April 15, 1999;
Opposite effects of FGF and BMP-4 on embryonic blood formation: roles of PV.1 and GATA-2.
In adult vertebrates, fibroblast
growth factor (FGF) synergizes with many hematopoietic cytokines to stimulate the proliferation of hematopoietic progenitors. In vertebrate development, the FGF signaling pathway is important in the formation of some derivatives of ventroposterior mesoderm
. However, the function of FGF in the specification of the embryonic erythropoietic lineage has remained unclear. Here we address the role of FGF in the specification of the erythropoietic lineage in the Xenopus embryo
. We report that ventral
injection of embryonic FGF (eFGF
) mRNA at as little as 10 pg at the four-cell stage suppresses ventral blood
) formation, whereas expression of the dominant negative form of the FGF receptor in the lateral mesoderm
, where physiologically no blood tissue
is formed, results in a dramatic expansion of the VBI
. Similar results were observed in isolated ventral
marginal zones and animal caps. Bone
morphogenetic protein-4 (BMP-4) is known to induce erythropoiesis in the Xenopus embryo
. Therefore, we examined how the BMP-4 and FGF signaling pathways might interact in the decision of ventral mesoderm
to form blood
. We observed that eFGF
inhibits BMP-4-induced erythropoiesis by differentially regulating expression of the BMP-4 downstream effectors GATA-2 and PV.1. GATA-2, which stimulates erythropoiesis, is suppressed by FGF. PV.1, which we demonstrate to inhibit blood
development, is enhanced by FGF. Additionally, PV.1 and GATA-2 negatively regulate transcription of each other. Thus, BMP-4 induces two transcription factors which have opposing effects on blood
development. The FGF and BMP-4 signaling pathways interact to regulate the specification of the erythropoietic lineage.
[+] show captions
FIG. 1. Effect of FGF signal on blood island formation in ventral mesoderm. (A) Whole-mount immunostaining results. The VMZs of 4-cell-stage embryos were each injected with 10 pg RNA encoding b-gal (A, C) or eFGF (B, D). In a separate experiment, the LMZs of some embryos were injected with 2 ng RNA encoding b-gal (E) or XFD (F). VBI was detected at tailbud stage by whole-mount immunostaining assay with the anti-Ta-globin antibody L5.41 (A, B, E, F), and somite muscle was detected by the antibody 12/101 (C, D). (G) RT-PCR results. The above RNAs were injected into the two animal blastomeres of the 2-cell-stage embryos or into the VMZ of the 4-cell-stage embryos. Animal caps were dissected from the injected embryos at stages 8.5 to 9, while VMZ tissues were dissected at stage 10. Both explants were cultured until tailbud stage and harvested for RT-PCR analysis. Total RNAs from sibling embryos were subject to the RT procedure with reverse transcriptase (named mbryofor positive control) or without (named o RTfor negative control) followed by PCR. The EF-1a transcript was detected as an internal control for equal RNA loading.
FIG. 4. (A, B) Ventral overexpression of PV.1 inhibits blood island formation. The VMZs of 4-cell-stage embryos were each injected with RNA encoding b-gal (1 ng) (A) or PV.1 (0.5 ng) (B). The injected embryos were allowed to develop until tailbud stage followed by whole-mount immunostaining to detect the VBI with the anti-Ta-globin antibody L5.41. (C) Effect of PV.1 signal on expression of erythroid genes. For 4-cell-stage embryos, RNA encoding b-gal or PV.1 (0.5 ng each) was injected into the ventral marginal area and the VMZ tissues were explanted at stage 10. For 2-cell stage-embryos, RNAs encoding b-gal (4 ng), BMP-4 (2 ng) plus b-gal (2 ng), or BMP-4 (2 ng) plus PV.1-EnR (2 ng) were injected into the animal pole and the animal cap (AC) tissues were dissected at stages 8.5 to 9. Both explants were cultured until early gastrula (stage 11) or tailbud (stage 30) stage and harvested for RT-PCR analysis.