Increased activity of mesenchymal ALK2-BMP signaling causes posteriorly truncated microglossia and disorganization of lingual tissues

Summary

Proper development of taste organs including the tongue and taste papillae requires interactions with the underlying mesenchyme through multiple molecular signaling pathways. The effects of bone morphogenetic proteins (BMPs) and antagonists are profound, however, the tissue-specific roles of distinct receptors are largely unknown. Here, we report that constitutive activation (ca) of ALK2-BMP signaling in the tongue mesenchyme (marked by Wnt1-Cre) caused microglossia—a dramatically smaller and misshapen tongue with a progressively severe reduction in size along the anteroposterior axis and absence of a pharyngeal region. At E10.5, the tongue primordia (branchial arches 1–4) formed in Wnt1-Cre/caAlk2 mutants while each branchial arch responded to elevated BMP signaling distinctly in gene expression of BMP targets (Id1, Snai1, Snai2, and Runx2), proliferation (Cyclin-D1) and apoptosis (p53). Moreover, elevated ALK2-BMP signaling in the mesenchyme resulted in apparent defects of lingual epithelium, muscles, and nerves. In Wnt1-Cre/caAlk2 mutants, a circumvallate papilla was missing and further development of formed fungiform papillae was arrested in late embryos. Our data collectively demonstrate that ALK2-BMP signaling in the mesenchyme plays essential roles in orchestrating various tissues for proper development of the tongue and its appendages in a region-specific manner.

1 |. INTRODUCTION

The mammalian tongue is an organ comprised of multiple highly organized tissues that enable its critical functions, including speaking, food processing, and taste sensation. Developmental defects in the tongue such as aglossia, macroglossia, microglossia, and ankyloglossia (Chandrashekar, Kashinath, & Suhas, 2014) can cause serious disease conditions like dysphasia, dysphagia, and dysgeusia. The development of the tongue requires tissue–tissue interactions governed by multiple signaling pathways (Iwata, Suzuki, Pelikan, Ho, & Chai, 2013; Liu et al., 2009; Liu, Grosse, et al., 2012; Liu, Komatsu, Mishina, & Mistretta, 2012; Liu, MacCallum, Edwards, Gaffield, & Mistretta, 2004; Millington et al., 2017). Thus, studies on the organogenesis and molecular regulation of the tongue and its appendages will help to understand the mechanisms underlying the tongue anomalies.

In mice, emergence of the tongue starts with the appearance of three lingual swellings from the first branchial arch (BA 1) at gestational day (E) 11.5. These swellings fuse and form the anterior two thirds of the tongue—the oral tongue at E12.5. Concurrently with the formation of the oral tongue, a small midline swelling (the copula) develops from BA 2 and fuses with another midline swelling originating from BAs 3–4 to form the posterior third of the tongue—the pharyngeal tongue (Cobourne et al., 2018). At this stage, myoblast cells migrate into the developing tongue and differentiate to become intrinsic glossal muscles (Cobourne et al., 2018; Millington et al., 2017). Meanwhile, taste papilla placodes emerge on the dorsal surface (Mbiene, Maccallum, & Mistretta, 1997; Mistretta, 1972) and will develop into taste papillae that host taste buds. The proper formation of the tongue and its appendages is an integrated process that requires a tight coordination among BAs 1–4 and various types of tissues/cells.

