Direct transcriptional repression of Zfp423 by Zfp521 mediates a bone morphogenic protein-dependent osteoblast versus adipocyte lineage commitment switch

Abstract

Osteoblasts and adipocytes arise from a common mesenchymal precursor cell. The cell fate decision of a mesenchymal precursor cell is under the influence of molecular cues and signaling pathways that lead to the activation or repression of lineage-specific transcription factors. The molecular mechanisms determining osteoblast versus adipocyte lineage specificity in response to bone morphogenic protein (BMP) remain unclear. In this study, we describe the mechanism through which Zfp521 (ZNF521), a regulator of lineage progression in multiple immature cell populations, regulates lineage specification of mesenchymal progenitor cells during BMP-induced differentiation events. In vivo deletion or in vitro knockdown of Zfp521 in mesenchymal precursors resulted in increased expression of the adipocyte determinant factor Zfp423 (ZNF423). This was concurrent with the loss of histone H3K9 methylation and an increase in histone H3K9 acetylation at the Zfp423 promoter, which together are indicative of decreased gene repression. Indeed, we found that Zfp521 occupies and represses the promoter and intronic enhancer regions of Zfp423. Accordingly, conditional deletion of Zfp521 inhibited heterotopic bone formation in response to local injection of BMP2. In contrast, marrow adiposity within BMP2-induced bone was markedly enhanced in Zfp521-deficient mice, suggesting that precursor cells lacking Zfp521 differentiate preferentially into adipocytes instead of osteoblasts in response to BMP2. Consistent with a cell-autonomous role of Zfp521 in mesenchymal precursors, knockdown of Zfp521 in stromal cells prevented BMP2-induced osteoblast marker expression and simultaneously enhanced lipid accumulation and expression of adipocyte-related genes. Taken together, the data suggest that Zfp521 is a cell fate switch critical for BMP-induced osteoblast commitment and identify Zfp521 as the intrinsic repressor of Zfp423 and hence of adipocyte commitment during BMP-induced mesenchymal precursor differentiation.

FIG 1

Zfp521 represses Zfp423 promoter activity. (A) qRT-PCR of Zfp521 and Zfp423 gene expression in undifferentiated proliferating W20 cells (Prolif) and in W20 cells 6 days following the induction of osteoblast (Ob) or adipocyte (Adipo) differentiation. (B) qRT-PCR for Zfp521 and Zfp423 gene expression in W20 cells stably expressing scrambled shRNA (shScr) or shRNA targeting Zfp521 (shZfp521). (C) ChIP analysis of Zfp521 on the Zfp423 promoter and intron 5. A coding region in Zfp423 exon 3 served as a negative control. The schematic shows the locations of the primers (A, B, and C) on the Zfp423 locus. (D) DNA affinity precipitation assay to determine binding of Zfp521 to the Zfp423 promoter or enhancer sequence. Nuclear extracts from W20 cells were incubated with biotinylated Zfp423 oligonucleotide probes and streptavidin-magnetic beads, after which Zfp521 binding was determined by Western blotting (WB). (E and F) Knockdown of Zfp521 enhanced the relative luciferase activity of a Zfp423 promoter luciferase construct (E) or a Zfp423 intron 5 luciferase construct (F) in HEK293 cells. (G and H) ChIP of acetylated histone H3 lysine 9 (H3K9ac) (G) or trimethylated histone H3 lysine 9 (H3K9me3) (H) in W20 cells stably expressing the indicated shRNA constructs using specific primers for the Zfp423 promoter region. (I) W20 cells were transfected with Zfp423 promoter luciferase plasmid and plasmids containing cDNA for the indicated transcription factors. Luciferase activity was measured 24 h after transfection. The data are presented as means ± standard deviations (SD); n ≥ 3. *, P < 0.05; **, P < 0.01; ***, P < 0.001; Student's t test.

FIG 2

Reduced BMP2-induced bone formation in Zfp521^−/− mice. (A) Schematic of a BMP2 calvarial injection model. (B and C) μCT 3D reconstruction (B) and cross-sectional image (C) of wild-type (WT) and Zfp521^−/− calvarium following treatment with BMP2. (D) μCT quantification of newly formed bone in response to BMP2. (E) Fluorescence image of a 6-day-old Dermo1-Cre:mT/mG calvarium demonstrating Cre activity (green) in periosteal cells. (F) Expression of Zfp521 in calvaria of 6-day-old Dermo1-Cre:Zfp521^f/f (Zfp521^cko) mice. (G and H) μCT 3D (G) and cross-sectional (H) images of control (Dermo1-Cre; Zfp521^+/+) and Zfp521^cko mice following treatment with BMP2. (I) μCT quantification of newly formed bone in response to BMP2. The data are presented as means and SD; n ≥ 3. *, P < 0.05; **, P < 0.01; Student's t test. Scale bars, 1 mm (C and H); 10 μm (E).

