Multilineage embryonic hematopoiesis requires hypoxic ARNT activity

Abstract

Although most cells undergo growth arrest during hypoxia, endothelial cells and placental cytotrophoblasts proliferate in response to low O(2). We demonstrate that proliferation of embryonic multilineage hematopoietic progenitors is also regulated by a hypoxia-mediated signaling pathway. This pathway requires HIF-1 (HIF-1alpha/ARNT heterodimers) because Arnt(-/-) embryoid bodies fail to exhibit hypoxia-mediated progenitor proliferation. Furthermore, Arnt(-/-) embryos exhibit decreased numbers of yolk sac hematopoietic progenitors. This defect is cell extrinsic, is accompanied by a decrease in ARNT-dependent VEGF expression, and is rescued by exogenous VEGF. Therefore, "physiologic hypoxia" encountered by embryos is essential for the proliferation or survival of hematopoietic precursors during development.

Figure 1

In vitro differentiation of ES cells and replating of hematopoietic progenitors. (a) Hypoxia. ES cells were differentiated for 9 days under 20% O₂, 5% CO₂, and 75% N₂ (N) or 3% O₂, 5% CO₂, and 92% N₂ (H) and CFU scored 6–7 days after replating under normoxic conditions. A significant increase (* P < 0.05) was seen in Arnt^+/+ CFU-E, CFU-M, CFU-GEMM, and CFU-GM cultured under hypoxia. No increase was seen from Arnt^−/− EBs under hypoxia. (b) Representative CFU in methylcellulose (original magnification; 10x). (c) Normoxia. ES cells were differentiated for 9 days and CFU scored 6–7 days after replating. CFU-E, CFU-M, CFU-GEMM, and CFU-GM (** P < 0.005) and CFU-G (* P <0.05) were all reduced in Arnt^−/− EBs. Also, Arnt^+/− CFU-GEMM showed a significant decrease (* P < 0.05) compared to Arnt^+/+ (n = 4).

Figure 2

Yolk sac hematopoietic colony formation assays. (a) E9.5 yolk sac colonies from Arnt^+/+ (n = 12), Arnt^+/− (n = 22), and Arnt^−/− (n = 11) embryos are shown. All Arnt^−/− CFU showed significant decreases (** P < 0.005) when compared to Arnt^+/+ CFU, as did the CFU-GEMM and CFU-GM of Arnt^+/− yolk sacs (* P < 0.05). (b) E8.5 yolk sac colonies from Arnt^+/+ (n = 16), Arnt^+/− (n = 36), and Arnt^−/− (n = 12) embryos are depicted. Arnt^−/− CFU-E showed significant decreases (** P < 0.0005) when compared to Arnt^+/+ or Arnt^+/− CFU. Non-CFU-E contain cells of the myeloid lineages, and Mixed contain both erythroid and myeloid cells.

Figure 2

Yolk sac hematopoietic colony formation assays. (a) E9.5 yolk sac colonies from Arnt^+/+ (n = 12), Arnt^+/− (n = 22), and Arnt^−/− (n = 11) embryos are shown. All Arnt^−/− CFU showed significant decreases (** P < 0.005) when compared to Arnt^+/+ CFU, as did the CFU-GEMM and CFU-GM of Arnt^+/− yolk sacs (* P < 0.05). (b) E8.5 yolk sac colonies from Arnt^+/+ (n = 16), Arnt^+/− (n = 36), and Arnt^−/− (n = 12) embryos are depicted. Arnt^−/− CFU-E showed significant decreases (** P < 0.0005) when compared to Arnt^+/+ or Arnt^+/− CFU. Non-CFU-E contain cells of the myeloid lineages, and Mixed contain both erythroid and myeloid cells.

Figure 3

Assay of ES-cell-derived G418^r CFU in Arnt/wild-type chimeras. (a) G418 selection strategy to distinguish wild-type or ES-derived CFU from chimeric animals. (b) Southern blot analysis of Arnt^−/− chimera bone marrow CFU cultured in either 0, 1.0, or 1.5 mg/ml G418. Only the targeted Arnt allele (mt) remains after G418 selection, indicating the presence of solely Arnt^−/−-derived G418^r CFU.

Figure 4

Cytokine rescue experiments. (a) Northern analysis of poly(A)-selected RNA (5 μg per lane) derived from Arnt^+/+ and Arnt^−/− EBs, differentiated for 9 days under normoxia (20% O₂) or hypoxia (3% O₂). 3′-UTR probes for VEGF (recognizing multiple splice variants) and HIF-1α (as a loading control) were used. (b) Exogenous VEGF (5 ng/ml), when added to the primary differentiation medium, was able to rescue the number of CFU produced by Arnt^−/− EBs to Arnt^+/+ levels (n = 3). (c) Exogenous EPO (2 U/ml) was not able to rescue progenitor number in Arnt^−/− EBs, indicating its deficiency does not lead to the observed phenotype (n = 3).

Figure 4

Cytokine rescue experiments. (a) Northern analysis of poly(A)-selected RNA (5 μg per lane) derived from Arnt^+/+ and Arnt^−/− EBs, differentiated for 9 days under normoxia (20% O₂) or hypoxia (3% O₂). 3′-UTR probes for VEGF (recognizing multiple splice variants) and HIF-1α (as a loading control) were used. (b) Exogenous VEGF (5 ng/ml), when added to the primary differentiation medium, was able to rescue the number of CFU produced by Arnt^−/− EBs to Arnt^+/+ levels (n = 3). (c) Exogenous EPO (2 U/ml) was not able to rescue progenitor number in Arnt^−/− EBs, indicating its deficiency does not lead to the observed phenotype (n = 3).

Figure 4

Cytokine rescue experiments. (a) Northern analysis of poly(A)-selected RNA (5 μg per lane) derived from Arnt^+/+ and Arnt^−/− EBs, differentiated for 9 days under normoxia (20% O₂) or hypoxia (3% O₂). 3′-UTR probes for VEGF (recognizing multiple splice variants) and HIF-1α (as a loading control) were used. (b) Exogenous VEGF (5 ng/ml), when added to the primary differentiation medium, was able to rescue the number of CFU produced by Arnt^−/− EBs to Arnt^+/+ levels (n = 3). (c) Exogenous EPO (2 U/ml) was not able to rescue progenitor number in Arnt^−/− EBs, indicating its deficiency does not lead to the observed phenotype (n = 3).