- Top of page
- Materials and methods
- Author contributions
Aberrant JAK2 signalling plays an important role in the aetiology of myeloproliferative neoplasms (MPNs). JAK2 inhibitors, however, do not readily eliminate neoplastic MPN cells and thus do not induce patient remission. Further understanding JAK2 signalling in MPNs may uncover novel avenues for therapeutic intervention. Recent work has suggested a potential role for cellular cholesterol in the activation of JAK2 by the erythropoietin receptor and in the development of an MPN-like disorder in mice. Our study demonstrates for the first time that the MPN-associated JAK2-V617F kinase localizes to lipid rafts and that JAK2-V617F-dependent signalling is inhibited by lipid raft disrupting agents, which target membrane cholesterol, a critical component of rafts. We also show for the first time that statins, 3-hydroxy-3-methyl-glutaryl coenzyme A (HMG-CoA) reductase inhibitors, widely used to treat hypercholesterolaemia, induce apoptosis and inhibit JAK2-V617F-dependent cell growth. These cells are more sensitive to statin treatment than non-JAK2-V617F-dependent cells. Importantly, statin treatment inhibited erythropoietin-independent erythroid colony formation of primary cells from MPN patients, but had no effect on erythroid colony formation from healthy individuals. Our study is the first to demonstrate that JAK2-V617F signalling is dependent on lipid rafts and that statins may be effective in a potential therapeutic approach for MPNs.
Somatic mutations in the gene encoding the JAK2 tyrosine kinase are prevalent in myeloproliferative neoplasms (MPNs) (Baxter et al, 2005; James et al, 2005; Kralovics et al, 2005; Levine et al, 2005; Zhao et al, 2005; Scott et al, 2007), a group of haematopoietic stem cell diseases characterized in part by expansion of one or more lineages in the myeloid compartment (Levine et al, 2007). Classical MPNs include polycythaemia vera (PV), essential thrombocythaemia (ET), and primary myelofibrosis (PMF). Patients with PV have a defect in the erythroid lineage, leading to overproduction of red blood cells (RBCs). The main cellular defect in ET lies within the thrombocytic lineage, resulting in an overproduction of platelets. In PMF, excessive blood cell formation in the bone marrow results in fibrosis of the bone marrow, which can impede proper haematopoiesis (Levine et al, 2007). A recurrent activating mutation in JAK2, JAK2-V617F is found in c. 95% of PV patients and about 50% of ET and PMF patients (Oh & Gotlib, 2010). Some JAK2-V617F-negative MPN patients exhibit other mutations that alter JAK2 signalling. These include exon 12 mutations of JAK2, mutations of cytokine receptors that signal through JAK2, and mutations of other proteins that regulate JAK2 function. Importantly, many of these mutations can initiate an MPN-like syndrome in mouse models (Oh & Gotlib, 2010). Collectively, these data suggest that JAK2 dysregulation contributes to MPN formation.
While JAK2 inhibitors have had significant success in recent clinical trials due to their ability to reduce constitutional symptoms and thus relieve suffering of patients, they have not readily reduced the allele burden and thus do not induce remission in patients (Scherber & Mesa, 2011; Tefferi, 2012). Thus, alternative therapeutic approaches that enhance neoplastic cell killing are still needed for MPN patients. Further understanding the regulation of JAK2 signalling in MPN cells may uncover additional sites of potential therapeutic intervention that may be effective at treating MPNs.
