The V617F activating mutation of janus kinase 2 (JAK2), a kinase essential for cytokine signalling, characterizes Polycythemia vera (PV), one of the myeloproliferative neoplasms (MPN). However, not all MPNs carry mutations of JAK2, and in JAK2-mutated patients, expression of JAK2V617F does not always result in clone expansion. In the present study, we provide evidence that inflammation-linked cytokines are required for the growth of JAK2V617F-mutated erythroid progenitors. In a first series of experiments, we searched for cytokines over-expressed in PV using cytokine antibody (Ab) arrays, and enzyme-linked immunosorbent assays for analyses of serum and bone marrow (BM) plasma, and quantitative reverse transcription–PCRs for analyses of cells purified from PV patients and controls. We found that PV patients over-expressed anti-inflammatory hepatocyte growth factor (HGF) and interleukin-11 (IL-11), BM mesenchymal stromal cells (BMMSCs) and erythroblasts being the main producers. In a second series of experiments, autocrine/paracrine cytokine stimulation of erythroblasts was blocked using neutralizing Abs specific for IL-11 or c-MET, the HGF receptor. The growth of JAK2V617F-mutated HEL cells and PV erythroblasts was inhibited, indicating that JAK2-mutated cells depend on HGF and IL-11 for their growth. Additional experiments showed that transient expression of JAK2V617F in BaF-3/erythropoietin receptor cells, and invalidation of JAK2V617F in HEL cells using anti-JAK2 small interfering RNA, did not affect HGF and IL-11 expression. Thus, anti-inflammatory HGF and IL-11 are upregulated in PV and their overproduction is not a consequence of JAK2V617F. As both cytokines contribute to the proliferation of PV erythroblasts, blocking the c-MET/HGF/IL-11 pathways could be of interest as an additional therapeutic option in PV.
Myeloproliferative neoplasms (MPNs) constitute a group of three clonal diseases: Polycythemia vera (PV), essential thrombocythemia (ET) and primary myelofibrosis. About half of MPN patients present with activating mutations in the janus kinase 2 (JAK2) gene, which encodes for a tyrosine kinase essential for the signalling of many cytokines (Baxter et al., 2005; James et al., 2005; Kralovics et al., 2005). Among MPNs, PV is characterized by an excessive production of erythrocytes, resulting in an elevated hematocrit associated with variable leukocytosis and thrombocytosis. Activating mutations of JAK2, most often V617F (JAK2V617F), are found in >95% of PV cases. It is well established that erythroid differentiation depends on erythropoietin (Epo) and that JAK2 is the main signal transducer activated by Epo receptors (Epo-Rs). Yet, in murine models, JAK2V617F is associated with polycythemia only when expressed at high levels (>50% JAK2V617F-mutated alleles), whereas a significant proportion of PV patients carry <50% JAK2V617F (Lacout et al., 2006; Lippert et al., 2006; Wernig et al., 2006; Tiedt et al., 2008). We now know that the JAK2V617F mutation is frequently associated with other molecular genetic abnormalities and that the JAK2V617F mutation can occur several times in certain patients; moreover, the presence of JAK2V617F does not ensure expansion of mutated progenitors (Nussenzveig et al., 2007; Kralovics, 2008; Lambert et al., 2009; Schaub et al., 2009; Cleyrat et al., 2010). Hence, the role of JAK2V617F in MPN and the consequences of its expression in hematopoietic progenitors remain incompletely understood.
Recently, a particular haplotype of chromosome 9p including the gene JAK2, called the ‘46/1’ or ‘GGCC’ haplotype, was found to be associated with a predisposition to MPN, with or without mutation of JAK2 (Jones et al., 2009; Kilpivaara et al., 2009; Olcaydu et al., 2009). The ‘46/1’ haplotype is also associated with chronic inflammation, notably inflammatory bowel disease (Ferguson et al., 2010). Interestingly, although most evident in primary myelofibrosis, signs of chronic inflammation, such as increased production of inflammatory cytokines, interleukin (IL)-6 and vascular endothelial growth factor, are also observed in PV and in ET (Wickenhauser et al., 1999; Musolino et al., 2002; Panteli et al., 2005; Le Bousse-Kerdilés and Martyré, 1999). Inflammation-associated symptoms were long considered consequences of activating mutations of JAK2, but recent clinical trials of JAK2 antagonists showed that drugs not strictly specific for JAK2 were most efficient in reducing inflammation-linked cytokine levels and symptoms, indicating that molecules other than JAK2 were likely involved (Tefferi, 2010).
We begun to investigate the possible stimulation of PV progenitors by inflammation-linked, JAK2-activating cytokines because of previous observations of high messenger RNA (mRNA) levels of JAK2 in patients with erythrocytosis, either reactive or secondary (secondary erythrocytosis (SE)—high expression of JAK2 wild type (JAK2WT)) or primary (PV—high expression of both JAK2WT and JAK2V617F) (Lippert et al., 2006). In PV, high mRNA expression of JAK2 could be due to multiple copies (>2 per cell) of the JAK2 gene, but this concerns a small minority of patients (Najfeld et al., 2007). As JAK2V617F is not sufficient for the development of the PV phenotype and by analogy with SE with increased Epo secretion, we reasoned that increased JAK2 mRNA expression in PV could be due to chronic stimulation by proerythroid, JAK2-activating cytokines other than Epo, which is low or undetectable in PV. To investigate this hypothesis, we used cytokine arrays to compare profiles of blood serum, bone marrow (BM) plasma and culture supernatants of BM mesenchymal stromal cells (BMMSCs) from PV and SE patients. Several proerythroid molecules were found to be over-expressed in PV: all were known to be linked to inflammation and were produced by BMMSCs or/and by hematopoietic progenitors themselves. These molecules were studied at the protein and mRNA levels using enzyme-linked immunosorbent assays (ELISAs) and quantitative reverse transcription–PCRs in purified cells: BMMSCs, CD34+ progenitors, glycophorin A-positive (GPA+) erythroblasts and CD3+ lymphocytes. Relevance to the pathogenesis of PV was investigated by studying the effects on in vitro growth of JAK2V617F-mutated erythroblasts. In addition, transfection of JAK2V617F in BaF-3/Epo-R cells, and anti-JAK2 small interfering (si)RNA in JAK2V617F-mutated HEL cells, was used to determine whether their production was a consequence of JAK2V617F.
