Hoxa2 Inhibits Bone Morphogenetic Protein Signaling during Osteogenic Differentiation of the Palatal Mesenchyme

Abstract

Cleft palate is one of the most common congenital birth defects worldwide. The homeobox (Hox) family of genes are key regulators of embryogenesis, with Hoxa2 having a direct role in secondary palate development. Hoxa2^-/- mice exhibit cleft palate; however, the cellular and molecular mechanisms leading to cleft palate in Hoxa2^-/- mice is largely unknown. Addressing this issue, we found that Hoxa2 regulates spatial and temporal programs of osteogenic differentiation in the developing palate by inhibiting bone morphogenetic protein (BMP) signaling dependent osteoblast markers. Expression of osteoblast markers, including Runx2, Sp7, and AlpI were increased in Hoxa2^-/- palatal shelves at embryonic day (E) 13.5 and E15.5. Hoxa2^-/- mouse embryonic palatal mesenchyme (MEPM) cells exhibited increased bone matrix deposition and mineralization in vitro. Moreover, loss of Hoxa2 resulted in increased osteoprogenitor cell proliferation and osteogenic commitment during early stages of palate development at E13.5. Consistent with upregulation of osteoblast markers, Hoxa2^-/- palatal shelves displayed higher expression of canonical BMP signaling in vivo. Blocking BMP signaling in Hoxa2^-/- primary MEPM cells using dorsomorphin restored cell proliferation and osteogenic differentiation to wild-type levels. Collectively, these data demonstrate for the first time that Hoxa2 may regulate palate development by inhibiting osteogenic differentiation of palatal mesenchyme via modulating BMP signaling.

Keywords: Hoxa2; RUNX2; bone morphogenetic protein (BMP); cleft palate; osteoblast; osteoprogenitor; proliferation.

Figure 1

Loss of Hoxa2 leads to increased osteogenic differentiation of the palatal mesenchyme at E16.5. Position matched coronal sections of wild-type and Hoxa2^−/− embryos at E16.5 were stained for ALPI (A–H), RUNX2 (I–L), and SP7 (M–P). Sections in the anterior region (A,B,E,F,I,J,M,N) were through the middle of the first molar tooth bud to detect osteogenic condensation of the palatal process of the maxilla. Sections in the posterior region (C,D,G,H,K,L,O,P) were through the osteogenic centers of the developing palatal process of the palatine bone. (A–D) ALPI staining (blue) counterstained with nuclear fast red. Scale bar, 100 μm. Boxed regions in (A–D) highlighting the palate are enlarged (E–H). Scale bar, 50 μm. (E,F) In the anterior hard palate, ALPI staining in the two condensations of the palatal process of the maxilla (marked in black dotted lines) was evidently increased in the Hoxa2^−/− embryos (F) compared to wild-type (E). (G,H) In the posterior hard palate, ALPI stained developing palatal process of the palatine bone (marked in black dotted lines) in the Hoxa2^−/− embryos (H) was increased compared to the wild-type (G), n = 5 biological replicates. (I–P) Immunohistochemical analyses of RUNX2 (green; I–L) and SP7 (red; M–P) in wild-type and Hoxa2^−/− palate at E16.5. RUNX2 was increased in both anterior (J) and posterior regions (L) of the Hoxa2 null hard palate, whereas SP7 was increased only in the anterior hard palate (N), n = 4 biological replicates. Scale bar, 50 μm. M1, first molar; Mb, mandible; Mx, maxilla; NS, nasal septum; pppb, the palatal process of the palatine bone; ppmx, the palatal process of the maxilla; T, tongue.

Figure 2

Loss of Hoxa2 leads to increased expression of osteogenic markers in the developing palate at E13.5 and E15.5. Quantitative real-time PCR (qRT-PCR) analyses indicate that the gene expression profile of osteogenic markers such as Runx2 (A,E), AlpI (B,F), and Sp7 (C,G) were upregulated in the developing Hoxa2^−/− palatal shelves at E13.5 (A–C) and E15.5 (E–G). Gene expression of Bglap was upregulated at E15.5 (H) but not at E13.5 (D). qRT-PCR data (n = 5 biological replicates) were normalized to β-actin and represented relative to wild-type (mean ± S.E.M; unpaired t-test, ^*p < 0.05; ^**p < 0.01; ^***p < 0.001). Western blot analyses of RUNX2 (I,J) and SP7 (I,K,L) were carried out using the microdissected palatal shelves from wild-type and Hoxa2^−/− mice at E13.5 and E15.5. RUNX2 protein expression was upregulated in the Hoxa2^−/− palate at E13.5 and E15.5 (I,J), whereas SP7 isoforms were upregulated at E15.5 (I,K,L). Densitometric analyses (n = 4 biological replicates) were normalized to β-ACTIN and represented relative to wild-type (mean ± S.E.M; unpaired t-test, ^**p < 0.01).

