Induction of complete remission of acute myeloid leukaemia by pegylated interferon-α-2a in a patient with transformed primary myelofibrosis


Type I interferons (IFN) – IFN-α and IFN-β, which share the same receptor (Platanias, 2005) – have antiproliferative activity in vitro against acute myeloid leukaemia (AML) cell lines, e.g. K562, HL-60, U937 (Geng et al, 1995; Benjamin et al, 2007; Zhang et al, 2007). IFN-α has been investigated for the treatment of AML, where it had modest effects (Rohatiner et al, 1983). Preclinical studies revealed that a continuous exposure to type I IFN was required to durably suppress AML. Enforced constitutive expression of type I IFN suppressed the malignant behaviour of K562 (Geng et al, 1995). IFN-α inhibited proliferation and induced apoptosis of U937 in a dose- and time-dependent fashion (Zhang et al, 2007). In a mouse xenograft model, proliferation of HL-60 and of primary human AML cells was inhibited only when there was continuous stable expression of type I IFN (Benjamin et al, 2007). Subcutaneous or intravenous injections had no antileukaemic effect and were not associated with detectable or stable plasma levels of type I IFN. This can be accounted for by the short half-life of IFN and provides an explanation for the dichotomy between the marked antileukaemic effects of IFN-αin vitro and the disappointing effects in AML patients in vivo.

The polyethylene glycol-conjugated (pegylated) formulation of IFN-α allows for stable continuous release of IFN, resulting in improved therapeutic efficacy and fewer side effects (Matthews & McCoy, 2004). Pegylated IFN-α-2a (Peg-IFN-α-2a) has a high therapeutic activity in the chronic phase of the myeloproliferative neoplasms polycythemia vera (Kiladjian et al, 2008) and primary myelofibrosis (PMF) (Ianotto et al, 2009), higher than its non-pegylated counterpart. We report here on the induction of complete remission by Peg-IFN-α-2a of AML occurring in transformed PMF.

The diagnosis of a PMF type myeloproliferative neoplasm was made in 1994 in a 49-year-old man. In 2006, a control marrow examination demonstrated worsening of myelofibrosis, but no signs of AML. There were no BCR/ABL1 fusion transcripts or JAK2 V617F mutation. In 2008, blood and marrow demonstrated evolution to AML (Figs 1 and 2A) with a CD45low CD117CD34+/− CD33CD13+ immunophenotype. Intensive polychemotherapy and allogeneic stem cell transplantation were turned down by the patient. After informed consent, a treatment with Peg-IFN-α-2a (Pegasys®; Roche, Basel, Switzerland) was initiated.

Figure 1.

 Clinical overview of the treatment of AML by Peg-IFN-α-2a in transformed PMF. (A) Time course of the evolution to AML in a patient with pre-existing PMF and the subsequent control of it by Peg-IFN-α-2a therapy. The leukaemic transformation (thin vertical grey line) was evidenced by 24.0% myeloblasts in peripheral blood (blue line) and by a sharp increase in blood expression of the leukaemia tumour marker WT1 mRNA (green line). Therapy with Peg-IFN-α-2a was rapidly initiated (upper figure panel). The red line reflects the dose regimen adjustments throughout the treatment course (weekly doses of 90, 135, 180 or 315 μg). Peg-IFN-α-2a dose was temporarily increased from 180 to 315 μg/week, because the blood myeloblast percentage had not decreased significantly and there was evidence from the preclinical xenograft model that exposure to higher concentrations of type I IFN was more successful in controlling AML (Benjamin et al, 2007). After 6 months of therapy, a marked decline in WT1 expression levels was noted as well as a persistent decrease of blood myeloblasts to pre-AML levels. WT1 expression was assessed on blood leucocytes by real-time quantitative reverse transcriptase-polymerase chain reaction (Ogawa et al, 2003). (B) Effect of Peg-IFN-α-2a on bone marrow myeloblast percentage (blue line). Leukaemic transformation was demonstrated by 33.5% blasts in the bone marrow (point 2), as compared to 3% in a preceding marrow examination performed more than 2 years earlier (point 1). Subsequent marrow studies showed decrease of myeloblasts 2 months after initiating Peg-IFN-α-2a treatment (point 3; 12.5% blasts) and complete remission after 5 months (point 4; 2.0% blasts). The inserts show the findings of cytogenetic, fluorescence in situ hybridization (FISH) and molecular analysis on serial marrow examinations. They document that under Peg-IFN-α-2a treatment, the AML-associated D835 FLT3 mutation was suppressed and the PMF-associated del(20)(q12), a partial deletion of the long arm of chromosome 20, was decreased, suggesting inhibition of the myeloproliferative haematopoiesis in addition to suppression of the leukaemic transformation (point 4). (C) Corresponding evolution of peripheral blood counts. Transient drops in haemoglobin were treated with packed red blood cell transfusions (yellow arrowheads; U: number of units transfused). A normalisation of the platelet count was observed 1 week after initiation of Peg-IFN-α-2a therapy, making it possible to stop anagrelide (upper figure panel). Treatment with Peg-IFN-α-2a was accompanied by a marked decrease in white blood cell (WBC) and absolute neutrophil (ANC) counts. Low-dose hydroxycarbamide (between 3500 and 5500 mg per week) was stopped after reaching the ANC nadir (upper figure panel).

Figure 2.

