreated patients, and by protocol design, response and NCI common toxicity criteria (version 2.0) were no further accruals were allowed beyond dose level 3 for either used. patient stratum. A total of 16 previously untreated patients were Molecular Correlative Studies. Blood samples were treated at this expanded MTD level. Attributable grade 3 or 4 collected in EDTA tubes (purple tops) and stored on ice Յ1h hematological and nonhematological toxicities were as follows: before centrifugation (1200 rev/min for 10 min at 4°C). Plasma (a) grade 3 or 4 neutropenia (without infection) in 4 patients was collected, aliquoted, and stored at Ϫ80°C. Samples were (25%); (b) grade 3 nausea/emesis in 2 patients (12%); (c) grade thawed on ice 1 h before analysis. ELISAs were performed 3 dehydration in 2 patients (12%); and (d) one patient each using commercially available kits following the manufacturer’s developed grade 3 thrombocytopenia, grade 3 infection, grade 3 instructions: PAI-1 Imulyse kit (Biopool International, Ventura, hypotension, grade 3 rash, and grade 3 myalgia. In the previ- CA) and human VEGF (BioSource International, Inc., Ca- ously treated stratum, 3 of 6 evaluable patients had DLTs at dose marillo, CA). The PAI-1 kit recognizes active and latent PAI-1, level 3, grade 3 fatigue in the first patient, grade 3 myalgia with as well as uPA/PAI-1 complexes, with less efficiency. The grade 4 infection in the second, and grade 3 infection in the VEGF kit recognizes the natural and recombinant hVEGF-165 third. Dose level 2 was identified as the MTD level for this form, the most abundant of the five isoforms of VEGF. Plasma stratum where none of the six patients enrolled had a treatment- samples were analyzed undiluted or diluted up to 1:16 to fall related DLT (please refer to Tables 3 and 4 for toxicity sum- within the range of the standard curve. ELISA plates were read maries). The median number of cycles was 3 for both the

Table 4 Organ systems involved for attributable grade 3 or 4 toxicities Organ system Grade 3 Grade 4 Total Blood/bone marrow 14 25 39 Cardiovascular (Arrhythmia) 1 0 1 Cardiovascular (General) 1 1 2 Constitutional symptoms 8 0 8 Dermatology/skin 1 0 1 Gastrointestinal 7 0 7 Infection/febrile neutropenia 9 1 10 Liver 1 0 1 Metabolic/laboratory 3 1 4 Neurology 9 0 9 Pain 5 0 5

previously treated (range 1–6) and previously untreated cohorts (range 1–8). Efficacy. Evaluation of response rate was not a principal objective of this study, but objective responses were observed. Of the 38 patients assessable for response, complete responses were seen in 3 patients (one previously treated NSCLC patient and two chemotherapy-naı¨ve patients with an unknown primary cancer), and a partial response was observed in another patient with carcinoma of unknown primary, for an overall response rate of 10%. One other patient had an unconfirmed partial response. Twenty-three patients exhibited stable disease (including patients clinically stable after one radiographic disease as- Fig. 1 Patient plasma levels of PAI-1 and VEGF before and after sessment). therapy. Blood samples were collected from patients before initial Molecular Correlates. The goal of these correlative treatment (Pre-treat) and after one cycle of therapy (Post-treat); plasma was isolated, and levels of PAI-1 and VEGF were measured by ELISA. studies was to assess the feasibility of measuring changes in Specimens were analyzed in duplicate on at least two independent runs. hypoxia-induced gene products. Levels of PAI-1 and VEGF Results are an average of all data points. A, PAI-1 levels. Thirty-four were measured by ELISA in pretreatment plasma specimens pretreatment specimens and 14 post-treatment specimens were available