Pathogenetic mechanisms in primary myelofibrosis

INTRODUCTION — Primary myelofibrosis [1-3] (PMF, chronic idiopathic myelofibrosis, agnogenic myeloid metaplasia) is one of the chronic myeloproliferative neoplasms, which are collectively characterized by clonal proliferation of myeloid cells with variable morphologic maturity and hematopoietic efficiency (table 1).

The primary disease process in PMF is a clonal hematopoietic stem cell disorder that results in chronic myeloproliferation and atypical megakaryocytic hyperplasia [4]. The secondary process of bone marrow fibrosis (BMF) is the result of nonclonal fibroblastic proliferation and hyperactivity induced by growth factors abnormally shed from clonally expanded megakaryocytes [5]. BMF is the hallmark of PMF and contributes to the impaired hematopoiesis that leads to severe anemia. In addition to BMF and anemia, patients with PMF suffer from marked splenomegaly, extramedullary hematopoiesis, and severe constitutional symptoms. (See "Clinical manifestations and diagnosis of primary myelofibrosis".)

The potential to arrest or reverse the BMF in PMF may offer an alternative approach to palliative therapy. With this goal in mind, the pathogenesis of PMF will be discussed here. The prognosis and treatment of PMF are discussed separately. (See "Myelofibrosis (MF): Management of primary MF and secondary MF".)

ETIOLOGY — The exact cause of primary myelofibrosis (PMF) is unknown. PMF, along with the other chronic myeloproliferative disorders, chronic myeloid leukemia, polycythemia vera, and essential thrombocythemia, is considered to arise from a somatic mutation of a pluripotent hematopoietic progenitor cell [6,7]. A defective stem cell "niche" within the bone marrow has been postulated for PMF [8]. (See "Overview of the myeloproliferative neoplasms".)

The occurrence of PMF has, in a minority of cases, been linked to exposure to thorium dioxide, petroleum manufacturing plants (especially toluene and benzene), and ionizing radiation [9-11]. A very high incidence of PMF has been noted in patients given thorium-based radiographic contrast material and in individuals exposed to atomic bombs at Hiroshima [12,13].

Mice with a mutation of the GATA-1 transcription factor gene develop a hematologic picture similar to myelofibrosis at 15 months of age, and may represent a suitable animal model for the human disease [5,14,15].

CELLULAR ABNORMALITIES

Chromosomal and other genetic abnormalities — Approximately 50 to 60 percent of patients with primary myelofibrosis (PMF) have clonal karyotypic abnormalities at diagnosis [5,16-21]. However, none of these abnormalities is specific to this disorder. The most common findings, accounting for 50 to 65 percent of the karyotypic changes, are deletion of a segment of the chromosome bearing the retinoblastoma gene (13q-), 20q-, and partial trisomy 1q (table 2). Patients with these cytogenetic abnormalities are reported to have a worse prognosis [16,17,22].

In our series of 884 patients with PMF, unfavorable karyotype has been one of the factors associated with reduced overall survival, and includes such changes as +8, -7/7q-, i(17q), -5/5q-, 12p-, inv(3), and 11q23 rearrangement [23]. The detrimental effect of unfavorable karyotype was mostly attributed to monosomal karyotype or inv(3)/i(17q).

Gains of chromosome 9 or 9p have also been noted in our studies (13 percent) and those of others (50 percent), suggesting that genes on 9p may play a crucial role in the pathogenesis of PMF [24]. Of interest, mutations of JAK2 (Janus kinase 2), a gene found on 9p, have been found in approximately 50 percent of patients with PMF. (See 'JAK2 mutations' below and "Overview of the myeloproliferative neoplasms", section on 'Mutations in PV, ET, and PMF'.)

A number of genetic mutations and single nucleotide polymorphisms have been identified in patients with PMF and other myeloproliferative neoplasms associated with leukemic transformation [25-29]. As an example, a A3669G polymorphism of the glucocorticoid receptor was found to contribute to the phenotype of excess myeloproliferation in PMF (eg, higher white blood cell count, larger spleen, higher frequency of CD34+ cells at diagnosis) and, in cooperation with the JAK2V617F mutation, increasing the risk of blastic transformation [30].