Over-expression of HGF and IL-11 in biological samples from PV patients
First, cytokine arrays were used to analyse cytokine expression in pools of serum from PV patients, compared with SE patients and healthy donors (HDs), and in pools of BM plasma from PV patients, compared with SE patients (Supplementary Figure 1). SE and PV differed from HD in having high levels of IL-8, leptin and macrophage chemotactic protein-1 (MCP-1) in serum. PV differed from SE in having high levels of hepatocyte growth factor (HGF) in serum, and high levels of tissue inhibitor of metalloproteases-1 (TIMP-1) in BM plasma. IL-8, leptin, MCP-1, HGF and TIMP-1 are molecules linked to inflammation, as was IL-11, previously reported elevated in PV (Hermouet et al., 2002; Corre-Buscail et al., 2005).
Second, IL-8, leptin, MCP-1, HGF, TIMP-1, IL-11 and also IL-6, a major pro-inflammatory cytokine, were measured in PV and SE using ELISAs and levels were compared with those of HD. As shown in Table 1, analysis of serum of SE and PV patients confirmed high levels of IL-8, leptin, HGF and MCP-1; TIMP-1 levels were low. PV differed from SE by lower levels of leptin in serum (male patients only) and higher levels of IL-11 and HGF. For comparison, IL-11 and HGF measured in the serum of 15 ET patients were either not detected (IL-11) or comparable to values observed for control SE patients: median serum HGF values, in pg/ml, were 1380 in SE (n=34) and 1594 in ET (n=15) vs 4673 in PV (n=49). TIMP-1 levels were confirmed as elevated in PV compared with SE only in BM plasma. Although serum IL-6 was elevated for some PV patients, overall IL-6 levels were not found to be significantly high in PV.
Of the three molecules confirmed as elevated in PV by ELISA, HGF and IL-11 seemed to be most interesting. Indeed, it was established that HGF could induce IL-11, an established stimulant of erythropoiesis (Quesniaux et al., 1992; Schwertschlag et al., 1999; Matsuda-Hashii et al., 2004; Ishii et al., 2007). In PV patients, serum levels of HGF and IL-11 were correlated with blood neutrophil counts (n=21, r=0.71, P<0.005) and hematocrit (n=45, r=0.43, P<0.01), respectively. Consequently, the rest of the study focused on HGF and IL-11.
Sequential production of HGF, IL-11, IL-6 and IL-8 by PV BMMSCs
According to the literature, HGF induces the production of IL-11 and TIMP-1, IL-11 induces expression of IL-6 and IL-8, and TIMP-1 acts on the cleavage and activation of HGF. Accordingly, BM plasma levels of IL-11, IL-6 and IL-8, as well as serum levels of TIMP-1 and HGF (Table 1), were correlated in PV (IL-11/IL-6: n=29, r=0.51, P<0.01; IL-11/IL-8: n=60, r=0.43, P<0.01; TIMP-1/HGF: n=32, r=0.42, P=0.02).
Regulation of IL-11 by HGF and that of IL-6 and IL-8 by IL-11 were studied in BMMSCs, an established source of cytokines linked to inflammation. ELISA performed on supernatants of 31 BMMSC cultures (12 PV, 5 idiopathic erythrocytosis (IE), 14 SE) confirmed that IL-11 and IL-8 were present at high levels in supernatants of PV BMMSCs, but detected little or no HGF (Table 1). IL-11 and IL-8 levels were correlated in BMMSC supernatants (n=31, r=0.74, P<0.01). BMMSC cultures (4 PV, 2 IE, 9 SE) were then stimulated either with HGF to study production of IL-11 and TIMP-1, or with IL-11 to study production of IL-6 and IL-8. Exposure to HGF increased IL-11 and TIMP-1 produced by PV BMMSCs (Figures 1a and b, respectively), whereas exposure to IL-11 significantly increased the production of both IL-8 and IL-6 (Figures 1c and d, respectively).
To verify that PV BMMSC cultures did not contain clonal, JAK2V617F-mutated cells, BMMSCs of five PV patients with blood granulocytes positive for JAK2V617F were tested for the presence of JAK2V617F: all expressed only JAK2WT.
High mRNA expression of HGF and IL-11 in purified erythroblasts from PV patients
We also studied CD3+ lymphocytes, previously described as producers of IL-11 in PV (Ishii et al., 2007); CD34+ progenitor cells; and GPA+ erythroblasts, all purified from the BM of six PV patients and seven SE controls. The %JAK2V617F measured in granulocyte DNA of the six PV patients ranged from 10 to 77%. Expression of HGF, IL-11 and IL-6 mRNAs was studied using reverse transcription–quantitative PCRs (Figure 2). In PV and in SE, BMMSCs were high producers of HGF and IL-6 (Figure 2a). Although protein levels of IL-11 measured by ELISA in supernatants of cultured BMMSCs were clearly higher in PV than in SE, no significant difference was detected in IL-11 mRNA levels in BMMSCs from PV and SE patients (Figure 2a). In BM CD3+ lymphocytes (Figure 2b) and in CD34+ progenitors (Figure 2c), mRNA expression of HGF, IL-11 and IL-6 was negligible. The BM origin of CD3+ lymphocytes in our series may explain the discrepancy with the results of Ishii et al. (2007), who studied blood T-lymphocytes. In contrast to other cell types, PV GPA+ erythroblasts expressed high mRNA levels of both HGF and IL-11 (Figure 2d). In PV and SE erythroblasts, mRNA levels of HGF, IL-11 and IL-6 were correlated (r=1.00, P<0.02 each). No correlation was found between JAK2V617F (% or mRNA levels), measured in each cell type, and HGF, IL-11 or IL-6 mRNA levels.