 Bone marrow histology at diagnosis of AML transformation (A) and 5 months after initiating Peg-IFN-α-2a treatment (B). Light micrographs of haematoxylin eosin (HE)- and anti-CD34- (inserts) stained paraffin sections of trephine biopsies from iliac crest. (A) Hypercellular bone marrow of PMF which has evolved into AML, as evidenced by excess of myeloblasts, part of which express CD34. (B) Hypocellular bone marrow, with a strong decrease of CD34+ myeloid cells, showing complete remission of AML. Bar: 100 μm (50 μm inserts).

Figure 1 summarises the clinical evolution. After 5 months of treatment, marrow analysis showed complete remission of AML (Fig 1B). This finding was confirmed by marrow histology, which revealed suppression of myeloid cell proliferation and CD34+ myeloblast infiltration (Fig 2). The AML-associated FLT3 D835 point mutation, which appeared at leukaemic transformation, was no longer demonstrable after starting Peg-IFN-α-2a (Fig 1B). Wilms tumour WT1 mRNA, a tumour marker in AML (Ogawa et al, 2003), that sharply increased at leukaemic transformation, decreased significantly during treatment (Fig 1A), confirming the antileukaemic activity of Peg-IFN-α-2a. At more than 10 months of follow-up and while on a maintenance dose of 180 μg/week, the patient is alive and well with peripheral blood counts indicative of continued complete remission of AML.

The dose of Peg-IFN-α-2a was determined on a pragmatic basis by taking into account blood and marrow findings. A weekly dose of 180 μg of Peg-IFN-α-2a usually results in stable serum IFN-α concentrations of around 25 ng/ml, corresponding to 6000 iu/ml (Matthews & McCoy, 2004). An antileukaemic effect is reached in vitro at concentrations of 1000–4000 iu/ml (Zhang et al, 2007) and in vivo at 3000 iu/ml (Benjamin et al, 2007). In general, the treatment was well tolerated, there were no infections and there was no need for hospitalisation. Three weeks after starting Peg-IFN-α-2a, the patient had mild transient arthralgias. Sporadic complaints of fatigue and palpitations were related to anaemia as they disappeared after transfusions. The loss of around 10% of the body weight, that started 3 months before the diagnosis of AML, presumably as a result of the leukaemic transformation, stopped 2 weeks after initiating Peg-IFN-α-2a.

As AML occurred while the patient was on hydroxycarbamide and this drug was stopped before reaching remission, there is little doubt that antileukaemic activity was brought about by Peg-IFN-α-2a. The antileukaemic effect of Peg-IFN-α-2a was considerably more pronounced than the limited clinical results of non-pegylated IFN-α in a small minority of AML patients (Rohatiner et al, 1983). This is congruent with the preclinical xenograft model, where control of AML was not obtained with subcutaneous or intravenous injections of type I IFN, which did not lead to stable plasma levels of IFN (Benjamin et al, 2007). They are also in sharp contrast to the failure to reach complete remission in any of the PMF patients in AML transformation treated with standard conventional chemotherapy (Mesa et al, 2005). Median survival time from diagnosis was 2·6 months and chemotherapy in this group of patients was associated with a 33% mortality rate.

The mechanisms by which IFN controls AML have been studied. Type I IFN induced JAK-1 and STAT-1 phosphorylation in HL-60 cells in vitro (Benjamin et al, 2007), compatible with IFN-mediated signalling (Platanias, 2005). There was increased apoptosis of AML cell lines in vitro and in the xenograft model in vivo following continuous exposure to type I IFN (Geng et al, 1995; Benjamin et al, 2007; Zhang et al, 2007). In general, type I IFNs induce apoptosis in different cell types through activation of IFN-stimulated response elements (ISRE) found in promoters of genes linked to apoptosis and/or antiviral immunity (Platanias, 2005). IFN-α treatment of leukaemic cells led to decreased cyclin E expression (Zhang et al, 2007) and to cell cycle arrest in S-phase (Geng et al, 1995), inhibiting cell proliferation.

In conclusion, this is the first report on the successful treatment of a poor prognosis AML using Peg-IFN-α-2a, without major side effects, hospital admission and/or administration of aggressive chemotherapy. It is the clinical proof-of-principle that Peg-IFN-α-2a not only has activity in the chronic phase of myeloproliferative neoplasms but also in acute myeloid malignancy. The low toxicity seen in this study supports conducting clinical trials in order to assess the place of Peg-IFN-α-2a in the treatment of primary AML and of AML transformation of myeloproliferative neoplasms.


The authors thank Dr Marc Van der Planken and Ms Francine Vertessen for expert bone marrow cytology; Dr Katrien Pieters and Dr Katrien Vermeulen for expert flow cytometry and molecular assays; Prof Jan Wauters and Ms Elvire Van Assche for expert cytogenetics; Prof Eric Van Marck for support; Prof Jean-Jacques Kiladjian for pre-print information regarding the use PEG-IFN-α-2a in non-transformed PMF; and the patient for insisting on trying an alternative non-toxic treatment. SA is a PhD research fellow of the Fund for Scientific Research-Flanders (FWO-Vlaanderen).

Statement of authors’ contribution

ZNB designed the treatment and follow-up, treated the patient, analysed the data and wrote the manuscript; SA made an equal contribution in writing the manuscript, had conceptual input and created Fig 1; VVM created Fig 2 and helped to write the manuscript; WAS and VFVT had conceptual input and helped to write the manuscript.