JAK2 normally functions in signal transduction initiated by cytokine receptor activation. JAK2 associates with cytokine receptors and becomes activated following cytokine receptor stimulation by ligand (Jatiani et al, 2010). Cytokine binding to its receptor causes a conformational change in receptor-associated JAK2 proteins, which then trans-phosphorylate each other leading to their full activation (Lu et al, 2008). Activated JAKs then phosphorylate the cytokine receptor, creating binding sites for downstream mediators like signal transducer and activator of transcription (STAT) molecules. STATs are then phosphorylated on tyrosines by activated JAKs (Jatiani et al, 2010). Phosphorylated STATs function as transcription factors, promoting expression of genes involved in growth, survival, differentiation, etc. In the case of JAK2-V617F, the valine to phenylalanine substitution at amino acid residue 617, allows for dysregulated kinase activity through loss of an autoregulatory function of the pseudokinase domain (Lee et al, 2009). This mutation effectively leaves JAK2 primed for activation by circumventing the need for a conformational change of the kinase induced by cytokine receptor stimulation. However, even though JAK2-V617F does not require cytokine stimulation to be activated, a cytokine receptor is still necessary for JAK2-V617F-mediated signalling and cell transformation (Lu et al, 2005, 2008). Thus, it is thought that cytokine receptors provide a scaffolding function for JAK2-V617F-initiated signalling.
Erythropoietin receptor (EpoR) uses JAK2 to transduce signals initiated by erythropoietin (Epo) to promote RBC production (Witthuhn et al, 1993). We have recently shown that wild-type EpoR/JAK2 signalling requires lipid rafts (McGraw et al, 2012). Lipid rafts are microdomains of the plasma membrane that are enriched in cholesterol and sphingolipids (Galbiati et al, 2001). These microdomains are more rigid than the majority of the plasma membrane and have been shown to function in membrane trafficking, cytoskeletal arrangement (Simons & Ikonen, 1997), virus entry (Scheiffele et al, 1999; Waheed & Freed, 2009), and cellular signalling (Simons & Toomre, 2000). Protein compartmentalization in membrane rafts facilitates protein interactions that regulate signal transduction activation, especially for some receptor-initiated signals at the cell surface (Simons & Toomre, 2000).
While we have shown that wildtype EpoR/JAK2 signalling requires membrane rafts for proper signalling (McGraw et al, 2012), the role of cholesterol and membrane rafts in pathological signalling by JAK2-V617F in MPNs has never been reported. We show for the first time that JAK2-V617F is localized to lipid rafts and JAK2-V617F-dependent signalling requires membrane cholesterol. By utilizing JAK2-V617F-dependent MPN model cell lines as well as primary cells from JAK2-V617F-positive MPN patients, we also show that JAK2-V617F-mediated transformation is sensitive to statins, inhibitors of the cholesterol-producing mevalonate pathway. Our data showing the requirement of cholesterol for JAK2-V617F-mediated signalling and the sensitivity of MPN cells to statins suggests that statins could potentially be incorporated into a therapeutic strategy for MPNs.
- Top of page
- Materials and methods
- Author contributions
JAK inhibitor therapy was recently approved for the treatment of MF patients. JAK inhibitors have proven to be effective at improving constitutional symptoms and reducing spleen size in MPN patients. However, they do not appreciably decrease disease allele burden and thus do not induce remission in patients (Pardanani et al, 2011; Tefferi, 2012). JAK inhibitors can block the aberrant JAK2 and JAK1 signalling induced by the cytokine storm associated with MPNs, and this may be the basis for improvement in MPN patients' constitutional symptoms (Pardanani et al, 2011; Tefferi, 2012). With the inability of JAK inhibitors to decrease the allele burden in MPN patients, exploration of alternative therapeutic approaches for MPN patients continues.
