Translation of anticancer efficacy from nonclinical models to the clinic.

Mouse cancer models have provided critical insights into tumor biology; however, clinical translation of these findings has been challenging. This perspective posits that factors impacting on successful translation start with limitations in capturing human cancer pathophysiology and end with challenges in generating robust translatable preclinical end points. A comprehensive approach that considers clinically relevant mouse models with both an integrated biomarker strategy and a complementary modeling and simulation effort will strengthen the current oncology drug development paradigm.

Mechanism-based pharmacokinetic/pharmacodynamic model for troxacitabine-induced neutropenia in cancer patients.

PURPOSES: The objective of this study was to develop a mechanism-based population pharmacokinetic/pharmacodynamic (PK/PD) model in describing troxacitabine-induced neutropenia in patients with cancer. METHODS: A total of 727 PK/PD samples from 31 patients with cancer were included in the analysis. A mechanism-based population PD model was developed to describe neutropenia and the final model consisted of (1) a drug-sensitive uncommitted progenitor cell compartment (2) three transit compartments, and (3) a circulating neutrophil compartment with feedback mechanism. The troxacitabine affected the proliferation of sensitive progenitor cells through an inhibitory E (max) model. The model parameters were estimated using the MCPEM algorithm that was implemented in a parallel computing platform consisting of a single computer equipped with a quad-core INTEL central processor unit. RESULTS AND CONCLUSIONS: The mechanism-based PK/PD model developed using parallelized MCPEM method adequately describes the complex relationship between the exposure and absolute neutrophil counts in troxacitabine-treated patients with cancer. The simulation results suggested that the less frequent dosing schedule of troxacitabine used currently in clinical studies was associated with less incidence of neutropenia compared to more frequent dosing schedule.

Mechanism-based receptor-binding model to describe the pharmacokinetic and pharmacodynamic of an anti-α5ß1 integrin monoclonal antibody (volociximab) in cancer patients.

PURPOSE: Volociximab is a chimeric IgG(4) that is being developed as a novel first-in-class anti-angiogenic, α(5)ß(1) integrin inhibitor for the treatment of solid tumors. A mechanism-based pharmacokinetic (PK)/pharmacodynamic (PD) model was developed to investigate the dynamic interaction between volociximab concentrations and free monocyte α(5)ß(1) integrin levels in cancer patients. METHODS: Twenty-one cancer patients from six dose cohorts (0.5, 1.0, 2.5, 5.0, 10, and 15 mg/kg) were included in the analysis. The fully integrated receptor-binding PK/PD model was developed and fit simultaneously to the PK/PD data. A Monte-Carlo parametric expectation-maximization method implement in S-ADAPT program was used to obtain estimates of population parameters and inter- and intra-subject variability. RESULTS: The PK/PD time profiles were well described by the model and the parameters were estimated with good precision. The model was used to simulate PK/PD time profiles for multiple dose regimens at various dose levels, and the results suggested that the monocyte α(5)ß(1) integrin binding was saturated (≤5% free) at week 16 in the majority of patients treated with volociximab doses ≥10 mg/kg IV every 2 weeks. CONCLUSIONS: The developed model is useful for anticipating the drug exposures and extent of volociximab binding to peripheral monocyte α(5)ß(1) integrin in untested regimens and for optimizing the design of future clinical trials.

A phase I and pharmacologic study of 9-aminocamptothecin administered as a 120-h infusion weekly to adult cancer patients.

