Over the past several decades, improvements in chemotherapeutic agents and supportive care have resulted in significant progress in treating patients with acute myeloid leukemia (AML). More recently, advances in understanding the biology of AML have resulted in the identification of new therapeutic targets. The success of all-trans-retinoic acid in acute promyelocytic leukemia and of imatinib mesylate in chronic myeloid leukemia have demonstrated that targeted therapy may be more effective and less toxic when well defined targets are available. At the same time, understanding mechanisms of drug resistance and means to overcome them has led to modification of some of the existing cytotoxic agents. Rational design and conduct of clinical trials is necessary to ensure that the full potential of these new agents is realized. Cancer 2004. © 2003 American Cancer Society.
Better descriptions of the molecular abnormalities that occur in leukemic cells have led to the identification of several new agents with potential therapeutic activity in patients with acute myeloid leukemia (AML) and other neoplastic myeloid disorders, such as myelodysplastic syndrome (MDS). Although traditional cytotoxic agents largely have been successful in inducing disease remission in many patients with AML, the majority eventually develop recurrent disease, and new therapeutic strategies are needed. Better definition of myeloid disorders based on biology and prognostic indicators may allow stratification of therapy. Biologic and targeted agents may be used to treat more indolent disorders and may be integrated with intensive regimens for more aggressive entities. Novel design of clinical trials may allow better allocation of different agents to different subgroups.1
This article is a review of the new agents undergoing evaluation for the treatment of AML and other myeloid malignancies. Most of these drugs have multiple modes of action, and the headings used in this review are arbitrary and by no means assume knowledge of the exact mechanisms of action, which, for some of the agents described, remain largely unknown.
New Nucleoside Analogs
Nucleoside analogs have been used widely in the treatment of hematologic malignancies.(Table 1) Cytosine arabinoside (Ara-C) is the mainstay of therapy in AML. After uptake by nucleoside transporters and metabolism by intracellular enzymes, nucleoside analogs are incorporated into newly synthesized DNA, leading to chain termination and inhibition of DNA synthesis. The cytotoxic activity of nucleoside analogs is dependent on their conversion to the active phosphorylated metabolite by intracellular enzymes, such as deoxycytidine kinase (dCK).2 Termination of their activity is through inactivation by enzymes, such as deoxycytidine deaminase (dCD), to nontoxic metabolites. Other enzymes, such as cytoplasmic 5′-nucleotidase, dephosphorylate the 5′-monophosphate derivatives, thereby opposing dCK activity and preventing the synthesis of active forms.2 Resistance to nucleoside analogs, through suppression of dCK expression or over-expression of the inactivating enzymes, has been reported. To overcome resistance, modifications of structure of nucleoside analogs have been attempted, and new agents have been developed.
Table 1. Some of the Purine and Pyrimidine Nucleoside Analogs Used in Cancer Therapy
| Compound 506|
Gemcitabine, either alone or in combination with other agents, has been used in the treatment of patients with solid and lymphoid malignancies.3 Experience in myeloid malignancies has been limited. Grunewald et al. reported a Phase I study administering gemcitabine at a dose of 10 mg/m2 per minute to maximize the accumulation of the drug in leukemia cells.4 The maximum tolerated dose (MTD) among 22 patients with recurrent and refractory acute leukemia or blast phase chronic myeloid leukemia (CML) was 4800 mg/m2 infused over 480 minutes.4 More recently, Gandhi et al. reported a Phase I study in patients with recurrent/refractory AML using gemcitabine at a fixed dose of 10 mg/m2 per minute. Prolongation of the duration of administration as a means of dose escalation was evaluated.5 Using that above infusion rate, the drug could be administered for longer than 12 hours without significant toxicity. Among 19 patients who were treated, 1 patient had a partial response (PR), and 2 patients had hematologic improvement.5
Rizzieri et al. administered an escalating duration of infusion of gemcitabine given at a dose of 10 mg/m2 per minute in combination with 12 mg/m2 per day of mitoxantrone for 3 days.6 The maximum recommended duration of gemcitabine infusion was 12 hours (total dose of gemcitabine, 7200 mg/m2). An overall response rate of 42% was reported in 26 patients with recurrent/refractory leukemia (complete response [CR] rate, 25%).6 Severe myelosuppression and mucositis were the most common dose-limiting toxicities. It was reported recently that a similar regimen induced a CR in 2 of 18 patients with advanced leukemia.7
The structure of clofarabine (2-chloro-2′-fluro-deoxy-9-β-D-arabinofuranosyladenine) incorporates some of the favorable antitumor properties of fludarabine and cladribine.8 The chloride chain in the adenosine ring provides resistance to deamination by adenosine deaminase; a fluoride in the 2′ position of the arabinofuranosyl ring hinders phosphorolysis by purine nucleoside phosphorylase. Clofarabine is phosphorylated by dCK to its active metabolite, clofarabine triphosphate, resulting in the inhibition of DNA polymerases and ribonucleotide reductase, enzymes important in nucleoside analog metabolism.8 In Phase I studies in patients with acute leukemia, reversible dose-limiting toxicity was hepatic dysfunction, which was observed at the clofarabine dose of 55 mg/m2 administered intravenously over 1 hour daily for 5 days every 3–6 weeks. The recommended dose for acute leukemia studies in adults was 40 mg/m2 daily for 5 days.8 Among 32 patients who were treated for refractory or recurrent acute leukemia, the response rate was 16%, which included 2 CRs and 3 patients who had bone marrow remission without platelet recovery.8 In the subsequent Phase II study of clofarabine given at a dose of 40 mg/m2 daily for 5 days, 62 patients with recurrent or refractory acute leukemia, high-risk MDS, or CML in blast phase (CML-BP) were treated.9 Seventeen of 31 patients (55%) with AML, 4 of 8 patients (50%) with MDS, and 7 of 11 patients (64%) with CML-BP achieved a CR. Another 9 patients (15%) achieved a CR without full platelet recovery. Grade 3–4 toxicities included skin rashes, transient liver dysfunction, and nausea/emesis.9 Ongoing studies are investigating clofarabine in combination with ara-C and anthracyclines for patients with AML as front-line and salvage therapy.
