Polycomb-mediated suppression of CEBPα is required for maintenance of various hematopoietic lineages in mouse. (A,I) Experimental procedure for conditional inactivation of polycomb function. BM cells of Ert2Cre-CAG^{flox-stop-GFP}-Cdkn2a^−/−Ring1a^−/−Ring1b^fl/fl mice or Ert2Cre-CAG^{flox-stop-GFP}-Cdkn2a^−/−Ring1a^−/−Ring1b^fl/+ mice were transplanted without (A) or with (I) competitor cells by IV injection into sublethally irradiated Rag2^−/− mice. Six weeks later, the transplanted mice were administrated tamoxifen intraperitoneally to delete Ring1b in blood cells. Two (A) or 8 (I) weeks after Ring1b deletion, mice were sacrificed and analyzed. (B-D) Flow cytometric profiles of GFP⁺ thymocytes (B) and GFP⁺ BM cells (C-D). Upper and lower panels show data of control (Δ/+; Cdkn2a^−/−Ring1a^−/−Ring1b^Δ/+, n = 5 in panels B and D, and n = 3 in panel C) and Ring1a/b KO (Δ/Δ; Cdkn2a^−/−Ring1a^−/−Ring1b^Δ/Δ, n = 6 in panel B, n = 3 in panel C, and n = 5 in panel D) mice, respectively. (E-G) Number of GFP⁺ DN3 cells (E), pro-B cells (F), and ErPs and MkPs (G) of control (black) and Ring1a/b KO (red) mice. (H) Survival curve with Kaplan-Meier plots after BM transplantation to sublethally irradiated Rag2^−/− mice. Blue and red lines show survival curve of control (Cdkn2a^−/−Ring1a^−/−Ring1b^Δ/+, n = 4) and Ring1a/b KO (Cdkn2a^−/−Ring1a^−/−Ring1b^Δ/Δ, n = 3) mice, respectively. Statistical significance of differences between the survival rates were calculated with the log-rank test. (J) Flow cytometric profiles of whole BM cells of control (n = 4), Ring1a/b KO in Cdkn2a^−/− background (n = 4), and Ring1a/b KO in Cdkn2a^+/− background (n = 3) mice with competitor cells. (K) Percentage of myeloid cells, RBCs, T cells, and B cells among GFP⁺ BM cells of control (blue) and Ring1a/b KO (red) mice with competitor cells. (L) Wright-Giemsa stain of BM smears obtained from control and Ring1a/b KO mice with competitor cells. Data are mean ± standard error of the mean. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001.