PURPOSE: To define the toxicity profile and the recommended phase II doses of 9-aminocamptothecin (9-AC) administered as a weekly 120-h infusion. METHODS: 9-AC was administered over 120 h weekly to 55 adult cancer patients with solid tumors over doses ranging from 0.41 to 0.77 mg/m2 per day in a phase I and pharmacologic study. 9-AC formulated in dimethylacetamide/polyethylene glycol (DMA) was administered on a 3 of 4-week schedule, and the newer colloidal dispersion (CD) formulation was given on a 2 of 3-week schedule. RESULTS: Overall, 193 courses of therapy were administered over 122 dose levels. On the 3 of 4-week schedule, 9-AC DMA infused at > or = 0.6 mg/m2 per day for 120 h weekly produced dose-limiting neutropenia, thrombocytopenia, and diarrhea, or resulted in 1-2-week treatment delays. Shortening treatments to 2 of 3 weeks resulted in dose-limiting neutropenia and fatigue at infusion rates > 0.72 mg/m2 per day. The ratio of 9-AC lactone to total (carboxylate + lactone) drug plasma concentrations at steady-state was 0.15 +/- 0.07. Clinical toxicities and drug pharmacokinetics were not substantially different between the DMA and CD formulations. One objective response was observed in a patient with bladder cancer and minor responses were observed in patients with lung and colon cancers. Plasma area under the concentration versus time curve for 9-AC lactone modestly correlated with the degree of thrombocytopenia (r=0.51) using a sigmoid Emax pharmacodynamic model. CONCLUSION: The recommended phase II dose for the 9-AC DMA formulation is 0.48 mg/m2 per h over 120 h for 3 of 4 weeks and for the 9-AC CD formulation is 0.6 mg/m2 per day over 120 h for 2 of 3 weeks. Both regimens were well tolerated and feasible to administer.

Measurement of the novel antitumor agent 17-(allylamino)-17-demethoxygeldanamycin in human plasma by high-performance liquid chromatography.

A sensitive HPLC assay has been developed to determine the concentration of 17-(allylamino)-17-demethoxygeldanamycin (AAG) in human plasma over the concentration range of 12.5 to 2,500 nM (7.33 to 1,465 ng/mL). After the addition of 1,000 nM geldanamycin as the internal standard, 1 mL samples of human plasma were subjected to solid-phase extraction, via Bond-Elut C18 cartridges, followed by analysis using an isocratic reversed-phase HPLC assay with UV detection. A Phenomenex Kingsorb, 3 micron, C18, 150x4.60 mm column and a Phenomenex Security Guard pre-column, C18 (ODS, Octadecyl), were used to achieve separation. AAG and GM were monitored at 334 and 308 nm, respectively, on a Hewlett-Packard 1050 Diode-Array Detector. The mobile phase, run at a flow-rate of 1 mL/min, was composed of 50% (v/v) 25 mM sodium phosphate (pH 3.00) with 10 mM triethylamine and 50% acetonitrile. HPLC effectively resolved AAG with retention times of 14.60 +/- 0.54 min and the internal standard geldanamycin at 10.72+/-0.38 min (n = 15). This assay was able to measure plasma concentrations of AAG, the lower limit of quantitation being 12.5 nM, at a starting dose of 10 mg/m2 infused intravenously over 1 h in a Phase I clinical trial in adult patients with solid tumors.

Human plasma carboxylesterase and butyrylcholinesterase enzyme activity: correlations with SN-38 pharmacokinetics during a prolonged infusion of irinotecan.

PURPOSE: To characterize the relationships between human plasma irinotecan carboxylesterase-converting enzyme activity, caboxylesterase-mediated hydrolysis of p-nitrophenyl acetate (pNPA), and the butyrylcholinesterase-mediated hydrolysis of butyrylthiocholine in human plasma and to test the ability of these in vitro tests to predict the variability in SN-38 pharmacokinetics in adult patients during a prolonged infusion of irinotecan. METHODS: Individual plasma-converting enzyme activity was measured in 20 adult cancer patients participating in a pharmacokinetic and phase I clinical trial of a prolonged 96-h intravenous infusion of irinotecan. The pNPA and butyrylthiocholine hydrolysis in patient plasma was also assayed. RESULTS: The irinotecan carboxylesterase-converting enzyme in human plasma had a Vmax of 89.9 +/- 22.7 pmol/h per ml plasma and a Km of 207 +/- 56 microM (mean +/- SD, n = 3). The mean value of the specific activity of this enzyme in 20 adult cancer patients was 10.08 +/- 2.96 pmol/h per ml plasma ranging from 5.43 to 15.39 pmol/h per ml. The area-under-the-concentration-versus time curve (AUC) ratio of SN-38 to irinotecan (AUCSN-38/AUCCPT-11) was used to assess the relative SN-38 exposure to the active metabolite in individual patients. Pharmacokinetic variations in the relative exposure to SN-38 did not correlate with the measured carboxylesterase-converting enzyme activity nor with plasma butyrylcholinesterase activity in our patient population. However, it did correlate with the measured pNPA hydrolysis activity in patient plasma (r2 = 0.350, P = 0.0124, n = 18). CONCLUSIONS: Determination of patient plasma pNPA hydrolysis activity may have utility in predicting SN-38 pharmacokinetics during prolonged infusions of irinotecan.

