Blood From a Stone; Erythropoietin From the Bone

In healthy people, erythropoietin produced in renal interstitial cells maintains the normal hematocrit. Aside from the kidney, only the liver and possibly brain glial cells are known to have the capacity to produce erythropoietin. This study from the laboratory of Amato Giaccia at Stanford University reveals that osteoblasts in the bone marrow also express erythropoietin. This remarkable observation expands the list of diverse functional properties of subendosteal osteoblasts, known essential components of the hematopoietic stem cell niche.

Erythropoietin production in the kidney is regulated by hypoxia inducible factor (HIF) signaling. In renal cells, hypoxia leads to increased levels of HIF proteins, which are transcription factors that promote expression of many genes including erythropoietin. HIF protein levels are tightly regulated post-transcriptionally by a rheostat system based on protein stabilization versus protein degradation. The proteins that promote HIF degradation are Von Hippel-Landau (VHL) protein and prolyl hydroxylase domain (PHD) enzymes. Although VHL disease is known primarily for causing renal cell carcinoma, hemangioblastomas, and pancreatic cysts, 15 to 20 percent of patients with congenital erythrocytosis, including those with Chuvash polycythemia, have mutations in the VHL gene. In these patients, the VHL protein is functionally abnormal in its role in the HIF degradation pathway, leading to increased erythropoietin production as a consequence of persistent expression of HIF proteins.

The novel finding in this paper is that VHL-HIF-erythropoietin signaling also occurs in osteoblasts, localizing this process to the bone marrow and, intriguingly, to the hematopoietic stem cell niche. Using a mouse model in which VHL and HIF protein levels could be genetically manipulated uniquely in osteoblasts, Rankin et al. showed that mice with knockout of VHL only in the osteoblasts had a phenotype consistent with that observed in patients with congenital erythrocytosis including elevated circulating levels of erythropoietin, polycythemia, and splenomegaly with extensive extramedullary hematopoiesis. By knocking out different HIF genes, they show that this effect is dependent upon stabilization of HIF2α (but not HIF1α) in osteoblasts. And by blocking erythropoietin signaling using a soluble erythropoietic receptor antagonist, the authors proved that the polycythemia in these mice is erythropoietindependent. Notably, the erythrocytosis caused by erythropoietin upregulation in osteoblasts completely suppressed erythropoietin production in renal cells.

Deletion of VHL in osteoblasts had additional effects on bone and bone marrow: Mice had both increased trabecular bone and a relative increase in hematopoietic stem cells. However, analysis of the committed progenitor subpopulations determined that osteoblast-derived erythropoietin only increased transferrin receptor and glycophorin-A-expressing erythroid progenitors responsible for the erythrocytosis. As HIF2α is the downstream mediator of the effects of VHL deletion in osteoblasts, Rankin et al. deleted HIF2α in osteoblasts to understand the physiologic role of VHL-HIFerythropoietin signaling in osteoblasts in steady-state conditions. Interestingly, this manipulation led to a statistically significant decrease in erythroid precursors in the marrow, but it did not cause anemia. When anemia was induced with phenylhydrazine to “stress” the animals, recovery from the anemia did not require osteoblastderived erythropoietin. While it is not clear why osteoblast-derived erythropoietin does not adequately compensate for the loss of renal erythropoietin in patients with kidney failure, the investigators next tested whether pharmacologic manipulation of osteoblast-derived erythropoietin production could potentially benefit patients with anemia caused by renal insufficiency. When they stabilized HIF proteins by injecting directly into the bone marrow small molecules (currently in clinical trials) that inhibit PHD, the mice developed the same erythrocytosis as that caused by deletion of VHL.