Considering that one EPO-hyFc molecule contains two EPO molecules, ED50 of an individual EPO is 55.2 pM, which is comparable to that of r-EPO, indicating that there was no significant loss of bioactivity of EPO by hyFc-fusion. Also, colony formation assays were performed to determine the effect of EPO-hyFc on the proliferation of bone marrow erythroid progenitor cells by quantifying CFU-Es and BFU-Es. One micromole of r- EPO and darbepoetin alfa treatment resulted in 131.8 and 105.3 CFU-E colonies, respectively. EPO-hyFc led to the highest number of CFU-E colonies, which is higher than those by EPO-IgG1 Fc treatment. Consistently, r-EPO and EPO-IgG1 Fc treatment induced comparable BFU-E colonies. In contrast, darbepoetin alfa and EPO-hyFc treatment generated the lowest and the highest numbers of BFU-E colonies, respectively. Taken together, these results suggest that EPO-hyFc is more effective in generating both CFU-E and BFU-E than EPO-IgG1 Fc, indicating the importance of a flexible hinge region. To compare the pharmacokinetic profiles of the two Fc-fused EPO proteins, we delivered a single 400-pmol/kg dose of r-EPO, EPO-hyFc, or EPO-IgG1 Fc into SD rats via the IV route. The serum half-life of r-EPO, used as a control, was 5.5 hours, consistent with previous reports. Notably, both EPO-IgG1 Fc and EPO-hyFc were cleared about five-times more slowly than r-EPO; moreover, the serum half-life of EPO-hyFc was longer than that of EPO-IgG1 Fc. Additionally, the AUClast values for EPOIgG1 Fc and EPO-hyFc were about 9- and 10-fold larger than that for r-EPO, indicating that the fusion of either hyFc or IgG1 Fc increased the in vivo residence time of EPO. Concomitant with the pharmacokinetic study, the percentage of reticulocytes in blood, a marker of in vivo bioactivity of EPO, was measured daily by flow cytometry. As shown in Fig. 4B and 4C, the number of reticulocytes in EPO-IgG1 Fc-injected rats was increased at 4 and 5 dpi compared to r-EPO-injected rats, peaking at 4 dpi. In contrast, EPO-hyFc significantly increased reticulocyte numbers from 2 to 5 dpi to an even greater extent. PBS-treated group has a tendency to slightly increase reticulocytes, presumably due to the frequent bleeding for PK study, which is consistent with a previous report. The AUClast for the increase in reticulocytes generated by EPO-hyFc reached 29.8 d6%, which was approximately 2.3- and 1.7-fold higher than that generated by r-EPO and EPO-IgG1 Fc, respectively. Taken together, our results indicate that the flexibility of the hyFc-hinge region may be mainly responsible for the better in vivo bioactivity of EPO-hyFc compared to EPOIgG1 Fc. Because sialic acid content plays a role in the in vivo half-life of proteins, as shown in darbepoetin alfa, we separated purified EPO-hyFc into high and low sialic acid forms by Q sepharose chromatography.