Mdm4 is required for fetal definitive, but not for adult, erythropoiesis. (A) Mdm4^lox/lox embryos that are either positive (EpoR^GFP-Cre/+) or negative (EpoR^+/+) for the EpoR-GFP-cre transgene at various stages of embryonic development. Mdm4^lox/lox; EpoR^GFP-Cre/+ embryos are significantly paler than control littermates at E14.5, whereas Mdm4^lox/lox; EpoR^GFP-Cre/+; p53^−/−embryos appear normal. Scale bar equals 15 μm. (B) Number of CFU-Es and CFC-m's detected in methylcellulose cultures supplemented with Epo or a mixture of IL-3, IL-6, and SCF, respectively. Numbers are expressed per 6 × 10⁵ nucleated fetal liver cells for the CFU-E count and 4.5 × 10⁴ cells for the CFC-m count. Data from controls (blue indicates Mdm4^lox/l+; EpoR^+/+; and red, Mdm4^lox/+; EpoR^GFP-Cre/+) and mutants (yellow indicates Mdm4^lox/lox; EpoR^GFP-Cre/+) represent the means (± SD) of 6 independent experiments. There is a significant difference between the mean numbers of CFU-Es in the Mdm4 mutants compared with their control littermates by the Student test (P < .001; S). (C) Hematocrits of Mdm4^lox/+; EpoR^+/+ (n = 15), Mdm4^lox/lox; EpoR^GFP-Cre/+ (n = 7), and Mdm4^lox/lox; EpoR^GFP-Cre/+; p53^−/− (n = 3) E14.5 littermates. □, ♦, and ▴ represent the data from the Mdm4^lox/lox; EpoR^GFP-Cre/+ mice, control mice, and Mdm4^lox/lox; EpoR^GFP-Cre/+; p53^−/− mice, respectively. Differences between control and Mdm4^lox/lox; EpoR^GFP-Cre/+ mice are statistically significant by the Student t test (P < .001; S). (D-E) inactivation of Mdm4 in erythroid progenitors leads to decreased cell proliferation. (D) Immunohistologic staining of phosphorylated histone H3 in the lateral ventricle (L. Ventricle; used here as a control), fetal liver (F. Liver), and dorsal aorta (D. Aorta) of E14.5 embryos with the indicated genotypes. Magnifications are indicated. (E) Percentage of BrdU-positive cells among 200 erythroid cells located in the dorsal aorta of E14.5 embryos. Cells (3 × 200) were analyzed in 2 different embryos for each genotype. Data from controls (Mdm4^lox/+; EpoR^+/+) and mutants (Mdm4^lox/lox; EpoR^GFP-Cre/+) are represented by the black and gray bars, respectively. Error bars denote standard error. There is a significant difference between the mean numbers by the Student t test (P < .001; S). (F) Five adult control (Mdm4^lox/+; EpoR^+/+) and Mdm4^lox/lox; EpoR^GFP-Cre/+ mice were injected with phenylhydrazine on days 0, 1, and 3. The corrected reticulocyte count allows assessment of erythropoietic rate, and was calculated assuming a normal hematocrit of .45, as follows: Corrected reticulocyte count (%) = reticulocyte count (%) × (hematocrit × 0.01/0.45). □ and ♦ represent the data from the Mdm4^lox/lox; EpoR^GFP-Cre/+ mice and control mice, respectively.