Bone morphogenetic proteins (BMP) are secreted proteins that may act in a paracrine manner to activate intracellular signaling in the surrounding cells and play crucial roles in cell growth, differentiation, and apoptosis during the development of many organs (Bragdon et al., 2011; Wang et al., 2014). BMPs bind to the Type II BMP receptors (BMPR2, ACVRII), activate Type I BMP receptors (ALK2, ALK3, and ALK6) and phosphorylate downstream signaling component Smad1, 5, and 8 to promote the expression of BMP target genes (e.g., Id1, Snai1, Snai2, and Runx2). Type I BMP receptors are known to be the main determinants of the downstream signaling activity (Wang et al., 2014). It has been reported that BMP ligands (BMP 2, 3, 4, 6, and 7), antagonists (noggin and follistatin) and receptors (ALK2, ALK3, and ALK6) are present in both epithelium and mesenchyme of the developing tongue (Beites et al., 2009; Jung, Oropeza, & Thesleff, 1999; Kawasaki et al., 2012; Suga et al., 2007; Zhou, Liu, & Mistretta, 2006). The effects of BMP ligands and their antagonists in the tongue and taste papilla development are profound (Beites et al., 2009; Zhou et al., 2006). However, the specific roles of mediating receptors and involved tissue compartments of BMP signaling in tongue organogenesis are largely unknown.

Constitutive activation of ALK3 in the neural crest and derived mesenchyme causes a slightly defected tongue in a previous study focusing on cleft palate formation and delayed tooth differentiation (Li et al., 2013). In the present study, we report that a proper level of ALK2-mediated BMP (hereafter ALK2-BMP) signaling in neural crest and derived mesenchymal cells (marked by Wnt1-Cre) is essential in regulating proliferation and apoptosis of these cells at an early embryonic stage for the proper tongue shape and size. Also, ALK2-BMP signaling plays important regulatory roles in mesenchymal interactions with other lingual tissues for their organization, and the development of taste papillae and taste buds.

2 |. RESULTS

2.1 |. Constitutive activation of ALK2-BMP signaling in neural crest and derived mesenchymal cells results in posteriorly truncated microglossia

To elevate the activity of type I BMP receptor ALK2-mediated BMP (hereafter ALK2-BMP) signaling in the tongue mesenchyme, transgenic mouse model with a constitutively activated (ca) form of Alk2 (caAlk2) (Fukuda et al., 2006) was used to cross with Wnt1-Cre that mark neural crest-derived mesenchymal cells in the tongue (Liu, Komatsu, et al., 2012; Thirumangalathu, Harlow, Driskell, Krimm, & Barlow, 2009). Compared to Cre−/caAlk2 littermate controls (Figure 1a), deficiencies were evident in multiple craniofacial regions of E12.5 Wnt1-Cre/caAlk2 mouse embryos including brain (Figure 1b, arrow), eyes (Figure 1b, arrowhead), and orofacial region (Figure 1b, open arrowhead). Mutant mice had microglossia (Figure 1d,f), with a significantly smaller tongue in mutants than that in littermate controls (526.51 ± 14.18 μm vs. 1,163.09 ± 13.78 μm in length of oral tongue, p < .01 in Figure 1g; 352.43 ± 6.5 μm vs. 749.23 ± 4.25 μm in width, p < .01 in Figure 1h). In the majority of Wnt1-Cre/caAlk2 mouse embryos, tongue swellings fused at the midline for a unitary tongue tip (24 mutants out of 32, Figure 1d,f) while some had a bifid anterior tongue (8 mutants out of 32, image not shown). A progressive decrease in tongue size was consistently observed along the anteroposterior axis in all of the mutants (tongues with a unitary or bifid tip). Although the Wnt1-Cre/caAlk2 mutant tongue was smaller, fungiform taste papilla placodes were observed on the anterior tongue region (Figure 1d,f, arrowheads) similarly to those in the littermate control (Figure 1c,e, arrowheads). On the smaller anterior tongues, fewer rows of fungiform taste papillae were seen and there was not an apparent change of density of papilla placodes. In the position where a distinct circumvallate papilla placode was observed in the littermate control tongues (Figure 1e, arrow), an enlarged Shh⁺ patch of tissue, presumably circumvallate papilla placode, was observed in the tongues of Wnt1-Cre/caAlk2 mutants (Figure 1f, arrow). A pharyngeal part of the tongue posterior to the presumable Shh⁺ circumvallate was missing in Wnt1-Cre/caAlk2 mutants (Figure 1f).

Figure 1.