Nucleoside analogs used in cancer therapy and naturally occurring nucleosides are D-enantiomers. Traditionally, it was believed that the L-enantiomers were not active anticancer agents due to the perceived inability of the activating enzymes to recognize them.10 Troxacitabine, which was developed from the antiviral drug lamivudine, was the first nucleoside L-enantiomer that showed significant activity in myeloid leukemias.10–12 Like ara-C, dCK catalyses the phosphorylation of troxacitabine to its active metabolite, whereas dCD is unable to inactivate it by deamination due to its chiral specificity.13 Giles et al. administered escalating doses of troxacitabine to 42 patients with refractory/advanced hematologic malignancies, including 31 patients with AML and 6 patients with MDS. Stomatitis and hand-foot syndrome were dose-limiting; the MTD was 8 mg/m2 per day for 5 days. Three CRs and 1 PR were observed in 30 evaluable patients with AML.10 In a Phase II study, the same investigators used an intravenous infusion of troxacitabine at a dose of 8 mg/m2 per day in 42 patients with refractory or recurrent hematologic malignancies and reported 2 CRs and 1 PR among 16 evaluable patients with AML.11 They also reported a 37% response rate in 16 patients with CML-BP. Combinations of troxacitabine with other cytotoxic agents (Ara-C, idarubicin, or topotecan) were evaluated in patients with refractory AML, advanced MDS, or CML-BP.12 Among 74 evaluable patients with AML or MDS, 10 patients (13%) achieved CRs.12 These included 7 CRs in 66 patients with AML: 4 with troxacitabine and Ara-C, 2 with troxacitabine and idarubicin, and 1 with troxacitabine and topotecan. Four additional patients with AML achieved hematologic improvement: All had received troxacitabine and Ara-C.12 Then, an adaptively randomized, prospective study of troxacitabine plus Ara-C (TA), troxacitabine plus idarubicin (TI), and Ara-C plus idarubicin (IA) was conducted in newly diagnosed older patients (age ≥ 50 years) with unfavorable karyotype AML.14 The CRs rates were 55% (10 of 18 patients) for IA, 45% (5 of 11 patients) for TA, and 20% (1 of 5 patients) for TI. Patients who received IA had a better cytogenetic profile, lower incidence of antecedent hematologic disorder, and worse performance status compared with patients who received TA and TI. No definite conclusions regarding the possibility of imbalances in the distribution of important prognostic covariates could be made due to the small number of patients randomized.14 Current research with troxacitabine is examining other schedules, such as low-dose infusional regimens, in patients with recurrent AML.
Hypomethylating agents (decitabine and 5-azacytidine)
Aberrant DNA methylation and other epigenetic events are important in the progression of a number of human neoplasms.15 Hypermethylation of promoter-associated CpG-rich regions (CpG islands) can lead to silencing and inactivation of tumor suppressor genes, acting as an alternative mechanism to deletions and mutations.16 Hypermethylation of promoters of genes, such as p15, have been associated with disease progression and with a worse outcome in patients with myeloid malignancies.17 Decitabine (5-aza-2′-deoxycitidine) is a pyrimidine analogue with significant antileukemic activity.18 Once incorporated into DNA, decitabine irreversibly inhibits DNA methyltransferases, enzymes that methylate newly synthesized DNA. In leukemia cells, this leads to hypomethylation of the promoters of tumor suppressor genes, their activation, and cell differentiation.19
Several studies have demonstrated the efficacy of decitabine in patients with recurrent or refractory leukemia.20 The dose-limiting toxicity of decitabine in early Phase I trials was prolonged myelosuppression with minimal extramedullary toxicity.21, 22 In a Phase I study, decitabine was given at a dose of 50–100 mg/m2 every 12 hours for 5 days to 130 patients with CML.23 Eighteen of 64 patients (28%) with CML-BP achieved objective responses, including 6 CRs, 2 PRs, 7 patients who had hematologic improvement, and 2 patients with return to chronic phase. Among 51 patients with accelerated-phase CML and 8 patients with CML-BP, 28 patients (55%) and 5 patients (63%), achieved objective responses, respectively. The estimated 3-year survival rate was < 5% for patients with CML-BP and 27% for patients with accelerated-phase CML.23 The only significant toxicity was myelosuppression, which was delayed, dose dependent, and prolonged. Other investigators have reported significant responses with decitabine combinations.24
An optimal hypomethylating effect can be achieved in vitro with low concentrations of decitabine. Increasing concentrations of decitabine have a direct cytotoxic effect and do not induce more hypomethylation. Issa et al. conducted a Phase I study to determine the optimal biologic dose in patients with recurrent/refractory myeloid malignancies.25 Fifty patients were treated with decitabine at a dose of 5 mg/m2 daily for 10 days with escalation to 10 mg/m2, 15 mg/m2, and 20 mg/m2 daily for 10 days. Other cohorts of patients received a dose of 15 mg/m2 daily for 15 days and 20 days. The drug was tolerated well, with two deaths reported from neutropenic sepsis and eight asymptomatic elevations of liver indices. The authors reported 9 CRs and 7 PRs with an overall response rate of 32% (unpublished observations). Eight of 35 patients (23%) with AML, 4 of 7 patients (57%) with MDS, and 4 of 5 patients (80%) with CML-BP achieved a response. Although responses were noted at all dose levels, the recommended dose for Phase II and combination studies was 15 mg/m2 for 10 days, because this regimen was associated with the most responses.
