Follistatin modulates a BMP autoregulatory loop to control the size and patterning of sensory domains in the developing tongue

Abstract

The regenerative capacity of many placode-derived epithelial structures makes them of interest for understanding the molecular control of epithelial stem cells and their niches. Here, we investigate the interaction between the developing epithelium and its surrounding mesenchyme in one such system, the taste papillae and sensory taste buds of the mouse tongue. We identify follistatin (FST) as a mesenchymal factor that controls size, patterning and gustatory cell differentiation in developing taste papillae. FST limits expansion and differentiation of Sox2-expressing taste progenitor cells and negatively regulates the development of taste papillae in the lingual epithelium: in Fst(-/-) tongue, there is both ectopic development of Sox2-expressing taste progenitors and accelerated differentiation of gustatory cells. Loss of Fst leads to elevated activity and increased expression of epithelial Bmp7; the latter effect is consistent with BMP7 positive autoregulation, a phenomenon we demonstrate directly. We show that FST and BMP7 influence the activity and expression of other signaling systems that play important roles in the development of taste papillae and taste buds. In addition, using computational modeling, we show how aberrations in taste papillae patterning in Fst(-/-) mice could result from disruption of an FST-BMP7 regulatory circuit that normally suppresses noise in a process based on diffusion-driven instability. Because inactivation of Bmp7 rescues many of the defects observed in Fst(-/-) tongue, we conclude that interactions between mesenchyme-derived FST and epithelial BMP7 play a central role in the morphogenesis, innervation and maintenance of taste buds and their stem/progenitor cells.

Fig. 1.

Fst expression and Fst ^-/- phenotypes in the developing tongue. (A) In situ hybridization (ISH) analysis of Fst expression in wild-type mouse tongue. Asterisk, Fst expression in epithelium of tongue anlage; m, mandibular component of first branchial arch; IE, intermolar eminence; j, developing jaw. (B) P0 tongues stained with Bromophenol Blue. CV, circumvallate papilla. (C) SEM of e17.5 tongue from wild-type and Fst ^-/- mice. Red asterisks indicate fungiform papillae in the anterior tongue (Ant). Post, posterior tongue. (D) Hematoxylin and Eosin (H&E)-stained (P0) and BrdU-immunostained (e17.5) sections through IE and Ant regions of wild-type and Fst^-/- tongue. Arrow and arrowhead in H&E-stained sections indicate regions of dysplastic epithelium (see text). Dashed lines in insets of A indicate apical and basal borders of lingual epithelium. Magnification is the same in all panels. Scale bars: 100 μm in A; 50 μm in D and in C, Post, highest magnification inset; 10 μm in C, Ant, highest magnification inset.

Fig. 2.

Aberrant expression of Sox2 in Fst^-/- tongue. (A) Developmental expression of Sox2-βgeo reporter detected by X-Gal staining in wild-type tongue. Insets are high magnifications of taste placodes/buds. (B-M′) Sox2-βgeo expression detected by X-Gal staining of whole-mount (B-G,H-M) and sagittal sections (C′-G′,I′-M′) of tongues from e14.5 (B-G′) and e17.5 (H-M′) Sox2-βgeo (control) and Fst^-/-; Sox2β-geo (mutant) mice. Arrowheads point to Sox2-βgeo expression in developing fungiform papillae; at e14.5, X-Gal-stained taste bud cell clusters near the surface are significantly smaller in mutants [182.56±2.6 μm² (±s.d.)] than in controls [269.5±4.4 μm² (±s.d.)] (P=0.0017, Student's t-test). Arrows indicate areas of aberrant invaginated Sox2-βgeo-expressing cells. CV, circumvallate papilla. Scale bars: 20 μm.

Fig. 3.

