Pivotal phase II studies in acute myeloblastic leukemia (AML) patients in first relapse have used gemtuzumab ozogamicin (GO) (Mylotarg) at a dose of 9 mg/m2 on days 1 and 14. These studies showed a 26% response rate (13% complete remission (CR) and 13% CRp (complete remission with incomplete platelet recovery)) but with high degree of hematological and liver toxicities. Based on in vitro studies showing a re-expression of CD33 antigenic sites on the cell surface of blasts cells after exposure to GO, we hypothesized that fractionated doses of GO may be efficient and better tolerated. Fifty-seven patients with AML in first relapse received GO at a dose of 3 mg/m2 on days 1, 4 and 7 for one course. Fifteen patients (26%) achieved CR and four (7%) CRp. Remission rate correlated strongly with P-glycoprotein and MRP1 activities. The median relapse-free survival was 11 months, similar for CR or CRp patients. Median duration of neutropenia <500/μl and thrombocytopenia <50 000/μl were, respectively, 23 and 21 days. No grade 3 or 4 liver toxicity was observed. No veno-occlusive disease occurred after GO or after hematopoietic stem cell transplantation given after GO in seven patients. Mylotarg administered in fractionated doses demonstrated an excellent efficacy/safety profile.
Gemtuzumab ozogamicin (GO) (Mylotarg) is an antibody targeted chemotherapy agent, composed of recombinant humanized antibody conjugated to a derivative of the antitumor antibiotic calicheamicin that binds to the CD33 antigen. The CD33 antigen is a cell surface protein expressed on the surface of leukemic blasts in more than 90% of patients with acute myeloid leukemia (AML). The antigen is not expressed on pluripotent stem cells.1, 2 Pharmacodynamic studies during phase 1 trial showed that the dose of 9 mg/m2 of GO was expected to be the best dose to provide optimal saturation of CD33 binding sites regardless of leukemic burden, although lower doses showed efficacy.3 The dose of 9 mg/m2 was selected for phase 2 pivotal studies and, as pharmacokinetic studies showed that the elimination half-lives of total and unconjugated calicheamicin were about 41 and 143 h, respectively, the dose was repeated at day 14.
The overall response rate (OR) for the 277 adult patients with AML (mean age: 58 years) in first relapse treated in phase 2 pivotal studies was 26% with 13% complete remission (CR) and 13% CRp (complete remission with incomplete platelet recovery).4 However, this regimen led to long duration of neutropenia and thrombocytopenia, and grade 3 and 4 liver function test abnormalities in 39% of the patients with a significant risk of developing veno-occlusive disease (VOD), especially in the setting of previous or subsequent transplantation.4 In further studies, lower doses of Mylotarg such as 6 mg/m2 were used in children in a phase 2 study, while dose of 4 or 6 mg/m2 was used in combination with chemotherapy.5, 6
During phase 2 studies, analyses of saturation and internalization of Mylotarg on CD33 antigenic sites were performed ex vivo on fresh blast cells from patients. A continuous renewed expression of CD33 antigens on the cell surface after exposure to Mylotarg was observed. This observation led to the hypothesis that repeated administrations of GO may be able to enhance internalization process and thereby the intracellular accumulation of the drug.7, 8 We hypothesized that fractionating the dose of 9 mg/m2 may enhance safety without hampering efficacy.
Thus, we designed a prospective multicenter phase 2 uncontrolled sequential trial to assess in adult patients with AML in first relapse the efficacy and safety of fractionated doses of GO, given at a dose of 3 mg/m2 on days 1, 4 and 7 as one course. Simultaneously, complete assessment of multidrug resistance (MDR) status was made. Measures of ABCB1 (P-glycoprotein (Pgp)), ABCC1 (multi-drug resistance protein 1 (MRP1)), ABCC2 (MRP2), ABCC3 (MRP3), ABCC5 (MRP5) and ABCG2 (breast cancer resistance protein (BCRP)) protein expression and function on fresh viable cells from patients were performed and the results were correlated with clinical response. We present here results of the final analysis, based on the 57 patients enrolled.
Materials and methods
Inclusion criteria were: patients 18 years of age with confirmed diagnosis of CD33-positive AML as determined by analysis of bone marrow aspirate and immunophenotyping. Patients were in first untreated relapse with a duration of first remission 3 months and 18 months. In addition, patients were required to have an Eastern Cooperative Oncology Group performance status of 0–2, a serum creatinine <180/μl, aspartate aminotransferase and alanine aminotransferase levels less than twice the upper normal limits. Patients with secondary leukemia, AML3, AML following diagnosed myelodysplastic syndrome or myeloproliferative disease or having received hematopoietic stem cell transplantation (HSCT) as part of their first treatment were excluded.