Increased expression of gp130 and STAT3 mRNAs in purified PV erythroblasts
Expression of c-MET (the HGF receptor), glycoprotein (gp)130 (the receptor chain shared by IL-11 and IL-6), JAK2WT, JAK2V617F, signal transducer and activator of transcription (STAT)3 and STAT5 was analysed in BMMSCs, in CD3+ lymphocytes and in purified BM progenitors using reverse transcription–quantitative PCRs. No significant difference in expression of total JAK2, STAT3 or STAT5 was observed between PV and SE in BMMSCs, CD3+ lymphocytes and CD34+ progenitors (not shown). In these cell types, there was no correlation between mRNA levels of HGF, c-MET, IL-11, IL-6, gp130, total JAK2, STAT3 or STAT5. In purified GPA+ erythroblasts, mRNA levels of c-MET, total JAK2 and STAT5 were similar in PV and in SE (Table 2). In contrast, mRNA levels of gp130 and STAT3, both activated by IL-11 and IL-6, were significantly elevated in PV erythroblasts. Moreover, in PV erythroblasts mRNA levels of c-MET, HGF, IL-11 and IL-6 were correlated, as were mRNA levels of HGF, IL-6, gp130 and STAT5 (Supplementary Table 3). No correlation was found with JAK2V617F mRNA levels. Taken together, correlation between HGF, IL-11 and IL-6 mRNA levels and increased HGF, IL-11, gp130 and STAT3 mRNA expression was consistent with activation of an HGF/c-MET/IL-11/IL-6/gp130/STAT3 cascade in PV erythroblasts.
Inhibition of in vitro growth of PV erythroid progenitors and HEL cells by neutralizing Abs directed againstc-MET and IL-11
We previously reported that neutralizing anti-IL-11 antibody (Ab) could block BFU-E and CFU-E growth in cultures of BM mononuclear cells (BMMCs) from PV patients, indicating partial dependence of PV erythroid progenitors on IL-11 (Corre-Buscail et al., 2005). HGF was blocked using neutralizing Ab directed against c-MET in cultures of HEL cells, which are homozygous for JAK2V617F and in liquid cultures of purified GPA+ erythroblasts from four patients with JAK2V617F-mutated PV (JAK2V617F allelic ratio: 37, 91, 95 and 98%) (Figure 3). Compared with control cultures without Ab, the number of viable cells in PV cultures incubated for 24 h with anti-c-MET Ab (2 μg/ml) was consistently decreased by more than 40%. Decreased viability in the presence of anti-c-MET Ab concerned JAK2V617F-mutated cells, as it was observed for HEL cells (100% JAK2V617F-mutated) and for purified erythroblasts of PV patients, three with >90% cells bearing the JAK2V617F mutation. A similar inhibition of HEL and PV erythroblast growth was also observed with anti-IL-11 Ab (2 μg/ml). Anti-IL-8 Ab (2 μg/ml)—used as a control Ab—had no effect in these conditions. Incubation of erythroblasts from SE patients with anti-c-MET Ab and anti-IL-11 Ab resulted in no inhibition (two patients) or a weak (−20%) inhibition (one patient) of erythroblast growth.
JAK2V617F independence of HGF and IL-11 expression
First, HGF level was measured in blood serum of three PV patients who were found to be positive for endogenous erythroid colonies, but who carried no mutation of JAK2 (negative for V617F and exon 12 mutations). For 2/3 of these ‘JAK2WT’ patients, serum HGF was found to be elevated (3216 and 4795 pg/ml) and comparable to the average serum HGF level of JAK2V617F-positive PV patients (4409 pg/ml, vs less than 1285 pg/ml in HDs).
Second, HGF, IL-11 and IL-6 mRNA expression was studied in HEL and UKE-1, two human cell lines homozygous for the JAK2V617F mutation. As shown in Table 2, high expression of IL-6, IL-11 or HGF was observed in HEL, but not UKE-1, indicating that homozygosity for JAK2V617F is not sufficient to induce high expression of HGF, IL-11 and IL-6 (of note, UKE-1 cells were derived from a patient initially diagnosed with ET).
Third, we used anti-JAK2 siRNA to block expression of JAK2V617F in HEL cells. As shown in Figure 4a, JAK2V617F mRNA expression was almost abolished. JAK2 expression was also inhibited at the protein level, by up to 70%, and a modest inhibition of cell survival was observed (Supplementary Figure 2). RT–qPCR and ELISA analysis of mRNA and protein expression showed that levels of HGF, IL-11 and IL-6 were not affected by invalidation of JAK2V617F (Figures 4b–d, Supplementary Figure 2).
Fourth, the effects of JAK2V617F were studied in murine BaF-3/Epo-R cells (Figure 5). Both WT and V617F human JAK2 were successfully expressed in BaF-3/Epo-R cells, as assessed by allele-specific RT–qPCR and western blotting analysis, which showed a clear increase in tyrosine phosphorylation of JAK2 in JAK2V617F transfectants (Figure 5a). In comparison, the increase in tyrosine phosphorylation of STAT5 and STAT3 in BaF-3/Epo-R cells expressing JAK2V617F was modest. Expression in BaF-3/Epo-R cells of wild-type and V617F-mutated JAK2, as well as another JAK2 mutant, JAK2L611V/V617F (Cleyrat et al., 2010), had no effect on mRNA expression of murine IL-11 or HGF, the latter remaining below detection level (Figure 5b). However, all forms of JAK2 induced a moderate increase in mRNA and protein levels of murine IL-6. Thus, JAK2WT and, to a lesser extent, JAK2V617F and other JAK2 mutants may induce expression of pro-inflammatory IL-6 in certain cell types, but have no effect on anti-inflammatory HGF and IL-11.
Last, we analysed HGF serum levels in PV patients in relation to JAK2V617F allelic burden; SE and ET patients negative for JAK2V617F were used as controls (Table 3). JAK2V617F-negative ET patients showed a moderate elevation of serum HGF levels, as did SE patients. HGF serum levels were further and similarly elevated for ET and PV patients with 1–50% JAK2V617F, with no correlation between serum HGF level and %JAK2V617F. PV patients with the highest serum HGF levels (>3500 pg/ml) had the highest JAK2V617F allelic burden: 58.7%±27 vs 41.0%±28.9 for PV patients with <3500 pg/ml (P=0.044), also with no correlation between serum HGF level and %JAK2V617F. Consistent with the upregulation of HGF being independent of JAK2V617F, a high secretion of HGF (>3500 pg/ml HGF) was typical of PV with >50% JAK2V617F, but was also observed with a lesser frequency, in other categories of patients, including JAK2V617F-negative PV, ET and SE. Hence, similar to very high hematocrit levels, high serum HGF was found to be associated with a high %JAK2V617F, but JAK2V617F was neither required nor sufficient (UKE-1 cells; BaF-3/Epo-R transfectants) to induce HGF production. Rather, a high level of HGF in serum may be considered as a marker of the size of the MPN clone and disease severity.
