Hypoxia regulates BMP4 expression in the murine spleen during the recovery from acute anemia

Abstract

Background: Bone marrow erythropoiesis is primarily homeostatic, producing new erythrocytes at a constant rate. However at times of acute anemia, new erythrocytes must be rapidly produced much faster than bone marrow steady state erythropoiesis. At these times stress erythropoiesis predominates. Stress erythropoiesis occurs in the fetal liver during embryogenesis and in the adult spleen and liver. In adult mice, stress erythropoiesis utilizes a specialized population of stress erythroid progenitors that are resident in the spleen. In response to acute anemia, these progenitors rapidly expand and differentiate in response to three signals, BMP4, SCF and hypoxia. In absence of acute anemic stress, two of these signals, BMP4 and hypoxia, are not present and the pathway is not active. The initiating event in the activation of this pathway is the up-regulation of BMP4 expression in the spleen.

Methodology/principal findings: In this paper we analyze the regulation of BMP4 expression in the spleen by hypoxia. Using stromal cell lines, we establish a role for hypoxia transcription factor HIFs (Hypoxia Inducible Factors) in the transcription of BMP4. We identified putative Hypoxia Responsive Elements (HREs) in the BMP4 gene using bioinformatics. Analysis of these elements showed that in vivo, Hif2alpha binds two cis regulatory sites in the BMP4 gene, which regulate BMP4 expression during the recovery from acute anemia.

Conclusions and significance: These data show that hypoxia plays a key role in initiating the BMP4 dependent stress erythropoiesis pathway by regulating BMP4 expression.

Figure 1. Hypoxia induces the transcription of BMP4.

(A) RT-PCR analysis of BMP4, Hif1α, Hif1β, Hif2α or β-actin (control) expression in the MSS31 spleen stromal cell line (top); BMP4 and 28S rRNA (control) in the W2017 osteoblast cell line (middle) and BMP4 and 28S rRNA (control) in the AFT024 fetal liver stromal cell line (bottom) at 20% O₂ or 1% O₂. BMP4 expression relative to 18S rRNA in these cell lines was also examined by quantitative RT-PCR. (B) MSS31 cells were treated with Actinomycin D (ActD) (1 ug/ml) for 30 minutes prior to shifting the culture to 1% O₂. At the indicated times cells were harvested, RNA was isolated and the expression of BMP4 was determined by RT-PCR.

Figure 2. HIF1α and HIF2α are required for BMP4 induction under hypoxia.

(A) RT-PCR analysis of the expression of HIF1α or HIF2α in HEK293T cells transfected with the indicated shRNA constructs. (B) RT-PCR analysis of BMP4 expression at 20% or 1% O₂ in HEK293T cells where HIF1α (left) or HIF2α (right) were knocked down with shRNAs.

Figure 3. Identification of putative HREs in the murine BMP4 gene.

(A) Schematic diagram of the location of the 5 putative HREs in BMP4 locus with extra 5000 bp both upstream and downstream, where the transcription start site is located at 6102 bp. The acquired sequence data were from September 2005. (B) The sequences of the 5 putative HREs are aligned among mammals. Sequences in grey box are conserved among 6 species. Below each alignment is the ESPERR 7 species RP score for each nucleotide.

Figure 4. HIF and co-activator p300 are associated with BMP4 under hypoxia in MSS31 cells.

The binding of Hif1α, Hif2α, and p300 to these sites was investigated by ChIP assay in MSS31 cells cultured at 20% O₂ or 1% O₂ for 16 hours. IgG or other antibodies served as controls. PCR were performed with primers amplify HRE-2 (A) and HRE-4(B). Representative gel data are shown. The relative intensity of PCR products normalized to the input lanes is shown by combining data from three to five individual ChIP assays. (C) Positive and specificity controls of ChIP with Hif1α were examined at Glut-1 locus.

Figure 5. Binding of HIF and p300 to the BMP4 locus is induced by anemia.

(A) RT-PCR analysis of BMP4 or β-actin (control) expression in primary spleen cells isolated from mice at 24 hours post induction of acute hemolytic anemia with phenylhydrazine (PHZ) and untreated mice. (B) Spleen cells were analyzed by ChIP assay using anti- HIF1α, HIF2α, p300 or IgG antibodies. PCR was performed with primers amplify HRE-2 (B) and HRE-4(C). Representative gel data are shown. The relative intensity of PCR products normalized to the input lanes is shown by combining data from three individual ChIP assays.

Figure 6. HRE-2 and HRE-4 can confer hypoxia inducible expression to a luciferase gene.

(A) Sequence of wildtype and mutated versions of HRE-2 and HRE-4. (B) Luciferase activity in HEK293T cells transfected with the indicated plasmids at 20% or 1% O₂. The firefly luciferase activity is normalized with renilla luciferase expression. Fold increase in luciferase activity was normalized with the increase observed in the control plasmid set to 1.0. p value was calculated to compare the luciferase activity under 20% versus 1% O₂. Each bar represents the average of three to twelve individual transfections.