Fatal Hepatic Veno-Occlusive Disease in a Phase I Study of Mylotarg and Troxatyl in Patients with Refractory Acute Myeloid Leukemia or Myelodysplastic Syndrome

Abstract

Mylotarg is a calicheamicin-conjugated humanized anti-CD33 monoclonal antibody. It is approved for therapy in patients older than 60 years with a first relapse of acute myeloid leukemia (AML) who are considered to be unsuitable for standard cytotoxic therapy [1]. Calicheamicin is an enediyne antibiotic antibiotic that binds to DNA and causes sequence-selective oxidation of deoxyribose. Both free and conjugated calicheamicin caused liver toxicity in preclinical testing of Mylotarg [1]. The approved Mylotarg regimen is 9 mg/m2 infused intravenously over 4 h on days 1 and 14 [1]. Mylotarg is associated with an incidence of approximately 20% of grade 3 or 4 hyperbilirubinemia and liver transaminitis [2]. Hepatic veno-occlusive disease (VOD) has been reported both in patients undergoing stem cell transplantation (SCT) following prior Mylotarg therapy [3], and in those receiving Mylotarg following SCT [4]. While VOD in the non-SCT setting was previously a vary rare event, it has become associated with Mylotarg therapy [5, 6]. No therapy has been approved for the prevention or treatment of VOD [7].Troxatyl is a nucleoside L-enantiomer with significant antileukemic activity in patients with refractory myeloid leukemias [8, 9, 10]. Skin rashes, stomatitis, and hand-foot syndrome are the major extramedullary toxicities with either single-agent administration or in combinations of Troxatyl with cytarabine, idarubicin, or topotecan [8, 10, 11].Based on the single-agent antileukemic activity of both Mylotarg and Troxatyl, differing mechanisms of action, and apparently nonoverlapping extramedullary toxicities, a phase I study of Mylotarg and Troxatyl in patients with refractory AML or MDS was conducted.Patients 55 years of age or older, in first relapse of AML or advanced MDS (RAEB, RAEB-T) with CD33 expression on 20% or more of bone marrow blasts were eligible if they had an ECOG performance status of <2, total bilirubin ≤2 mg/dl, AST or ALT <3 times the upper limit of normal, and serum creatinine ≤1.5 mg/dl. Patients were excluded if they had had prior treatment with Mylotarg or Troxatyl, or prior SCT. No minimal duration of complete remission (CR) was required. The Institutional Review Board approved the study. All patients gave written informed consent. The initial regimen was Troxatyl 4 mg/m2 infused intravenously over 30 min daily for 5 days, and Mylotarg 9 mg/m2 infused intravenously over 4 h on days 1 and 14. VOD was diagnosed using the Seattle and Baltimore standard criteria as hyperbilirubinemia (bilirubin >2 mg/dl) with one or more of the following: painful hepatomegaly or fluid retention, e.g. ascites or sudden weight gain (>5% of pretreatment weight).Details of patients participating in the study are shown in table 1. The first patient treated died of VOD on day 35 of therapy. Five patients were then treated on an altered regimen of Mylotarg 9 mg/m2 on day 1 only and Troxatyl 4 mg/m2 on days 1–5. Three of 4 evaluable patients (1 MDS, 2 AML) achieved complete remissions without platelet recovery (CRp), mean Crp duration was 6 weeks (range 4–8). The 6th patient treated during the study died of VOD on day 14 of therapy. Two patients developed skin rashes (1 each grade 2 and 3), which rapidly responded to oral steroid therapy. One patient developed grade 4 hepatic transaminitis, which resolved completely.Mylotarg is associated with the development of fatal VOD in patients who have not received prior SCT [5]. In a recent report, the incidence of VOD in 119 patients with myeloid leukemia receiving Mylotarg-containing non-SCT regimens was assessed [5]. The cohort included 92 AML patients and 25 advanced MDS patients; 61 had had no prior treatment and 56 had relapsed. Fourteen (12%) patients developed VOD. The diagnosis of VOD was supported by histology obtained by transjugular liver biopsy in 2 patients and radiological examinations in a further 10 patients. Four (29%) of 14 patients with VOD had received no prior cytotoxic therapy, including 2 patients who received single-agent Mylotarg therapy.Sepsis may both cause some of the diagnostic components of VOD and accompany the hepatic failure and ascites associated with VOD. As in a previous series, the cardinal clinical feature in both patients who developed overt VOD in this study was the abrupt onset of very significant weight gain, associated with right upper quadrant pain, ascites, and abdominal distension. Alternative explanations for this symptom complex such as acute hepatitis, progressive systemic or abdominal sepsis, pancreatitis, and bowel obstruction were not found in either patient at the time of VOD diagnosis. VOD was considered to be the major cause of death in both patients; VOD was considered to be the major cause of death where there was no evidence of concomitant other, possibly lethal factors. VOD was confirmed by liver biopsy in the second patient (fig. 1).In a phase I study of Troxatyl, 3 CRs and 1 partial remission were observed in 30 patients with advanced AML, a CRp in 1 patient with MDS, and a return to chronic-phase disease in 1 patient with chronic myelogenous leukemia (CML-BP) [10]. The recommended phase II single-agent dose for further studies in these patients was defined as 8 mg/m2 daily for 5 days [10]. In a subsequent phase II study, 6 of 16 evaluable patients (37%) with CML-BP achieved a return to chronic-phase disease [8]. In these studies, stomatitis, hand-foot syndrome and skin rash were observed as extramedullary grade 3 or 4 toxicities [8, 10]. No grade 3 or 4 hepatitis or VOD has been observed in patients receiving single-agent Troxatyl therapy for either hematological malignancies or solid tumors [8, 10, 11]. No VOD has been observed in a cohort of 80 patients with refractory AML, MDS, or CML-BP receiving Troxatyl combined with ara-C, topotecan or idarubicin in a multicenter phase I/II randomized study. As neither VOD nor severe hepatotoxicity has been associated with Troxatyl therapy and both are associated with Mylotarg, whether given alone or in combination with other agents [2, 4, 5, 6], it seems likely that Mylotarg was primarily responsible for these toxicities in the current study. Mylotarg consists of a humanized anti-CD33 antibody (hP67.6) linked to N-acetyl-gamma calicheamicin DMH, a potent enediyne antitumor antibiotic [1]. Mylotarg targets CD33-expressing cells by binding the CD33 antigen. After internalization, the calicheamicin derivative is released from the antibody by acid hydrolysis, and a reactive intermediate of the calicheamicin derivative is formed through reduction by glutathione. After binding to the minor groove of the DNA duplex, thereby positioning the enediyne ring so as to abstract hydrogen atoms from partner strands, the calicheamicin metabolite makes double-strand breaks in DNA at the oligopyrimidine-oligopurine tracts. Due to its cytotoxicity (at least 1,000 times that of doxorubicin), no human studies with calicheamicin or its metabolites have been performed. Calicheamicin metabolites are detectable in the serum of patients receiving Mylotarg therapy [12]. As VOD has not been reported in association with treatments using other anti-CD33 antibodies, calicheamicin probably mediates this toxicity, possibly by causing damage to hepatic sinusoidal endothelial cells [7]. Selective binding of Mylotarg to hepatic stromal cells and/or local concentration of Mylotarg within the liver may occur. Blakolmer et al. [13]have investigated the expression of hematopoietic stem cell markers, including CD33, in fetal liver during various stages of development and reported that CD33 is expressed in both ductal plate cells and in bile duct cells. Further studies of Mylotarg therapy should incorporate close monitoring for VOD.