MGP Interacts with BMP-4 and BMP-2 and Supports Normal and Malignant Hematopoietic Stem/Progenitor Cells

The general transcriptional coregulatory complex Mediator, subcomplex of the RNA polymerase II holoenzyme, is the endpoint convergence of a variety of intracellular signals, and initiates transcription through recruitment of general initiation factors and formation of a functional preinitiation complex. Among the Mediator subunits, MED1 is crucial for hematopoiesis: it is known to be employed in transcription of genes involved in hematopoietic niche function, as well as in differentiation of hematopoietic precursor cells (e.g., RARα- and VDR-mediated myelomonopoiesis, GATA1-mediated erythromegakaryopoiesis, and iNKT cell development). Among the attenuated genes in Med1^-/- mesenchymal stromal cells was matrix Gla protein (MGP), modulator of the BMP-SMAD signals. As MGP is abundantly expressed in bone tissues and reportedly interacts with niche factors BMP-4 and BMP-2, we hypothesized that MGP might modulate niche function in hematopoiesis through the interaction with these BMPs. We tested this possibility by an in vitro niche model composed of a coculture of bone marrow (BM) mesenchymal stromal cells and BM hematopoietic cells. When mouse BM hematopoietic cells were cocultured with OP-9 or MS-5 mouse BM stromal cells in the presence of the blocking antibody against MGP, the growth and DNA synthesis of BM hematopoietic cells were attenuated and the number of long-term culture-initiating cells (LTC-ICs), which corresponded to the supported hematopoietic stem/progenitor cells (HSPCs), decreased. We next asked if MGP also functioned as a niche factor for malignant HSPCs, To this end, we utilized MB-1 cells, blast crisis CML-derived myeloblastic leukemia cells. MB-1 cells are dependent on cocultured BM stromal cells and possess characteristics of leukemia stem cells (LSCs). When MB-1 cells were cocultured with OP-9 or MS-5 BM stromal cells in the presence of the anti-MGP antibody, the number of MB-1 cells, as well as the number of cobblestone formation, i.e., LSC nature, of MB-1 cells, was profoundly attenuated. Therefore, MGP appears to have a trophic effect on hematopoietic cells and to support normal and malignant HSPCs in our niche model. MGP was, both at mRNA and protein levels, expressed abundantly in BM stromal cells but scarcely in normal hematopoietic cells. However, MB-1 leukemia cells (presumably ectopically) expressed a meaningfully high level of MGP. During the coculture of normal BM hematopoietic cells with OP-9 or MS-5 BM stromal cells, MGP was induced prominently and transiently after one day. MGP, expressed in MS-5 or OP-9 BM stromal cells, was likewise transiently induced after 1 day of culture with BM hematopoietic cells that were physically dissociated from these BM stromal cells by the transwell apparatus, indicating that secreted humoral factor(s) induced the MGP expression in the BM stromal cells. As for a malignant situation, the expressions of MGP in both MB-1 leukemia cells and OP-9 or MS-5 BM stromal cells were fluctuated in an oscillated days period. Since the support of HSPCs required their association with BM stromal cells, the induced MGP expression was apparently insufficient for HSPCs support, and the action of induced MGP may be indirect. Therefore, we next asked if BMP-4 and BMP-2, which reportedly associate with MGP, were also induced by the coculture and were employed in HSPCs support jointly with MGP. Indeed, during the coculture, BMP-4 was rapidly and transiently induced within a few days, simultaneously with MGP, followed by a subsequent and sustained induction of BMP-2 that lasted for over a week. However, neither BMP-4 nor BMP-2 was produced by BM stromal cells when cultured with dissociated BM hematopoietic cells using the transwell apparatus. Therefore, the induction of BMP2/BMP4 was dependent on physical association of hematopoietic cells and stromal cells. GST-pulldown assays and mammalian two-hybrid assays confirmed that MGP specifically interacted with both BMP-4 and BMP-2. To elucidate the mechanism of action of the blocking antibody against MGP, we tested if the antibody inhibited the interaction between MGP and BMP-4. Indeed, serial GST pulldown assays disclosed that the addition of the antibody specifically abolished the interaction between GST-MGP and BMP-4. These results indicate that MGP might act as a niche factor for normal and malignant HSPCs, at least in part, through interacting with, and modulating the action of, BMP-4 and BMP-2.