Stromal cell derived factor (SDF)-1/CXCL12 has been shown to promote the survival of embryonic retinal ganglion cells, peritoneal B1a (PerBla) lymphocytes and chronic lymphocytic leukemia B cells. Our previous studies implicated SDF-1 as an important factor in enhancing survival of murine bone marrow (BM) hematopoietic stem cells and, human cord blood and adult human BM myeloid progenitors. Since pluripotent embryonic stem (ES) cells can give rise to differentiated cell types derived from all three primary germ layers (endoderm, mesoderm, and ectoderm) and adult stem cells are generated during embryoid body (EB) formation, we investigated whether SDF-1 has effects on survival of ES cells and EB generation of hematopoietic progenitor cells. In order to establish SDF-1 expression patterns during EB formation, we screened supernatants during day 1–5 EB formation for SDF-1 production by three murine ES cell lines (E14, R1 and CCE). We observed low but detectable SDF-1 secreted in cultures of ES cells and day 1 stage EBs. SDF-1 was increased in the media from day 2 stage EBs and continued to increase through days 4–5. CXCR4, SDF-1 receptor, expression was also analyzed. CXCR4 mRNA expression was low in ES cells and day 1–3 EBs, and increased significantly from Day 4 EBs, reaching maximum levels at day 5, and decreasing after day 6. Surface CXCR4 expression was consistent with mRNA data. To determine if SDF-1 had an effect on ES cell survival, we cultured ES cells without serum, and added serum at either 0, 24, 48 or 96 hrs to each of the following groups: A) Control, B) SDF-1 (100ng/ml) or C) AMD3100 (1 μM), an SDF-1 receptor (CXCR4) antagonist. Colonies were scored 7 days after the addition of serum. SDF-1 enhanced survival of ES cells, while AMD3100 decreased survival. We also checked the apoptosis of ES cells after withdrawing serum for 24, 48, 72 and 96 hours in four groups: A) control, B) SDF-1 (100ng/ml), C) AMD3100 (1 μM) or D) AMD3100 (1 μM) and SDF-1 (100ng/ml). SDF-1 decreased apoptosis and AMD3100 blocked the SDF-1 effect. AMD3100 alone increased apoptosis compared to control. This suggests that AMD3100 blocked endogenous SDF-1 effects. To determine if SDF-1 had an effect on differentiation of hematopoietic progenitor cells, we added SDF-1 (100 ng/ml), AMD3100, or AMD3100 plus SDF-1 (100 ng/ml) at the beginning of EB formation, immediately after removal of LIF, and quantitated primitive erythroid (p-BFU-E), definitive erythroid (d-BFU-E), granulocyte-macrophage (CFU-GM) and multipotential Granulocyte/Erythroid/Macrophage/Megakaryocyte (CFU-GEMM) colony formation. In comparison to control cells (cultured without SDF-1 and AMD3100), SDF-1 increased numbers of p-BFU-E, d-BFU-E, CFU-GM, and CFU-GEMM colonies. Addition of AMD3100 with SDF-1 blocked the enhancing effect of SDF-1. In addition, significantly decreased numbers of colonies were also observed in the presence of AMD3100 alone. This suggests that AMD3100 blocks endogenous SDF-1 actions, consistent with our data on SDF-1 production during EB formation. In order to determine when SDF-1 starts affecting hematopoiesis, hemangioblast colony assays were used. Neither SDF-1 nor AMD3100 influenced hemangioblast colony formation or expression of Flk-1 mRNA, a marker of hemangioblasts. The results suggest a role for SDF-1 in ES cell growth and differentiation.