Baseline studies included bone marrow examination with CD33 immunophenotype, cytogenetics and determination of MDR status.
Cytogenetics results were classified according to that of the South West Oncology Group.9
ABC proteins expression
Pgp, MRP1, MRP2, MRP3, MRP5 and BCRP protein expression was measured by labelling fresh viable cells with UIC2, QCRL3, M2I4, M3II9, M5I1 and BXP-21 monoclonal antibodies, respectively (concentration 5 × 103 μg/ml), and phycoerythrin-labeled second antibody as described elsewhere.10, 11 Fluorescence was analyzed on a Beckman Coulter (EPICS Altra) flow cytometer. ABC-Protein expression was determined by Kolmorogov–Smirnov (KS) test, which measures the difference between two distribution functions and generates a value ranging from 0 to 1.0. A higher D-value indicates a wider difference between the two functions, thus a strong expression of ABC proteins.
ABC proteins function
We have used JC1 (7.5 μM)±cyclosporin A (CsA) (2 × 10−6 M) to assess Pgp activity, calcein-AM (10−6 M)±MK571 (20 × 10−6 M) to assess MRP1 activity, mitoxantrone (MITO) (20 × 10−6 M)±MK571 to assess MRP3 activity and MITO±fumitremorgin C (10−5 M) to assess BCRP activity.11 Intensity of probe fluorescence in the presence or the absence of modulator was compared using the KS test. A higher D-value (ranging from 0 to 1) indicates a wider difference between the two functions, thus a more resistant group of cells. For each sample, 4000 events were collected. The expression and function of ABC proteins were established only in blasts selected by a weak intensity of CD45 antibody.
The protocol was approved and monitored by central and local institutional review boards, and the study was conducted in a manner consistent with the Declaration of Helsinki Principle and Good Clinical Practice Guidelines. All patients provided written informed consent before treatment.
Induction therapy: Patients were scheduled to receive three doses of Mylotarg administered as monotherapy intravenously (i.v.) as a 2-h infusion of 3 mg/m2/day on days 1, 4 and 7.
Consolidation therapy: It was recommended that patients in CR or CRp received consolidation courses with high doses of cytarabine. Patients less than 55 years received cytarabine at a dose of 3 g/m2 i.v. every 12 h for 3 days and patients older than 55 years and/or patients with a creatinine clearance less than 50 ml/min received cytarabine at a dose of 1 g/m2 i.v. every 12 h for 3 days.
Eligible patients may receive HSCT, but a minimum delay of 90 days was recommended between Mylotarg treatment and HSCT.12
Evaluation of response and safety
Blood counts, biochemistry including transaminases and bilirubin levels were determined three times each week during the follow-up period.
A bone marrow aspiration was performed at day 15 to assess clearance of blasts.
Final determination of remission status was assessed by blood and bone marrow examination as soon as normalization of blood counts was observed and/or at a maximum of 43 days after the first dose of GO treatment. Response criteria were those of the criteria of National Cancer Institute (NCI) revised by IWG.13 Complete response was defined as follows: absence of any tumour and <5% bone marrow blasts with polymorphonuclear cells (PMN) >1000/μl, platelets >100 000/μl and independence of transfusions. CRp was defined according to the same criteria, except that platelets were <100 000/μl. Patients who did not meet the criteria for CR or CRp were categorized as no remission (NR).
The incidence of treatment-emergent adverse events (TEAEs) was used to assess safety. The United States NCI Common Toxicity Criteria (version 2) was used to assess the severity of all TEAEs.
Sample size and statistical analysis
The trial was planned to detect a 15% benefit in response rate with Mylotarg, from the reference rate of literature of 15%, controlling for type I and II error rates at 5 and 15%, respectively.
Sequential analyses according to a triangular test was scheduled every 10 patients. At the second interim analysis, based on the first 26 patients (performed in June 2005), the trial was stopped because the upper boundary was crossed and thus we conclude Mylotarg is beneficial. We present here results of the final analysis, based on 57 patients.
Overall survival was measured from the first GO administration to death. Relapse-free survival (RFS) was measured from the first documentation of CR or CRp to the date of relapse or death before relapse. All survival data were analyzed using Kaplan–Meier estimates, with comparison based on log-rank tests. Cumulative incidence of relapse was estimated using non-parametric curves, and then compared by Gray's test.
All P-values were two sided, with 0.05 or less denoting statistical significance. Statistical analyses were performed on SAS (SAS Inc., Cary, NC, USA) and R software packages.