Representative images (a–f) and measurement of tongue size (g,h) to illustrate the deficiencies of craniofacial region and tongue in E12.5 Wnt1-Cre/caAlk2 mutants compared to Cre⁻/caAlk2 littermate controls. (a, b) Light microscopy images of the heads of a littermate control (a) and Wnt1-Cre/caAlk2 mutant (b). Arrows point to the brain; arrowheads point to the eye; open arrowheads point to the orofacial region. Scale bars: 1 mm. (c, d) Scanning electron microscopy (SEM) images of littermate control (c) and Wnt1-Cre/caAlk2 mutant (d) tongues. Scale bars: 200 μm. (e, f) Light microscopy images of whole mount immunoreaction for Shh (blue), a developing taste papilla marker, in littermate control (e) and Wnt1-Cre/caAlk2 mutant (f) tongues. Arrowheads in c–f point to fungiform taste papilla placodes. The arrow in e and f points to the Shh+ circumvallate papilla placode in the posterior oral tongue. Dashed lines in c and e separate the oral tongue from the pharyngeal tongue. IE, intermolar eminence. Scale bars: 200 μm. (g, h) Histograms (X ± SD; n =3) represent the length (g) and width (h) of the oral tongue in Wnt1-Cre/caAlk2 and littermate controls. **p ≤ .01 Student’s t test compared to littermate control tongue

2.2 |. Fungiform papillae were well developed in the smaller E14.5 Wnt1-Cre/caAlk2 mutant tongue

At E14.5 when tongue and taste papillae attain their stereotypical shape and spatial distribution (Figure 2a), Wnt1-Cre/caAlk2 mutants continued to exhibit microglossia (Figure 2b) compared to the Cre−/caAlk2 littermate control (Figure 2a). Developing fungiform taste papillae were observed in the tongue of E14.5 Wnt1-Cre/caAlk2 mutants (Figure 2b,d, arrowheads, inset in 2d) in similar structure to those of littermate controls (Figure 2a,c, arrowheads, inset in 2c). Similarly to E12.5, there was not an apparent change of papilla density on the smaller anterior tongues. Further, the pharyngeal tongue region was missing and a Shh⁺ circumvallate papilla seen in littermate controls (Figure 2c, arrow) was absent in E14.5 Wnt1-Cre/caAlk2 mutants (Figure 2d). Similar to E12.5, both unitary (14 out of 18, Figure 2f) and bifid (4 out of 18, Figure 2b,d) anterior tips was observed in the smaller tongues of E14.5 Wnt1-Cre/caAlk2 mutants.

Figure 2.

Development of tongue and taste papillae in E14.5 and E16.5 Wnt1-Cre/caAlk2 mutant and Cre⁻/caAlk2 littermate control mice. (a–f) Whole mount images (scanning electron microscopy in a, b, light microscopy in c–f) of E14.5 littermate control (a, c, e) and Wnt1-Cre/caAlk2 (b, d, f) tongues. Whole mount immunoreactions were performed against Shh (c, d) or βIII-tubulin (e, f). Insets in c and d show the restriction of Shh immunoproducts to fungiform taste papilla epithelium on tongue sections. Arrowheads in a–f point to developing fungiform taste papillae. Arrows in a and c point to the circumvallate papilla. Open arrowheads in e and f point to βIII-tubulin+ nerve fibers in littermate control (e) and Wnt1-Cre/caAlk2 mutant (f) tongues. Scale bars: 100 μm. (g–j) Single-plane laser scanning confocal images of sagittal sections of littermate control (g) and Wnt1-Cre/caAlk2 mutant (i) tongues immunostained for βIII-tubulin (green). h and j are high magnification images of the anterior tip of littermate control (h) and Wnt1-Cre/caAlk2 mutant tongues (j). Arrowheads in g–i point to βIII-tubulin+ nerve fibers. Dotted lines in h and j separate the epithelium from the underlying mesenchyme. Scale bars: 100 μm in g, i; 50 μm in h, j. (k–n) SEM images of E16.5 littermate control (k, l) and Wnt1-Cre/caAlk2 mutant (m, n) tongues to illustrate the left anterior tongue tip (k, m) and a fungiform papilla (l, n) in a littermate control (k, l) and Wnt1-Cre/caAlk2 mutant (m, n) tongue. Arrowheads point to fungiform papillae. Scale bars: 100 μm in k, m; 10 μm in l, n

