BMP4/Smad5 dependent stress erythropoiesis is required for the expansion of erythroid progenitors during fetal development

Abstract

The rapid growth of the embryo places severe demands on the ability of the cardiovascular system to deliver oxygen to cells. To meet this need, erythroid progenitors rapidly expand in the fetal liver microenvironment such that by E14.5, erythropoiesis predominates in the fetal liver. In this report we show that the BMP4/Smad5 dependent stress erythropoiesis pathway plays a key role in the expansion of erythroid progenitors in the fetal liver. These data show that the fetal liver contains two populations of erythroid progenitors. One population resembles the steady state erythroid progenitors found in the adult bone marrow. While the second population exhibits the properties of stress erythroid progenitors found in adult spleen. Here we demonstrate that defects in BMP4/Smad5 signaling preferentially affect the expansion of the stress erythroid progenitors in the fetal liver leading to fetal anemia. These data suggest that steady state erythropoiesis is unable to generate sufficient erythrocytes to maintain the rapid growth of the embryo leading to the induction of the BMP4 dependent stress erythropoiesis pathway. These observations underscore the similarities between fetal erythropoiesis and stress erythropoiesis.

Figure 1. Expansion of Stress BFU-E in the fetal liver of f/f and control mice

Fetal liver cells were isolated on the indicated days during fetal development and plated in methylcellulose media containing only Epo. BFU-E were scored after five days in culture. (A) Relative number of BFU-E per 5×10⁵ fetal liver cells. (B) Total number of stress BFU-E per fetal liver. The significance of the differences between f/f and control discussed in the text are indicated on the figure. If no other p value is given then the difference was not significant.

Figure 2. Expansion of steady state BFU-E in the fetal liver of f/f and control mice

Fetal liver cells were isolated on the indicated days during fetal development and plated in methylcellulose media containing Epo + IL-3. BFU-E were scored after five days in culture. The number of Epo+BPA BFU-E was calculated by subtracting the number of Epo-only BFU-E from the number of Epo+IL-3 BFU-E. (A) Relative BFU-E per 5×10⁵ fetal liver cells. (B) Total number of steady state BFU-E per fetal liver. The significance of the differences between f/f and control discussed in the text are indicated on the figure. If no other p value is given then the difference was not significant.

Figure 3. Expression of BMP4 in the fetal liver of f/f and control mice

(A) (top) Fetal livers from control and f/f embryos isolated at the indicated days were sectioned and stained with anti-BMP4 antibodies. (Bottom) Negative control staining with isotype control antibodies of fetal liver sections from C57BL/6-f/f and C57BL/6 control embryos. (B) Western blots of whole fetal liver lysates from the indicated days of development probed with anti-BMP4 antibodies and anti-actin antibodies as a loading control.

Figure 4. Effect of BMP4 on Epo only and Epo + IL-3 BFU-E formation

(A) (left) Fetal liver cells were isolated on the indicated days during fetal development and plated in methylcellulose media containing Epo alone or Epo + BMP4. The fold increase in the number of BFU-E (Epo+BMP4/Epo-only) was calculated at each time point. Significant increases in the number of Epo-only BFU-E following BMP4 treatment are indicated on the figure. (right) Western blot of cell lysates from E14.5 f/f and control fetal liver cells probed with anti-Phospho-Smad1, 5, 8, anti-Smad5 and β-actin as a loading control. (B) Fetal liver cells were isolated on E12.5 and plated in the indicated media ± BMP4 to test the effect of BMP4 on the formation of Epo-only or Epo + IL-3 BFU-E. Significant differences are indicated on the figure. If no other p value is given then the difference was not significant.

Figure 5. f/f embryos do not exhibit excessive apoptosis or defects in erythroid terminal differentiation

(A) Fetal liver sections from E14.5 C57BL/6-f/f and C57BL/6 control embryos were stained for apoptotic cells by TUNEL assay. Bright field views are shown to the left. (B) Fetal liver cells from E14.5 C57BL/6-f/f and C57BL/6 control embryos were incubated in erythroid differentiation media for 48 hours. The cells prior to culture (top) and after culture (bottom) were stained with Wright-Giemsa and o-Dianisidine stain to identify hemoglobinized cells.

Figure 6. CD31+Kit+Sca1-Lin- population contains both Epo-only and Epo+IL-3 BFUE

(A) Fetal liver cells were isolated from E14.5 control embryos and sorted for CD31+Kit +Sca1-Lin-, CD31-Kit+Sca1-Lin- or MEP cells. 25,000 cells were plated in methylcellulose media containing Epo or Epo+IL3. BFU-E were scored after five days in culture. (B) Representative flow diagrams from E14.5 control and f/f fetal liver cells gated on Sca1-Lin- cells. (right) Graph showing the average percentage of fetal liver cells that are CD31+Kit+Sca1-Lin- in from E14.5 control and f/f mice. (C) Stress BFU-E and total BFU-E present in the CD31+Kit+Sca1-Lin- population of cells from the fetal livers of control or f/f mice were determined. Significant differences are indicated on the figure. If no p value is given then the difference was not significant.

Figure 7. Yolk sac contains stress BFU-E and progenitor cells that develop into stress BFU-E when Yolk sac cells are plated on fetal liver stromal cells

(A) Schematic of the direct plating of yolk sac cells and co-culture with AFT024 cells prior to plating. (B) Total number of BFU-E (stress BFU-E+ Steady State BFU-E) present in the yolk sac of C57BL/6-f/f and C57BL/6 control mice on the indicated days of embryonic development. *p<0.05 when E10.5 control yolk sacs were compared with E10.5 mutant and E11.5 mutant or control yolk sacs. (C) Left panel: Yolk sac cells from E10.5, E11.5 and E12.5 control and f/f embryos were directly plated in media containing Epo alone or Epo+IL-3. BFU-E were scored 5 days later. Right panel: Yolk sac cells from E10.5 and E11.5 control and f/f embryos were co-cultured on AFT024 fetal liver stromal cells for 48 hours and then plated in methylcellulose media containing Epo alone or Epo+IL-3. BFU-E were scored 5 days later. Epo + IL-3 BFU-E were calculated by subtracting the number of Epo-only BFU-E from the total number of BFU-E formed in Epo+IL-3. The significance of the differences between different time points in f/f and control yolk sacs as discussed in the text are indicated on the figure. If no other p value is given then the difference was not significant. (D) The fold increase in stress BFU-E observed when E10.5 and E11.5 yolk sac cells were co-cultured AFT024 cells.