Combination therapies in Myeloproliferative Neoplasms: why do we need them and how to identify potential winners?

The myeloproliferative neoplasms (MPN) are clonal myeloid disorders characterized by proliferation of mature myeloid cells, such that in polycythaemia vera (PV), the red cell proliferation dominates, platelets in essential thrombocythaemia (ET) and in myelofibrosis (MF), there may be cytopenia or proliferation, but the characteristic feature is the strikingly abnormal bone marrow stroma. These entities have a tendency to show phenotypic mimicry and may transform from one to another, for example, 20–30% of patients with PV are likely to develop MF. The significant event in this field was the recognition that Janus Kinase-2 (JAK2) activation was highly prevalent, followed by the description of the JAK2V617F mutation in 2005 (vide infra), which stimulated renewed interest in disease biology. Janus Kinase-2-targeted therapies have led to marked improvements for patients with this condition. However, it is obvious that the pathogenesis of these complex disorders reaches beyond this mutation; only 50–60% of patients with ET, for example, have the JAK2 mutation and several additional mutations have been described, which are of relevance in both the pathogenesis and clinical phenotype of these conditions. The therapeutic benefits seen with the effect of JAK2 inhibitors are striking, including reduction of massive splenomegaly, resolution of constitutional symptoms and prolongation of survival as seen with Ruxolitinib, the first of the class JAK1/2 inhibitors [1, 2]. Such improvements, however, are not of the same magnitude as the magnitude of benefits associated with BCR/ABL inhibition in chronic myeloid leukaemia, for example. Most likely, this reflects a number of issues; firstly, that none of the inhibitors yet developed is specific for mutant JAK2, and secondly, that JAK2 activation or its consequence is not the only pathogenic mechanism operating in these intriguing disorders. This has several important implications for this field: we need to better understand the disease biology and develop systems for testing other novel therapies, either alone or in combination. In this issue, two papers (Choong et al. and Bartalucci et al. [3, 4]) describe different approaches to the issue of testing and evaluating drug combinations in MPN and demonstrate benefits in vitro, which probably merit clinical investigation. What are the potential pathways of interest? Constitutive activation of JAK2 in MPN drives multiple downstream pathways, including the signal transducer and activator of transcription (STAT) 3/STAT5, mitogen-activated protein kinases (MAPK), extracellular signal–regulated kinase (ERK) and Phosphatidylinositol 3-kinase (PI3)/Akt/Mammalian target of Rapamycin (mTOR) pathways [5–8]. Augmented ‘genetic instability’ and epigenetic modification may ensue (as described in Fig. 1). This genetic instability may subsequently facilitate disease progression. Activation of the PI3K pathway leads to pleiotropic effects on cellular proliferation, metabolism, intracellular trafficking and survival [9, 10]. Recent work has demonstrated elevated levels of phospho-STAT5, phosphoAkt and phospho-mTOR in progenitor cells derived from MPN patients [11]. Studies of both ‘knock-in’ JAK2V617F murine models and MPN patient samples have documented increased reactive oxygen species (ROS) generation secondary to Akt-mediated downregulation of catalase activity, with an accompanying increase in double-stranded breaks (DSB) [12]. Furthermore, the PI3K/Akt/mTOR pathway is an important mediator of cellular drug resistance in both solid and haematological malignancies [13–15]. The PI3K/Akt/mTOR pathway is therefore an attractive therapeutic target in patients with MPN. Three main classes of PI3Ks, class I, II and III, exist within cells. Class I PI3Ks, heterodimers consisting of a regulatory subunit (p85) and catalytic subunit (p110), are activated by cell surface receptors, including key receptor Tyrosine Kinases (RTKs) such as JAK, and are well characterized. Three isoforms of the p110 catalytic subunit exist: p110a, b and d. Activation of class I PI3Ks induces phosphorylation of the 3′-hydroxyl group of the inositol ring of phosphatidylinositol-4, 5-triphosphate, PI [4, 5] P2 (PIP2), leading to the generation of phosphatidylinositol-3, 4,5-triphosphate, PI [3–5] P3 (PIP3) [9, 10]. PIP3 promotes Akt translocation to the plasma membrane and subsequent phosphorylation of threonine 308 (Thr308) by 3′phosphoinositidedependent kinase-1 (PDK1) occurs, leading to activation [9, 10, 16–18]. In general, full activation of Akt requires not only phosphorylation of Thr308 but also Serine 473 (Ser473) within the hydrophobic C-terminal domain by PDK2 [19]. Akt activation inhibits Tuberose Sclerosis (TSC) 1/TSC2 complex function and leads to enhanced mTOR activity. Moreover, diverse effects outside the PI3K/Akt/mTOR ‘vertical pathway’ can also occur following Akt activation, predominantly driving a proliferative and anti-apoptotic state. These include inhibition of both procaspase-9 and BAD, a pro-apoptotic BCL-2