Decitabine and the related drug 5-azacytidine (AZA) have shown significant activity in MDS. Silverman et al. reported on 191 patients with MDS who were randomized to receive either AZA at a dose of 75 mg/m2 per day subcutaneously for 7 days every 28 days or supportive care.26 Responses were seen in 60% of patients who were treated with AZA (7% CR rate, 16% PR rate, and a 37% hematologic improvement rate) compared with a 5% response rate in patients who received supportive care (P < 0.001). The median time to transformation to AML or death was 21 months with AZA versus 13 months with supportive care (P = 0.007). A landmark analysis after 6 months showed a significant improvement in survival for the AZA arm (median survival, 18 months vs. 11 months; P = 0.03).26 There also was a significant improvement in quality-of-life factors, including fatigue and psychological state, with AZA.27
Decitabine has been investigated in elderly patients with MDS.28 In a study of 66 patients with MDS, decitabine was administered at a dose of 45 mg/m2 per day for 3 days every 6 weeks for a maximum of 6 cycles.28 The overall response rate was 49%. Patients who had an International Prognostic Scoring System (IPSS) high-risk score had a response rate of 64%. The median response duration was 31 weeks, the median survival from diagnosis was 22 months, and the median survival from the start of therapy was 15 months.28 Disappearance of chromosomal abnormalities that were present before the initiation of therapy and reversal of hypermethylation were observed in responding patients.29, 30 A recent update of this study in 169 patients (median age, 70 years) with intermediate-risk or high-risk MDS reported a response rate of 49% and a median response duration of 40 weeks (95% confidence interval [95% CI], 34.9–45.1 weeks) and a median survival of 15 months (95% CI, 12.3–17.7 months).31 Ongoing studies include a randomized, multicenter study comparing decitabine with supportive care in patients with MDS and other studies investigating alternative schedules and subcutaneous route of administration in patients with MDS, AML, and CML after failure to imatinib mesylate.
Agents that Reverse Multidrug Resistance
Chemotherapeutic agents that are active against myeloid malignancies include anthracyclines and epipodophyllotoxins. Resistance to these agents is correlated with expression of P-glycoprotein (Pgp), multidrug resistance protein (MRP), and the major vault protein (lung resistance-related protein).32 P-gp is a highly conserved, 170-kilodalton plasma membrane protein that functions as an ATP-dependent, multidrug exporter with broad specificity for natural product-derived agents.33 Its overexpression decreases intracellular accumulation of anthracyclines and is associated with relative in vitro resistance to them. Cyclosporine A, verapamil, and PSC-833 inhibit Pgp and can overcome this resistance.34 Expression of the multidrug resistance (MDR1) gene coding for Pgp is high in elderly patients with AML and is associated with worse CR rates.35
In a recent study by the South West Oncology Group, a total of 226 patients with AML were randomly assigned to sequential therapy with cytarabine and infusional daunorubicin with or without intravenous cyclosporine A.36 Patients who were randomized to receive cyclosporine A had a slightly higher CR rate (40% vs. 33%; P = 0.14) but a significantly better recurrence free rate (34% vs. 9% at 2 years; P = 0.031) and overall survival rate (22% vs. 12%; P = 0.046).36 The cyclosporine effect was apparent in both MDR1 positive and MDR1 negative patients but was more significant in MDR1 positive patients (median survival, 12 months vs. 4 months). The steady-state daunorubicin serum concentrations were higher for cyclosporine-treated patients, possibly due to inhibition of hepatic metabolism of daunorubicin.36
PSC-833 and other MDR modulators
