The Contact with MDS Endothelial Cells Alters the Pattern of Lineage-Specific Gene Expression During Normal Hematopoietic Differentiation

Abstract 1718

In the embryo, hematopoiesis and angiogenesis occur concurrently, since endothelial and hematopoietic cells derive from a common cell precursor, the hemangioblast. In the post natal life, the bone marrow (BM) vascular niche has an important role in supporting hematopoietic stem cells (HSC) proliferation and differentiation. Moreover, during the processes of homing and mobilization, BM vascular niche regulates HSC migration. In BM failure, these functions are promoted by vascular niches located outside of the BM, such as in the spleen, thus confirming that endothelial cells are relevant to the normal hematopoietic differentiation. The dominant feature of myelodyspslatic syndromes (MDS) is the impaired capacity of HSC to give rise to a terminally differentiated normal cell population in one or more lineages. We hypothesized that in MDS patients the vascular niches within and eventually outside of the BM might be altered and that they do not support adequately the physiologic hematopoietic differentiation, thus contributing to the pathogenesis of the disease. To investigate this hypothesis, we isolated and expanded endothelial colony forming cells (ECFCs) (Ingram et al. Blood. 2004;104: 2752) from the peripheral blood of patients with MDS (RCMD) and we used ECFCs as a surrogate of the vascular niche to differentiate hematopoietic cells. Normal CD34+ were seeded over a layer of endothelial cells and cultured for 8 days in different cytokine- containing media able to induce selective granulocytic-macrophage, erythroid or megakaryocyte differentiation. At baseline and after 5 and 8 days of co-culture, hematopoietic cells were recovered by gentle repeated washings, and then they were analyzed with RT-PCR for the expression of lineage specific genes. For each lineage, two genes, mainly involved either in the early or in the late phase of the differentiation, respectively, were analyzed. Basically, we found that co-culturing CD34+ cells in contact with a normal endothelial layer the gene expression observed in the corresponding no-layer cultures was amplified. This effect was strictly dependent on the physical contact between hematopoietic and endothelial cells, since it was abolished when cells were grown separated from a porous membrane. We observed that in co-cultures with normal hematopoietic cells and MDS ECFCs, the modulatory effect exerted by the endothelial layer was partially or completely lost. For example, CD34+ cells undergoing granulocyte differentiation showed a progressive decline of PU.1 gene expression, followed by the rise of the expression of MPO gene; these effects were much more evident if cells were cultured over normal ECFCs. In contrast, this pattern of gene expression was deeply perturbed when normal CD34+ cells were cultured with MDS ECFCs: actually, at day 7, the PU.1 RNA level was six folds higher and the MPO RNA level was three folds lower than that observed in co-cultures with normal ECFCs. Similar figures were observed for RUNX1 and GP1b gene expression in cells undergoing megakaryocyte differentiation. Moreover, CD34+ cells undergoing erythroid differentiation on normal ECFCs exhibited the progressive up-regulation of FOG-1 and of transferrin receptor (TFR) gene expression: interestingly, the co-culture with MDS ECFCs resulted in a lower expression of FOG -1 gene but also in a surprisingly higher expression of the TFR gene. Overall, these data suggest that in MDS does exist an impairment of talking between endothelial and hematopoietic cells. To understand which molecular pathways are involved could represent the basis for a novel therapy approach.