We initiated our studies to further our understanding of the requirements for JAK2-V617F-mediated signal transduction in an effort to uncover novel avenues for therapeutic intervention for MPNs. We recently demonstrated that EpoR/JAK2 signalling requires lipid raft formation (McGraw et al, 2012) and thus wanted to determine the potential role of lipid rafts in deregulated JAK2 signalling in MPNs. While previous studies support the notion that JAK2 functions in lipid rafts (Sehgal, 2003; McGraw et al, 2012), our studies are the first to demonstrate that the MPN driver JAK2-V617F co-localizes with lipid rafts (Figs 1 and 2). Localization of this tyrosine kinase to lipid rafts is not seen in all cells, largely because not all cells exhibit raft staining (Fig 1). This may be due to the dynamic nature of lipid rafts, which is influenced by factors such as variability in raft size and half-life (Harder & Simons, 1997; Kurzchalia & Parton, 1999; Pralle et al, 2000; Edidin, 2001; Anderson & Jacobson, 2002). In addition, only a minor fraction of JAK2 was associated with DRMs. This is not surprising for multiple reasons. First, rafts are dynamic in nature and all cells did not display raft staining. Second, JAK2 is a cytoplasmic protein and more recently has been found in the nucleus of cells, including MPN cells (Dawson et al, 2009; Rinaldi et al, 2010). Third, our hypothesis is that JAK2-V617F is associated with a transmembrane receptor, such as a cytokine receptor (e.g. EpoR). Therefore, JAK2-V617F is not physically present in rafts per se, but rather associated with a protein in rafts. DRM isolation experiments utilized an overnight ultracentrifigation spin and it is likely that some JAK2 protein would not maintain its interaction with raft-associated proteins during this protocol and thus fractionate with the remainder of the JAK2, which is non-raft associated.
Using agents that disrupt cholesterol in the plasma membrane, we found that JAK2 and STAT5 activation in JAK2-V617F-dependent cells were dependent on cholesterol in the plasma membrane, while JAK2 and STAT5 activation in K562 cells, which express wildtype JAK2, were not (Fig 3). JAK2-V617F requires cytokine receptors for activation (Lu et al, 2005, 2008) while wildtype JAK2 activation in K562 cells is probably induced by the BCR-ABL1 tyrosine kinase (Xie et al, 2001; Hantschel et al, 2012). We believe that JAK2 activation by mechanisms that involve a cell surface receptor may be more sensitive to lipid raft disruption than activation of JAK2 by non-receptor mechanisms, such as BCR-ABL1. The BCR-ABL1-induced constitutive JAK/STAT signalling may not rely on lipid rafts because the cytoplasmic BCR-ABL1 tyrosine kinase may activate or signal to these molecules directly (Xie et al, 2001; Hantschel et al, 2012). Lipid rafts may play an integral role in the receptor scaffolding function for JAK2-V617F activation by coordinating the proper molecular complexes at the cell surface (Simons & Toomre, 2000; Lu et al, 2005, 2008).
MPN model cell lines are also more sensitive to statin treatment than BCR-ABL1 positive K562 cells. We find MPN cells are sensitive to single digit micromolar statins, which is similar to certain acute myeloid leukaemia cell lines, but significantly less than cells from a variety of solid tumours (Dimitroulakos et al, 1999). This may, in part, be due to the inherent driving oncogenic lesions in these cells, compared to other cancers. Statin treatment also inhibited the growth of primary MPN cells. Importantly, the growth of primary MPN cells is more sensitive to statins than cells from healthy controls (Fig 6). This is in agreement with other studies looking at the effect of statins on normal and neoplastic haematopoietic cell growth, where normal haematopoietic cells are not sensitive to statins until high doses are achieved (Newman et al, 1997; Dimitroulakos et al, 1999; Dai et al, 2007). This suggests statins may be considered as a potential therapeutic agent for MPNs, although the effect of statins on JAK2-V617F-negative MPNs needs to be determined.
Although a requirement for lipid rafts in JAK2 signalling could provide a mechanistic rationale for the use of statins to inhibit MPN cells, cholesterol is not the only end product of the mevalonate pathway (Demierre et al, 2005). While we have not obtained evidence that statins inhibit lipid raft formation in our cell systems, we showed that statins do appear to inhibit the localization of JAK2-V617F to lipid rafts (Fig 4D), which in effect also targets the requirement of rafts for signalling. Importantly, statins also inhibit protein prenylation by inhibiting the production of farnesyl pyrophosphate and geranygeranyl pyrophosphate, two other end products of the mevalonate pathway downstream of HMG-CoA reductase. Interestingly, it has been shown that EpoR cell surface expression requires protein geranylgeranylation (Hamadmad & Hohl, 2007). It is possible that the effects of statins in MPN cells may be mediated through protein prenylation, perhaps through inhibition of a requisite cytokine receptor for JAK2-V617F-mediated signalling. In our efforts to ascertain further details regarding the mechanism of statin affects on MPN cells, we have determined that adding back geranylgeranyl pyrophosphate to cells can reduce the statin-induced loss of viability of cells, but does not significantly restore proliferation of cells (not shown). Thus, while the mechanistic details by which statins inhibit MPN cell growth are probably complex and remain to be elucidated, our data suggest that statins may be a candidate to be used as a potential therapeutic strategy to target MPN cells. While these details will be the focus of future studies, it is important to note that statins induce MPN cell apoptosis (Fig 5A and B). This is significant because JAK2 inhibitors fail to decrease allele burden in patients and additional therapeutic approaches to complement JAK2 inhibitors are needed, especially ones that can contribute to an apoptotic response in MPN cells.