PSC-833 is a nonimmunosuppressive cyclosporine analog that increases daunorubicin retention in MDR cells 20-fold more than cyclosporine.37 PSC-833 lacks renal toxicity, and concentrations sufficient to block Pgp function in vitro are achieved clinically at relatively nontoxic doses.38 PSC-833 can retard hepatic clearance of cytotoxic drugs, which often need to be reduced when they are coadministered with PSC-833. In an early study, the addition of PSC-833 to mitoxantrone, etoposide, and ara-C (PSC-MEC) was safe and was associated with a CR rate of 29% in patients with recurrent/refractory AML.39
The Cancer and Leukemia Group B (CALGB) investigators randomized 120 elderly patients with AML to receive either cytarabine (100 mg/m2 per day for 7 days), daunorubicin (60 mg/m2 per day for 3 days), or etoposide (100 mg/m2 per day for 3 days) (ADE) or to receive the same agents at lower doses (daunorubicin at a dose of 40 mg/m2 per day and etoposide at a dose of 60 mg/m2 per day for 3 days) with PSC-833 (ADEP).40 The excessive early mortality in the ADEP arm led to its early closure. The rates of CR, nonresponse, and induction death were 46%, 34%, and 20%, respectively, for ADE compared with 39%, 17%, and 44%, respectively, for ADEP (P = 0.008). The median disease free survival (7 months vs. 8 months; P = 0.38) and overall survival (approximately 33% at 1 year for both arms) were not different.40 The CALGB investigators currently are studying younger patients (age ≤ 60 years) with a reduction in the dose of etoposide (40 mg/m2 per day for 3 days) to avoid the early induction toxicity. In a randomized trial by the Eastern Cooperative Oncology Group in 113 patients with recurrent/refractory AML, no advantage was reported for PSC-MEC over MEC alone.41 It is noteworthy that the CR rates for both arms were independent of Pgp expression. Solary et al. reported the results of a randomized trial that examined the addition of quinine to reverse Pgp-mediated drug resistance in 415 adult patients (ages 15–60 years) with de novo AML.42 The addition of quinine had no influence on survival, regardless of Pgp or MDR1 expression.42 More selective and specific MDR modulators, such as the pipecolinate derivatives VX-710 and VX-853, which target Pgp, multidrug resistance protein (MRP-1), and breast cancer resistance protein (BCRP) are under development.43
Agents that Modulate Cellular Signaling Pathways
Intercellular interactions mediated by neighboring cell surface antigens, cytokines, and growth factors ultimately influence gene transcription either directly or indirectly through intracellular signaling pathways. These stimuli, in turn, modulate cellular processes, such as proliferation, differentiation, and death.44 In the normal hematopoietic system, orderly function of these pathways provides an appropriate constellation of hematopoietic cells. It is believed that leukemogenesis results from disorders of these pathways, leading to proliferation, increased cell survival, maturation arrest, and malignant transformation.
Kinases, adaptor or docking proteins, and transcription factors are three classes of proteins involved in cellular signaling. Inappropriate function of members of each group has been linked to a number of hematologic malignancies.44 Constitutive activation of a number of kinases as well as dysfunction of adaptor proteins and transcription factors (such as Ras and Myc) have been linked to leukemic transformation.
A number of oncogenes with constitutive kinase activity have been described as attractive therapeutic targets. Mutations that remove inhibitory domains of the molecule or that induce the kinase domain to adopt an activated configuration lead to the constitutive activation of the protein product. Aberrations of genes, including c-ABL, c-FMS, FLT3, c-KIT, platelet-derived growth factor receptor α (PDGFRα), and PDGFRβ, that normally are involved in the regulation of hematopoiesis have been described.45 Consequently, a number of downstream signaling cascades, such as the Jak-Stat pathway, the Ras/Raf/MAPK pathway, and the PI3K pathway, are activated, leading to inappropriate cell proliferation and survival. Despite the pivotal role of these pathways in normal cellular function, their inhibition may not be associated with significant clinical toxicity, because they are activated inappropriately in leukemic cells.