Upregulation of inflammation-linked cytokines—IL-6, IL-10, vascular endothelial growth factor and tumor necrosis factor alpha—is now well established and is generally assumed to be a consequence of JAK2V617F. Yet, symptoms of chronic inflammation are observed both in JAK2V617F-mutated MPN and in MPN with no mutation of JAK2, particularly primary myelofibrosis (Geissler et al., 1998; Le Bousse-Kerdilés and Martyré, 1999; Wickenhauser et al., 1999; Hermouet et al., 2002; Musolino et al., 2002; Corre-Buscail et al., 2005; Panteli et al., 2005). The findings of the present study of PV patients support the opposite reasoning, that is, that upregulation of certain cytokines linked to inflammation is independent of the acquisition of the JAK2V617F mutation. This is consistent with recent reports that certain anti-JAK2 drugs, particularly those that are not strictly specific for JAK2, efficiently decrease inflammation-associated symptoms—splenomegaly, fever, weight loss—without affecting the JAK2V617F burden (Hitoshi et al., 2010; Tefferi, 2010; Verstovsek, 2010).
Our study demonstrates that anti-inflammatory cytokines HGF and IL-11, as well as gp130, the signalling receptor chain common to IL-11 and IL-6, are upregulated in PV independently of JAK2V617F. In addition, TIMP-1, a molecule that controls HGF activity and has been reported to be over-expressed in primary myelofibrosis, was also found to be elevated in PV (Ho et al., 2007; Kopitz et al., 2007). HGF, IL-11 and IL-6 are produced by JAK2WT BMMSCs and JAK2V617F-mutated erythroid progenitors; they act in cascade, activate the STAT3 pathway and stimulate myeloid hematopoiesis. Thus, two autocrine/paracrine HGF/IL-11 loops appear to cooperate in PV BM: one concerns BMMSCs, main producers of HGF in vitro; the other concerns erythroblasts, main producers of IL-11 (Figure 6). Importantly, hematocrit and serum IL-11 were correlated in PV and Abs blocking IL-11 or c-MET/HGF inhibited the growth of JAK2V617F-mutated erythroblasts. No correlation was found between JAK2V617F and HGF/IL-11 levels in patients, and in vitro JAK2V617F had no effect on HGF and IL-11 mRNA expression.
Autocrine/paracrine HGF/IL-11 loops do not seem to be activated in ET as levels of HGF were only moderately increased and IL-11, an established stimulant of erythropoiesis inducible by HGF, was not detected at all in the serum of ET patients. Altogether, our data are consistent with the upregulation of both HGF and IL-11 in MPN being characteristic of PV and independent of JAK2V617F.
Consistent with independence from JAK2V617F, deregulation of HGF, IL-11 and IL-6 has been reported in other haematological malignancies, notably multiple myeloma (Derksen et al., 2003; Giuliani et al., 2004). In both PV and multiple myeloma, BMMSCs were found to express high levels of IL-11 and IL-6, possibly in response to paracrine stimulation by HGF and IL-11 produced by malignant progenitors (Uchiyama et al., 1993; Zdzisinska et al., 2008). In PV erythroblasts as in malignant plasma cells, over-expression of HGF and IL-11 was associated with high expression of the receptors for these cytokines, c-MET and gp130. Interestingly, the receptor for HGF, c-MET, belongs to a family of so-called ‘dependence receptors’, which induce cell death unless they are activated by their ligand (Shinomiya et al., 2004; Bernet and Mehlen, 2007). Inversely, disturbing the equilibrium between a dependence receptor and its ligand, either by over-expression of ligand or by deletion or inactivation of the receptor, leads to inappropriate cell survival and proliferation, accumulation of genetic abnormalities and, eventually, cell transformation—a succession of events found in solid tumors and in chronic haematological malignancies, such as MPN and multiple myeloma. This suggests that the HGF/c-MET pathway may exert a fundamental role in the maintenance of normalcy in dividing cells. Indeed, increased production of HGF is a common cellular response to hypoxia, frequent in the context of cancer and chronic inflammation (Tacchini et al., 2001; Kitajima et al., 2008). In addition to PV erythroblasts and myeloma cells, many solid tumors depend on sustained HGF/c-MET activity for their growth and survival. In consequence, HGF and c-MET have become novel targets in the therapy of solid tumors and multiple myeloma (Du et al., 2007; Sattler and Salgia, 2007; Stellrecht et al., 2007; Comoglio et al., 2008). Our results suggest that blocking the HGF/c-MET and IL-11 pathways could be a novel therapeutic option in PV, perhaps for those patients who are no longer controlled by phlebotomy or/and hydroxurea (Girodon et al., 2008). In this regard, the established ability of interferon-α to suppress expression of c-MET and IL-11 may contribute to the disappearance of JAK2V617F-mutated clonal cells in PV patients treated with interferon-α (Aman et al., 1996; Radaeva et al., 2002; Kiladjian et al., 2006).
In PV erythroblasts, over-expression of HGF and IL-11 was associated with increased expression of STAT3 and gp130, the receptor chain common to IL-11 and IL-6 responsible for intracellular signalling. According to Jenkins et al. (2005, 2007), a definite threshold of STAT3 signalling is critical for the regulation of normal hematopoiesis. These authors showed that abnormal activation of STAT3 through receptors of cytokines from the IL-6 family induced MPN-like disorders in mice, as well as an increase in expression of IL-6 and IL-11; HGF was not studied. The excellent correlations found in erythroblasts between mRNA levels of STAT3-activating molecules gp130, IL-6, HGF and those of STAT5, presumably activated in PV because of the expression of JAK2V617F, are consistent with a tight regulation of the equilibrium between the STAT3 (cell survival) and STAT5 (proliferation) pathways in these cells. By providing a growth advantage, HGF/IL-11-induced excessive activation of the STAT3 pathway likely influences phenotype in MPN, as shown recently in chronic myeloid leukemia for BCR—ABL-induced phenotypes in animal models (Coppo et al., 2009).