from 34 and 36 patients, respectively (Fig. 1). PAI-1 levels for analysis. PAI-1 levels ranged from 8.3 to 455 ng/ml with a mean value of 112 ng/ml and a median value of 83.35 ng/ml. Normal range is ranged from 8.3 to 455 ng/ml with a mean value of 112 ng/ml Ͻ Յ considered 43 ng/ml. B, VEGF levels. Thirty-six pretreatment speci- and a median value of 83.35 ng/ml (normal value: 43 ng/ml). mens and 15 post-treatment specimens were available for analysis. VEGF levels ranged from undetectable to 291 pg/ml, with a VEGF levels ranged from undetectable to 291 pg/ml with a mean value mean value of 39.6 pg/ml and a median value of 17.35 pg/ml of 39.6 pg/ml and a median value of 17.35 pg/ml. Normal range is Յ (normal value: Յ120 pg/ml). In this limited and heterogeneous considered 120 pg/ml. sample set, no statistically significant correlations could be demonstrated between PAI-1 or VEGF plasma levels and clinical outcome or status of previous therapy. A positive correla- tion between PAI-1 and VEGF plasma levels was observed whereas the other patient with a complete response had a de- (P Ͻ 0.0001, Pearson r ϭ 0.5514), indicating that, in an indi- crease in PAI-1 but an increase in VEGF. The patient with a vidual patient, if one marker was elevated, the other marker partial response showed a decrease in PAI-1 with a minimal tended to be elevated as well. increase in VEGF. Of all of the patients, the one with progres- Plasma was also obtained after one cycle of therapy for sive disease had the highest proportional increase in PAI-1 measurement of levels of PAI-1 (in 14 patients) and VEGF (in levels and the second highest increase in VEGF levels. 15 patients). Post-treatment changes in PAI-1 and VEGF levels were assessed and compared with clinical responses. Compari- DISCUSSION sons of plasma PAI-1 and VEGF levels in 13 individual patients Hypoxic tumor cells have been shown to be resistant to before and after one cycle of therapy are shown in Fig. 2, A and radiotherapy and are variably resistant to conventional chemo- B, respectively. In this sample set, one patient had progressive therapeutic agents (23). The difference in radiation sensitivity disease while on therapy, one patient had a partial response, and between aerobic and hypoxic conditions, referred to as the 2 patients had a complete response. All others had stable disease oxygen enhancement ratio, is typically in the range of 2.5 to 3 during therapy. For both PAI-1 and VEGF, the patient with in human tumors. In the absence of oxygen, radiation-induced progressive disease showed a sharp increase in plasma concen- radicals which can cause DNA damage, may be reversed by trations after one cycle of therapy. One patient with a complete donation of hydrogen from nonprotein sulfhydryls, providing a response had a decrease in both PAI-1 and VEGF levels, mechanism for hypoxia-induced radioresistance (23). Although

animals, based on the minimal impact on body weight and the absence of any toxic deaths in the mice treated with tirapazamine-containing regimens. These encouraging preclinical data led to the development of the current study. This Phase I trial demonstrated the feasibility and tolerability of the combination of carboplatin and paclitaxel, both given at full cytotoxic doses, in combination with tirapazamine. The toxicities observed were those that would be expected with the individual agents, including stomatitis, emesis, neutropenia and infection, myalgia, and transient hearing loss. No unexpected toxicities occurred, and some objective responses were seen. The tirapazamine dose and toxicities observed in this study at the MTD level achieved compares favorably to those seen in other trials combining tirapazamine with chemotherapy