Shh expression and Wnt-β-catenin activity are decreased in Fst^-/- tongue. (A) ISH for Shh in tongues from wild-type and Fst^-/- littermates at different developmental ages. Insets show higher magnifications of boxed regions (e12.5) and developing fungiform papillae (e14.5). J, jaw; CV, circumvallate papilla. (B) X-Gal staining of tongues of BAT-Gal^+/- (control) and Fst^-/-; BAT-Gal mice at the same ages as in A. Arrowheads indicate developing fungiform papillae. Wnt-β-catenin activity is normally present in filiform papillae (Iwatsuki et al., 2007). Scale bars: e12.5, 100 μm (low power), 20 μm (insets); e14.5, 100 μm (low power), 50 μm (insets); e17.5, 50 μm.

Fig. 4.

Ectopic development and innervation of gustatory cell domains in Fst^-/- tongue. (A) ISH for Sox2 and Shh, and X-Gal staining for Wnt-β-catenin activity (BAT-Gal), in the IE of wild-type and Fst^-/- mouse tongues at e18.5. (B) (Top) Immunofluorescence for PGP9.5 in sagittal sections of tongue to detect ingrowing nerve fibers in the IE of wild-type and Fst^-/- tongues at e17.5. (Bottom) Combined X-Gal staining (to detect the Sox2-βgeo reporter) and HRP immunohistochemistry (to detect PGP9.5-immunoreactive nerve fibers) in sagittal sections of tongue at the same age. (C) Immunofluorescence for β-galactosidase (to detect the Sox2- βgeo reporter) and gustducin in sections through the IE and anterior (Ant) regions of control and Fst^-/-; Sox2-βgeo tongues at e18.5. To ensure that comparable sections of IE were analyzed in control and mutant tongues, the CV was used as a landmark. (D) ISH for Foxa2 in wild-type and Fst^-/- tongues at e18.5. Insets show higher magnifications of boxed regions. Scale bars: 50 μm in A,B; 20 μm in D and in IE in C; 10 μm in Ant in C.

Fig. 5.

Upregulation of Bmp7 expression and signaling in Fst^-/- tongue. (A,B) ISH for Bmp7 (blue) and anti-phospho-SMAD1,5,8 (P-SMAD, green) immunostaining in sagittal sections of e17.5 wild-type and Fst^-/- mouse tongues. Bmp7 ISH insets show higher magnifications of boxed regions. P-SMAD immunofluorescence micrographs show comparable regions of anterior (A) or IE (B) tongue at higher magnification. Scale bars: 20 μm.

Fig. 6.

Rescue of the Fst^-/- phenotype by loss of Bmp7 and demonstration of BMP7 auto-activation. (A) Tongues from P0 mice of indicated genotypes stained with Bromophenol Blue. (B) ISH analysis for Shh in e16.5 tongues of indicated genotypes. (C) Tongues from e12.0 Bmp7^lacZ/+ mice were cultured for 2 days in the absence (Ctrl) or presence (+BMP7) of BMP7, cryosectioned and stained with X-Gal. n=4 tongues per condition. (D) SEM of e17.5 wild-type and Bmp7^-/- tongues. Fungiform papillae (red asterisks) in anterior (Ant) tongue are of normal size and spacing in Bmp7^-/- mice. Filiform papillae in posterior tongue (Post) and IE are also normal. Asterisks denote artifacts from the ISH procedure. Scale bars: 20 μm in B,C; 10 μm in D, Ant high magnification, 50 μm in Post high magnification.

Fig. 7.

FST and BMP7 form a regulatory circuit that regulates taste papilla formation and patterning in the developing mouse tongue. (A) Model illustrating how FST and BMP7 act to regulate SOX2 expression, Wnt-β-catenin signaling and SHH expression in developing fungiform papillae of the anterior tongue. (B) Representation of results from computational simulation showing how BMP7 and FST act to regulate the fidelity of taste papillae patterning in the developing anterior tongue. The complete simulation is given in Fig. S9 in the supplementary material. See main text for a detailed explanation.