Phase 2 trial of imatinib mesylate in myelofibrosis with myeloid metaplasia

Myelofibrosis with myeloid metaplasia (MMM) is a clonal stem cell disorder that is characterized by anemia, marked splenomegaly, bone marrow fibrosis, and extramedullary hematopoiesis.1 Median survival time is estimated at 5 years, and causes of death include transformation to acute myeloid leukemia and consequences of progressive cytopenia and cachexia. Conventional treatment provides symptomatic improvement in less than one third of the patients and does not improve survival. Promising investigational treatments include allogeneic hematopoietic stem cell transplantation,2 autologous hematopoietic stem cell transplantation,3 and drug therapy with thalidomide4 or etanercept.5 

Although the clonal nature of MMM is well established, the pathogenesis of the florid bone marrow stromal reaction that includes collagen fibrosis, neo-angiogenesis,6 and osteosclerosis is poorly understood. Current information suggests a pivotal role for megakaryocyte–monocyte-derived cytokines, such as platelet-derived growth factor (PDGF), transforming growth factor-β, and basic fibroblast growth factor, all of which are angiogenic and fibrogenic.7 This cytokine-mediated bone marrow stromal reaction may contribute to clinical phenotype and disease progression.

Imatinib mesylate is an orally bioavailable, 2-phenylaminopyrimidine derivative that inhibits bcr/abl and related kinases.8This drug has resulted in impressive hematologic and cytogenetic response rates in chronic-phase chronic myeloid leukemia.9In vitro, imatinib mesylate inhibits the tyrosine kinase activities of c-kit,10 ARG kinase,11 and the PDGF receptor.12 In addition, imatinib mesylate inhibits c-kit or PDGF receptor–dependent cell proliferation ex vivo and in vivo.13,14 Because PDGF and c-kit15 have been implicated in the pathogenesis of MMM, we evaluated the efficacy of imatinib mesylate in this disorder.

Previously published criteria were used for the diagnosis of MMM.1,16 In addition to a pretreatment history and physical examination, all study patients underwent pretreatment bone marrow examination with cytogenetic and fluorescence in situ hybridization studies. Oral imatinib mesylate was administered as single-agent therapy according to a protocol approved by the Mayo Clinic Institutional Review Board. All patients were started at a dose of 400 mg/d. The National Cancer Institute–designed Common Toxicity Criteria Version 2.0 were used to grade toxicity. Imatinib mesylate was held in the event of grade 3 or greater toxicity in neutropenia or thrombocytopenia and grade 2 or greater toxicity in other parameters. On complete resolution of side effects, the drug was restarted at 200 mg/d. Recurrence of side effects at the reduced drug dose level required withdrawal from the study.

The effects of imatinib mesylate on peripheral blood myeloid and erythroid colony formation were evaluated in 19 of the 23 study patients. Twenty milliliters peripheral blood (in EDTA) was obtained from each patient before starting protocol treatment. Peripheral blood mononuclear cells (PBMCs) were obtained through Ficoll-Hypaque density centrifugation, washed in Dulbecco phosphate-buffered saline, and resuspended in Iscoves modified Dulbecco medium. Isolated PBMCs were then plated in growth factor–containing methylcellulose medium (Methocult; StemCell Technologies; Vancouver, BC, Canada) in diluent (0.1% dimethyl sulfoxide) containing 0, 3.125, 6.25, 12.5, or 25 μM imatinib mesylate. After incubation for 14 days at 37°C in 5% CO₂, myeloid (CFU-GM) and erythroid (CFU-E, BFU-E) colonies were counted by light microscopy using morphologic criteria established by the manufacturer.

