Targeted drug delivery by gemtuzumab ozogamicin: mechanism-based pharmacokinetic model for treatment strategy improvement and prediction of individual responses.
E. Jager(1), V. H.J. van der Velden(2), J. G. te Marvelde(2), R. B. Walter(3,4), Z. Agur(1,6), V. Vainstein(1,5).
(1) Institute for Medical BioMathematics, Bene Ataroth, Israel; (2) Department of Immunology, Erasmus MC, University Medical Center Rotterdam, The Netherlands; (3) Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA; (4) Department of Medicine, Division of Hematology, University of Washington, Seattle, WA, USA; (5) Department of Hematology, Hadassah Medical Center, Jerusalem, Israel; (6) Optimata Ltd, Ramat Gan, Israel.
Objectives: In this work, we present the analysis of a general mechanism-based PK model for a conjugated mAb-based drug using experimental and clinical data of Gemtuzumab Ozogamicin (GO) interactions with leukemic blasts. GO is a anti-CD33 monoclonal antibody conjugated to chemotherapeutic agent calicheamicin. The main objectives of the study were, firstly, to evaluate individual parameter values of blast-drug interactions in AML patients and determination of their relative significance for the response to treatment, and, secondly, to propose optimized strategies of GO combined with other cytoreductive chemotherapeutics for future clinical trials.
Methods: We implemented a general mechanism-based PK model of monoclonal antibody-target cell population interaction to the data from relapsed AML patients who participated in phase II clinical trials of GO monotherapy [1,2,3]. All computer simulations were run on a PC computer with Matlab 7.0 software. Peripheral blood was drawn at 0, 3 and 6 hour points after initiation of a 2-hour intravenous infusion of GO at a dose of 9 mg/m2. Additionally, AML193 cells were incubated in vitro with various GO doses for 15, 60, 180 and 360 minutes. Patients' blasts and AML193 cells were examined by flow cytometry to determine numbers of free and GO-bound CD33 molecules on the cell surface. Efflux of GO from leukemic blast cells was previously estimated using dye efflux essay [3]. The resultant data were used for estimation of the PK model parameters in the population individual patients. The model allowed for estimation of the intracellular drug content, as a surrogate for drug efficacy. Different GO treatment schedules were then simulated.
Results: In vitro data on AML193 cells enabled determination of CD33-GO association and dissociation rates. Forty seven of 276 patients who participated in the GO clinical trial had both data on drug efflux and CD33 saturation and therefore could be analyzed by the model to determine parameters of CD33 production and internalization rates of free and bound CD33. The model can successfully retrieve the differences in blood concentration profiles of GO in the leukemic patients during the first versus the second infusions as well as to estimate non-invasively the leukemic blast burden. Parameter sensitivity analysis show that low blast burden, intermediate CD33 antigen production rate, and low drug efflux are key characteristics that determine high intracellular GO exposure according to model calculations, all other parameters being much less influential. Moreover, even a modest blast burden reduction may significantly increase intracellular GO exposure and allow reduction in the dose of GO. EORTC-GIMEMA AML-19 Trial (published at 2008 ASH meeting) showed that the reduced GO dose (6 mg/m2) is effective, as well as administration at days 1 and 7 is more efficacious than at days 1, 3 and 5. These results match fully our model simulations, constituting additional line of its validation.
Conclusions: Our results suggest that GO efficacy could be enhanced when used after the leukemic tumor burden was lowered, e.g by alternative cytoreductive agents. Furthermore, the presented model proposes optimization of GO administration schedule as verified in an independent clinical trial. Additionally, we suggest that estimation of CD33 production and drug efflux in blasts of individual AML patients can better define the population with the most susceptible disease.
References:
[1] van Der Velden et al. Targeting of the CD33-calicheamicin immunoconjugate Mylotarg (CMA-676) in acute myeloid leukemia: in vivo and in vitro saturation and internalization by leukemic and normal myeloid cells. Blood 2001;97:3197-204
[2] van der Velden et al. High CD33-antigen loads in peripheral blood limit the efficacy of gemtuzumab ozogamicin (Mylotarg) treatment in acute myeloid leukemia patients. Leukemia 2004;18:983-8
[3] Walter, R.B. et al. CD33 expression and P-glycoprotein-mediated drug efflux inversely correlate and predict clinical outcome in patients with acute myeloid leukemia treated with gemtuzumab ozogamicin monotherapy. Blood 2007;109:4168-70.