Last, as deregulation of HGF and/or IL-11, gp130 and STAT3 is common to many tumoral processes, the genetic events resulting in over-expression of these molecules in PV are unlikely to be primary events causing PV, but rather, frequent parallel events that occur independently of JAK2 mutation(s) and are also found in other malignancies, as described in MPN for several cytogenetic abnormalities (Kralovics, 2008; Schaub et al., 2009). Parallel or ‘passenger’ genetic events responsible for the deregulation of HGF and IL-11 would be consistent with the observations that not all PV patients show high levels of HGF, and that JAK2V617F alone does not ensure a PV phenotype nor the expansion of progenitors bearing this mutation. Candidate genes for deletion or mutation include c-MET, often mutated in cancer and leading to inappropriate cell proliferation, thus predisposing cells to acquire mutations in other genes. Other possible candidates are mutations of genes encoding for IL-11 receptors, or DNA demethylation of the promoters of the genes encoding for HGF and IL-11. One may also consider the possibility of acquired mutations of genes of the oxygen-sensing pathway, leading to increased HGF expression.
In summary, as described in multiple myeloma, autocrine/paracrine anti-inflammatory HGF and IL-11 are over-expressed in PV and likely contribute to the expansion of clonal erythroblasts, in a JAK2V617F-independent manner. Consequently, the value of blocking the c-MET/HGF pathway, tested as an additional therapeutic option in multiple myeloma and in solid tumors, might be of interest in PV. Finally, the genetic events responsible for over-expression of HGF and, subsequently, of IL-11 and IL-6, deserve investigation in PV, as well as in multiple myeloma and other hematopoietic malignancies.
Materials and methods
With consent, serum and/or BM plasma samples were obtained from 122 patients at the time of diagnosis (58 PV, 37 SE, 12 IE, 15 ET). Diagnosis of PV and ET was made according to the World Health Organization criteria 2002 (Vardiman et al., 2002). All PV had low serum Epo and endogenous erythroid colonies; all SE had elevated serum Epo, no endogenous erythroid colonies and an identified cause of SE; IE patients had normal serum Epo and no endogenous erythroid colonies. JAK2V617F status was established for 107/122 patients; 34/34 SE and 8/8 IE tested were negative; 48/51 PV and 5/15 ET tested were found to be positive for JAK2V617F. Serum was also obtained from 17 HD. Owing to occasional insufficient sample collection, serum and BM assays were not performed for all patients.
UKE-1 and BaF-3/Epo-R cells were kindly provided by Dr Walter Fiedler (Hamburg, Germany) and by Dr Radek Skoda (Basel, Switzerland), respectively. Human cell lines HEL and UKE-1 were grown in RPMI and 10% foetal calf serum (FCS). Murine BaF-3/Epo-R cells were grown in RPMI, 2% FCS and 1 IU/ml Epo.
BM samples were centrifuged at 150 g for 10 min without brake. Supernatant (=BM plasma) was pipetted off and cell pellets were suspended in RPMI; BMMCs were separated on Ficoll medium (1.077 g). After washing, aliquots of BMMCs were either used for immediate cultures or cell purification, or frozen in dimethyl sulfoxide in liquid nitrogen. Adherent BMMSCs were obtained after 12 days of culture as described previously (Hermouet et al., 2002; Corre-Buscail et al., 2005). Blood granulocytes were isolated from the lower interphase of a Ficoll density gradient (Dobo et al., 2004; Lippert et al., 2006). CD34+ cells (purity >95%), CD3+ lymphocytes (purity >98%) and GPA+ erythroblasts (purity >98%) were purified from frozen BMMCs using Microbead kits (Miltenyi Biotec, Germany). Cell pellets were kept in 500 μl Trizol (Invitrogen, Frederick, MD, USA) and stored at −80 °C until mRNA extraction.
Quantification of JAK2WT and JAK2V617F
Genomic DNA was prepared using QiaAmp DNA mini-kit (Qiagen, Valencia, CA, USA), and JAK2WT and JAK2V617F allele-specific quantitative PCRs were performed in genomic DNA and in cDNA as described (Lippert et al., 2006, 2009).
Cytokine profiles of serum and BM plasma, kept frozen at −80 °C, were analysed using cytokine arrays VI and VII (RayBiotech, Inc., Norcross, GA, USA; the list of cytokines tested is shown in Supplementary Table 1), and ELISAs: human and murine IL-6 and IL-11 Quantikine kits (R&D Systems, Abingdon, UK); IL-8, leptin and HGF EASIA kits (BioSource Europe); and MCP-1 and TIMP-1 instant kits (Bender Medsystems, Vienna, Austria).
Quantitative mRNA studies of cytokines and signalling molecules
Cell pellets kept in Trizol were submitted to mRNA extraction. After DNAse treatment and reverse transcription, cDNAs were kept frozen at −80 °C. Quantitative PCRs for TIMP-1, HGF, IL-11, IL-6, c-MET, gp130 (sequence common to variant forms), STAT3 and STAT5 (sequence common to α and β forms), as well as β-actin and RPLP0, used as control genes, were performed with specific primers and probes (Supplementary Table 2) on Rotorgene 6000 (Corbett Research, Mortlake, NSW (Inner West), Australia). Dilutions of DNA preparations of pCR4-TOPO-TA constructs (Invitrogen) containing the relevant amplicons were used as copy number standards for qPCR assays. As less variation in mRNA expression was observed for RPLP0 than for β-actin, results were expressed as mean number of copies per 1000 copies of RPLP0.