Pharmacokinetics and pharmacodynamic effects of 5-fluorouracil given as a one-hour intravenous infusion.

PURPOSE: Clinical toxicity associated with 5-fluorouracil (5-FU) is related to the area under the plasma concentration-time curve (AUC). Recently, short-term infusions of 5-FU given over 30 or 60 min have been substituted for conventional "bolus" 5-FU given over 3-5 min in randomized clinical trials, but there are only limited pharmacokinetic data for these altered infusion durations. We therefore wished to determine the pharmacokinetics and toxicity associated with 5-FU given as a 1-h intravenous (i.v.) infusion. METHODS: A group of 22 adults with advanced gastrointestinal tract cancers and no prior systemic chemotherapy for advanced disease received interferon alpha-2a (5 MU/m2 s.c., days 1-7), leucovorin (500 mg/m2 i.v. over 30 min, days 2-6) and 5-FU (370 mg/m2 i.v. over 1 h, days 2-6). The doses of 5-FU and interferon-alpha were adjusted according to individual tolerance. The pharmacokinetics and clinical toxicity were retrospectively compared with patients receiving the same regimen under the same treatment guidelines except that 5-FU was given over 5 min. RESULTS: The regimen was well tolerated, and 41% of the patients tolerated 5-FU dose escalations to 425-560 mg/m2 per day. Grade 3 or worse diarrhea and fatigue ultimately occurred in 14% of the patients each. Granulocytopenia, mucositis, and diarrhea appeared to be appreciably milder in the present trial compared with our prior phase II experience in colorectal cancer. The peak 5-FU plasma levels and AUC with 370 mg/m2 5-FU given over 1 h were 7.3-fold and 2.4-fold lower than previously measured in 31 patients who received 5-FU over 5 min. CONCLUSION: Increasing the length of 5-FU infusion to 1 h seemed to substantially reduce the clinical toxicity with this modulated 5-FU regimen, likely due to markedly lower peak 5-FU plasma levels and AUC. Changes in the duration of a short infusion of 5-FU clearly affects the clinical toxicity, but raises the concern of a potentially adverse impact on its antitumor activity. These results suggest the importance of including precise guidelines concerning the time over which 5-FU is given in clinical trials. Having a specified duration of 5-FU infusion is also important if 5-FU dose escalation is considered.

N-(phosphonacetyl)-L-aspartate and calcium leucovorin modulation of fluorouracil administered by constant rate and circadian pattern of infusion over 72 hours in metastatic gastrointestinal adenocarcinoma.