While our work is the first to directly investigate the role of lipid rafts and cholesterol in MPN cells, there is additional evidence that suggests cellular cholesterol levels could play a role in MPN cell biology. Mice deficient in cholesterol efflux transporters ABCA1 and ABCG1 display an MPN-like phenotype (Yvan-Charvet et al, 2010). This suggests that an increase in cellular cholesterol in haematopoietic cells can lead to an MPN-like phenotype. In fact, Yvan-Charvet et al (2010) demonstrated that haematopoietic stem and progenitor cells from these mice displayed aberrant proliferation, and that removal of cholesterol from these cells restored a normal proliferative phenotype. These studies clearly indicate that cellular cholesterol can regulate growth control pathways of haematopoietic stem and progenitor cells, and that increasing cholesterol levels can lead to aberrant myeloproliferation. Thus, cellular cholesterol may play an important role in the development of human MPNs. Our work, showing that alteration of membrane cholesterol with lipid raft disrupting agents inhibits JAK2-V617F signalling, together with the results reported by Yvan-Charvet et al (2010), suggests that altering cholesterol in haematopoietic stem and progenitor cells may affect cell signalling that leads to JAK2-V617F-driven myelopoiesis. Thus, altering cellular cholesterol or inhibiting localization of JAK2-V617F to lipid rafts, perhaps through the use of statins, may be an effective approach to target the aberrant myelopoiesis associated with MPNs.
The use of statins to treat MPN patients has been previously rationally suggested (Hasselbalch & Riley, 2006; Hasselbalch, 2012). This hypothesis is based on the antithrombotic, antiproliferative, proapoptotic, and antiangiogenic effects of statins and the role thrombohaemorrhagic complications play in MPNs. The use of statins in the treatment of MPNs has been discussed in the context of the potential role of chronic inflammation in the development of MPNs. The antiinflammatory effects of statins may be advantageous to MPN patients as chronic inflammation may be a driving force toward clonal evolution as well as a deadly myelofibrotic state (Hasselbalch & Riley, 2006; Hasselbalch, 2012). For example, tumour necrosis factor α (TNFα) may contribute to clonal expansion of MPN cells (Fleischman et al, 2011) and simvastatin lowers TNFα expression in myeloid cells in patients (Ferro et al, 2000). Also, MPN patients have an increased risk of developing both haematological and non-haematological secondary cancers and this may be due to the elevated inflammation associated with MPNs (Vannucchi et al, 2009; Frederiksen et al, 2011). Thus, in addition to the potential direct effects of statins on MPN cells, statins may also contribute to the amelioration of disease through their antiinflammatory effects.
In summary, we found that JAK2-V617F is associated with lipid rafts and that signalling by this constitutively activated kinase is dependent on proper lipid raft formation. Statins reduce JAK2 localization to lipid rafts, induce apoptosis of MPN cells, and inhibit colony formation of primary cells from MPN patients. Given that JAK inhibitors have not had success at reducing allele burden in MPN patients, additional therapeutic approaches are needed in order to induce remission in these patients. Our work suggests that statins might be an effective component of a therapeutic strategy for MPN patients. Additional studies are needed to investigate the potential efficacy of statins, alone and in combination with JAK inhibitors, as a potential therapeutic option for MPNs.