Mutagenesis and transient expression of JAK2 constructs in BaF-3/Epo-R cells
Plasmid pCR2.1 containing the cDNA of JAK2WT (gift from Dr Jan Cools) was used as matrix for directed V617F mutagenesis. PCR was performed using Pfu Ultra DNA polymerase (Stratagene, Amsterdam, The Netherlands) with the following primers: forward, 5′-IndexTermCACAAGCATTTGGTTIndexTermGTAAATTATGGAGTATG-3′; reverse, 5′-IndexTermCATACTCCATTATTTAIndexTermCAACCAAATGCTTGTG-3′. PCR products were amplified into TOP 10 chemically competent bacteria (Invitrogen) and purified. Another PCR using Pfu Ultra was performed to extract WT and V617F JAK2 cDNA from plasmid pCR2.1 for cloning into pcDNA3.1. Mutagenesis was checked by sequencing. In all, 25 μg of pcDNA3.1 containing either JAK2WT or JAK2V617F cDNA was used to transfect 107 BaF-3/Epo-R cells using the Amaxa Nucleofector device (Amaxa, Cologne, Germany). Cells were then grown in 4 ml of RPMI medium supplemented with serum and Epo for 24 h, then harvested for immunoblotting and RT–qPCR studies. For cell signalling studies, transfected BaF-3/Epo-R cells were washed, deprived of Epo for 5 h in RPMI medium supplemented with 10% FCS, then stimulated for 10 min with Epo (25 IU/ml) in RPMI and 2% FCS, at 37 °C. Cells were harvested, washed twice in cold PBS, lysed in 60 μl RIPA buffer, then proteins (25 μg) were loaded on 10% SDS–PAGE and transferred to polyvinylidene fluoride membranes (Millipore Corporation, Billerica, MA, USA) (Corre-Buscail et al., 2005; Cleyrat et al., 2010). After blocking with 5% non-fat dry milk, membranes were incubated with Abs specific for p-STAT5 (Zymed, San Francisco, CA, USA), p-JAK2 and p-STAT3 (Cell Signalling, Danvers, MA, USA) or β-actin (Millipore Corporation), then stripped and re-probed with Ab specific for total JAK2 (Millipore Corporation), STAT3 and STAT5 (BD Bioscience, San Jose, CA, USA). Revelation was made with BM Chemiluminescence Blotting Substrate (Roche, Mannheim, Germany).
Silencing of JAK2 expression in HEL cells
5 × 104 HEL cells were incubated without, or with 10−4, 10−3, 10−2 or 10−1 μM of human JAK2 siRNA (Accell SMARTpool E-003146-00-0005, Thermo Fisher Scientific, Waltham, MA, USA) for 72 h and JAK2, IL-6, IL-11, HGF and RPLP0 mRNA, and protein levels were measured using the RT–qPCR and immunoblotting assays described above.
Semi-solid and liquid cultures of erythroid progenitors
BFU-E clonogenic assays were performed in collagen-based medium without cytokines or with stem cell factor (25 ng/ml) and in methylcellulose (medium H4531, StemCell Inc., Vancouver, Canada) (Dobo et al., 2004; Corre-Buscail et al., 2005). Purified BM erythroblasts were grown for 24 h in IMDM liquid medium and serum, with or without Epo, with neutralizing polyclonal Ab, used at 2 μg/ml: anti-HGF, anti-c-MET and anti-IL-11 Ab were from R&D. IL-8 Ab was from Endogen (Cambridge, MA, USA).
Pearson's and Spearman's rank correlation tests were used to investigate potential relationship between cytokine levels, %JAK2V617F and blood cell counts. The Mann–Whitney test was used to investigate differences in mRNA and protein levels, or groups of patients. P<0.05 was statistically significant.
Aman MJ, Bug G, Aulitzky WE, Huber C, Peschel C . (1996). Inhibition of interleukin-11 by interferon-alpha in human bone marrow stromal cells. Exp Hematol 24: 863–867.
Baxter EJ, Scott LM, Campbell PJ, East C, Fourouclas N, Swanton S, Cancer Genome Project et al. (2005). Acquired mutation of the tyrosine kinase JAK-2 in human myeloproliferative disorders. Lancet 365: 1054–1061.
Bernet A, Mehlen P . (2007). Dependence receptors: when apoptosis controls tumor progression. Bull Cancer 94: E12–E17.
Cleyrat C, Jelinek J, Girodon F, Boissinot M, Ponge T, Harousseau J-L et al. (2010). JAK2 mutation and disease phenotype: A double L611V/V617F in cis mutation of JAK2 is associated with isolated erythrocytosis and increased activation of AKT and ERK1/2 rather than STAT5. Leukemia 24: 1069–1073.
Comoglio PM, Giordano S, Trusolino L . (2008). Drug development of MET inhibitors: targeting oncogene addiction and expedience. Nat Rev Drug Discov 7: 504–516.
Coppo P, Dusanter-Fourt I, Vainchenker W, Turhan AG . (2009). BCR-ABL induces opposite phenotypes in murine ES cells according to STAT3 activation levels. Cell Signal 21: 52–60.
Corre-Buscail I, Pineau D, Boissinot M, Hermouet S . (2005). Erythropoietin-independent erythroid colony formation by bone marrow progenitors exposed to interleukins 11 and 8. Exp Hematol 33: 1299–1308.
Derksen PW, de Gorter DJ, Meijer HP, Bende RJ, van Dijk M, Lokhorst HM et al. (2003). The hepatocyte growth factor/Met pathway controls proliferation and apoptosis in multiple myeloma. Leukemia 17: 764–774.
Dobo I, Donnard M, Girodon F, Mossuz P, Boiret N, Boukhari R et al. (2004). Standardization and comparison of endogenous erythroid colony assay performed with bone marrow or blood progenitors for the diagnosis of polycythaemia vera. Hematol J 5: 161–167.
Du W, Hattori Y, Yamada T, Matsumoto K, Nakamura T, Sagawa M et al. (2007). NK4, an antagonist of hepatocyte growth factor (HGF), inhibits growth of multiple myeloma cells: molecular targeting of angiogenic growth factor. Blood 109: 3042–3049.
Ferguson LR, Han DY, Fraser AG, Huebner C, Lam WJ, Morgan AR et al. (2010). Genetic factors in chronic inflammation: Single nucleotide polymorphisms in the STAT-JAK pathway, susceptibility to DNA damage and Crohn’ disease in a New Zealand population. Mutat Res 690: 108–115.
Geissler K, Ohler L, Fodinger M, Kabrna E, Kollars M, Skoupy S et al. (1998). Interleukin-10 inhibits erythropoietin-independent growth of erythroid bursts in patients with polycythemia vera. Blood 92: 1967–1972.