BACKGROUND: We have reported that N-(phosphonacetyl)-L-aspartic acid (PALA) 1266 mg/m2 can safely be given 24 hours prior to the start of a 72-hour infusion of fluorouracil (FUra) and leucovorin (LV) at doses of 2000 and 500 mg/m2/day. Since inhibition of aspartate carbamoyltransferase (ACTase) activity was evident 4 hours post PALA, we wished to evaluate PALA given 1 hour prior to FUra. Further, we studied the toxicity and pharmacokinetics with FUra given by either fixed- or variable-rate infusion. PATIENTS AND METHODS: Twenty-seven patients with gastrointestinal tract adenocarcinomas were treated with PALA 1266 mg/m2/15 min followed by a 72-hour infusion of FUra and LV (1750 & 500 mg/m2/day) given by fixed- or variable-rate (peak at 4:00 A.M.). RESULTS: Clinical toxicity was similar in two consecutive cycles in 17 patients receiving fixed- and variable-rate infusion at the same FUra dose. Overall, grade 3 stomatitis and hand-foot syndrome occurred in 12% and 4% patients receiving fixed- and in 16% and 10.5% of patients receiving variable-rate infusions. Six of 24 evaluable patients (25%) had a partial response. The profile of FUra plasma levels (Cp) over a 24-hour period during fixed- and variable-rate infusions were strikingly different, but the average Cp and area under the concentration-time curves were comparable. ACTase activity was significantly decreased at 4 and 24 hours after PALA (12% and 18% of baseline; P < 0.001), but enzyme activity had recovered to 40% by 72 hours. CONCLUSIONS: This regimen was active and well tolerated with similar toxicities with FUra given by either fixed- or variable rate infusion. PALA 1266 mg/m2 significantly inhibited ACTase activity for at least 24 hours.

Phase I and pharmacokinetic trial of weekly oral fluorouracil given with eniluracil and low-dose leucovorin to patients with solid tumors.

PURPOSE: Fluorouracil (5-FU) given as a weekly, high-dose 24-hour infusion is active and tolerable. We evaluated an oral regimen of eniluracil (which inactivates dihydropyrimidine dehydrogenase [DPD]), 5-FU, and leucovorin to simulate this schedule. PATIENTS AND METHODS: Patients received a single 24-hour infusion of 5-FU (2,300 mg/m(2) on day 2) with leucovorin (15 mg orally [PO] bid on days 1 through 3) to provide reference pharmacokinetic data. Two weeks later, patients began treatment with eniluracil (20 mg) and leucovorin (15 mg) (PO bid on days 1 through 3) and 5-FU (10 to 15 mg/m(2) PO bid on day 2). RESULTS: Dose-limiting toxicity (diarrhea, neutropenia, and fatigue) was seen with 5-FU 15 mg/m(2) PO bid on day 2 given weekly for either 6 of 8 weeks or 3 of 4 weeks, whereas five of seven patients tolerated 5-FU 10 mg/m(2) PO bid given weekly for 3 of 4 weeks. Eniluracil led to a 35-fold reduction in 5-FU clearance. Fluoro-beta-alanine, a 5-FU catabolite, was not detected in plasma during oral 5-FU-eniluracil therapy. DPD activity was markedly suppressed in all patients during eniluracil therapy; the inactivation persisted after the last eniluracil dose; percentages of baseline values were 1.8% on day 5, 4.5% on day 12, and 23.6% on day 19. CONCLUSION: The recommended oral dosage of 5-FU (10 mg/m(2) PO bid) given with eniluracil and leucovorin is approximately 115-fold lower than the reference dosage for 24-hour infusional 5-FU. This difference is greater than expected given the reduction in 5-FU clearance. DPD inactivation persisted for several weeks after completion of eniluracil therapy.

Phase II trial of paclitaxel by 96-hour continuous infusion in combination with cisplatin for patients with advanced non-small cell lung cancer.