Between May and October 2001, 23 patients (median age, 63 years; range, 37-78 years) with MMM were treated with imatinib mesylate at an initial dose of 400 mg/d. Table 1outlines the pretreatment clinical and laboratory characteristics of the patients. Seven patients were in a high-risk prognostic category,17 9 were red blood cell transfusion–dependent, 11 had substantial constitutional symptoms, and 8 had been previously treated with chemotherapy. None of the patients underwent splenectomy, and the median palpable spleen size was 10 cm (range, 0-25 cm) below the left costal margin. Fifteen (65%) patients had clonal cytogenetic abnormalities. Fluorescence in situ hybridization studies did not reveal an abnormal bcr/abl fusion signal in any patient.

Initial treatment with imatinib mesylate at 400 mg/d was held in 16 (70%) patients after 1 to 12 weeks because of side effects. Severe to moderate neutropenia (absolute neutrophil count less than 1 × 10⁹/L) occurred in 6 patients at a median of 25 days (range, 15-75 days) from the initiation of treatment. Interestingly, this side effect was observed in 6 of 10 patients with pretreatment white blood cell counts (WBCs) of less than 5 × 10⁹/L but not in any of 13 patients with baseline WBCs of more than 5 × 10⁹/L. Drug-induced neutropenia resolved in all patients after treatment discontinuation. In contrast to CML, in which thrombocytopenia is observed with imatinib mesylate,9 a greater than 50% increase in platelet count was documented in 11 (48%) patients. Of these patients, 4 had thrombocytosis, with platelet counts (× 10⁹/L) increasing from 436 to 1086 in patient 2, 246 to 610 in patient 4, 848 to 1517 in patient 12, and 343 to 604 in patient 20. Unfortunately, none of the platelet increases occurred in patients with pretreatment platelet counts below 100. In addition to these hematologic side effects, nonhematologic side effects included limb pain or exacerbation of pre-existing joint and muscle pain (5 patients [patients 7, 15, 16, 18, 22]), peripheral edema (3 patients [patients 3, 12, 13]), diarrhea (1 patient [patient 6]), and hyperbilirubinemia (1 patient [patient 6]). Fluid retention and musculoskeletal pain occurred in only those patients with a history of the same.

Protocol treatment was restarted at the reduced dose of 200 mg/d in 12 of the 16 patients whose treatment was held because of side effects. The aforementioned side effects recurred and necessitated permanent cessation of treatment in 9 patients. Overall, 11 (48%) patients, including those in whom treatment rechallenge at 50% dose reduction was possible, were able to continue treatment beyond 3 months (7 at 400 mg/d and 4 at 200 mg/d). Of these, 5 patients discontinued treatment after 3 months because of side effects or personal decision. The remaining 6 patients either have completed the scheduled 6 months of treatment or are currently on therapy (3 patients in each group). To date, none of the patients has experienced a response in anemia, and only 2 patients have obtained a greater than 50% reduction in spleen size (patients 4 and 10). Posttreatment bone marrow specimens were available in 5 patients and showed no change from pretreatment findings.

To determine whether the hematologic side effects of imatinib mesylate therapy were predicted from in vitro assays, we examined the effects of imatinib mesylate on colony formation by circulating myeloid (Figure1A) or erythroid (Figure 1B) progenitor cells. As a basis for comparison, the peak and trough imatinib mesylate concentrations are estimated to be 4.6 and 1.5 μM, respectively, in patients receiving 400 mg/d.9 In these assays, the drug demonstrated variable degrees of growth suppression of myeloid and erythroid progenitors that were more pronounced at higher than therapeutic concentrations. There was no correlation between in vitro drug activity and in vivo clinical effects.

The current study reveals a high incidence of imatinib mesylate toxicity in MMM and identifies a patient population at risk for specific side effects. The prevalence of side effects and the strict protocol requirements may have masked potential activity at higher dose levels. The observed effect of the drug on platelet count was inconsistent with the stimulatory effect of PDGF on bone marrow stromal production of thrombopoietin18 and the expansion of megakaryocyte progenitors.19 Unfortunately, these effects of imatinib mesylate were not accompanied by any discernible clinical benefit. However, the current study does not rule out the possibility of a favorable drug effect in the context of longer treatment or imatinib mesylate–based combination therapy.