Girodon F, Schaeffer C, Cleyrat C, Mounier M, Lafont I, Dos Santos F et al. (2008). Frequent reduction or absence of detection of the JAK2-mutated clone in JAK2V617F-positive patients within the first years of hydroxurea therapy. Haematologica 93: 1723–1727.
Giuliani N, Colla S, Morandi F, Rizzoli V . (2004). The RANK/RANK ligand system is involved in interleukin-6 and interleukin-11 up-regulation by human myeloma cells in the bone marrow microenvironment. Haematologica 89: 1118–1123.
Hermouet S, Godard A, Pineau D, Corre I, Raher S, Lippert E et al. (2002). Abnormal production of interleukin (IL)-11 and IL-8 in polycythemia vera. Cytokine 20: 178–183.
Hitoshi Y, Lin N, Payan DG, Markovtsov V . (2010). The current status and the future of JAK2 inhibitors for the treatment of myeloproliferative diseases. Int J Hematol 91: 189–200.
Ho CL, Lasho TL, Butterfield JH, Tefferi A . (2007). Global cytokine analysis in myeloproliferative disorders. Leuk Res 31: 1389–1392.
Ishii T, Zhao Y, Shi J, Sozer S, Hoffman R, Xu M . (2007). T cells from patients with polycythemia vera elaborate growth factors which contribute to endogenous erythroid and megakaryocyte colony formation. Leukemia 21: 2433–2441.
James C, Ugo V, Le Couedic JP, Staerk J, Delhommeau F, Lacout C et al. (2005). A unique clonal JAK2 mutation leading to constitutive signalling causes polycythaemia vera. Nature 434: 1144–1148.
Jenkins BJ, Roberts AW, Najdovska M, Grail D, Ernst M . (2005). The threshold of gp130-dependent STAT3 signalling is critical for normal regulation of hematopoiesis. Blood 105: 3512–3520.
Jenkins BJ, Roberts AW, Greenhill CJ, Najdovska M, Lundgren-May T, Robb L et al. (2007). Pathologic consequences of STAT3 hyperactivation by IL-6 and IL-11 during hematopoiesis and lymphopoiesis. Blood 109: 2380–2388.
Jones AV, Chase A, Silver RT, Oscier D, Zoi K, Wang YL et al. (2009). JAK2 haplotype is a major risk factor for the development of myeloproliferative neoplasms. Nat Genet 41: 446–449.
Kiladjian JJ, Cassinat B, Turlure P, Cambier N, Roussel M, Bellucci S et al. (2006). High molecular response rate of polycythemia vera patients treated with pegylated interferon alpha-2a. Blood 108: 2037–2040.
Kilpivaara O, Mukherjee S, Schram AM, Wadleigh M, Mullally A, Ebert BL et al. (2009). A germline JAK2 SNP is associated with predisposition to the development of JAK2(V617F)-positive myeloproliferative neoplasms. Nat Genet 41: 455–459.
Kitajima Y, Ide T, Ohtsuka T, Miyazaki K . (2008). Induction of hepatocyte growth factor activator gene expression under hypoxia activates the hepatocyte growth factor/c-Met system via hypoxia inducible factor-1 in pancreatic cancer. Cancer Sci 99: 1341–1347.
Kopitz C, Gerg M, Bandapalli OR, Ister D, Pennington CJ, Hauser S et al. (2007). Tissue inhibitor of metalloproteinases-1 promotes liver metastasis by induction of hepatocyte growth factor signaling. Cancer Res 67: 8615–8623.
Kralovics R, Passamonti F, Buser AS, Teo SS, Tiedt R, Passweg JR et al. (2005). A gain-of-function mutation of JAK-2 in myeloproliferative disorders. New Engl J Med 352: 1779–1790.
Kralovics R . (2008). Genetic complexity of myeloproliferative neoplasms. Leukemia 22: 1841–1848.
Lacout C, Pisani DF, Tulliez M, Gachelin FM, Vainchenker W, Villeval JL . (2006). JAK2V617F expression in murine hematopoietic cells leads to MPD mimicking human PV with secondary myelofibrosis. Blood 108: 1652–1660.
Lambert JR, Everington T, Linch DC, Gale RE . (2009). In essential thrombocythemia, multiple JAK2-V617F clones are present in most mutant-positive patients: a new disease paradigm. Blood 114: 3018–3023.
Le Bousse-Kerdilés MC, Martyré MC . (1999). Dual implication of fibrogenic cytokines in the pathogenesis of fibrosis and myeloproliferation in myeloid metaplasia with myelofibrosis. Ann Hematol 78: 437–444 . Review.
Lippert E, Boissinot M, Kralovics R, Girodon F, Dobo I, Praloran V et al. (2006). The JAK2-V617F mutation is frequently present at diagnosis in patients with essential thrombocythemia and polycythemia vera. Blood 108: 1865–1867.
Lippert E, Girodon F, Hammond E, Reading NS, Jelinek J, Badbaran A et al. (2009). Concordance of assays designed for the quantitation of JAK2V617F (1849G>T): a multi-centre study. Haematologica 94: 38–45.
Matsuda-Hashii Y, Takai K, Ohta H, Fujisaki H, Tokimasa S, Osugi Y et al. (2004). Hepatocyte growth factor plays roles in the induction and autocrine maintenance of bone marrow stromal cell IL-11, SDF-1 alpha, and stem cell factor. Exp Hematol 32: 955–961.
Musolino C, Calabro L, Bellomo G, Martello F, Loteta B, Pezzano C et al. (2002). Soluble angiogenic factors: implications for chronic myeloproliferative disorders. Am J Hematol 69: 159–163.
Najfeld V, Cozza A, Berkofsy-Fessler W, Prchal J, Scalise A . (2007). Numerical gain and structural rearrangements of JAK2, identified by FISH, characterize both JAK2617V>F-positive and -negative patients with Ph-negative MPD, myelodysplasia, and B-lymphoid neoplasms. Exp Hematol 35: 1668–7166.
Nussenzveig RH, Swierczek SI, Jelinek J, Gaikwad A, Liu E, Verstovsek S et al. (2007). Polycythemia vera is not initiated by JAK2V617F mutation. Exp Hematol 35: 32–38.
Olcaydu D, Harutyunyan A, Jäger R, Berg T, Gisslinger B, Pabinger I et al. (2009). A common JAK2 haplotype confers susceptibility to myeloproliferative neoplasms. Nat Genet 41: 450–454.