Our purpose was to determine the antitumor efficacy and safety profile of the combination of paclitaxel administered by 96-h continuous i.v. infusion followed by bolus cisplatin in patients with untreated advanced non-small cell lung cancer (NSCLC). Fifty-eight patients with untreated advanced or recurrent NSCLC were enrolled between October 1995 and December 1998. The median patient age was 60 years (age range, 34-75 years). Twenty-four patients were female. The majority of patients (n = 52) had an Eastern Cooperative Oncology Group performance status of 0/1. Twelve patients had stage IIIB NSCLC, 43 had stage IV disease, and 3 had recurrent disease after prior resection. Seven patients had received cranial irradiation for brain metastases, and 5 patients had received bone irradiation before enrollment. Patients were treated with paclitaxel (120 mg/m2/96 h) by continuous i.v. infusion followed by cisplatin (80 mg/m2) on day 5. Therapy was administered every 3 weeks as tolerated until disease progression or a maximum of six cycles. A total of 264 cycles of therapy were administered. Twenty-nine patients received all six cycles. Forty-six patients had measurable disease, with 20 patients achieving a partial response, and no complete responses were seen (overall response rate, 43%; 95% confidence interval, 29-60%). The median progression-free survival was 5.5 months. At a median potential follow-up of 27.2 months, the median survival for all 58 enrolled patients was 8.5 months, and the actuarial 1-year survival was 37% (95% confidence interval, 25.9-50.5%). This is the most extensive evaluation of prolonged continuous infusional paclitaxel in patients with advanced-stage cancer. In contrast to predictions from in vitro cytotoxicity models, the regimen does not appear to be obviously superior to shorter infusion times in the clinical setting. Additional trials of this regimen in patients with NSCLC are therefore of low priority.

Measurement of plasma uracil using gas chromatography-mass spectrometry in normal individuals and in patients receiving inhibitors of dihydropyrimidine dehydrogenase.

A sensitive gas chromatographic-mass spectrometric method is described for reliably measuring endogenous uracil in 100 microl of human plasma. Validation of this assay over a wide concentration range, 0.025 microM to 250 microM (0.0028 microg/ml to 28 microg/ml), allowed for the determination of plasma uracil in patients treated with agents such as eniluracil, an inhibitor of the pyrimidine catabolic enzyme, dihydropyrimidine dehydrogenase. Calibration standards were prepared in human plasma using the stable isotope, [15N2]uracil, to avoid interference from endogenous uracil and 10 microM 5-chlorouracil was added as the internal standard.

Phase I and pharmacologic study of irinotecan administered as a 96-hour infusion weekly to adult cancer patients.

PURPOSE: We conducted a phase I and pharmacologic study of a weekly 96-hour infusion of irinotecan to determine the maximum-tolerated dose, define the toxicity profile, and characterize the clinical pharmacology of irinotecan and its metabolites. PATIENTS AND METHODS: In 26 adult patients with solid tumors, the duration and dose rate of infusion were escalated in new patients until toxicity was observed. RESULTS: In 11 patients who were treated with irinotecan at 12.5 mg/m(2)/d for 4 days weekly for 2 of 3 weeks, dose-limiting grade 3 diarrhea occurred in three patients and grade 3 thrombocytopenia occurred in two patients. The recommended phase II dose is 10 mg/m(2)/d for 4 days given weekly for 2 of 3 weeks. At this dose, the steady-state plasma concentration (Css) of total SN-38 (the active metabolite of irinotecan) was 6.42 +/- 1.10 nmol/L, and the Css of total irinotecan was 28.60 +/- 17.78 nmol/L. No patient experienced grade 3 or 4 neutropenia during any cycle. All other toxicities were mild to moderate. The systemic exposure to SN-38 relative to irinotecan was greater than anticipated, with a molar ratio of the area under the concentration curve (AUC) of SN-38 to irinotecan of 0.24 +/- 0.08. One objective response lasting 12 months in duration was observed in a patient with metastatic colon cancer. CONCLUSION: The recommended phase II dose of irinotecan of 10 mg/m(2)/d for 4 days weekly for 2 of 3 weeks was extremely well tolerated. Further efficacy testing of this pharmacologic strategy of administering intermittent low doses of irinotecan is warranted.

The clinical development of 9-aminocamptothecin.

