This site is intended for health professionals only
Division of Hematology Oncology
The Radhay Khanna MDS Center
University of Massachusetts Medical Center
The current activity in the rapidly expanding myelodysplastic syndromes (MDS) arsenal was ignited by the recognition that cytopenias in these patients could be the result of tumour necrosis factor (TNF)-induced apoptosis.(1,2) Concurrent biological insights into genetic mutations, epigenetics and angiogenesis were also being identified as potential therapeutic targets. A review of the present therapeutic landscape shows that while a number of targeted therapies are being explored in small clinical trials (flt3, histone deacetylase [HDAC], matrix metalloproteinases [MMP] inhibitors, suberoylanilide hydroxamic acid [SAHA], anti-angiogenic agents, bone marrow stimulants or immunosuppressive therapies), options for most MDS patients are restricted to four drug strategies.
Trials of agents specific for their anti-TNF effects started with pentoxifylline, etanercept (Enbrel) and infliximab (Remicade), all of which benefited a small subset of MDS patients. Thalidomide was introduced because of its anti-angiogenic, anti-TNF and immuno‑modulatory activities, and produced haematological improvement in about 20% of patients, mainly restricted to the erythroid lineage.(3) The potential for teratogenecity with thalidomide led to the development of the more potent and less toxic analogue lenalidomide. A pilot study reported this drug to be effective in the vast majority of transfusion-dependent MDS patients with the del(5q) chromosome abnormality and a smaller subset of patients without this karyotype.(4) Two subsequent national trials have now confirmed these results. Among 148 patients with del(5q), 93 patients (64%) became transfusion- independent and in 215 non-del(5q) MDS patients with transfusion-dependent low/Int-1 risk disease, transfusion independence was achieved in 46/215 (21%).(5,6) Exciting as these results are in treating subsets of patients defined on the basis of cytogenetic abnormalities, the myelosuppressive side-effects and the inability to predict non-del(5q) responders highlight the need for investigations into the dosage and the mechanism of action of this agent.
5-Azacytidine (Aza-C) and decitabine
Silencing of tumour suppressor genes by hyper‑methylation has been recognised as a therapeutic target for over three decades. Aza-C is a DNA methyltransferase inhibitor that can also replace the normal cytosine in the CpG islands of S-phase cells. The pivotal phase III study was a randomised controlled Cancer and Leukemia Group B (CALGB) trial carried out in 191 patients with MDS of all subtypes.(7) A 60% overall response was found in patients on the treatment arm (7% complete response [CR], 16% partial response [PR] and 37% with haematological improvement [HI]) compared with 5% of those on the observation arm. Median time to transformation to acute leukaemia was significantly longer in patients receiving treatment compared with those on the observation arm (21 versus 12 months). This study led to FDA approval for Aza-C in all categories of MDS patients. 5-Aza-2′-deoxycitadine (decitabine), a pyrimidine nucleoside analogue of Aza-C is at least five times more potent in its ability to inhibit the enzyme DNA methyltransferase I in vitro. In a phase III multicentre trial, 170 MDS patients were randomised to either receive decitabine 15mg every eight hours for three days every six weeks or best supportive care.(7) The response rate by IWG criteria was 35% (10% CR, 15% PR, 10% HI) for decitabine versus 0% for the supportive care arm. A more recent decitabine trial using lower doses has produced 47% CRs.(8)
Zarnestra and Lonafarnib
Farnesyl transferase inhibitors (FTIs) were specifically developed to target the ras intracellular pathway, which can be in hyperdrive due to activating mutations found in up to 20% of MDS patients. A multi-institutional phase III trial of Zarnestra (R115777) at 300mg bid for 21 of every 28 days accrued 81 patients. Responses were seen in 28/81 patients, with 7 CRs, 2 CRs with incomplete platelet recovery, 2 PR and 17 patients showing HI.(9) The median time to leukaemic transformation was 15 months. Lonafarnib (SCH66336) is another FTI that is of interest because of its apparent benefit for patients with chronic myelomonocytic leukaemia (CMMoL) and/or thrombocytopaenia(10) and is presently being investigated in a phase III trial.
Two phase II studies of arsenic trioxide (ATO) as a single agent have been completed in MDS patients, one in Europe and one in the USA. The dosing for the US study was 0.25mg/kg, five days/week for two weeks followed by no dosing for two weeks,(11) while the schedule for the European study was 0.30mg/kg for five days and 0.25mg/kg twice weekly.(12) Responses were seen in approximately 25%, with some patients showing trilineage and major responses. A trial combining thalidomide and ATO produced 7/28 responders, with both the trilineage responders showing an inv(3)(q21q26.2) abnormality.(13) This observation prompted measurement of the myeloid malignancy-associated proto-oncogene EVI1 levels due to its location at 3q26. Four patients had high pretherapy EVI1 levels, as measured by quantitative polymerase chain reaction (PCR). This included two patients with inv(3)(q21q 26.2). Three of four patients with high pretherapy EVI1 levels showed unexpectedly good responses, while one died early in the first cycle. In-vitro studies using 32Dcl3 cells forced to express Evi-1 and then incubated with ATO showed a striking decrease in proliferation compared with cells without Evi-1 expression, thus confirming an increased sensitivity of these cells to ATO. Thus, both in-vivo and in-vitro data indicate heightened sensitivity to the effects of arsenic in cells with dysregulated and overexpressed Evi-1.
The choice for treating MDS patients is restricted by the availability of drugs outside of experimental trials; Zarnestra/lonafarnib/lenalidomide are only available on experimental studies. For lower-risk patients, the choice is limited to thalidomide and Aza-C, the latter not being as frequently used because of its potential toxic side-effects. For higher-risk disease, Aza-C is usually tried first, the failures being treated with ATO. Except for lenalidomide, each of the strategies described above produces about 20% responses; however, it is not known whether it is the same 20% of patients who respond regardless of the agent used. Effective therapies would in fact already exist for a majority of MDS patients if only it were possible to match the right drug to the right patient. New genomic-based technologies provide an opportunity to develop preselection criteria based on unique gene expression profiles. This is the work that now needs to be accomplished with alacrity.