and/or radiation, e.g., the MTD of tirapazamine in combination with cyclophosphamide or with cisplatin was 260 mg/m2 (24, 25). Subsequently, several Phase II and III studies have used the combination of tirapazamine with cisplatin (75 mg/m2 every 3 weeks) at doses of 260–390 mg/m2 with good tolerance (26–29). When used as a single agent in combination with radiation, tirapazamine has been well tolerated at doses of Յ260 mg/m2 thrice weekly (30). Tirapazamine has also been administered in combination with cisplatin and radiation at doses of Յ290 mg/m2/week during the weeks when cisplatin was given and 160 mg/m2 three times weekly during noncisplatin weeks (31). This study also examined the role of measurement of potential plasma markers of tumor hypoxia as surrogates for response. Preclinical microarray studies in a head and neck cancer cell line have demonstrated that a group of four genes with secreted products exhibited increased expression under Fig. 2 PAI-1 and VEGF plasma levels in matched pre and post- treatment samples. PAI-1 and VEGF levels before and after one cycle of hypoxic conditions: (a) VEGF; (b) IGFBP-3; (c) endothelin-2; therapy were charted in individual patients to determine whether and (d) PAI-1 (11). Interestingly, PAI-1 exhibited the greatest changes in plasma levels were predictive for clinical response. Bold level of induction in response to hypoxia. PAI-1 is a critical dotted lines, the patient with progressive disease; bold solid lines, the component of the normal hemostatic system and has already patients with objective clinical responses. All other patients had stable been associated with tumor invasion and metastasis (32). Fur- disease during treatment. A, PAI-1 levels; B, VEGF levels. thermore, elevated PAI-1 levels have been correlated with an inferior outcome in a variety of solid tumors, including NSCLC. IGFBP-3 is a pro-apoptotic gene that is transcriptionally up- these observations provided the basis for early studies attempt- regulated in response to hypoxia. However, studies by Koong et ing to improve radioresponsiveness by the use of hyperbaric al. (11) have failed to demonstrate evidence of apoptosis despite oxygen or radiosensitizing agents, such as misonidazole, neither induction of IGFBP-3, suggesting alternate mechanisms of approach has achieved general clinical application. In contrast, growth regulation in response to hypoxia. VEGF is a well- clinical studies with the novel hypoxic cytotoxin tirapazamine in characterized hypoxia-inducible, angiogenic cytokine that plays NSCLC have yielded promising results (18). a central role in tumor-related angiogenesis, a critical process in Weitman et al. (20) studied the interaction between tumor growth, invasion, and metastases (33). VEGF is abnor- tirapazamine and paclitaxel-carboplatin using the MV-522 mally overexpressed in a variety of tumor types, detectable in human lung carcinoma xenograft model in nude mice. As a patient serum or plasma, and correlates with worse prognosis single agent, tirapazamine did not yield any substantial anti- (34–36). Recently, it has also been demonstrated that osteopon- tumor activity in this human lung tumor model. However, a tin, an acidic glycoprotein important in bone formation, was substantial increase in tumor growth inhibition was seen in induced 10-fold in the serum of patients with pancreatic cancer, animals treated with the tirapazamine-paclitaxel-carboplatin a highly hypoxic tumor.3 Osteopontin induces endothelial cell triplet as compared with animals treated with paclitaxel- migration and also up-regulates the endothelial cell migration carboplatin alone. Specifically, the triplet yielded a 50% induced by VEGF (37). In one study, patients with stage I complete tumor response rate, whereas no complete responses were observed with the nontirapazamine-containing doublet. Interestingly, tirapazamine did not appear to increase the toxicity of the paclitaxel-carboplatin doublet in the 3 A. Koong, personal communication.

