GATA4 represses RANKL in osteoblasts via multiple long-range enhancers to regulate osteoclast differentiation

Abstract

GATA4 is a transcription factor that is responsible for tissue-specific gene regulation in many tissues, and more recent studies showed that it is necessary for osteoblast differentiation. Previously, we showed that in vivo deletion of Gata4 using Cre-recombinase under the control of the Col1a1 2.3 kb promoter, showed significantly reduced trabecular bone properties. To understand the role of GATA4 in more differentiated cells, GATA4^fl/fl mice were crossed with mice expressing Cre-recombinase under the control of the osteocalcin promoter. MicroCT analysis of trabecular bone properties of the femur and tibia from 14-week-old female osteocalcin-Cre/GATA4^fl/fl (OCN-cKO) mice showed a significant reduction in percentage bone volume, a decrease in trabecular number and an increase in trabecular spacing. In vivo, histomorphometric analysis revealed a decrease in the number of osteoblasts and an increase in the number of osteoclasts in the tibiae of OCN-cKO mice. In vivo and in vitro systems correlated a decrease in Gata4 mRNA with increased RANKL gene expression. To determine if RANKL is a direct target of GATA4, chromatin immunoprecipitation (ChIP)-sequencing was performed, and it demonstrated that GATA4 is recruited to seven enhancers near RANKL. Furthermore, when Gata4 is knocked down, the chromatin at the RANKL region is further opened, as detected by a reduction in histone 3 lysine 27 trimethylation (H3K27me3) and an increase in histone 3 lysine 4 dimethylation (H3K4me2) in the RANKL locus. In vitro, TRAP staining of cells from bone marrow cultures from Gata4 knockout cells show that the increased levels of RANKL are sufficient for osteoclast formation. Together, the data suggest that GATA4 directly represses RANKL expression via seven cis-regulatory regions and plays an important role in maintaining proper bone development and osteoclast formation.

Keywords: Bone; GATA4; Osteoblast; Osteoclast; RANKL.

Fig. 1.

Trabecular bone volume is decreased in GATA4 OCN-cKO mice. (A, B) μCT reconstruction of trabecular bone in the distal tibia metaphysis of 14-week-old female mice. (C) Mean percentage bone volume (BV/TV), (D) trabecular number (Tb.N), (E) trabecular separation (Tb.Sp), (F) trabecular thickness (Tb.Th), (G) cortical thickness and (H) cortical porosity of WT and OCN-cKO mice. Black bars indicate wildtype (WT) mice; white bars indicate GATA4 OCN-cKO mice. Data are mean ± standard deviation from 12 WT and 12 cKO mice. Student’s t-test: *P < 0.05; **P < 0.01 compared with WT.

Fig. 2.

Bone formation is decreased and resorption is increased due to a reduced number of osteoblasts and an increased number of osteoclasts, respectively. (A) Serum from WT and OCN-cKO mice was assayed using RatLaps (C-terminal telopeptides of type I collagen, CTX-I) EIA. (B) Serum from WT and OCN-cKO mice was assayed using P1NP EIA. (C) Bone histomorphometric analysis was performed on WT and OCN-cKO tibiae. Osteoblast surface, (D) osteo-blast number, (E) osteoclast surface and (F) osteoclast number. (G–H) Paraffin-embedded WT and OCN-cKO femurs were stained for TRAP (pink) and counter-stained with methyl green to identify nuclei. Student’s t-test: *P < 0.05. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3.

GATA4 OCN-cKO mice have an increased expression of RANKL. (A, B). Immunohistochemistry for RANKL in WT (A) and OCN-cKO (B) mice.

Fig. 4.

GATA4 expression is inversely correlated with RANKL mRNA. (A, B, C) Bone marrow from WT and OCN-cKO mice was differentiated for two weeks with β-glycerophosphate and ascorbic acid and then mRNA was obtained. (D, E, F) Calvarial osteoblasts were isolated and the silenced with lentivirus to Gata4 (shG4) or control (shC) lentivirus. The cells were differentiated for two weeks with β-glycerophosphate and ascorbic acid, in vitro. (G, H, I) GATA4^fl/fl bone marrow treated with Adenovirus GFP (adv-GFP) or Adenovirus Cre recombinase (Adv-CRE). qPCR was performed with primers to Gata4 (A, D, G), RANKL (B, E, H) and osteoprotegerin (OPG: C, F, I). qPCR data was normalized to β-actin (Actb) mRNA. Student’s t-test: *P < 0.05; **P < 0.01, ***P < 0.001 compared with WT.

Fig. 5.

GATA4 regulates RANKL at eight different loci. Schematic diagram of the RANKL genomic locus. The arrows indicate the direction of gene transcription. The GATA4 binding sites are denoted A-H. VDR sites are denoted D1-D5. The TCR (T cell control region) and CTCF binding sites are indicated [8]. The distances from the transcriptional start site are marked. (B) ChIP was performed using streptavidin-coated beads with wildtype or Flag-biotin-GATA4 calvarial osteoblasts. qPCR was performed with primers to the indicated regions. (C) Calvarial osteoblasts were transduced with either shC or shGATA4 lentivirus. ChIP was performed with an antibody to H3K27me3 or normal IgG. qPCR was performed to the indicated regions. Each PCR was normalized to input and represented as fold enrichment over Actb. Intergenic sequence 7 (IS7) is a negative control. An additional negative control (NC) and positive control (PC) region were used. N = 3, Student’s t-test: * = P < 0.05. (D) Calvarial osteoblasts were transduced with either shC or shGATA4 lentivirus. ChIP was performed with an antibody to H3K4me2 or normal IgG. qPCR was performed to the indicated regions. Each PCR was normalized to input and represented as fold enrichment over a negative intergenic sequence (IS8). Intergenic sequence 7 (IS7) is a negative control. N = 3, Student’s t-test: *P < 0.05; **P < 0.01, ***P < 0.001.

Fig. 6.

Loss of GATA4 leads to a sufficient level of RANKL for osteoclastogenesis. Bone marrow from wildtype or OCN-cKO was differentiated in the osteogenic differentiation media with M-CSF and RANKL (A, B) or M-CSF without RANKL (C, D). The number of TRAP-positive cells from five wells of A–D was enumerated (E). RANKL cDNA was quantified by qPCR (F) from cells treated identically to those in A–D. *P < 0.05.