RUNX1 Expression Characterizes the Endothelial Cells from the Spleen and Bone Marrow of Patients with Primary Myelofibrosis

Abstract

Background.Primary myelofibrosis (PMF) is a Philadelphia-negative (Ph−) myeloproliferative disorder characterized by extramedullary haematopoiesis and abnormal neoangiogenesis in both the bone marrow (BM) and the spleen. We previously provided evidence that endothelial cells (ECs) from either the spleen or the splenic vein of PMF patients frequently share the JAK2V617F mutation with the hematopoietic malignant cells. More recently, we confirmed this observation also in BM-derived ECs of PMF patients. The mechanism underlying this phenomenon remains, however, not yet clarified. RUNX1 is a critical regulator of hematopoiesis, required for hematopoietic stem cell (HSC) generation and function. In human embryo, it is expressed in all emerging HSCs and progenitors and it is a necessary transcription factor for endothelial to hematopoietic transition. In the adult humans it is expressed in all blood cells, in decreasing intensity according to the maturation status, except erythrocytes. In angiogenesis, it induces endothelial differentiation and maturation as well as vascular network formation by promoting expression of VE-cadherin. Finally, it is involved in retinal aberrant neoangiogenesis.Aim. To assess if neoangiogenetic activity observed in spleen and BM of PMF patients is associated with RUNX1 expression in ECs.Patients and Methods. Paraffin-embedded BM sections from patients with MF (n=3), and from patients with lymphomas (named as CTRLs), who underwent BM biopsy for disease staging (n=2), were used for immunostaining analysis. Spleens samples were collected from 3 patients with PMF, who received splenectomy for clinical reasons, and from 2 healthy subjects (HS), who underwent surgery after traumatic damage. Fresh spleen samples were embedded in OCT, and then snap-frozen and stored in liquid nitrogen. Endothelial colony forming cells (ECFCs) were obtained from peripheral blood (n=1 PMF, n=1 HS), PMF spleen tissue (n=1) and cord blood (n=1), according to Ingram et al (Blood 2004;104:2752) and cytospun on slides at confluence. ECFCs, BM and spleen sections were stained with antibodies directed against RUNX1 and VE-cadherin. The images were obtained by confocal laser scanning microscopy (Olympus Fluoview FV10i, 60x objective) and processed by IMAGEJ software. We evaluated 10 fields for each BM and spleen section and measured the number of RUNX1-positive vessels/total vessels. We considered positive a vessel that contained at least one RUNX1-positive cell. The results are shown as mean ± SD.Results. Immunofluorescence staining of spleen and BM sections confirmed the presence of increased neoangiogenetic processes in samples from PMF patients with respect to CTRLs. The staining of BM and spleen sections obtained from PMF patients with antibodies anti-RUNX1 and anti-VE-cadherin (endothelial marker) allowed the detection of RUNX1+VE-cadherin- cells in the parenchima of BM and spleen (Figure 1A, 1C), occasionally in perivascular position (Figure 1A, arrowhead). Interestingly, in all patients analyzed, we could detect RUNX1+VE-cadherin+ cells in both the parenchima and in the vessels both in the BM and the spleen (Figure 1A, 1C, arrow). In BM, 22 ± 12% of the vessels had at least one double positive cell in the microvessel wall; in spleen tissue the percentage of RUNX1+ ECs increased to 60 ± 13%.We detected RUNX1+VE-cadherin- cells only in the BM parenchima of CTRLs but not in the spleen of HS, whereas RUNX1+VE-cadherin+ cells were never observed in the vessels of neither the spleen nor the BM (Figure 1B, 1D) of HS and CTRLs, respectively.The circulating ECFCs obtained from both PMF and HS were, as expected, VE-cadherin+ but did not express RUNX1; however, 30% of spleen-derived-ECFCs of the PMF patient were RUNX1+. In cord blood-derived ECFCs a small percentage (3%) of RUNX1+ cells was observed.Conclusions.Our data show a selective expression of RUNX1 in splenic- and BM-ECs of PMF patients, suggesting that activation of RUNX1 expression could be associated with the neoangiogenetic processes that characterize the disease. The expression of RUNX1 in ECFCs from spleen but not in their circulating counterpart suggests a role for the splenic microenvironment in determining RUNX1 expression in ECs. Further studies are ongoing to assess the genotype of RUNX1+ ECs of PMF patients. DisclosuresNo relevant conflicts of interest to declare.