adenocarcinoma of the lung whose tumors expressed osteopon- 7. Grau, C., and Overgaard, J. Effect of cancer chemotherapy on the tin and VEGF were found to have a significantly worse prog- hypoxic fraction of a solid tumor using a local tumor control assay. nosis compared with those whose tumors did not express these Radiother. Oncol., 13: 301–309, 1988. markers (37). Unfortunately, changes in the plasma levels of 8. Kennedy, K. A., Teicher, B. A., Rockwell, S., and Sartorelli, A. C. The hypoxic tumor cell: a target for selective cancer chemotherapy. these secreted proteins after cytotoxic chemotherapy have not Biochem. Pharmacol., 29: 1–8, 1980. yet been fully investigated. 9. Vaupel, P., and Kallinowski, F. Tissue oxygenation of primary and In this trial, we measured plasma levels of two important xenotransplanted human tumors. Radiat. Res., 2: 707–712, 1987. markers of tumor hypoxia, VEGF and PAI-1, before and after 10. Semenza, G. L. Hypoxia inducible factor-1: oxygen homeostasis treatment. Meaningful statistical correlations between the levels and disease pathophysiology. Trends Mol. Med., 7: 351–435, 2001. of these markers and tumor response could not be obtained from 11. Koong, A. C., Denko, N. C., Hudson, K. M., Schindler, C., Swiersz, this limited data set. However, it was intriguing to note that all L., Koch, C., Evans, S., Ibrahim, H., Le, O. T., Terris, D. J., and Giaccia, 3 of the patients who had a measurable response to therapy had A. J. Candidate genes for the hypoxic tumor phenotype. Cancer Res., a decrease in the level of PAI-1, whereas the one patient with 60: 883–887, 2000. progressive disease for whom the levels of these markers are 12. Graeber, T. G., Osmanian, C., Jacks, T., Housman, D. E., Koch, C. J., Lowe, S. W., and Giaccia, A. J. Hypoxia mediated selection of cells available had a marked increase in both PAI-1 and VEGF after with diminished apoptotic potential in solid tumors. Nature (Lond), 379: one cycle of treatment. The possibility that changes in levels of 88–91, 1996. VEGF and PAI-1 after only one cycle of treatment may corre- 13. Rofstad, E. K., and Danielsen, T. Hypoxia induced metastasis late with eventual clinical response is a hypothesis that deserves of human melanoma cells: involvement of vascular endothelial exploration in future studies. growth factor-mediated angiogenesis. Br. J. Cancer, 80: 1697–1707, This Phase I trial formed the basis for SWOG 0003, a 1999. recently completed Phase III trial randomizing patients with 14. Brizel, D. M., Sibley, G. S., Prosnitz, L. R., Scher, R. L., and metastatic NSCLC to carboplatin (AUC 6) and paclitaxel (225 Dewhirst, M. W. Tumor hypoxia adversely affects the prognosis of 2 2 carcinoma of the head and neck. Int. J. Radiat. Oncol. Biol. Phys., 38: mg/m ) with or without tirapazamine (330 mg/m ). This trial 285–289, 1997. will determine, in a cooperative group setting, whether the 15. Brizel, D. M., Scully, S. P., Harrelson, J. M., Layfield, L. J., Bean, addition of tirapazamine to standard palliative chemotherapy in J. M., Prosnitz, L. R., and Dewhirst, M. W. Tumor oxygenation predicts advanced NSCLC will improve the dismal median survival (ϳ8 for the likelihood of distant metastases in human soft tissue sarcoma. months) associated with this disease. In addition, serial plasma Cancer Res., 56: 941–943, 1996. specimens will be obtained from patients accrued to this study to 16. Hockel, S., Schlenger, K., Vaupel, P., and Hockel, M. Association