Next, we examined outgrowth and branching of nerve fibers that traverse through the tongue mesenchyme and innervate the developing taste papillae. At E14.5, in contrast to extensive βIII-tubulin⁺ nerve fibers approaching taste papillae in the littermate control (Figure 2e, open arrowhead), nerve fibers were absent in the vicinity of Wnt1-Cre/caAlk2 mutant fungiform taste papillae (Figure 2f). Obviously, βIII-tubulin⁺ nerve fibers were present in Wnt1-Cre/caAlk2 mutant mandible (Figure 2f, open arrowheads). On sections, compared to the extensively branched nerve fibers in littermate control tongues (Figure 2g, arrowhead), βIII-tubulin⁺ nerve fibers were loosely distributed in the Wnt1-Cre/caAlk2 mutant tongue mesenchyme (Figure 2i, arrowhead). Moreover, these fibers did not branch out into taste papillae nor penetrate the epithelium (Figure 2j) in contrast to the dense distribution of nerve fibers in the papilla core in littermate controls (Figure 2h, arrowhead).

To evaluate the further development of fungiform papillae, E16.5 littermate control (Figure 2k,l) and Wnt1-Cre/caAlk2 mutant (Figure 2m,n) tongues were examined using scanning electron microscopy. Similar to E14.5, developing fungiform papillae were observed in the E16.5 Wnt1-Cre/caAlk2 mutants (Figure 2m,n, arrowheads) similarly in structure to those of littermate controls (Figure 2k,l, arrowheads).

2.3 |. Development and organization of multiple lingual tissues were deficient in the E18.5 Wnt1-Cre/caAlk2 mutant tongue

Wnt1-Cre/caAlk2 mutant mice die shortly after birth, thus E18.5 embryos were harvested as the last stage for phenotypic analyses. Compared to the Cre⁻/caAlk2 littermate controls (Figure 3a), microglossia in E18.5 Wnt1-Cre/caAlk2 mutants was evident (Figure 3b). Neither pharyngeal tongue nor circumvallate papilla (Figure 3a, arrow) was seen in the E18.5 Wnt1-Cre/caAlk2 mutants (Figure 3b). In the littermate control tongues spine-like filiform papillae (Figure 3a,c,d, open arrowheads) and flaking of the superficial epithelial cells were profound. In contrast, filiform papillae were less profound and epithelial cell flaking was not noticed on the surface of Wnt1-Cre/caAlk2 mutant tongue (Figure 3b,e,f).

Figure 3.

Further advancement of taste papilla development and formation of early taste buds in E18.5 Wnt1-Cre/caAlk2 mutant and Cre⁻/caAlk2 littermate control tongues. (a–f) SEM images of E18.5 littermate control (a, c, d) and Wnt1-Cre/caAlk2 mutant (b, e, f) tongues. The arrow in a points to the circumvallate papilla. In c–f, arrowheads point to fungiform papillae, open arrowheads in c, d point to filiform papillae. High-magnification images of the left anterior tongue tip (c, e) and a fungiform papilla (d, f) in a littermate control (a, c, d) and Wnt1-Cre/caAlk2 mutant (b, e, f) tongue. Scale bars: 500 μm in a, b; 100 μm in c, e; 10 μm in d, f. (g, i) Littermate control (g) and Wnt1-Cre/caAlk2 mutant (i) tongue epithelial sheets immunostained with pan-taste cell marker K8 (magenta). Arrowheads point to K8+ fungiform taste buds. Scale bars: 200 μm. (h, j) Single-plane laser scanning confocal images of sagittal sections of littermate control (h) and Wnt1-Cre/caAlk2 mutant (j) tongues immunostained with K8 (magenta) and βIII-tubulin (green). Dotted lines in h and j encircle K8+ taste buds; open arrowheads point to βIII-tubulin+ nerve fibers. Scale bars: 20 μm