In summary, mutations in PMF involve primarily JAK2 but also to a smaller extent MPL, LNK, CBL, TET2, ASXL1, IDH1, IDH2, IKZF1, EZH2, DNMT3A, TP53, SF3B1 and SFSR2. None of these mutations has been consistently traced back to the ancestral clone, and their disease-initiating potential is uncertain. Some (eg, JAK2 and SF3B1 mutations) might contribute to specific phenotypes such as erythrocytosis and ring sideroblasts, respectively. Available data suggest inferior survival in PMF associated with nullizygosity for JAK2 46/1 haplotype, low JAK2V617F allele burden, and the presence of IDH, ASXL1, SRSF2 or EZH2 mutations. Some of these mutations (eg, IDH, SRSF2) have also been associated with inferior leukemia-free survival. In contrast, the presence or absence of JAK2V617F, MPL or TET2 mutations do not appear to affect survival.

Cytokine abnormalities — Fibroblasts do not share the chromosomal abnormalities found in hematopoietic cells [31]. As a result, it is thought that bone marrow fibrosis, as well as many of the other signs and symptoms of PMF, are secondary reactions to the clonal hemopathy mediated by cytokines released from the neoplastic megakaryocytes and other clonally expanded hematopoietic cells, such as T and B cells [32,33].

Spontaneous hematopoietic colony growth in vitro — One of the hallmarks of polycythemia vera (PV) is in vitro erythropoietin- independent growth of red blood cell (RBC) colonies. Spontaneous growth of both RBC colonies and megakaryocytes (in the absence of exogenous cytokines) has also been described in a minority of patients with PMF and ET.

A number of findings support the involvement of the trophic hormone thrombopoietin (TPO) and/or its receptor (c-MPL) in the pathogenesis of PMF, particularly in the pathogenesis of bone marrow fibrosis (see 'Bone marrow fibrosis' below). There is also evidence that the TPO system is involved in the spontaneous megakaryocyte growth in PMF. As examples, both preparations of the soluble murine TPO receptor and antisense oligonucleotides against the TPO receptor markedly inhibit spontaneous megakaryocyte growth in PMF and in ET and PV [34,35].

JAK/STAT pathway

JAK2 mutations — One suggested mechanism of intrinsic growth factor hypersensitivity in PMF involves megakaryocyte overexpression of FKBP51 associated with JAK2/STAT5 activation, consistent with the discovery of an activation mutation of JAK2 in patients with PV, ET, and PMF [5]. (See "Overview of the myeloproliferative neoplasms", section on 'Mutations in PV, ET, and PMF' and "Molecular pathogenesis of congenital erythrocytoses and polycythemia vera", section on 'JAK2 mutations'.)

Constitutive activation of the JAK/STAT pathway appears to be an important pathogenetic event in patients with the myeloproliferative disorders. However, only approximately 50 percent of patients with PMF have the JAK2 mutation, suggesting the presence of other modes of JAK/STAT activation, such as activating mutations in hematopoietic-specific cytokine receptors (see below) [36].

Calreticulin gene mutations — Mutations in the calreticulin gene (CALR) have been reported in approximately 35 percent of patients with ET [37,38]. Thus far CALR mutations have been seen only in patients with ET or PMF and have not been observed in patients with PV or in those with JAK2 or MPL mutations. (See "Overview of the myeloproliferative neoplasms", section on 'Mutations in PV, ET, and PMF'.)

MPL mutations — Mutations of the thrombopoietin receptor (MPL) are also capable of activation of JAK/STAT signaling in patients with PMF [36]. The finding of novel somatic activating mutations in the thrombopoietin receptor (MPL mutations W515L and W515K) in approximately 5 percent of patients with PMF, 1 percent of those with ET, and in none of the patients with PV suggests that these MPL mutations favor megakaryocyte fate while the JAK2 mutation favors erythroid fate [36,39,40].

BONE MARROW FIBROSIS — Bone marrow fibrosis (BMF) in primary myelofibrosis (PMF) results from the abnormal deposition of excess collagen derived from fibroblasts. Although several types of collagen provide the reticular matrix support in normal bone marrow, types III, IV, and I are the major components of BMF associated with PMF [41].

The collagen-producing fibroblasts in PMF are functionally and physically similar to normal fibroblasts and are polyclonal [42,43]. They are secondarily stimulated to proliferate and overproduce collagen by growth factors secreted by the neighboring megakaryocytes [44,45]. Thus, BMF in PMF is a reactive process accompanying the underlying clonal stem cell disorder.