Constitutively active kinases as targets of therapy
Several translocations have been described that lead to constitutive activation of tyrosine kinases in a manner similar to the Philadelphia chromosome: 1) The translocation t(5;12)(q33;p13) generates the fusion protein TEL-PDGFRβ and leads to the constitutive activation of the platelet-derived growth factor, PDGFRβ, in some patients with chronic myelomonocytic leukemia (CMML).46 TEL-PDGFRβ can confer cytokine independent growth to Ba/F3 cells.47 Imatinib mesylate inhibits the kinase PDGFRβ and is effective therapy in patients with this translocation.48 Other PDGFRβ fusion partners also have been described, and it has been reported that imatinib is effective in some patients with these fusion genes.49
2) The fusion gene FIP1L1-PDGFRα has been demonstrated in 9 of 16 (56%) patients with idiopathic hypereosinophilic syndrome (HES), leading to the constitutive activation of PDGFRα kinase in the neoplastic clone. In vitro studies of the effect of imatinib mesylate on the growth of Ba/F3 cells that expressed FIP1L1-PDGFRα showed that they were inhibited efficiently by much lower concentrations of imatinib than Ba/F3 cells that expressed BCR-ABL (50% inhibitory concentration, 3.2 nM vs. 582 nM, respectively).50 It is noteworthy that a fraction of patients with HES respond to therapy with imatinib mesylate.50, 51 Responses can be achieved at a lower dose of imatinib (100 mg per day) than the dose used in patients with CML and can occur in the absence of a detectable FIP1L1-PDGFRα fusion, suggesting that other, as yet unidentified molecular abnormalities/mechanisms also may exist.50
3) Imatinib mesylate inhibits the catalytic activity of the receptor tyrosine kinase, c-kit, which is expressed in > 90% of patients with AML. Anecdotal responses to this agent in patients with AML have been reported.52 However, larger series have shown minimal or no activity.53 Forty-eight patients with AML (n = 10 patients), MDS (n = 8 patients), myelofibrosis (n = 18 patients), atypical CML (n = 7 patients), CMML (n = 3 patients), or polycythemia vera (n = 2 patients) were treated with imatinib at a dose of 400 mg daily.53 None of the patients with AML or MDS responded.53
4) Membrane-bound receptor tyrosine kinases are enzymes with an extracellular ligand-binding domain, a transmembrane domain, and a highly conserved intracellular domain. Through tyrosine phosphorylation, they mediate the activation of a number of downstream signaling proteins.54 Ligand binding as well as cell-cell interactions through cell adhesion molecules activate these enzymes.55 Phosphorylated tyrosine residues in specific domains of these receptors serve as high-affinity docking sites for SH2-containing adaptor and effector proteins.56 Receptor tyrosine kinases include diverse molecules that are considered members of several distinct classes. Of interest in myeloid malignancies are members of Class III, including platelet-derived growth factor receptor (PDGFR), macrophage colony-stimulating factor (FMS-R or CSF-1R), stem cell factor receptor (KIT), and FMS-like tyrosine kinase 3 receptor (FLT3R).57, 58 The role of these receptors in leukemic transformation and their potential as therapeutic targets are under investigation. Constitutive activation of FLT3R, leading to stimulation of multiple signaling pathways and malignant transformation, has been demonstrated.59 Two well described molecular events reported in > 30% of patients with AML result in such constitutive activation of the receptor. Internal tandem duplication mutations of the FLT3R gene occur at exons 11 and 12 of the gene that code for the juxtamembrane domain of receptor.60, 61 Similarly, point mutations of codon 835 of the FLT3R gene, located in the activation loop of its tyrosine kinase domain, have been reported in 7% of patients with AML.62 These activating mutations of FLT3R, particularly a homozygous activation, confer a worse outcome.63, 64 Inhibitors of such aberrant activation are undergoing clinical evaluation.65 Early clinical trials have demonstrated activity of FLT3 inhibitors, particularly in patients who harbor FLT3R mutations in their leukemic blasts.(Table 2) Their role in leukemia, particularly in combination with other agents, is undergoing further investigation.
Table 2. Recently Reported Trials of FLT3R Inhibitors
|Stone et al.136||PKC412||8||Yes||75 mg orally 3 times per day||6 with ↓ PB blasts||—|
|Foran et al.137||SU11248||32||Yes||25 mg→50 mg→75 mg→100 mg orally everyday; 2 weeks on/2 weeks off||13/16 Evaluable with > 50% ↓ PB blasts||1 Grade 3 fatigue each at 50-mg and 75-mg dose level|
|Foran et al.138||SU11248||29||Yes||Single dose orally, 50–350 mg||> 50% Inhibition of FLT3R in all patients receiving ≥ 200 mg||—|
|Smith et al.139||CEP701||5||No||40 mg→60 mg→ 80 mg orally twice daily||1 CRp at 60 mg||—|
|Heinrich et al.140||MLN518||6||Yes||50 mg orally every 12 hrs||2/3 Evaluable with > 50% ↓ PB blasts||—|
Ras farnesylation as target for therapy
Inhibition of Ras signal transduction is of interest in AML and MDS. Ras-mediated signaling can be inhibited by prevention of its membrane localization, inhibition of Ras protein expression using antisense nucleotides, and inhibition of its downstream targets.66 The C-terminal prenylation of Ras, which is required for its association with the cell membrane and its transforming activity, is mediated by the enzyme farnesyltransferase (Ftase). Several pharmacologic inhibitors of Ftase have been developed.67 Four such Ftase inhibitors are in clinical trials: R115777 (Zarnestra), SCH66336, L778123, and BMS214662.68–70 These agents may have Ras independent effects on other cellular signaling components that may contribute to their antileukemic action.71 Karp et al. conducted a Phase I study of oral Zarnestra at doses ranging from 100–1200 mg twice daily for 21 days every 4 weeks for up to 4 full cycles in 35 patients with poor-risk, acute leukemias.72 Inhibition of Ftase activity was noted at or above the 300-mg twice-daily dose level. Dose-limiting, reversible neurotoxicity was observed at the 1200-mg twice-daily dose level. A response rate of 29% (2 CRs and 8 PRs) was reported. Responses occurred at all dose levels (100–900 mg twice daily) and occurred independent of Ras mutational status (none of the 34 leukemic samples demonstrated an N-Ras mutation).72 Kurzrock et al. recently reported results of a Phase I trial using Zarnestra administered twice daily (3 weeks on/1 week off) for 8 weeks to 21 patients with MDS.73 Maintenance therapy at the dose and schedule tolerated during induction could be continued until toxicity or lack of benefit. The dose-limiting toxicity was fatigue, which occurred at a total dose of 900 mg daily. A response rate of 30%, including 1 CR, was reported. Responses were independent of the baseline Ras mutational status.