Fungiform papillae were observed in the oral tongue of E18.5 Wnt1-Cre/caAlk2 mutants (Figure 3b,e,f), however these fungiform papillae (Figure 3e,f, arrowheads) were underdeveloped compared to those of E18.5 littermate controls (Figure 3c,d, arrowheads) as they were similar in appearance to those in E16.5 littermate control (Figure 2k,l, arrowheads) and Wnt1-Cre/caAlk2 (Figure 2m,n) mutant tongues. Further, early taste buds were detected in the tongue epithelial sheets (Figure 3i, arrowheads) and sagittal tongue sections (Figure 3j) of Wnt1-Cre/caAlk2 mutants with a pan taste cell marker Keratin 8 (K8+) as those in littermate control epithelial sheets (Figure 3g, arrowheads) and sagittal sections (Figure 3h). Among the taste buds detected in the E18.5 Wnt1-Cre/caAlk2 mutants, 41% had a typical morphology as those in littermate controls (Figure 3h), while the majority (59% as in Figure 3j) appeared to have an increased height from base to apex with most K8+ cells located apically. Such shape of taste buds was also observed in E17.5 littermate control and Wnt1-Cre/caAlk2 tongues (data not shown). In contrast to the dense innervation of taste buds in the littermate controls (Figure 3h, open arrowhead), βIII-tubulin⁺ nerve fibers were absent in taste buds and dramatically reduced in the mesenchymal core of fungiform papillae in Wnt1-Cre/caAlk2 mutants (Figure 3j, open arrowhead).

Neural crest-derived mesenchyme/connective tissue serves as a scaffold that is important for tissue differentiation and organization (Rinon et al., 2007). To identify the effects of altered ALK2-BMP signaling in the connective tissue, main lingual tissues were examined. Vimentin⁺ stromal cells were extensively distributed in the E18.5 tongue mesenchyme of both littermate controls (Figure 4a) and Wnt1-Cre/caAlk2 mutants (Figure 4b). A dense layer was obvious in the lamina propria—a zone of loose connective tissue under the epithelium, in both littermate controls (Figure 4a and inset, arrowheads) and Wnt1-Cre/caAlk2 mutants (Figure 4b and inset, arrowheads). E-Cadherin⁺ epithelium was progressively thinner along the anteroposterior axis in the dorsal surface of oral tongue of E18.5 Wnt1-Cre/caAlk2 mutants (Figure 4d and inset, arrowhead) compared to that of littermate controls (Figure 4c and inset, arrowhead). Fungiform taste papillae were not as well developed in the E18.5 Wnt1-Cre/caAlk2 mutant tongues (arrow in inset of Figure 4d) as those in the littermate controls (arrow in inset of Figure 4c). Further, Spinelike filiform papillae were not as evident in the Wnt1-Cre/caAlk2 mutant tongues (open arrowhead in inset of Figure 4d) as that of littermate control tongues (open arrowhead in inset of Figure 4c). Defects in muscle differentiation and organization were examined using muscle specific marker Desmin. All four intrinsic glossal muscles were disorganized in the E18.5 Wnt1-Cre/caAlk2 mutants (Figure 4f,f₁) compared to littermate controls (Figure 4e,e₁). Superior and inferior longitudinal muscles that are close to the dorsal and ventral tongue surface (Figure 4e₁, sLG, iLG) were not apparent in the Wnt1-Cre/caAlk2 mutants (Figure 4f₁). Transverse muscles (Figure 4e₁, Tr) were sparse and largely empty in the space for their distribution in Wnt1-Cre/caAlk2 mutants (Figure 4f₁, Tr). In contrast, vertical muscles (Figure 4e₁, Ve) were increased in bundle size in the middle region of the tongue in Wnt1-Cre/caAlk2 mutants (Figure 4f, Ve).

