TGF-β Signaling in Cranial Neural Crest Affects Late-Stage Mandibular Bone Resorption and Length

Abstract

Malocclusions are common craniofacial malformations which cause quality of life and health problems if left untreated. Unfortunately, the current treatment for severe skeletal malocclusion is invasive surgery. Developing improved therapeutic options requires a deeper understanding of the cellular mechanisms responsible for determining jaw bone length. We have recently shown that neural crest mesenchyme (NCM) can alter jaw length by controlling recruitment and function of mesoderm-derived osteoclasts. Transforming growth factor beta (TGF-β) signaling is critical to craniofacial development by directing bone resorption and formation, and heterozygous mutations in TGF-β type I receptor (TGFBR1) are associated with micrognathia in humans. To identify what role TGF-β signaling in NCM plays in controlling osteoclasts during mandibular development, mandibles of mouse embryos deficient in the gene encoding Tgfbr1 specifically in NCM were analyzed. Our lab and others have demonstrated that Tgfbr1^fl/fl;Wnt1-Cre mice display significantly shorter mandibles with no condylar, coronoid, or angular processes. We hypothesize that TGF-β signaling in NCM can also direct later bone remodeling and further regulate late embryonic jaw bone length. Interestingly, analysis of mandibular bone through micro-computed tomography and Masson's trichrome revealed no significant difference in bone quality between the Tgfbr1^fl/fl;Wnt1-Cre mice and controls, as measured by bone perimeter/bone area, trabecular rod-like diameter, number and separation, and gene expression of Collagen type 1 alpha 1 (Col1α1) and Matrix metalloproteinase 13 (Mmp13). Though there was not a difference in localization of bone resorption within the mandible indicated by TRAP staining, Tgfbr1^fl/fl;Wnt1-Cre mice had approximately three-fold less osteoclast number and perimeter than controls. Gene expression of receptor activator of nuclear factor kappa-β (Rank) and Mmp9, markers of osteoclasts and their activity, also showed a three-fold decrease in Tgfbr1^fl/fl;Wnt1-Cre mandibles. Evaluation of osteoblast-to-osteoclast signaling revealed no significant difference between Tgfbr1^fl/fl;Wnt1-Cre mandibles and controls, leaving the specific mechanism unresolved. Finally, pharmacological inhibition of Tgfbr1 signaling during the initiation of bone mineralization and resorption significantly shortened jaw length in embryos. We conclude that TGF-β signaling in NCM decreases mesoderm-derived osteoclast number, that TGF-β signaling in NCM impacts jaw length late in development, and that this osteoblast-to-osteoclast communication may be occurring through an undescribed mechanism.

Keywords: Transforming Growth Factor-beta Type I Receptor; bone remodeling; bone resorption; jaw; mandible; maxillofacial development; neural crest; osteoclasts.

Figure 1.. TGF-β signaling in neural crest does not affect mandibular bone quality.

Lateral views of micro-CT scanned E18.5 (A) control and (B) Tgfbr1^fl/fl;Wnt1-Cre mouse mandibles show differences in overall size and shape. Transverse sections were analyzed from three different regions in the mandible: molar, incisor, and ventral. Sections from the mandibles were stained with Masson’s Trichrome, which stains bone blue, from (C) control and (D) Tgfbr1^fl/fl;Wnt1-Cre littermate E18.5 embryos. Higher magnification images from Masson’s Trichrome stained ventral sections (white boxes) are shown for the (E) control and (F) Tgfbr1^fl/fl;Wnt1-Cre littermate E18.5 embryos. The sections were analyzed to determine if there were differences in bone stereology. There were no significant differences in (G) bone perimeter (B.Pm) per bone area (B.Ar), (H) trabecular-like width, (I) trabecular-like number, and (J) trabecular-like separation between the controls (blue) and Tgfbr1^fl/fl;Wnt1-Cre mice (cKO, yellow). Using RT-qPCR to assay for (K) Col1α1 and (L) Mmp13 mRNA as a marker for bone deposition shows no significant difference in Tgfbr1^fl/fl;Wnt1-Cre (yellow) at E18.5 compared to controls (blue).