between tumor hypoxia and malignant progression in advanced cancer determine circulating osteopontin, PAI-1, and VEGF levels and of the uterine cervix. Cancer Res., 56: 4509–4515, 1996. to correlate changes in these levels with tumor response. This 17. Zeman, E. M., Brown, J. M., Lemmon, M. J., Hirst, V. K., and Lee, study should help clarify the role of determining plasma levels W. W. SR-4233: a new bioreductive agent with high selective toxicity of tumor hypoxia markers in predicting response to therapy. for hypoxic mammalian cells. Int. J. Radiat. Oncol. Biol. Phys., 12: 1239–1242, 1986. 18. Weitman, S., Mangold, G., Marty, J., Dexter, D., Hilsenbeck, S., ACKNOWLEDGMENTS Rake, J., Juniewicz, P., and Von Hoff, D. Evidence of enhanced in vivo We thank Tracey Baratta (City of Hope) for data management activity using tirapazamine with paclitaxel and paraplatin regimens against the MV-522 human lung cancer xenograft. Cancer Chemother. support and Christopher Ruel for statistical assistance. We also thank Pharmacol., 43: 402–408, 1999. Dr. Percy Ivey of the Cancer Therapy Evaluation Program of the NCI 19. Kelly, K., Crowley, J., Bunn, P. A. Jr., Presant, C. A., Grevstad, P. for her support. K., Moinpour, C. M., Ramsey, S. D., Wozniak, A. J., Weiss, G. R., Moore, D. E., Israel, V. K., Livingston, R. B., and Gandara, D. R. Randomized Phase III trial of paclitaxel plus carboplatin versus vinorel- REFERENCES bine plus cisplatin in the treatment of patients with advanced non–small- 1. Gray, L. H., Conger, A. D., and Ebert, M. The concentration of cell lung cancer: A Southwest oncology group trial. J. Clin. Oncol., 19: oxygen dissolved in tissues at the time of irradiation as a factor in 3210–3218, 2001. radiotherapy. Br. J. Radiol., 26: 638–648, 1953. 20. Von Pawel, J., Von Roemeling, R., Gatzemeier, U., Boyer, M., 2. Denekamp, J., Fowler, J., and Dische, S. The proportion of hypoxic Elisson, L. O., Clark, P., Talbot, D., Rey, A., Bulter, T. W., Hirsh, V., cells in a human tumor. Int. J. Radiat. Oncol. Biol. Phys., 2: 1227–1228, et al. Tirapazamine plus cisplatin versus cisplatin in advanced non– 1977. small-cell lung cancer: A report of the International CATAPULT I 3. Moulder, J. E., and Rockwell, S. Tumor hypoxia: its impact on Study Group. J. Clin. Oncol., 18: 1351–1359, 2000. cancer therapy. Cancer Metastasis Rev., 5: 313–341, 1987. 21. Ng, K., Treat, J., O’Dwyer, P., Friedland, D., and Miller, V. A. A 4. Vaupel, P. W., and Hockel, M. Oxygenation status of human tumors: phase I trial of the addition of paclitaxel to tirapazamine (tpz) and a reappraisal using computerized pO2 histography. In: P. W. Vaupel, cisplatin in patients with advanced non small cell lung cancer. Proc. Am. D. R. Kelleher, and M. Gunderoth (eds.), Tumor Oxygenation, pp. Soc. Clin. Oncol., 17: 496, 1998. 220–232. New York: Gustav Fischer, 1995. 22. Gandara, D. R., Lara, P. N., Lau, D. H. Raschko, J., Lenz, H. J., and 5. Brown, J. M. The hypoxic cell: a target for selective cancer therapy– Doroshow, J. H. Tirapazamine (TPZ) plus taxol/carboplatin (carbo): a eighteenth Bruce F. Cain Memorial Award lecture. Cancer Res., 59: California Cancer Consortium phase I trial and implications for SWOG 5863–5870, 1999. 0003. Proc. Am. Soc. Clin. Oncol., 20: 334a, 2001. 6. Gatenby, R. A., Coia, L. R., Richter, M. P., Katz, H., Moldofsky, P. 23. Gandara, D. R., Lara, P. N., Jr., Goldberg, Z., Le, Q. T., Mack, P. J., Engstrom, P., Brown, D. O., Brookland, R., and Broder, G. J. Oxygen C., Lau, D. H., and Gumerlock, P. H. Tirapazamine: prototype for a tension in human tumors: in vivo mapping using CT guided probes. novel class of therapeutic agents targeting tumor hypoxia. Semin. On- Radiology, 156: 211–214, 1985. col., 29: 102–109, 2002.