From 12 March to 8 May, 57 patients from 13 institutions were included. Of these, seven (12%) did not fulfill the inclusion criteria: five had a first remission duration longer than 18 months, one patient at 22 months, three patients at 23 months and one patient at 108 months; one patient had neurological symptoms without meningeal blast cells involvement, but received intrathecal chemotherapy few days before treatment with Mylotarg and one patient had secondary AML. According to the intent-to-treat principle, they were all included in the analysis.
Baseline characteristics of the patients are shown in Table 1.
All patients received the three doses of Mylotarg infusions.
Blast clearance at day 15: Forty-six patients had a bone marrow aspiration obtained at day 15 after the first dose of GO. Twenty-five (54%) of these patients had less than 5% leukemic blast cells, of whom 13 reached CR or CRp and 12 did not. Of the 21 patients who had more than 5% blasts on bone marrow aspiration at day 15, three had CRp. This was considered to be statistically significant (P=0.018).
Response at day 43: Nineteen (33%) patients achieved overall remission, including 15 who had CR and four patients CRp.
Therapeutic failures were as follows: one patient had early death before day 15 bone marrow evaluation, one patient died at day 27 during treatment-induced bone marrow hypoplasia with no persistent leukemia and 36 patients had persistent leukemia (of whom two died before day 43, at days 23 and 30).
There was no significant difference in remission rates according to age 60 years, cytogenetic risk groups and duration of first remission 12 months (Table 2).
Overall survival, cumulative incidence of relapse and DFS
Of the 57 patients, a total of 40 (70.2%) deaths were observed. (Figure 1). The median overall survival was 8.4 months (8.3 and 8.9 months for patients <60 years of age and >60 years, respectively, P=0.15 by the log-rank test). Among the 19 patients in remission, 11 patients relapsed and nine of them died from leukemia: seven among the 15 patients who developed CR and two among the four in CRp (P=0.96 by the log-rank test).
Following remission, 11 (57.9%) patients relapsed, with an estimated cumulative incidence of relapse at 1 year of 57.4%. (Figure 2). The median RFS was estimated at 11.0 months, similar for patients who achieved CR (median, 11.6 months) and CRp (median, 8.6 months) (P=0.58 by the log-rank test). Three of the five patients in remission younger than 60 years experienced relapse compared to eight relapses in the 14 patients older than 60 years, but the difference was not statistically significant (P=0.47 by the log-rank test).
Results from MDR studies
Of the 57 patients, 40 were evaluable for expression and function of ABC proteins. Using the previously described cutoff criteria,11, 14 67 and 52% of patients presented Pgp expression and activity, respectively, 65 and 62%, respectively, for MRP1 and 50 and 45% for BCRP. MRP2 expressed in 30%, MRP3 in 37% and MRP5 in 15%.
The achievement of CR or CRp correlated with low dye efflux. The mean D-value for Pgp function from 14 patients with CR or CRp was 0.10±0.09, compared with 0.37±0.21 (P=0.0001) in 26 samples from non-responders (non-responders were the patients with persistent leukemia. Two patients who died were not included in MDR study) (Figure 3a). The corresponding mean D-values for MRP1 function were 0.15±0.14 and 0.35±0.20 (P=0.004). The same D-value differences were found when ABC protein expressions were analyzed. In contrast, expressions of MRP2, MRP3 and MRP5, and expression and function of BCRP were not significantly different among samples from patients with CR or CRp, compared with samples from non-responders (Figure 3b).
Therapy after GO treatment
Of the 19 patients in CR or CRp, 18 received additional treatment with high-dose cytarabine.
There were no toxic deaths after consolidation courses.
Four CR patients, one CRp and two NR patients received HSCT, three allogeneic and four autologous. Mean time between Mylotarg infusion and HSCT was 5 months (range, 112–220 days).
Hematologic adverse events: CR and CRp patients had recovery of their absolute neutrophil count to >500/μl in a median of 23 days from the first dose of GO. Patients who achieved CR or CRp had a recovery of platelet counts to 50 000/μl in a median of 20 days. Seven responding patients did not receive any red blood cell (RBC) transfusions, and six patients received 2–6 RBC transfusions. Five responder patients did not need any platelet transfusions, and seven patients received 2–4 platelet transfusions.
During the treatment period, grade 3 TEAEs that occurred in >1% patients included sepsis (31.5%), fever (15.8%), rash (10.5%), pneumonia (7%), bleeding (7%), mucositis (3.5%), diarrhea (1.75%), headaches (1.75%), tachycardia (1.75%) and edema (1.75%). No grade 4 toxicity was observed. No infectious deaths occurred.
Liver function abnormalities: Grade 1 or 2 hyperbilirubinemia (1.5–3 × upper limit of normal (ULN)) was reported in four patients. Grade 1 or 2 elevations of aspartate aminotransferase or alanine aminotransferase levels (N–5 × ULN) were observed, respectively, in 23 and 9 of the patients. No episodes of VOD occurred.