Figure 2.. TGF-β signaling in neural crest does not control the spatial pattern of mandibular bone resorption or tissue mineral density.

Micro-CT scanned mandibles from control and Tgfbr1^fl/fl;Wnt1-Cre E16.5 littermate embryos, (A, B) dorsal and (C, D) lateral views, and E18.5 littermate embryos, (E, F) dorsal and (G, H) lateral views. Tartrate resistant acid phosphatase (TRAP) stained mandibles from control and Tgfbr1^fl/fl;Wnt1-Cre E16.5 littermate embryos, (I, J) dorsal and (K, L) lateral views, and E18.5 littermate embryos, (M, N) dorsal and (O, P) lateral views. Micro-CT analysis of bone volume from (Q) E16.5 and (R) E18.5 control and Tgfbr1^fl/fl;Wnt1-Cre embryos and tissue mineral density from (S) E16.5 and (T) E18.5 control and Tgfbr1^fl/fl;Wnt1-Cre embryos.

Figure 3.. TGF-β signaling in neural crest controls the amount of mandibular bone resorption.

Adjacent transverse sections from mandibles were that stained with Masson’s Trichrome were stained with Tartrate Resistant Acid Phosphatase (TRAP), which stains bone resorption red, and methyl green from (A) control and (B) Tgfbr1^fl/fl;Wnt1-Cre E18.5 littermate embryos. Sections were stained and analyzed from three different regions in the mandible: molar, incisor, and ventral. Zoomed in views from TRAP and methyl green stained ventral sections (black boxes) are shown for the (C) control and (D) Tgfbr1^fl/fl;Wnt1-Cre littermate E18.5 embryos. The sections were analyzed to determine if there were differences in the amount of bone resorption. There were significant differences in bone resorption, as analyzed by (E) osteoclast number per bone area (Oc.N/B.Ar) and (F) osteoclast perimeter per bone area (Oc.Pm/B.Ar), of an approximately 3.2-fold decrease for both in the Tgfbr1^fl/fl;Wnt1-Cre mice when compared to control. (G) Using RT-qPCR to assay for Rank mRNA as a marker for number of osteoclasts shows a significant 2.5-fold decrease in Tgfbr1^fl/fl;Wnt1-Cre at E18.5 when compared to controls. (H) Levels of Mmp9 mRNA shows a 0.64-fold decrease in Tgfbr1^fl/fl;Wnt1-Cre at E18.5 when compared to controls.

Figure 4.. TGF-β signaling in neural crest does not alter osteoblast-to-osteoclast signaling pathways through Rankl, Opg, or M-csf.

Well-known signaling factors from the osteoblast lineage to osteoclasts are Rankl, Opg, and M-csf. Using RT-qPCR to assay for (A) Rankl, (B) Opg, (C) Rankl/Opg, and (D) M-csf mRNA expression revealed no significant difference in Tgfbr1^fl/fl;Wnt1-Cre at E18.5 compared to controls. To inhibit TGFBR1 after the apoptosis in Tgfbr1^fl/fl;Wnt1-Cre mice from E10-E14.5, we injected a TGFBR1inhibitor (SB431542) to quail embryos at HH33, which by Carnegie staging is similar to an E15 mouse and therefore after the E10-E14.5 apoptosis. Embryos were collected at HH38 and stained with alizarin red, which stains calcified bone red. (E) In vehicle-treated control quail, the upper and lower jaws align, but in (F) TGFBR1inhibitor-injected quail the lower jaw is significantly shorter than the upper jaw. (G) The ratio of lower to upper jaw length is significantly less in the TGFBR1-inhibitor injected quail compared to the control injected quail.