24. Hoff, P. M., Saad, E. D., Ravandi-Kashani, F., Czerny, E., and treatment of refractory solid tumors. Int. J. Radiat. Oncol. Biol. Phys., Pazdur, R. Phase I trial of i. v. administered tirapazamine plus cyclo- 44: 349–353, 1999. phosphamide. Anti-Cancer Drugs, 12: 499–503, 2001. 31. Craighead, P. S., Pearcey, R., and Stuart, G. A phase I/II evaluation 25. Johnson, C. A., Kilpatrick, D., von Roemeling, R., Langer, C., of tirapazamine administered intravenously concurrent with cisplatin Graham, M. A., Greenslade, D., Kennedy, G., Keenan, E., and and radiotherapy in women with locally advanced cervical cancer. Int. J. O’Dwyer, P. J. Phase I trial of tirapazamine in combination with Radiat. Oncol. Biol. Phys., 48: 791–795, 2000. cisplatin in a single dose every 3 weeks in patients with solid tumors. 32. Bajou, K., Noel, A., Gerard, R. D., Masson, V., Brunner, N., J. Clin. Oncol., 15: 773–780, 1997. Holst-Hansen, C., Skobe, M., Fusenig, N. E., Carmeliet, P., Collen, D., 26. Miller, V. A., Ng, K. K., Grant, S. C., Kindler, H., Pizzo, B., and Foidart, J. M. Absence of host plasminogen activator inhibitor 1 Heelan, R. T., von Roemeling, R., and Kris, M. G. Phase II study of the prevents cancer invasion and vascularization. Nat. Med., 4: 923–928, combination of the novel bioreductive agent, tirapazamine, with cispla- 1998. tin in patients with advanced non-small-cell lung cancer. Ann. Oncol., 8: 33. Pluda, J. M. Tumor associated angiogenesis: mechanisms, clinical 1269–1271, 1997. implications, and therapeutic strategies. Semin. Oncol., 24: 203–218, 27. Treat, J., Johnson, E., Langer, C., Belani, C., Haynes, B., Green- 1997. berg, R., Rodriguez, R., Drobins, P., Miller W Jr., Meehan, L., McKeon, A., Devin, J., von Roemeling, R., and Viallet, J. Tirapazamine with 34. Berger, D. P., Herbstrit, L., Dengler, W. A., Marme, D., Mertels- cisplatin in patients with advanced non-small cell lung cancer: a phase mann, R., and Fiebig, H. H. Vascular endothelial growth factor (VEGF) II study. J. Clin. Oncol., 16: 3524–3527, 1998. mRNA expression in human tumor models of different histologies. Ann. Oncol., 6: 817–825, 1995. 28. Bedikian, A. Y., Legha, S. S., Eton, O., Buzaid, A. C., Papado- poulos, N., Coates, S., Simmons, T., Neefe, J., and von Roemeling, 35. Nicol, D., Hi, S. I., Walsh, M., Teh, B., Thompson, L., Kennett, C., R. Phase II trial of tirapazamine combined with cisplatin in chemo- and Gotley, D. Vascular endothelial growth factor expression is in- therapy of advanced malignant melanoma. Ann. Oncol., 8: 363–367, creased in renal cell carcinoma. J. Urol., 157: 1482–1486, 1997. 1997. 36. Dosquet, C., Coudert, M. C., Lepage, E., Cabane, J., and Richard, 29. Bedikian, A. Y., Legha, S. S., Eton, O., Buzaid, A. C., Papadopou- F. Are angiogenic factors, cytokines, and soluble adhesion molecules los, N., Plager, C., McIntyre, S., and Viallet, J. Phase II trial of escalated prognostic factors in patients with renal cell carcinoma? Clin. Cancer dose of tirapazamine combined with cisplatin in advanced malignant Res., 3: 2451–2458, 1997. melanoma. Anticancer Drugs, 10: 735–739, 1999. 37. Shijubo, N., Uede, T., Kon, S., Maeda, M., Segawa, T., Imada, A., 30. Shulman, L. N., Buswell, L., Riese, N., Doherty, N., Loeffler, J. S., Hirasawa, M., and Abe, S. Vascular endothelial growth factor and von Roemeling, R. W., and Coleman, C. N. Phase I trial of the hypoxic osteopontin in stage I lung adenocarcinoma. Am. J. Respir. Crit. Care cell cytotoxin tirapazamine with concurrent radiation therapy in the Med., 160: 1269–1273, 1999.

Downloaded from clincancerres.aacrjournals.org on September 27, 2021. © 2003 American Association for Cancer Research. Tirapazamine Plus Carboplatin and Paclitaxel in Advanced Malignant Solid Tumors: A California Cancer Consortium Phase I and Molecular Correlative Study

Primo N. Lara, Jr., Paul Frankel, Philip C. Mack, et al.

Clin Cancer Res 2003;9:4356-4362.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/9/12/4356

Cited articles This article cites 35 articles, 9 of which you can access for free at: http://clincancerres.aacrjournals.org/content/9/12/4356.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/9/12/4356.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/9/12/4356. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.