Introduction
Recombinant interferon-
(IFN-), an immunomodulatory cytokine, is used to treat chronic hepatitis C virus infection and as a therapy for several cancers, such as melanoma and Kaposi's sarcoma. IFN- also has potent therapeutic activity in the myeloproliferative diseases (MPDs), such as chronic myeloid leukemia (CML), polycythemia vera and essential thrombocythemia. MPDs are clonal hematopoietic stem cell disorders characterized by overproduction of mature myeloid or erythroid cells, which share a common pathophysiology involving dysregulated tyrosine kinase signaling1.
Despite the effectiveness of IFN-, its mechanism of action in the MPDs is poorly understood. Interest in the topic is now rekindled by several recent clinical studies2, 3, 4, 5 that hint that IFN- may target leukemic stem cells.
Clinicians of a certain age will recall when medical therapy of CML was palliative. Busulfan and hydroxyurea, once the drugs of choice, suppressed production of myeloid cells but did not selectively target the malignant clone containing the Philadelphia chromosome, the CML-specific translocation product that creates the BCR-ABL fusion gene. Nor did these drugs interrupt the inexorable progression of CML from chronic phase, in which myeloid differentiation is preserved, to blast crisis, a terminal condition resembling acute leukemia.
In 1986, IFN- was tested in CML subjects and was found to induce cytogenetic and even molecular remissions in which BCR-ABL mRNA transcripts became undetectable. In some subjects, these remissions were maintained when treatment was discontinued6. IFN- was also found to normalize blood counts in people with other MPDs, including some subjects with polycythemia vera and essential thrombocythemia7. Cytogenetic studies of selected subjects suggested that IFN could specifically suppress malignant stem cells in MPDs that lack a Philadelphia chromosome, but it was impossible to prove such effects in the absence of a broadly applicable molecular marker.
In 1990, the demonstration that the product of the Philadelphia chromosome, the BCR-ABL fusion tyrosine kinase, could induce CML-like disease in mice accelerated the search for drugs that could block its enzymatic activity1. In 2001, the ABL kinase inhibitor imatinib mesylate abruptly supplanted IFN- as front-line therapy for patients newly diagnosed with CML. Treatment with imatinib results in vastly superior cytogenetic and molecular responses8, but imatinib and other second-line ABL kinase inhibitors are plagued by the problem of acquired resistance and their inability to eliminate quiescent BCR-ABL+ stem and progenitor cells. These drawbacks have revived the search for treatment strategies that can eradicate leukemic stem cells in CML and other MPDs9.
In 2005, the discovery of a somatic mutation (V617F) in the JAK2 tyrosine kinase in nearly every individual with polycythemia vera and about half of the individuals with essential thrombocythemia10 provided a molecular marker for these diseases that was analogous to BCR-ABL in CML. With this marker in hand, researchers have provided clinical evidence suggesting that IFN- may specifically target leukemic stem cells in these MPDs.
In a study published in Blood, Kiladjian et al.2 treated a cohort of individuals afflicted with polycythemia vera with a pegylated formulation of IFN-2a. They observed complete hematological remission (normalization of erythrocyte and leukocyte counts) in 83% of subjects. In 24 of 27 evaluable subjects, remission was accompanied by a decrease in the mutant JAK2 allele in granulocytes from a mean of 49% to 27%2, and, in one subject, the mutant JAK2 became undetectable, consistent with a molecular remission. Similar results have been obtained in an ongoing study by Quintás-Cardama et al.3 examining PEG–IFN-2a treatment in individuals with polycythemia vera and essential thrombocythemia.
A third report describes 12 subjects with CML who achieved molecular remission on imatinib and who subsequently discontinued kinase inhibitor therapy4. Half of the subjects promptly relapsed with detectable BCR-ABL mRNA transcripts, whereas the others remained in molecular remission without imatinib, with a median follow-up of 18 months. Interestingly, all six of the latter subjects had been previously treated with IFN-4. Another recent study found that imatinib can induce molecular remission in more than half of subjects who have prior cytogenetic remissions in response to IFN-5, a much higher rate than in subjects with CML that had not been previously treated with IFN-8.
Collectively, these studies provide strong but indirect evidence that IFN- preferentially targets the mutant clone in CML, polycythemia vera and essential thrombocythemia and might act to decrease or eliminate the malignant stem cell population in these MPDs.
The relative kinetics of the molecular responses to imatinib and IFN- in CML also support this hypothesis. BCR-ABL mRNA transcripts show an initial rapid decline in imatinib-treated patients that may be the result of elimination of committed progenitors. This decline is followed by a plateau—reflecting the persistence of resistant leukemic stem cells11, whereas molecular responses in IFN-treated people with CML require much longer treatment periods6. The fact that IFN- can induce molecular responses in both CML and polycythemia vera further suggests that leukemic stem cells expressing dysregulated tyrosine kinases might be uniquely sensitive to this cytokine.
These clinical 'bedside' findings argue for additional basic and translational 'bench' research into the molecular mechanisms of IFN action in the MPDs. Twenty years after the introduction of IFN therapy for MPD, little is known about how it operates or why some patients respond to it while others do not.
The possibilities are numerous: some actions of IFN may work directly on the malignant stem cell, such as induction of interferon regulatory factor-8 (ref. 11) and Fas and inhibition of BCR-ABL transcription13. IFN- also selectively impairs proliferation of primitive CML progenitors14. Moreover, both BCR-ABL and JAK2 V617F promote hematopoietic cell proliferation and survival through pathways involving the cell cycle regulator p27 and Foxo transcription factors—providing a potential common mechanism for IFN. In addition to direct effects, IFN- may also target malignant stem cells through its ability to restore normal 
1-integrin–mediated adhesion to the bone marrow niche. IFN- also has pleiotropic immunological actions, including increasing the cytotoxicity of T and NK cells and inducing cell-mediated and humoral immune responses to candidate MPD antigens15.
More work at the bench may illuminate the basic mechanisms of IFN- in CML and polycythemia vera and thereby offer new approaches to eradicate malignant stem cells in MPDs, resulting in permanent cure. On the clinical side, randomized studies of IFN- in combination with kinase inhibitors and with vaccination are warranted in people with CML who have not attained molecular remission on imatinib.