Lancet et al. reported on a Phase II study of R115777 (Zarnestra™; Johnson & Johnson, Warren, NJ) given at a dose of 600 mg twice daily for 21 days in 41 patients with untreated, poor-risk AML (defined as any of the following: age ≥ 65 years, age ≥ 18 years with adverse cytogenetics, secondary AML) and MDS with an IPSS score ≥ 1.5.74 Complete and partial hematologic responses were observed in 10 of 30 evaluable patients (33%), including 8 CRs and 2 PRs, 9 of which occurred after 1 cycle of therapy. An additional 12 patients had stable disease after 1 cycle of therapy. Three of four patients with trisomy 8 chromosomal abnormality achieved CRs after one cycle of therapy, the fourth patient achieved a PR. Grade 4 toxicity that occurred in 6 of 36 patients (17%) included mainly infections associated with neutropenia.74
Induction of apoptosis
Suppression of programmed cell death (apoptosis) has been demonstrated in myeloid as well as lymphoid malignancies.75 Dysregulation of apoptosis prolongs the survival of leukemic cells, resulting in their expansion independent of cell division. Defects of apoptotic machinery promote resistance to immune-based destruction, facilitate growth factor independent cell survival, and confer resistance to cytotoxic agents.75 Thus, down-regulation of antiapoptotic proteins, such as bcl-2, may reduce the threshold for chemotherapy resistance and restore chemosensitivity to leukemia cells.76 Genasense (G3139) is an 18-mer phosphorothioate oligodeoxynucleotide antisense that was designed to bind the first 6 codons of human bcl-2 mRNA. G3139 has been administered safely in combination with fludarabine, cytarabine, plus granulocyte-colony stimulating factor (FLAG) to 20 patients with refractory AML or acute lymphoblastic leukemia (ALL).77 Tolerable nausea, emesis, fever, fluid retention, and electrolyte imbalances were the main side effects encountered. bcl-2 mRNA levels were down-regulated in 9 of 12 patients evaluated (75%); the CR rate was 35%.77
Apoptosis also may be enhanced by activating the caspase cascade through blockade of the IAPs and through inhibition of the nuclear factor κB (NFκB) pathway (using IκB kinase inhibitors or proteasome inhibitors) and the PI3K pathway (using kinase inhibitors).78 IAPs, including the X-linked inhibitor of apoptosis (XIAP), can block apoptosis through interaction with members of caspase family.79 Dysregulation of the caspase cascade has been implicated in human neoplasia. IAPs, the natural inhibitors of caspases, are pivotal in this regulatory process.80 An association between increased XIAP levels and decreased survival has been reported in patients with AML.81 Down-regulation of XIAP protein by adenoviral antisense vector results in sensitization to radiation-induced cell death.82
The ubiquitin-proteasome pathway is central for degradation of intracellular proteins.83 Covalent attachment of a polyubiquitin chain targets proteins to the proteasome. The proteasome functions as a regulator of metabolic pathways, allowing up-regulation of degradation of certain proteins without affecting the proteolysis of other substrates.83 Proteasome inhibitors can affect cellular levels of oncogenic proteins differentially.83 Important substrates for proteasome degradation include cyclins and CDKIs, transcription factors (such as p53, NFκB, c-Myc, c-fos, and c-Jun), a number of apoptosis family of proteins, IAPs, and some caspases.83–85 Bortezomib (PS341) is a specific and potent inhibitor of proteasome.86 It induces apoptosis and overcomes resistance in a number of cell lines as well as primary cells from patients with CLL.87, 88 Sustained clinical activity was not demonstrated in a Phase I trial in 15 patients with acute leukemia and MDS, but transient decreases in blood or bone marrow blasts were reported.89 Preclinical synergy with other agents, such as topoisomerase I inhibitors, have led to its evaluation in combination with chemotherapy.
Other therapeutic targets are steps in the serine/threonine cascades.90 Geldanamycin and derivatives of radicicol destabilize Raf protein and interfere with Raf signaling.91, 92 Staurosporine derivatives (UCN-O1, CGP41251, and protein kinase C 412 [PKC412]) inhibit PKC signaling and have been examined in cell lines and clinical studies.93–95 Aberrant MEK and ERK activities have been demonstrated in AML and CML.96, 97 MEK inhibitors (PD098059, PD184352, and UO126) modulate cellular proliferation, differentiation, and apoptosis.98 PD184352 (also known as CI-1040) currently is undergoing Phase I evaluation in cancer patients.99 Pharmacologic inhibitors of PI3K, wortmannin and LY294002, have shown significant potency in preclinical studies.100 The mammalian target of rapamycin (mTOR) is a downstream effector of the PI3K/Akt (protein kinase B) signaling pathway that mediates cell survival and proliferation.101 Rapamycin analogues, including CCI-779 and RAD001, and small molecule mTOR inhibitors, such as AP23573, are under investigation.102
Monoclonal antibodies to CD33
Monoclonal antibodies against antigens expressed by tumor cells have produced encouraging results in several solid and hematologic cancers.103 The prototype murine monoclonal antibodies raised against human antigens were highly immunogenic. This obstacle was overcome through humanization of the parent mouse antibody, resulting in human-mouse chimeric antibodies with a human constant region and a mouse hypervariable region.104 These chimeric proteins do not elicit a human-antimouse antibody (HAMA)-neutralizing response. Thus, they can be given repeatedly without loss of effectiveness, and have a longer circulation time.
