Hematopoietic niches: targets of GVHD

In this issue of Blood, Shono and colleagues provide clear evidence that the disrupted hematopoiesis associated with graft-versus-host disease is not caused by the direct elimination of hematopoietic precursors, but instead is the product of an attack to the osteoblastic cell compartment by alloreactive CD4⁺ T cells. This destruction is mediated by Fas-Fas ligand interactions and results in the loss of integrity of the microenvironment able to support hematopoiesis.¹ 

Graft-versus-host disease (GVHD) is one of the main barriers to the success of allogeneic bone marrow transplantation. The main manifestation of GVHD is the immune attack to multiple organs, with the gastrointestinal tract, lungs, liver, and skin being the preferential targets.²  Although not usually described with the classic symptoms, marrow suppression with compromised hematopoiesis and poor immune reconstitution has long been appreciated in clinical GVHD. The cause of this phenomenon has not been carefully studied and up until now, it was not known whether this was the product of destruction of hematopoietic progenitors or an indirect effect of damage to the bone marrow microenvironment.

Using murine models of GVHD, Shono et al in this issue of Blood confirm the bone marrow (BM) suppression effect. After transplantation of allogeneic T cells, in a fully major histocompatibility complex (MHC)–mismatched BM model, there was a decrease in hematopoietic components in bone marrow and periphery. In bone marrow, these included compromise of hematopoietic stem cells, rapid decrease of early erythropoiesis, and a drastic loss of B-cell progenitors. Interestingly, the development of all these cells and their lineage progressions are dependent on the integrity of bone marrow microenvironment.

To identify the cell types responsible for mediating these alterations, the experimental groups received isolated fractions of T cells demonstrating that allogeneic CD4⁺ T cells, and not CD8⁺ T cells, were the predominant effectors. In addition, using transplantation protocols that included donor T cells isolated from Fas ligand–deficient mice and in separate experiments bone marrow from Fas-deficient donors, the authors clearly showed that Fas-Fas ligand interactions were necessary to mediate the hematopoietic loss, and that the targets for GVHD were not the repopulating hematopoietic progenitors. This last observation is very intriguing as it suggests that host cells with functions in hematopoietic support could be targets of the GVHD. To investigate this notion, hematopoietic progenitors were isolated from animals with ongoing GVHD and control chimeras and retransferred into myeloablated hosts. These experiments, done in a competitive reconstitution setting compared with untreated progenitors, showed that hematopoietic progenitors isolated from GVHD mice could reconstitute lethally irradiated recipients in a multilineage fashion to the same extent as progenitors isolated from control groups. As expected for host mice deficient in hematopoietic support, syngeneic hematopoietic progenitors failed to engraft efficiently in mice undergoing GVHD.

Advances in the field of bone marrow hematopoiesis have described multiple nonhematopoietic cells with the potential of contributing to the homing, maintenance, and renewal of early hematopoietic progenitors. These cells collectively define “hematopoietic niches” and include cells of the osteoblastic lineage, vascular endothelium, and cells of the sympathetic nervous system.^3-7  The relative importance of each of these components has been a point of debate, and new concepts are emerging on their specific roles supporting hematopoiesis at different levels and in different physiologic conditions. To identify which of these compartment(s) could be affected by the GVHD process, the authors made a comparative histologic analysis of the bone marrow in animals with GVHD and control counterparts. These studies showed a dramatic decrease in osteoblastic cells, identified by their endosteal location and morphology. These cells were almost absent as early as 7 days after transplantation. Their osteoblastic identity was established by strict complementary approaches including evaluation of expression of osteoblast lineage- specific genes, and measurement of rate of bone formation by dynamic histomorphometry. These 2 parameters were completely obliterated as product of the GVHD, documenting a dramatic compromise of the osteoblast compartment. Vascular parameters measured as microvascular permeability were also affected. However, there was not a parallel loss of endothelial cells. GVHD in the experimental model caused major disruptions in the bone marrow microenvironment components responsible for supporting hematopoiesis and explain the observed hematopoietic failure.

Once effector cell types, phenotypes, and targets of destruction were identified, the investigators used immunodepletion of effector cells as a therapeutic approach to reduce the deleterious effects of the process. Treatment of experimental groups with anti-CD4 antibodies allowed the recovery of bone marrow and peripheral hematopoiesis in the GVHD setting. This correlated with partial recovery of osteoblastogenesis.

In addition to identifying the nature of effector cells and mechanisms, this paper makes several new contributions. First, it establishes that the impairment of hematopoiesis induced by GVHD occurs very early and it is not an effect of systemic GVHD. Second, hematopoietic failure is not the product of direct destruction of hematopoietic progenitors, but instead is the product of alterations of the bone marrow microenvironment responsible for supporting hematopoietic engraftment and progression. Third, it clearly demonstrates that the osteoblastic niche is one of the targets of GVHD. This work also leaves several questions open. It was shown that T cells recognize osteoblasts in an MHC-independent fashion. Thus, apart from Fas-Fas ligand interaction, is there an additional specific type of recognition between these cells? Is the interaction between T cells and osteoblastic cells dependent on the methods of preconditioning? Is the loss of osteoblasts the main cause of destabilization of the hematopoietic process and/or the observed defects in vascular permeability? Future work answering these questions will be important to the design of combined therapies to ameliorate the damage induced by GVHD. In addition, the evaluation of hematopoietic niche integrity could have prognostic value to predict severity of GVHD. Finally, considering that T cells can alter bone remodeling in multiple pathologic conditions, these studies have the potential to provide important information on destruction of bone integrity associated to exacerbated inflammation and infections.⁸