Background and rationale: Identifying risk factors for bone loss is of utmost importance in the prevention of osteoporotic bone fractures. We and others have recently described the dramatic erythropoietin (EPO)-induced bone loss. Recombinant EPO and its long-acting derivative, darbepoetin alpha (DA), are frequently and interchangeably used in clinical practice for similar indications including diseases (e.g. multiple myeloma) in which bone mass may be compromised already, as a part of the disease process. Here, we examined in a murine model, the dose response to the erythroid stimulating agents (ESA) EPO and DA, in the erythroid and bone compartments.Methods:C57B6/JRcchsd 9-week old female mice were treated for two weeks with either EPO or DA using two dose intensities. EPO and DA were administered subcutaneously thrice or once weekly, respectively, for a total weekly dose of 27IU/mouse or 540IU/mouse, hereafter defined correspondingly as intermediate (ID) and high dose (HD) subgroups. Bone morphometry of the distal femoral metaphysis was assessed using microCT. Flow cytometry was used for immunophenotyping of bone marrow (BM) cells. One-way ANOVA was used to compare variables among the treatment groups.Results: First, we found that ID-DA induced bone loss to a similar extent as ID-EPO as reflected by a significant (p<0.05) reduction in the trabecular bone fraction (BV/TV, -31%), bone mineral density (BMD, -53%) and trabecular number (Tb.N, -17%). HD-EPO and HD-DA resulted in even higher degrees of bone mass reduction: -48% in BV/TV and -57% in Tb.N (p<0.01 for both). Noteworthy, despite modestly superior gain (~10%) in hemoglobin level (21 and 22.5 mg/dl p< 0.01 in HD-EPO and ID-EPO, respectively; mean hemoglobin levels in diluent-treated mice was 16.4), administration of HD-EPO resulted in substantially higher bone loss compared to ID-EPO. This was manifested in ~80% further reduction in BV/TV as well as 2.5-fold further drop in Tb.N as demonstrated by ΔBV/TV and ΔTb.N (i.e. reduction relative to the control) of 48% versus 27% and 57% versus 23%, respectively (p<0.02 for both parameters). By flow cytometry EPO, but not DA, induced a 35% reduction in alkaline phosphatase positive CD11b negative (ALPL+/CD11b-) preosteoblast population in the bone marrow (p<0.01). The percentage of CD115+ BM cells, containing osteoclast precursors, was increased by 26-30% following both EPO and DA treatments (p<0.05 for EPO/DA versus control and non-significant p>0.05 for EPO versus DA groups). In view of the established role of B cells in bone metabolism, we determined the effects of EPO and DA on BM B cell compartment. HD-EPO but not ID-EPO induced a "maturation arrest"-like state in the bone marrow B cell population, increasing the proportion of B220+/CD19-/CD43high/IgM- pre-pro-B two-fold and B220+/CD19+/CD43high/IgM- pro-B cells 2.8-fold (p<0.05) while reducing the proportion of B220+/CD19+/CD43low/IgM+immature B cells by 2.3-fold (p<0.01). Intriguingly, these alterations were not noted following DA treatment.Conclusion:Applying clinically comparable dosing regimen, both recombinant EPO and DA exert a similar, dose-dependent bone loss, not directly proportional to the erythroid response. However, DA, in contrast to EPO, spares the preosteoblast and bone marrow B cell populations. These finding suggest a different mechanism of action of these closely related ESA preparations on the skeletal and immune systems thus warranting further exploration. Finally, our findings advocate for initiating and maintaining patients with the lowest possible doses of EPO/DA to minimize untoward skeletal and other extra-erythropoietic effects, while achieving the desired correction of anemia.This study was supported by a grant from the Israel Cancer Association to DN, YG and MM. DisclosuresNo relevant conflicts of interest to declare.