Figure 4.

Representative photomicrographs of sagittal sections of E18.5 Wnt1-Cre/caAlk2 mutant and Cre⁻/caAlk2 littermate control tongues. (a–f) Single-plane laser scanning confocal images of littermate control (a, c, e) and Wnt1-Cre/caAlk2 mutant (b, d, f) tongues immunostained with mesenchymal cell marker Vimentin (a and b, green), epithelial cell marker E-Cadherin (c and d, green), or muscle cell marker Desmin (e and f, green). Insets in a–d show representative high-magnification images. Arrowheads in a and b point to the dense stromal cell layer under the epithelium. Dashed lines in insets of a and b separate the epithelium from the underlying mesenchyme. Arrows in insets of c and d point to fungiform papillae. Arrowheads in c and d point to the epithelium. Open arrowheads in insets of c and d point to filiform papillae. e₁. and f₁ are high-magnification images of e and f. sLG, superior longitudinal muscle; iLG, inferior longitudinal muscle; Ve, vertical muscle; Tr, transverse muscle. Scale bars: 100 μm in a–f; 50 μm in e₁, f₁

2.4 |. Tongue primordia (i.e., branchial arches) developed deficiently in E10.5 Wnt1-Cre/caAlk2 mutants

To understand the cause of microglossia in Wnt1-Cre/caAlk2 mutants, the development of tongue primordia (i.e., BAs 1–4) was examined. Wnt1-Cre marked RFP signals were distributed in BAs 1–4 (Figure 5a₁, arrowheads) at E10.5. RFP⁺ cells were under E-cadherin⁺ epithelium and exclusively present in the mesenchyme (Figure 5a₂–a₆).

Figure 5.

Representative photomicrographs of branchial arches (BAs) of E10.5 Wnt1-Cre/RFP, Wnt1-Cre/caAlk2, and Cre⁻/caAlk2 littermate controls. (a) Whole mount image of an E10.5 Wnt1-Cre/RFP embryo. Numerals 1–4 in a show individual BAs. a₂-a₆ are single-plane laser scanning confocal images of Wnt1-Cre/RFP BAs immunostained with E-Cadherin (green). a₃-a₆ are high-magnification images of individual BAs in a₂. Scale bars: 400 μm in a₁; 100 μm in a₂; 50 μm in a₃-a₆. (b) Western blot bands (left) of p-Smad1/5/8, p-Smad2/3, and GAPDH in the mesenchyme of BAs 1–4 of littermate controls and Wnt1-Cre/caAlk2 mutants. Histogram (right) represents the intensities of Western blot bands of p-Smad1/5/8, p-Smad2/3 as means ± standard deviation (X ± SD; n = 3). **p ≤ .01 Student’s t test compared to littermate control group. (c) Single-plane laser scanning confocal images of littermate control (c₁-c₃) and Wnt1-Cre/caAlk2 (c₄-c₆) BA transverse sections immunostained with p-Smad1/5/8 (green). Scale bars: 50 μm. (d) Histograms (X ± SD; n = 3) represent the total number of p-Smad1/5/8+ cells per mm2 (top) and total number of mesenchymal cells per mm2 (bottom) in individual BAs of Wnt1-Cre/caAlk2 and littermate controls. *p ≤ .05, **p ≤ .01 compared to littermate control BAs using two-way ANOVA followed by Fisher’s least significant difference (LSD) analyses. (e, f) SEM images to illustrate the side (e) and dorsal (f) view of BAs of a littermate control (e₁ f₁ and Wnt1-Cre/caAlk2 mutant (e₂, f₂). Numerals 1–4 in e and f mark individual BAs 1–4. Arrowheads in f point to the dorsal swelling of BA 1. Arrows in f point to the copula in BA 2. Scale bars: 100 μm

