Primitive myelofibrosis (PMF) is a lethal myeloproliferative disorder (MPD) biologically characterized by a spontaneous megakaryocytic growth and the development of bone marrow and spleen fibrosis. PMF and essential thrombocytemia (ET) differ by their clinical features despite similar oncogenic molecular events (i.e. JAK2V617F and MPLW515 mutations) and spontaneous megakaryocytic growth. We investigated whether a secondary oncogenic event re-inforcing the constitutive JAK/cytokine receptor signaling could explain phenotypic differences between ET and PMF. We hypothesized that SOCS proteins, known to inhibit JAK/STAT signaling, may be down-modulated in PMF but not in ET. Komura et al. have previously reported that SOCS1 inhibits spontaneous megakaryocytic growth in PMF (Komura, Chagraoui et al. 2003). Other groups have shown that SOCS3, SOCS1 and SOCS2 promoters are hypermethylated in PMF (Capello, Deambrogi et al. 2008; Fernandez-Mercado, Cebrian et al. 2008). We investigated whether SOCS3 promoter hypermethylation plays a role in PMF megakaryocte spontaneous growth. We first studied 40 RNA samples from MPDs. RNAs were extracted from purified neutrophils and analyzed for SOCS3 expression using RT-real time PCR (21 PMF samples: 10 JAK2V617F-positive PMF, 2 MPLW515L and K PMF and 9 JAK2- and MPL-negative PMF; 10 PV samples and 10 ET samples: 5 with JAK2V617F mutation and 5 without mutation). PMF samples harbored a reduction in SOCS3 expression (related to β-Actin) when compared with all other samples (median 13 vs. median 24, student test p<0.05). However, there was no correlation with the presence of JAK2V617F. We next hypothesized that the reduction in SOCS3 expression could be a consequence of SOCS3 promoter methylation. We sequenced SOCS3 promoters from 5 PMF samples identified as down-expressing SOCS3 and 5 without changes in SOCS3 expression, 5 ET samples and 3 healthy controls, after DNA bisulfite treatment. Methylation of at least 3 out of 45 CpG islands was only detected in the 5 down-expressing SOCS3 PMF samples (3, 4, 11, 20 and 31 CpG sites). To clearly demonstrate that such methylation was responsible for the decrease in SOCS3 expression, purified CD34+ cells from 2 PMF down-expressing SOCS3 and 2 controls were cultured with TPO, SCF and IL-3 and treated with or without a methylation inhibitor (5-azacytidine: 1 μM) for 2 days. Expression of SOCS3 and JAK2 were compared using RT-real time PCR. Absence of methylation resulted in a 2- to 3-fold increase in SOCS3 expression in the PMF samples only. This up-regulation was not observed for JAK2 expression. To determine the role of SOCS3 down-expression in PMF, we over-expressed SOCS3 (or a control vector) in PMF (n=3) or normal (n=2) CD34+ cells using a SOCS3-eGFP expression lentiviral vector. SOCS3-expressing CD34+ cells were sorted based on eGFP at one-cell per well (4,000 wells) and cultured with or without 10 ng/mL TPO for 10 days. SOCS3 over-expression reduced the cloning capacity of normal CD34+ cells by 50% in presence of TPO illustrating its role in the normal negative regulation of the JAK/STAT signaling induced by TPO. Similar results were observed for PMF-derived cells. However SOCS3 did not reduce the spontaneous growth of PMF samples, independently of their first oncogenic event (i.e. JAK2V617F or MPLW515 mutations) suggesting that SOCS3 cannot reduce the spontaneous signaling of JAK/STAT in JAK2V617F or MPLW515 mutation backgrounds. This is in agreement with results of Hookam et al. who have previously shown that SOCS3 did not inhibit JAK2V617F signalling (Hookham, Elliott et al. 2007). These experiments demonstrate thatepigenetic events are present in PMF,SOCS3 promoter is methylated in at least part of PMFs whatever the first oncogenic event is, andSOCS3 inhibits TPO-induced megakaryocytic proliferation but not spontaneous growth in PMF.