The BMP/BMPR/Smad pathway directs expression of the erythroid-specific EKLF and GATA1 transcription factors during embryoid body differentiation in serum-free media

Erythroid cell-specific gene regulation during terminal differentiation is controlled by transcriptional regulators, such as EKLF and GATA1, that themselves exhibit tissue-restricted expression patterns. Their early expression, already in evidence within multipotential hematopoietic cell lines, has made it difficult to determine what extracellular effectors and transduction mechanisms might be directing the onset of their own transcription during embryogenesis. To circumvent this problem, we have taken the novel approach of investigating whether the ability of embryonic stem (ES) cells to mimic early developmental patterns of cellular expression during embryoid body (EB) differentiation can address this issue. We first established conditions whereby EBs could form efficiently in the absence of serum. Surprisingly, in addition to mesoderm, these cells expressed hemangioblast and hematopoietic markers. However, they did not express the committed erythroid markers EKLF and GATA1, nor the terminally differentiated β-like globin markers. Using this system, we determined that EB differentiation in BMP4 was necessary and sufficient to recover EKLF and GATA1 expression and could be further stimulated by the inclusion of VEGF, SCF, erythropoietin and thyroid hormone. EBs were competent to respond to BMP4 only until day 4 of differentiation, which coincides with the normal onset of EKLF expression. The direct involvement of the BMP/Smad pathway in this induction process was further verified by showing that erythroid expression of a dominant negative BMP1B receptor or of the inhibitory Smad6 protein prevented induction of EKLF or GATA1 even in the presence of serum. Although Smad1, Smad5 and Smad8 are all expressed in the EBs, BMP4 induction of EKLF and GATA1 transcription is not immediate. These data implicate the BMP/Smad induction system as being a crucial pathway to direct the onset of EKLF and GATA1 expression during hematopoietic differentiation and demonstrate that EB differentiation can be manipulated to study induction of specific genes that are expressed early within a lineage.