CD33 is expressed on the surface of > 90% of AML cells with an average antigen density of 10,000 sites per cell.105 CD33 is expressed on normal granulocyte-monocyte colony-forming units and on some primitive erythroid progenitors. However, its expression on tissues other than the hematopoietic system and the normal pluripotent hematopoietic stem cells is not prominent.105 CD33 expression is universal in patients with acute promyelocytic leukemia (APL) with a high surface density of the antigen.
Studies of the naked antibody to CD33 (HuM195) have demonstrated minimal toxicity, infusion reactions, and minimal HAMA response.106 Using HuM195 in patients with APL, Jurcic et al. reported a higher PCR negative rate (compared with ATRA alone) when the antibody was given to PCR positive patients in hematologic CR after an induction course of ATRA with or without chemotherapy. Eleven of 22 patients (50%) became PCR negative with HuM195, compared with 7 of 34 historic controls (21%).107 Feldman et al. reported 2 CRs and 1 PR with HuM195 in 50 patients with recurrent/refractory AML.108 MEC plus HuM195 was compared with MEC alone in 191 adults with primary refractory AML or with a first recurrence of AML.109 An overall response rate of 43% (27 CRs and 13 pathologic CRs [CRps]) was reported in 94 patients who received the combination compared with an overall response rate of 26% (20 CRs and 5 CRps) in 97 patients who received MEC alone (P = 0.015). No increase in chemotherapy-related toxicity was reported for the combination.109 No differences in rates of CR, overall response, or survival were seen with the addition of HuM195.
HuM195 rapidly internalizes into target cells on binding the antigen.106 Because of this, it has been conjugated with toxins or radioisotopes to facilitate their delivery.110 Such conjugates would deliver the toxin or radiation selectively to the leukemic cells, resulting in improved efficacy/toxicity profiles. Gemtuzumab ozogamicin (GO), a conjugate of a CD33 antibody to calicheamicin, recently was approved by the Food and Drug Administration for the treatment of elderly patients (age > 60 years) with AML in first recurrence and a first CR duration > 3–6 months.111 A multicenter study of 142 patients with AML in first recurrence who were treated with GO 9 mg/m2 on Days 1 and 15 resulted in a CR rate of 16% and an overall response rate of 30%. Grade 3 or 4 hyperbilirubinemia was reported in 23% of patients (median time to onset, 8 days; median duration, 20 days).111 Venoocclusive disease (VOD) was reported in 14 of 119 patients (12%) who were treated with GO alone or in combinations.112 The incidence of VOD with GO in the setting of transplantation may be higher. Rajvanshi et al. reported liver toxicity in 11 of 23 patients (48%) who received GO for recurrent AML after undergoing stem cell transplantation.113 This was associated with VOD-like findings (weight gain, ascites, and jaundice) in 7 patients (30%). Liver sinusoidal injury with extensive fibrosis, centrilobular congestion, and hepatocyte necrosis was noted on pathology.113 In another retrospective analysis of 62 patients with AML and MDS who underwent allogeneic stem cell transplantation, 13 patients (21%) developed VOD, including 9 of 14 patients (64%) with prior GO exposure and 4 of 48 patients (8%) without prior GO exposure (P < 0.0001).114 Nine of 10 patients who underwent transplantation ≤ 3.5 months after they received GO developed VOD compared with 0 of 4 patients who underwent transplantation > 3.5 months after they received GO.114
GO was investigated in elderly patients (age ≥ 65 years) with newly diagnosed AML. The CR rates were 45% in 20 patients with normal karyotypes and 6% in 31 patients with unfavorable karyotypes. These results were inferior to the results from a group of historic controls who were treated with IA, with corresponding CR rates of 54% and 44%, respectively.115 Furthermore, survival was superior for IA compared with GO with or without the addition of interleukin-11 in an attempt to enhance platelet recovery (P = 0.03).115
A Phase I–II study combined cytarabine and daunorubicin with GO in patients age < 60 years with untreated AML.116 The MTD in the Phase I study was Ara-C at a dose of 100 mg/m2 per day on Days 1–7, daunorubicin at a dose of 45 mg/m2 per day on Days 1–3, and GO at a dose of 6 mg/m2 on Day 4. The CR rate was 83% in 18 evaluable patients. The regimen was tolerated well, with no VOD reported.116 Kell et al. administered GO 3 mg/m2 on Day 1 of induction chemotherapy (1 of 3 regimens: DAT, DA, or FLAG-IDa).117 Forty-one of 55 patients who were treated (85%) achieved a CR with the first course. Nonhematopoietic toxicity was hepatic; the regimen that contained both 6TG and GO resulted in a higher incidence of liver toxicity (22 of 39 patients who received 6TG developed Grade 3 or 4 liver toxicity vs. 1 of 16 patients who did not receive 6TG). GO also was administered with consolidation courses. A VOD-like syndrome was noted in 2 of 15 patients who received GO with Courses 1 and 2 and in 0 of 17 patients who received GO with Course 3.117