Panteli KE, Hatzimichael EC, Bouranta PK, Katsaraki A, Seferiadis K, Stebbing J et al. (2005). Serum interleukin (IL)-1, IL-2, sIL-2Ra, IL-6 and thrombopoietin levels in patients with chronic myeloproliferative diseases. British J Haematol 130: 709–715.
Quesniaux VF, Clark SC, Turner K, Fagg B . (1992). Interleukin-11 stimulates multiple phases of erythropoiesis in vitro. Blood 80: 1218–1223.
Radaeva S, Jaruga B, Hong F, Kim WH, Fan S, Cai H et al. (2002). Interferon-alpha activates multiple STAT signals and down-regulates c-Met in primary human hepatocytes. Gastroenterology 122: 1020–1034.
Sattler M, Salgia R . (2007). c-MET and hepatocyte growth factor: potential new targets in cancer therapy. Curr Oncol Rep 9: 102–108.
Schaub FX, Jäger R, Looser R, Hao-Shen H, Hermouet S, Girodon F et al. (2009). Clonal analysis of deletions on chromosome 20q and JAK2-V617F in MPD suggests that del20q acts independently and is not one of the pre-disposing mutations for JAK2-V617F. Blood 113: 2022–2027.
Schwertschlag US, Trepicchio WL, Dykstra KH, Keith JC, Turner KJ, Dorner AJ . (1999). Hematopoietic, immunomodulatory and epithelial effects of interleukin-11. Leukemia 13: 1307–1315.
Shinomiya N, Gao CF, Xie Q, Gustafson M, Waters DJ, Zhang YW et al. (2004). RNA interference reveals that ligand-independent MET activity is required for tumor cell signalling and survival. Cancer Res 64: 7962–7970.
Stellrecht CM, Phillip CJ, Cervantes-Gomez F, Gandhi V . (2007). Multiple myeloma cell killing by depletion of the MET receptor tyrosine kinase. Cancer Res 67: 9913–9920.
Tacchini L, Dansi P, Matteucci E, Desiderio MA . (2001). Hepatocyte growth factor signalling stimulates hypoxia inducible factor-1 (HIF-1) activity in HepG2 hepatoma cells. Carcinogenesis 22: 1363–1371.
Tefferi A . (2010). JAK inhibitors in myeloproliferative neoplasms. The Hematologist 7: 5 . Review.
Tiedt R, Hao-Shen H, Sobas MA, Looser R, Dirnhofer S, Schwaller J et al. (2008). Ratio of mutant JAK2-V617F to wild-type Jak2 determines the MPD phenotypes in transgenic mice. Blood 111: 3931–3940.
Uchiyama H, Barut BA, Mohrbacher AF, Chauhan D, Anderson KC . (1993). Adhesion of human myeloma-derived cell lines to bone marrow stromal cells stimulates interleukin-6 secretion. Blood 82: 3712–3720.
Vardiman JW, Harris NL, Brunning RD . (2002). The World Health Organization (WHO) classification of the myeloid neoplasms. Blood 100: 2292–2302.
Verstovsek S . (2010). Therapeutic potential of janus-activated Kinase-2 inhibitors for the management of myelofibrosis. Clin Cancer Res 16: 1988–1996.
Wernig G, Mercher T, Okabe R, Levine RL, Lee BH, Gilliland DG . (2006). Expression of Jak2V617F causes a polycythemia vera-like disease with associated myelofibrosis in a murine bone marrow transplant model. Blood 107: 4274–4281.
Wickenhauser C, Thiele J, Lorenzen J, Schmitz B, Frimpong S, Schramm K et al. (1999). Polycythemia vera megakaryocytes but not megakaryocytes from normal controls and patients with smoker's polycythemia spontaneously express IL-6 and IL-6R and secrete IL-6. Leukemia 13: 327–334.
Zdzisinska B, Bojarska-Junak A, Dmoszynska A, Kandefer-Szerszen M . (2008). Abnormal cytokine production by bone marrow stromal cells of multiple myeloma patients in response to RPMI8226 myeloma cells. Arch Immunol Ther Exp 56: 207–221.
We are indebted to Mrs Danielle Pineau for excellent technical help and to colleagues of the Clinical Hematology Departments of the hospitals of Nantes, La Roche-sur-Yon and Lorient for providing patient samples. The study was supported by grants from the Ligue Nationale contre le Cancer (Comité de Loire-Atlantique) and the Association pour la Recherche contre le Cancer. MB, CC and MV have been recipients of scholarships from the French Ministry of Research (MB: 2004–2007; CC: 2007–2010; MV: 2009–2010) and from the Association pour la Recherche contre le Cancer (MB, 2008). Authorship: MB, CC, MV, IC performed research, analysed data and helped to write the paper; MB and CC contributed equally. YJ analysed data and revised the paper. SH designed and performed research, analysed data and wrote the paper.
The authors declare no conflict of interest.
Supplementary Information accompanies the paper on the Oncogene website
About this article
Cite this article
Boissinot, M., Cleyrat, C., Vilaine, M. et al. Anti-inflammatory cytokines hepatocyte growth factor and interleukin-11 are over-expressed in Polycythemia vera and contribute to the growth of clonal erythroblasts independently of JAK2V617F. Oncogene 30, 990–1001 (2011). https://doi.org/10.1038/onc.2010.479
- Polycythemia vera
- hepatocyte growth factor (HGF)
- interleukin 11 (IL-11)
- interleukin 6 (IL-6)
Genetics and Pathogenetic Role of Inflammasomes in Philadelphia Negative Chronic Myeloproliferative Neoplasms: A Narrative Review
International Journal of Molecular Sciences (2021)
Hematology/Oncology Clinics of North America (2021)
Inflammatory Microenvironment and Specific T Cells in Myeloproliferative Neoplasms: Immunopathogenesis and Novel Immunotherapies
International Journal of Molecular Sciences (2021)
Blood Cells, Molecules, and Diseases (2021)
Anti-Glucosylsphingosine Autoimmunity, JAK2V617F-Dependent Interleukin-1β and JAK2V617F-Independent Cytokines in Myeloproliferative Neoplasms