The anti-inflammatory peptide early pregnancy factor/chaperonin 10 (cpn10) was identified by 2D-electrophoresis/mass spectrometry as one of the proteins increased in human umbilical cord endothelial cells (HUVEC) after treatment with erythropoietin (EPO). EPO increased the amount of cpn10 released into the medium of HUVEC cultures, despite the absence of a secretory signal peptide. Although immunosupressive agents would represent an indirect advantage for red cell formation under conditions of infection and inflammation, it is possible that cpn10 may have direct effects on erythroid cells. We show that the chaperonin decreased cell proliferation in cultures of the erythroleukemia cell line K562 and increased the amounts of the erythroid differentiation markers glycophorin A and hemoglobin in TF-1 cells. Nevertheless, cpn10 is not a specific erythroid cell differentiation factor, because monolayers of skin fibroblasts overexpressing cpn10 had significantly higher levels of the differentiation marker collagen I than normal fibroblasts. Nothing is known about the mechanism of action of cpn10 in its capacity as a general differentiation factor. An analysis of early changes taking place in K562 cells after incubation with cpn10 using antibody microarrays identified several phosphorylation events, including a decrease of GSK-3alpha phosphorylation. Further studies in TF-1 cells indicated that cpn10 decreased the phosphorylation of cofilin-1 while stimulating that of GSK-3beta. Furthermore, glycophorin A production decreased in the presence of a GSK-3 inhibitor in the same cells. These experiments support the idea that GSK-3-regulated signal transduction pathways are not only important for stem cell maintenance but may be involved in events controlling cell differentiation.