9-Aminocamptothecin (9-AC) is a topoisomerase I-targeting agent first synthesized by Wani and Wall in 1986. Because of its potent in vitro effects and promising preclinical activity in colorectal cancer animal models, it was designated a high-priority compound for further drug development by the NCI. In 1993, 9-AC first entered clinical trials as a 72-hour intravenous (i.v.) infusion. Predictable myelosuppression was the major dose-limiting toxicity, and pharmacokinetic studies showed a relatively short plasma half-life and unstable lactone ring. Unfortunately, phase II studies using this schedule showed minimal or no activity in tumors such as colorectal and lung cancer. Modest activity was observed in ovarian cancer and in refractory lymphomas. Efforts to improve systemic drug exposure by utilizing alternative schedules of administration of 9-AC such as prolonged, continuous intravenous infusions have also been tested. However, phase II studies of 120-hour weekly infusions of 9-AC have not shown improved activity against solid tumors such as colorectal cancer. More recently, a daily times 5 days i.v. administration schedule has been tested. Currently, further development of intravenously administered 9-AC for the treatment of colorectal cancer is not promising. Thus, topotecan and irinotecan remain the only two successfully developed topoisomerase I-targeting drugs in the United States. This experience with 9-AC raises important questions regarding how to best select new topoisomerase I-targeting drugs for future clinical development.

A Phase I study of raltitrexed, an antifolate thymidylate synthase inhibitor, in adult patients with advanced solid tumors.

The purpose of this study was to perform a Phase I trial of raltitrexed, a selective inhibitor of thymidylate synthase, and to determine the pharmacokinetic and toxicity profiles as a function of raltitrexed dose. Fifty patients with advanced solid tumors and good performance status were treated with raltitrexed as a 15-min i.v. infusion every 3 weeks, at doses escalating from 0.6 to 4.5 mg/m2. Asthenia, neutropenia, and hepatic toxicity were the most common dose-limiting toxicities in this largely pretreated patient population, but they occurred during the initial cycle in only one of nine patients treated with 4.0 mg/m2 and in two of nine patients treated with 4.5 mg/m2. Only 2 of 13 patients treated with 3.5 mg/m2 ultimately experienced unacceptable toxicity after three and seven cycles, compared with 42 and 56% of patients receiving 4.0 and 4.5 mg/m2 after medians of three and two cycles, respectively. The maximum raltitrexed plasma concentration and the area under the plasma concentration-time curve increased in proportion to dose. Raltitrexed clearance was independent of dose and was associated with the estimated creatinine clearance. Asthenia, neutropenia, and hepatic transaminitis were dose-related and tended to occur more frequently when patients received three or more cycles of therapy. A 3-week treatment interval was feasible in the majority of patients at all doses. Although 4.0 mg/m2 appeared to be a safe starting dose in this pretreated patient population, about half who received two or more courses ultimately experienced dose-limiting toxicity. A dose of 3.5 mg/m2 was well tolerated in most patients.

High inter- and intrapatient variation in 5-fluorouracil plasma concentrations during a prolonged drug infusion.

The purpose of the study was to examine inter- and intrapatient variation in 5-fluorouracil (5-FU) plasma concentrations in adult cancer patients receiving a 3-day drug infusion. Fourteen patients received 1266 mg/m2 N-(phosphonacetyl)-L-aspartate (PALA) infused i.v. over 15 min on day 1, followed immediately by a loading dose of 500 mg/m2 calcium leucovorin over 30 min. Then a prolonged infusion of leucovorin at 500 mg/m2/day and 5-FU at 1750 mg/m2/day was administered as either a constant rate or as a circadian infusion over 72 h. During constant rate infusions, 5-FU concentrations within individuals varied by 1.7-fold, but no uniform time of peak or trough concentration was observed. Transformation of these data by setting the time of peak to 0 h and by expressing concentrations as the percentage of the 24-h mean value revealed a nonrandom distribution of the time from peak to trough with a median time of 12 h (P = 0.027). These transformed data were also successfully fit to a circadian cosinor function (P < 0.001). During multiple constant rate 5-FU infusions, the intrapatient variability was high; the times of peak 5-FU concentration occurred at the same approximate sampling time 43% of the time, and troughs coincided 17% of the time. No difference in clinical toxicity was observed when matched constant rate and circadian infusions of 5-FU were compared. High inter- and intrapatient variability exists in 5-FU plasma concentrations in adult cancer patients during constant rate infusions with no evidence of a consistent circadian rhythm in untransformed data.