Ruxolitinib and DNA methyltransferase-inhibitors: a foray into combination regimens in myelofibrosis

A major breakthrough in understanding the pathobiology of myeloproliferative neoplasms (MPN) was made in 2005 when several groups independently detected a somatic point mutation in a highly conserved residue of the pseudokinase domain of the JAK2 (Janus kinase 2) tyrosine kinase (V617F) [1–3]. Mutated JAK activates a number of downstream pathways implicated in the proliferation and survival of malignant cells, including the STATs (signal transducers and activators of transcription), a family of important latent transcription factors. Intriguingly, even patients with MPN devoid of the JAK2V617F mutation demonstrate STAT activation. A search for alternative pathways of STAT activation resulted in the identification of myeloproliferative leukemia virus oncogene (MPL) and calreticulin (CALR) mutations [4–6]. The JAK2V617F, CALR and MPL mutations are mutually exclusive, with 97% of patients with MPN harboring at least one of these mutations. The pervasive activation of the JAK–STAT pathway in myelofibrosis (MF) lent credence to the development of a potent and selective JAK inhibitor ruxolitinib. The subsequent advent of ruxolitinib greatly reduced the stigmata associated with MF, from improvements in spleen size, weight, performance status and symptom control to prolonged survival, and in selected patients even reversal of marrow fibrosis [7–9]. However, a proportion of patients remain refractory to ruxolitinib, suggesting alternative molecular drivers of MF. Furthermore, the potential for therapy-related worsening of blood counts may restrict the ability to administer ruxolitinib in a subset of patients with high-risk MF with significant cytopenias at presentation. Alternative therapeutic approaches, including development of rational combinations with ruxolitinib as well as identification and targeting of non-JAK–STAT drivers of MF, are required to overcome these therapeutic hurdles. In this issue, Tabarroki et al. describe successful implementation of one such approach, wherein three patients with MF achieved control of splenomegaly and/or constitutional symptoms in addition to hematological improvements, with blast reduction, by combining a DNA methyltransferase (DNMT) inhibitor (5-azacytidine [AZA] or decitabine) with a JAK–STAT inhibitor (ruxolitinib) [10]. This article highlights the feasibility and potential benefit of such a combination in select patients with MF. The potential mechanisms of action of DNMT inhibitors and their role in patients with MF are discussed in this commentary.

Identification of factors that predict response to DNMT inhibitors would allow the selection of patients with MF with the highest likelihood of deriving benefit from such combinations. In addition to traditional risk factors such as age, cytogenetics, performance status and others, additional factors such as molecular mutations may modulate the response to these agents. A number of additional mutations have been identified in patients with MF, revealing a more complicated mutational landscape than was hitherto imagined. These include mutations in regulators of DNA methylation (TET2, IDH 1 and 2, DNMT3A), mutations that regulate chromatin structure (EZH2, ASXL1), mutations that activate signaling pathways (LNK, C-cbl, NF1, PTPN11) and mutations that regulate splicing (U2AF1, SF3B1, SRSF2) [5]. Of particular interest are mutations that regulate epigenetic pathways, including methylation and chromatin structure. These mutations seem to occur in 35–40% of patients with primary MF. Moreover, studies in acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) have reported higher response rates to DNMT inhibitors among patients harboring mutations involving the epigenetic regulators of DNA methylation [11–13]. However, the data supporting the use of these mutations as predictive biomarkers for response to DNMT inhibitors are not unanimous, with other studies failing to identify a direct correlation between the presence of various epigenetic mutations and response to DNMT inhibitors [14]. Ongoing studies may help to further define the role of epigenetic mutations as predictors of response to DNMT inhibitors.

Notwithstanding the potential association between the mutations in well-defined epigenetic regulators with clinical outcomes, DNMT inhibitors may produce responses via a more global mechanism of action. For example, hypermethylation of genes that are not traditional epigenetic regulators of DNA methylation such as the calcitonin gene 5’ area and RAR-beta² gene has been demonstrated in patients with MF using methylation-specific polymerase chain reaction (PCR) [15,16]. Similarly, hypermethylation of cell cycle control genes (p15INK4b and p16INK4a) has been demonstrated in patients with MF in transformation to AML [17].

In a clinical trial performed at M. D. Anderson Cancer Center, the DNA methyltransferase inhibitor AZA (used at a standard dose of 75 mg/m²/day × 7) produced a response rate of 24% (3% partial response and 21% clinical improvement) in a cohort of 34 patients with MF (76% previously treated) [18]. Responses were observed in patients both with and without the JAK2V617F mutation, but were short lived, and myelosuppression was a major problem. DNA methylation analysis revealed that clinical responders had a trend toward more rapid hypomethylation at day 7 of AZA therapy. Notwithstanding the fact that demethylation has been confirmed following treatment with AZA, particularly at low (non-directly cytocidal) doses of the agent, when using the chronic administration schedule developed for AZA in MDS, the extent of demethylation does not correlate with clinical/hematologic responses, suggesting other putative mechanisms of action.

There appears to be obvious potential to increase therapeutic benefit (without inducing limiting toxicities) by combining ruxolitinib with low dose DNMT inhibitors in patients with MF. A number of questions remain unanswered regarding the role of these agents in MF, including the mechanism of action, predictors of response, optimum dosage and schedule in combination with JAK inhibitors. Our current clinical trial (NCT01787487) investigating the combination of ruxolitinib and low dose AZA is under way, and will hopefully provide answers to these questions.

Potential conflict of interest

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.