The major megakaryocyte-derived cytokine implicated in the mediation of BMF in PMF is transforming growth factor (TGF)-beta [4,46], which may interact with thrombopoietin (figure 1). Other growth factors also may contribute to the fibrotic reaction, including platelet derived growth factor (PDGF), epidermal growth factor (EGF), basic fibroblast growth factor (bFGF, FGF-2), vascular endothelial growth factor (VEGF), calmodulin, matrix metalloproteinase-9, and lysyl oxidase [4,5,46-53]. (See 'Transforming growth factor-beta' below.)

This fibrous reaction with abnormal accumulation of extracellular matrix components is also dependent on matrix metalloproteinases (MMP) and tissue inhibitors of MMPs (TIMP) [54]. In a study of 25 patients with PMF, significantly decreased plasma levels of MMP-3 were found, which correlated inversely with the degree of bone marrow fibrosis [55]. Elevated levels of TIMP-1 were also found, suggesting that the balance between MMPs and TIMP may be essential in fibrosis formation.

It has been suggested, based upon a murine model and on bone marrow samples from patients with PMF, that there is a significant degree of entry of hematopoietic cells into megakaryocyte cytoplasm, a phenomenon called emperipolesis [56]. This is probably brought about by increased expression and abnormal localization of P-selectin by the megakaryocytes, leading to engulfment, activation, and damage to the cells, mainly neutrophils and eosinophils, passing through the megakaryocyte cytoplasm. The result is release of lytic granules from the engulfed cells, progressive destruction of megakaryocytes with degradation and lysis of their alpha granules, and release of growth factors, resulting in the marked fibroblast activation and infiltration characteristic of PMF.

Transforming growth factor-beta — The major megakaryocyte-derived cytokine implicated in the mediation of BMF in PMF is transforming growth factor (TGF)-beta, which may interact with thrombopoietin (figure 1) [47,57-62].

TGF-beta is a glycoprotein that is synthesized and secreted primarily by the monocyte-macrophage system and endothelial cells, and also by megakaryocytes [4,57,58]. It is capable of enhancing the production and secretion of extracellular matrix proteins, including collagen types III and I from fibroblasts [59,60], and may also interact with IL-1, fibronectin, and substance P [63]. The expression of TGF-beta has been demonstrated at the level of mRNA and secreted peptide in clonal megakaryocytes [57,61]. In addition, circulating concentrations of TGF-beta are increased in patients with PMF [64].

In vitro, anti-TGF-beta antibodies decrease the collagen synthesis mediated by megakaryoblast-conditioned media [57]. This type of inhibition can also be achieved by tumor necrosis factor-alpha and interferon-gamma, which suppress the activation of type I collagen gene expression by TGF-beta [62].

Thrombopoietin — A PMF-like syndrome can be induced in mice via exposure to high concentrations of thrombopoietin [65-70]. Thrombopoietin (TPO) is the major growth factor required for megakaryocyte growth and development. The TPO glycoprotein can be modified by deglycosylation and subsequent coupling to polyethylene glycol (PEG) to be less immunogenic and have an extended half-life.

Mice injected daily with high doses of PEG-TPO develop a reversible bone marrow fibrosis accompanied by thrombocytosis, megakaryocytic hyperplasia, splenomegaly, extramedullary hematopoiesis, and anemia [66,68]. A similar syndrome has been observed in TPO-transfected mice [67,70] and has been attributed to increased TGF-beta activity secondary to an increased megakaryocyte mass [67,68]. Although the process was reversible by bone marrow transplantation [67], chronic exposure led to a fatal myeloproliferative disorder [65].

The role of TPO has been further evaluated in immune-compromised mice. Overexpression of the TPO gene using adenovectors in SCID mice (severe combined immune deficient) resulted in thrombocytosis, increased marrow megakaryocytes, fibrosis, and extramedullary hematopoiesis that mimicked PMF [71]. However, similar overexpression of thrombopoietin in NOD-SCID mice (which have reduced monocyte and macrophage function in addition to the lymphocyte deficiency in SCID mice) produced thrombocytosis and megakaryocytosis but no fibrosis. These results imply that other monocyte/macrophage mediators are involved in causing the fibrosis.

Overview — The above discussion supports the pathogenetic role of TGF-beta in BMF. The pathogenetic interaction of TPO and TGF-beta in the human disease has not been elucidated. Serum TPO concentrations in patients with PMF are higher than those of normal volunteers or patients with reactive thrombocytosis and have been correlated with the degree of BMF [72,73]. However, there was no correlation with megakaryocyte mass, and the demonstration of low TPO-receptor (c-Mpl) expression by platelets and megakaryocytes from patients with PMF suggests that defective TPO clearance may contribute to the above normal TPO values in patients with PMF [74,75].

