Background: Primary myelofibrosis (PMF) is a myeloproliferative neoplasm characterized by myeloid cell proliferation and progressive bone marrow (BM) fibrosis. The degree of the BM fibrosis correlates with unfavorable clinical characteristics and short survival. Whereas it was thought that BM fibrosis in PMF is secondary, caused by mesenchymal stromal cells (MSC) stimulated by cytokine-producing neoplastic megakaryocytes and platelets, we have recently found that BM fibrosis in PMF is induced by neoplastic monocyte-derived fibrocytes which, like MSC, produce collagen-I, collagen-III and fibronectin. Using a PMF immunodeficient mouse model, we demonstrated that inhibition of monocyte differentiation into fibrocytes ameliorates BM fibrosis. Although an increased number of fibrocytes was detected in the BM of patients with PMF, the mechanisms that induce collagen overproduction are still unknown. Glioma oncogene homolog 1 (Gli1), a transcription factor of the hedgehog signaling pathway, activates genes affecting cell migration and tissue fibrosis. Gene expression analyses identified matrix metalloprotease 9 (MMP9) as a key effector of Gli1 that promotes fibrocyte migration into sites of tissue injury. A recent study demonstrated that knockout or inhibition of Gli1 alleviates BM fibrosis. Since it was thought that BM fibrosis is induced by MSC, the function of other BM-derived Gli1-expressingcells was not investigated. Because fibrocytes induce BM fibrosis in PMF, we sought to evaluate the levels of Gli1 in PMF BM cells and in BM-derived fibrocytes and MSC.

Methods: BM trephine biopsies and aspirates were obtained from randomly selected PMF patients and healthy controls. Opal multiplex fluorescence IHC was performed on tissue sections using fibrocyte (CD45, CD68, procollagen type I)- and MSC (CD90, CD105)-specific monoclonal antibodies in combination with validated anti-Gli1 antibodies. Single BM cells were classified based on image pattern analysis and their Gli1 signal intensity was assessed. In addition, Low-density cells from BM aspirates were cultured using fibrocyte and MSC culture assays. Following confirmatory immunostaining analysis, RNA was extracted and analyzed using quantitative RT-PCR.

Results: Gli1 intensity in fibrocytes of PMF BM biopsies was significantly higher than Gli1 intensity in fibrocytes of healthy control BM biopsies (median: 11.11 vs. 7.16 counts per cell, P=0.02). The median Gli1 signal intensity/cell was 4.3 counts higher in PMF fibrocytes than in PMF MSC (95%CI: 0.5-11.75; P=0.033), whereas Gli1 intensity/cell of PMF MSC was similar to that of normal BM MSC. Similarly, Gli1 mRNA levels were significantly higher in cultured PMF BM-derived fibrocytes than in cultured normal BM-derived fibrocytes (7.16±1.35 vs. 1±0.55, P=0.007). In addition, mRNA levels were higher in PMF BM-derived fibrocytes than in PMF BM-derived MSC (7.16±1.35 vs. 1.36±0.23, P=0.006). Downstream activity of Gli1 in cultured cells was assessed by quantitating the expression of the MMP9 gene. Remarkably, PMF BM-derived fibrocytes expressed significantly higher MMP9 mRNA levels compared to normal BM-derived fibrocytes (12.47±2.25 vs. 1±0.45, P=0.003). In contrast, in PMF BM-derived MSC MMP9 mRNA levels were only 0.03±0.01, 12-fold lower than those of PMF BM-derived fibrocytes (P<0.001) and similar to MMP9 mRNA levels in normal BM-derived MSC.

Conclusions: Because Gli1 protein is invariably overrepresented in fibrocytes but not in MSC and Gli1, while MMP9 mRNA levels are higher in PMF BM-derived fibrocytes than in MSC, and considering previous Gli1 gene knockout studies, it is reasonable to assume that Gli1+ BM fibrocytes rather than MSC are the major contributors to the induction of BM fibrosis in PMF.


Verstovsek:Celgene: Membership on an entity's Board of Directors or advisory committees; Novartis: Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; Incyte: Consultancy; Italfarmaco: Membership on an entity's Board of Directors or advisory committees.

Author notes


Asterisk with author names denotes non-ASH members.