FIG 3

Increased adiposity in Zfp521^−/− mice. (A) Toluidine blue and von Kossa staining of BMP-induced newly formed bone from WT and Zfp521^−/− mice injected according to the protocol shown in Fig. 2A. (B and C) Toluidine blue and von Kossa staining (B) and histomorphometric quantification of osteoclast numbers (N. Oc) and osteoblast numbers (N. Ob) per bone perimeter (B. Pm) or adipocyte numbers (C) in BMP-induced newly formed bone from control and Zfp521^cko mice. The data are presented as means ± SD; n ≥ 3. *, P < 0.05; ***, P < 0.001; NS, not significant; Student's t test. Scale bars, 500 μm (A); 270 μm (B).

FIG 4

Zfp521 mediates BMP2-induced osteogenesis. (A) Confirmation of Zfp521 knockdown by qRT-PCR. (B and C) W20 cells expressing Zfp521 (shZfp521) or scrambled (shScr) shRNA were cultured with or without 100 ng/ml BMP2 for 48 h, after which alkaline-phosphatase activity was visualized by staining (B) or quantified enzymatically (C). (D) qRT-PCR of osteogenesis-related BMP target gene expression in W20 cells cultured with or without 100 ng/ml BMP2 for 24 h. (E and F) W20 cells were cotransfected with a BMP-responsive luciferase reporter together with vectors expressing shZfp521, shScr, Zfp521 cDNA (pZfp521), or empty-vector control (pCMV6) and then treated with 100 ng/ml BMP2 for 16 h. (G) shZfp521 or shScr stable W20 cells were treated with 100 ng/ml BMP2 for 1 h, after which phosphorylation of Smad1/5 was analyzed by Western blotting. The data are presented as means ± SD; n ≥ 3. *, P < 0.05; **, P < 0.01; ***, P < 0.001 relative to shScr or pCMV6 cells within the same vehicle or BMP2 treatment group; Student's t test.

FIG 5

Modulation of BRE-dependent transcriptional activity by Zfp521. (A) Schematic demonstrating the zinc finger (ZF) motifs and Smad-binding and BRE binding domains of wild-type Zfp521 and Zfp521 deletion constructs. (B) BMP-responsive luciferase (IdWT4-luc) activity in W20 cells 24 h after transfection with wild-type Zfp521 or Zfp521 deletion constructs. (C) Coimmunoprecipitation analysis of Smad4 and deletion mutants of Zfp521. W20 cells were transfected with the indicated expression vectors, after which the cells were lysed and immunoprecipitated (IP) with anti-Flag antibody and blotted (IB) with anti-Zfp521 or anti-Flag antibody. Flag-Smad4 bound wild-type Zfp521 and the ΔZF1-8 mutant but did not bind ΔZF9-19 or Δ9-30. (D) Luciferase reporter activities of the BRE-dependent IDWT4-luc and a mutant reporter in which the four BMP response elements had been mutated (IDMutB-luc) to prevent Smad binding. Reporters were cotransfected into W20 cells together with Zfp521 or empty-vector control and then treated with 100 ng/ml BMP2 for 16 h. (E) Zfp521 works in concert with Smad1/4. W20 cells were transfected with IDWT4F-luc and the indicated expression vectors. Luciferase activity was measured 16 h after treatment with 100 ng/ml BMP2. The data are presented as means and SD; n ≥ 3. *, P < 0.05; **, P < 0.01; ***, P < 0.001 relative to the empty-vector control unless otherwise indicated.

FIG 6

Zfp521 suppresses adipocyte commitment. (A) Oil red O staining of W20 cells expressing scrambled or Zfp521 shRNA differentiated in adipogenic medium supplemented or not with BMP2 for 6 days. (B and C) qRT-PCR of adipocyte selective genes (B) or osteogenic selective genes (C) in the differentiated cultures shown in panel A. (D) In vivo gene expression of key adipocyte commitment factors within the BMP2-induced newly formed bone of wild-type and Zfp521^cko mice. Mice were injected with BMP2 as indicated, and qRT-PCR was performed on mRNA isolated directly from newly formed bone tissue 24 h after the final injection. The data are presented as means ± SD; n ≥ 3. *, P < 0.05; **, P < 0.01; ***, P < 0.001; NS, not significant relative to shScr within the same vehicle or BMP2 treatment group; Student's t test.

FIG 7

Model for the coordinated regulation of adipocyte and osteoblast commitment by Zfp521. BMP signals promote the differentiation of mesenchymal progenitor cells. In uncommitted precursors, Zfp521 acts as an intrinsic block to Zfp423 expression by maintaining the Zfp423 promoter and enhancer in a transcriptionally silent state. In the absence of Zfp521, Zfp423 expression is triggered, allowing BMP or proadipogenic stimuli to further stimulate Zfp423 expression, leading to exaggerated adipogenic lineage commitment.