Figure 3

Hoxa2 inhibits osteoblast differentiation of MEPM cells in vitro. Wild-type and Hoxa2^−/− MEPM cells were subjected to osteogenic differentiation in vitro for up to 21 days (d). The differentiated cells were stained for ALPI at d8 (A,B), ARS at d15 (C,D), and d21 (E,F). ARS stained osteocyte matrices from wild-type and Hoxa2^−/− MEPM cells were extracted and quantified at d15 (G) and d21 (H). Experiment was carried out three times and the data shown here are from a representative experiment with n = 3 biological replicates; mean ± S.E.M; unpaired t-test, ^*p < 0.05; ^***p < 0.001). Hoxa2^−/− MEPM cells displayed increased matrix deposition and mineralization at d15 (G) and d21 (H), respectively. (I–L) qRT-PCR analyses revealed that gene expression profile of osteogenic markers such as Runx2 (I), AlpI (J), Sp7 (K), and Bglap (L) were upregulated in the Hoxa2^−/− MEPM cells in a stage-specific manner during osteoblast differentiation. Data was normalized to β-actin and represented relative to wild-type at d0 (n = 3 biological replicates; mean ± S.E.M; two-way ANOVA with Bonferroni post-hoc test, ^*p < 0.05; ^**p < 0.01; ^***p < 0.001).

Figure 4

Hoxa2^−/− palatal shelves exhibit increased osteoprogenitor proliferation and commitment at E13.5. Osteoprogenitor cells in the developing palatal shelves of wild-type and Hoxa2^−/− embryos were evaluated using RUNX2 immunostaining (A,B) and RUNX2-positive cells were counted manually using ImageJ platform (C). Proliferation rate was assessed using Ki67 immunostaining (D–F) at E13.5. Scale bar, 50 μm; N, nasal; O, oral. Proliferating osteoprogenitor cells (cellspositive for both RUNX2 and Ki67) (Runx2 /Ki67) (G,H) relative to the total number of mesenchymal cells (DAPI-positive) from wild-type and Hoxa2^−/− palatal shelves were counted in the nasal side (I). Hoxa2^−/− embryos exhibited increased RUNX2-positive (C), Ki67-positive (F) and RUNX2/Ki67-double positive (I) cells in the nasal side of the palatal shelves (n = 5 biological replicates; mean ± S.E.M; unpaired t-test, ^*p < 0.05; ^**p < 0.01; ^***p < 0.001). Expression of cell cycle regulator Cyclin D1 (Ccnd1) mRNA was upregulated in the Hoxa2^−/− palatal shelves (J) from E12.5 to E14.5. qRT-PCR data was normalized to β-actin and represented relative to wild-type at respective embryonic stages (n = 6 biological replicates; mean ± S.E.M; unpaired t-test, ^*p < 0.05; ^***p < 0.001).

Figure 5

Hoxa2 regulates canonical BMP signaling in the developing palate. Gene expression of BMP ligands, Bmp2 (A) and Bmp4 (B) were upregulated in Hoxa2^−/− palatal shelves at E13.5 and E15.5. qRT-PCR data (n = 4 biological replicates) were normalized to β-actin (mean ± S.E.M; unpaired t-test, ^*p < 0.05; ^***p < 0.001). Representative immunoblot (C) of pSMAD 1/5/8 from the developing palate of wild-type and Hoxa2^−/− embryos at E15.5. (D) Densitometric analysis represents the relative expression of pSMAD 1/5/8 normalized to SMAD 1/5/8 and represented relative to wild-type (n = 4 biological replicates; mean ± S.E.M; unpaired t-test, ^*p < 0.05).

Figure 6

Blocking canonical BMP signaling with dorsomorphin rescues the aberrant osteoprogenitor cell proliferation and osteogenic differentiation in the Hoxa2^−/− MEPM cells. Dorsomorphin treatment restored gene expressions of Bmp2 (A), Bmp4 (B), and Runx2 (C) in the Hoxa2^−/− MEPM cells close to wild-type levels during osteogenic differentiation in vitro at d8. Data represented relative to wild-type MEPM cells treated with DMSO (n = 4 biological replicates; mean ± S.E.M; one-way ANOVA followed by Bonferroni post-hoc test, ^*p < 0.05; ^***p < 0.001; ns, not significant). (D) Representative immunoblots showing restoration of RUNX2 and pSMAD 1/5/8 in Hoxa2^−/− MEPM cells treated with dorsomorphin during osteogenic differentiation in vitro at d8 (n = 3 biological replicates). (E) Cell proliferation analysis in the wild-type and Hoxa2^−/− MEPM cells treated with DMSO or dorsomorphin during osteogenic differentiation at d3 (n = 5 biological replicates; mean ± S.E.M; one-way ANOVA followed by Bonferroni post-hoc test, ^***p < 0.001; ns, not significant). (F) ALPI staining revealed that treatment with dorsomorphin nullified the aberrant osteogenic differentiation in Hoxa2^−/− MEPM cells in vitro at d8.

Figure 7

Schematic diagram depicting the role of Hoxa2 in proliferation and osteogenic differentiation of the palatal mesenchyme. (A) Hoxa2 inhibits canonical BMP signaling in the developing palate, which in turn restricts the expression domain of osteogenic markers such as Runx2, AlpI, and Sp7. (B) In wild-type, Hoxa2 expression peaks during early palatogenesis to control cell proliferation and to maintain mesenchymal cells in an undifferentiated stage by regulating BMP signaling pathway. (C) Loss of Hoxa2 leads to upregulation of BMP signaling resulting in increased osteoprogenitor cell proliferation and osteogenic differentiation, possibly accounting for the failure in the elevation of palatal shelves resulting in manifestation of cleft palate.