Immunoconjugates and other monoclonal antibodies
Monoclonal antibodies conjugated to α-emitting radioisotopes, such as 213(Bi) and 211(At), have been constructed and have shown preclinical activity.118 A Phase I study of HuM195-213(Bi) in 18 patients with recurrent/refractory AML or CMML showed reductions of circulating and bone marrow blasts in 14 of 15 evaluable patients (93%) and in 14 of 18 evaluable patients (78%), respectively.119 Other radioimmunoconjugates, including β-particle-emitting agents (using 131I and 90Y) and radiolabeled antibodies against CD45, are being investigated in the transplantation setting.120
The receptor for human granulocyte–macrophage-colony stimulating factor (GM-CSF) is expressed strongly by human leukemic cells.121 DT388-GM-CSF is a fusion toxin that links the catalytic and translocation subunits of diphtheria toxin (DT388) to GM-CSF. Leukemic progenitors isolated from patients with AML were sensitive to DT388-GM-CSF irrespective of the clinical responsiveness of the patients to standard chemotherapeutic agents.122 In a Phase I study, Frankel et al. reported 1 CR and 2 PRs in 31 patients with resistant AML.123 DT(388)IL3, another fusion toxin composed of DT388 and human interleukin 3, is in preclinical development.124
Imbalance of proangiogenic and antiangiogenic factors (such as vascular endothelial growth factor [VEGF] and basic fibroblast growth factor) may be pathogenic in AML and MDS.125, 126 Thalidomide, a potential antiangiogenic agent, combined with chemotherapy did not show a therapeutic benefit in patients with AML.127 Newer agents, such as bevacizumab, an anti-VEGF monoclonal antibody, are under evaluation in patients with AML, MDS, and CML.126 A Phase II study of bevacizumab administered in a time-sequential combination with ara-C and mitoxantrone to 32 patients (3 untreated patients with high-risk AML, 16 patients with refractory AML, 13 patients with recurrent AML, and 1 patient with CML-BP) resulted in a response rate of 56% (CR rate, 44%).128 Small-molecule receptor tyrosine kinase inhibitors directed at the VEGF signaling pathway, such as SU5416, SU6668, ZD6474, ZK1222584, and CGP41251, also are in development.94, 129 Giles et al. investigated SU5416 in 55 patients (33 patients with AML and 22 patients with MDS): Three patients achieved PRs, and two patients had hematologic improvements.130 Antiangiogenic agents may be more beneficial in combination with cytotoxic agents (e.g., in patients with colon and renal cell carcinomas) or in the setting of minimal residual disease.
Histone deacetylase inhibitors
At low doses, hypomethylating agents (5-azacytidine and decitabine) inhibit DNA methyltransferases and influence chromatin remodeling.16, 19 Methylation of promoter regions of tumor suppressor genes and DNA silencing are some of the events in chromatin remodeling that may lead to carcinogenesis.15 Several translocations in AML involve abnormal recruitment of histone deacetylases (HDAC) to promoters, resulting in transcriptional repression.131, 132 By removing acetyl groups, HDACs allow tighter binding of histones to DNA and prevent gene transcription. Several HDAC inhibitors (phenylbutyrate, trichostatin A, suberoylanilide hydroxamic acid [SAHA], valproic acid, and depsipeptide) are under investigation.133, 134 A recent Phase I study of depsipeptide given at a dose of 13 mg/m2 infused on Days 1, 8, and 15 every 28 days135 in 9 patients with AML showed transient declines in blast counts and tumor lysis syndrome in 1 patient; none of those patients achieved a response. Histone acetylation measured in 6 patients 4 hours after depsipeptide infusion showed a median increase in acetylation for H3 histone of 40% (range, 10–160%) and a median increase in acetylation for H4 histone of 100% (range, 0–240%). Dose escalation to achieve the minimum effective pharmacologic dose was recommended. Because of a cumulative increase in gastrointestinal and constitutional side effects associated with the weekly schedule, an alternative dosing schedule was recommended.135 It has been demonstrated that SAHA, which was developed initially as an apoptosis-inducing and differentiation-inducing agent, inhibits the activity of both Class I and Class II HDACs, SAHA is currently under investigation in patients with advanced leukemias and MDS.
Progress in understanding the molecular biology of neoplastic transformation already has led to the discovery of a multitude of new agents, which currently are undergoing evaluation in patients with myeloid disorders. When specific and well defined targets have been available, single agents have demonstrated remarkable activity and limited toxicity. Better definition of the complex processes initiating and sustaining the leukemic change no doubt will lead to a better definition of targets for therapeutic intervention and to the eventual realization of the goal of curative therapies.