Seven patients received HSCT after GO treatment. Three patients received an allogeneic BMT (two NR, one CR): one patient relapsed 2 months after transplant, one patient died 5 months after transplant and one patient is alive in CR 14 months after transplant. None of these patients have developed VOD. Four patients (one CRp, three CR) were treated with autologous BMT: three relapsed after 4, 9 and 9 months. One is alive in CR 12 months after transplant.
Fifty-seven patients with AML in first relapse with a median age of 64 years were treated with fractionated doses of Mylotarg: 3 mg/m2 at days 1, 4 and 7 as one course. Patients included in our study satisfied the same inclusion criteria as AML patients included in pivotal phase 2 studies that used two doses of 9 mg/m2 of GO infusions administered 14 days apart.4 Thus, our results and those of pivotal studies may be compared.
We observed a 33% OR rate with 26% CR and 7% CRp, confirming the efficacy of GO as a monotherapy to induce a second remission in AML patients. This response rate compares favorably with the response rate observed in pivotal studies where the OR rate was 26 with 13% CR and 13% CRp.4 RFS and survival were not different between patients in CR or CRp.
Two prognostic factors were found for response: The rate of bone marrow blast cells at day 15 and Pgp and MRP1 activity.
Although a clearance of blasts observed on bone marrow examination at day 15 correlated with remission, it is of note that half of the patients with less than 5% blast cells on bone marrow examination at day 15 did not reach remission. This finding was also observed in pivotal phase 2 studies using 9 mg/m2.4 To our knowledge, no other study in the literature has correlated blast clearance at day 15, and remission or prognosis in AML patients treated for relapse. So it is not known if this finding is specific to Mylotarg treatment or related to the treatment of relapsed AML.
Previous studies have shown that ABC protein activities are prognostic factors in AML.10, 11, 15 The role of ABC proteins has been studied in patients treated with GO at a dose of 9 mg/m2, and it has been shown that blast cell Pgp expression correlated with treatment failure.16 It has also been shown that ABCG2 (BCRP) expression and activity were not associated with reduced GO activity in cell lines.17 In our study, Pgp and MRP1 activity were strongly associated with a poor clinical response, confirming previous report.16 In contrast, ABCG2 (BCRP) expression and activity were not correlated with clinical resistance. Thus we confirm, in AML samples, that BCRP does not confer resistance to calicheamicin. Moreover, we showed that other members of ATP binding cassette family, which confer (MRP3) or not (MRP2 and MRP5) resistance to conventional therapy in AML (anthracycline),14 are not implicated in resistance to GO.
Another main observation of our study was the excellent safety profile. Deaths before day 43 occurred in 4/57 patients (7%) and two of these patients died with resistant disease. PMN (>500/μl) and platelets recovery (>50 000/μl) occurred in 23 and 21 days, respectively, durations less than half the cytopenias duration observed in pivotal studies. No RBC transfusions, and no platelet transfusions were necessary in, respectively, seven and five responder patients. The low rate of grade 3 sepsis (31%) and the absence of deaths due to infections despite profound neutropenia observed in all patients were remarkable. This may be due to the transient myelosuppression observed and the absence of digestive membranes toxicity related to Mylotarg.
One of the major concerns using GO was the risk of liver toxicity. In studies using a dose of 9 mg/m2 of GO, grade 3 or 4 liver function tests abnormalities were reported in 39% of the patients. Also, VOD was often observed in the setting of HSCT and GO. VOD has also been reported in patients treated with GO alone.18 In our study, no grade 3 or 4 hyperbilirubinemia or rise in serum aminotransferase were observed. This favorable safety profile allows 18 of the 19 responder patients to receive additional courses of consolidation chemotherapy with high doses of cytosine arabinoside (AraC) in a short delay after the first infusion of GO, and seven patients to receive HSCT. No cases of VOD were observed after consolidation with HD AraC and after HSCT.
In conclusion, we demonstrated that the use of fractionated doses allowed to give a total dose of 9 mg/m2 with similar efficacy and highly better safety compared to the results observed in pivotal phase 2. This favorable safety profile encourages testing this new schedule in combination with conventional induction chemotherapy. A phase 1/2 study of fractionated GO plus escalated doses of daunorubicin and cytarabine is in progress.
Also our findings confirm that Pgp and MRP1 are the major predictive factors for response to GO and suggest that treatment trials combining GO with MDR reversal agents are warranted. Given the low liver toxicity observed with fractionated doses, combination protocols using CsA or PSC833 or new reversal agent as Zosuquidar that minimally alter the pharmacokinetics of chemotherapy could be designed.
About this article
Targeted Oncology (2018)