Finally, CD34+ hematopoietic progenitor cells from patients with PMF have increased expression of mRNA transcripts of bFGF and its receptors and decreased expression of TGF-beta type II receptors [76]. The pathogenetic and clinical relevance of these observations is currently unknown, although the myeloproliferation characteristic of this disease may result from abnormal proliferation of these precursors [77].

EXTRAMEDULLARY HEMATOPOIESIS — The mechanisms responsible for extramedullary hematopoiesis (EMH) in primary myelofibrosis (PMF) are not understood. The distribution of hematopoietic tissue approximates that in the fetus. In both a mouse model of marrow fibrosis and in marrow fibrosis associated with metastatic disease, there is abnormal release of marrow precursors into distorted bone marrow sinusoids and then into the circulation [78,79]. These marrow precursors may be responsible for EMH in PMF [79,80], although increased trafficking of CD34+ cells into the circulation may also play a role [81]. (See "Clinical manifestations and diagnosis of primary myelofibrosis", section on 'Circulating CD34+ cells'.)

Foci of EMH may also be found in soft tissues, body cavities, serosal surfaces, central nervous system, skin, and other locations. EMH may resemble tumors at these sites, often causing obstructive symptoms, especially in the central nervous system. Such foci may enlarge significantly postsplenectomy, perhaps due to the loss of filtering function of the spleen.

In some patients with PMF, hematopoiesis is only present in extramedullary sites [82]. EMH is rarely as effective as medullary hematopoiesis, thereby leading to the cytopenias so often found in this disease.

INCREASED BONE MARROW VASCULARITY — A number of studies have shown that bone marrow microvessel density (MVD) is increased in patients with primary myelofibrosis (PMF) and other BCR/ABL1-negative myeloproliferative neoplasms [83-85]. In one of the studies, MVD was significantly increased in 114 patients with PMF, when compared with bone marrows from 44 normal controls, 15 patients with polycythemia vera and 17 with essential thrombocythemia [83]. This phenomenon, as with bone marrow fibrosis, is thought to be secondary to release of several growth factors by megakaryocytes (eg, VEGF, PDGF, TGF-beta, bFGF). (See 'Bone marrow fibrosis' above.)

In this study, increased MVD significantly correlated with the presence of megakaryocyte clumping, increased splenic size, and, in a multivariate model, decreased overall survival. Based on these correlations, phase II studies employing thalidomide and lenalidomide, agents with antiangiogenic properties, are ongoing [5,83,86]. (See "Overview of angiogenesis inhibitors", section on 'Immunomodulatory drugs (IMiDs)'.)

SUMMARY

●Primary myelofibrosis (PMF) – PMF is characterized by chronic myeloproliferation, atypical megakaryocytic hyperplasia, increased bone marrow vascularity, and bone marrow fibrosis (BMF). Increased bone marrow vascularity and BMF are related to proliferation and hyperactivity of non-clonal fibroblasts that are induced by growth factors shed from clonally expanded megakaryocytes. (See 'Bone marrow fibrosis' above and 'Increased bone marrow vascularity' above.)

●Pathogenesis – The following features have been implicated in the pathogenesis of PMF/BMF (see 'Overview' above):

•Chromosomal abnormalities – Half of patients with PMF have clonal chromosomal abnormalities, but none is specific to this disorder. (See 'Chromosomal and other genetic abnormalities' above.)

•Abnormal megakaryocyte growth – Growth of red blood cell (RBC) colonies and/or megakaryocytes with little or no exogenous cytokines is seen in some patients with PMF. (See 'Spontaneous hematopoietic colony growth in vitro' above.)

•JAK/STAT activation – Constitutive activation of the JAK/STAT pathway, in association with mutations of JAK2, MPL, or CALR, contributes to the pathogenesis of PMF. (See 'JAK/STAT pathway' above.)

•Bone marrow fibrosis – BMF may be related to abnormal expression of transforming growth factor-beta (TGF-beta), thrombopoietin (TPO), and/or the TPO receptor, MPL. (See 'Transforming growth factor-beta' above and 'Thrombopoietin' above and 'Cytokine abnormalities' above.)

ACKNOWLEDGMENT — The editors of UpToDate acknowledge the contributions of Stanley L Schrier, MD as Section Editor on this topic, his tenure as the founding Editor-in-Chief for UpToDate in Hematology, and his dedicated and longstanding involvement with the UpToDate program.