In Primary Myelofibrosis, several lines of evidence suggest that pleiotropic cytokine TGF-β1, released by clonal proliferation of pathological megakaryocytes and/or monocytes, plays a prominent role in reticulin fibers deposition. This cytokine is synthesized as a biologically inactive molecule that needs to be activated in order to trigger biological responses. However, the mechanisms involved in local TGF-β1 activation within the hematopoietic environment remain unclear. Since TGF-β1 and thrombospondin-1 (TSP-1) are synthesized and stored within the same organelles in megakaryocytes, one can speculate that the abnormal release of both molecules leads to pathological local TGF-β1 activation that becomes ultimately responsible of fibrosis development in the vicinity of these cells. To investigate the role of TSP-1 in local TGF-β1 activation, we used the TPOhigh murine model of bone marrow (BM) fibrosis. BM cells from wild-type (WT) or Tsp-1-null male littermates were infected with a retrovirus encoding the murine TPO protein and engrafted into lethally irradiated WT or Tsp-1-null female hosts, respectively, leading to the following engraftment combinations, WT/WT (WT TPOhigh mice, n=21) and Tsp-1-null/Tsp-1-null (Tsp-1-null TPOhigh mice, n=17). Lethally-irradiated hosts were engrafted with 4 to 8 × 106 cells in 3 independent experiments. Peripheral blood was analyzed every 4 weeks during 3 months and mice were killed for histological analysis at week 8 and 12 post-engraftment. The magnitude of plasma TPO level increase was comparable regardless of the TPOhigh mice groups. Chimerism levels, analyzed in recipients by FISH on the presence of the donor Y chromosome in whole nucleated BM cells, were more than 90% in either WT or Tsp-1-null TPOhigh mice. We report here that all TPOhigh mice developed a similar myeloproliferative syndrome associated with TGF-β1 overproduction. Surprisingly, we were able to detect the active form of TGF-β1 in BM and spleen extracellular fluids in all mice, including Tsp-1-null TPOhigh mice, suggesting that alternative mechanisms are mainly responsible for local TGF-β1 activation in this murine model of myelofibrosis. We then confirmed that Tsp-1-null platelets are able to activate TGF-β1 in vitro in response to thrombin. As predicted by the detection of the active form of TGF-β1, Tsp-1-null TPOhigh mice developed BM and spleen fibrosis which appears, intriguingly, to be of a greater grade than the one displayed by WT TPOhigh mice. Since TSP-1 is a potent inhibitor of angiogenesis, we investigate whether this increased fibrosis could be correlated with an augmentation of neoangiogenesis. The microvascular density (MVD) in control Tsp-1-null BM were higher than in control WT one (10±4.7 vs 0.6±0.2; p<0.001), as expected. However, MVD displayed by Tsp-1-null TPOhigh mice (8.3±4.4) did not rise above the one displayed by control Tsp-1-null mice and was similar to MVD observed in WT TPOhigh mice (5.7±2.9). Thus, the increase of myelofibrosis in Tsp-1-null TPOhigh mice cannot be explained by an augmentation of neoangiogenesis. Since TGF-β1 levels were similar in both TPOhigh groups, we hypothesized that this increase could be related to an enhanced TGF-β1-mediated response by Tsp-1-null BM fibroblasts. Indeed, we could show that Tsp-1 deficiency is associated with sustained phospho-Smad3 levels and a 10-fold increase in collagen III transcription level by BM fibroblasts in response to TGF-β1. Together, our results

  • show that TSP-1 is not the major activator of TGF-β1 in this in vivo model of myelofibrosis;

  • suggest that other mechanisms are involved in this activation;

  • shed light on a possible new mechanism of TGF-β1 regulation by one of its own activator.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author