An elevated activation of ALK2-BMP signaling was validated by detecting the phosphorylation of Smads. Phosphorylated (p)-Smad1/5/8 known to mediate BMP signaling was detected at a higher level in BAs of E10.5 Wnt1-Cre/caAlk2 mutants compared to Cre⁻/caAlk2 littermate controls (595.02 ± 17.56 vs. 155.29 ± 6.26 in mean gray scale intensity, p < .01 in Figure 5b). In contrast, the level of p-Smad2/3, that is, known to mediate TGF-β signaling (Grönroos et al., 2012) was not apparently altered in Wnt1-Cre/caAlk2 mutants (82.56 ± 11.65 vs. 71.98 ± 4.49 in mean gray scale intensity, p > .05 in Figure 5b). On sections, mesenchymal cells that were positive for p-Smad1/5/8 immunosignals were significantly increased (p < .01; Figure 5d) in each of BAs 1–4 of E10.5 Wnt1-Cre/caAlk2 mutants (Figure 5c₄–c₆,d) compared to littermate controls (Figure 5c₁–c₃,d). Further, an alteration of mesenchymal cell density in BA 1 (increase) and BA 4 (decrease) was observed in the E10.5 Wnt1-Cre/caAlk2 mutants (p < .05; Figure 5d).

All four BAs (n = 18) were developed, and from the side view of BAs 1–4 there were no obvious differences between Wnt1-Cre/caAlk2 mutants (Figure 5e₂) and littermate controls (Figure 5e₁). However, dorsal view showed that BA 1 had a flat dorsal surface in E10.5 Wnt1-Cre/caAlk2 mutants (Figure 5f₂, arrowhead) in contrast to the profound swelling in littermate controls (Figure 5f₁, arrowhead). An enlarged copula was observed in Wnt1-Cre/caAlk2 mutant BA 2 (Figure 5f₂, arrow) compared to that of the Cre⁻/caAlk2 littermate control (Figure 5f₁, arrow).

2.5 |. BAs 1–4 exhibited distinct gene expression and changes in Wnt1-Cre/caAlk2 mutants

To understand the downstream factors that are mediating the tongue region-specific effects of elevated ALK2-BMP signaling, individual BAs were separated under a microscope at the clear boundaries between BAs, and transcripts of BMP target genes for cell growth (Id1, Snai1, Snai2, and Runx2) were analyzed in E10.5 C57BL/6J wild type (WT), Wnt1-Cre/caAlk2 mutants and Cre⁻/caAlk2 littermate controls. Overall, individual BAs showed different gene expression levels and alterations (Figure 6a). Out of the four tested BMP target genes, levels of Snai2 and Runx2 were significantly different (p < .01) among BAs 1–4 in E10.5 C57BL/6J WT embryos (Figure 6a). No significant changes in Id1 and Snai1 expression levels were observed within any of the E10.5 WT BAs. In contrast to those in littermate controls, gene expression levels were altered in Wnt1-Cre/caAlk2 mutant BAs (Figure 6b). Specifically, compared to Cre⁻/caAlk2 littermate controls, Wnt1-Cre/caAlk2 mutants had: (a) no detected level of Snai1 and Snai2 expression in BA 1 of Wnt1-Cre/caAlk2 mutants; (b) increased Snai1 and Snai2, and decreased Runx2 in BA 2 (p < .05; Figure 6b); (c) significantly higher mRNA levels of all the BMP target genes tested in BA 4 (p < .05); (d) no alterations of gene expression in BA 3 (Figure 6b; p > .05).

Figure 6.

Quantitative RT-PCR of bone morphogenetic protein (BMP) target genes in branchial arches of E10.5 C57BL/6J wild type (WT, a); Wnt1-Cre/caAlk2 mutant, Cre⁻/caAlk2 littermate control and C57BL/6J WT/caAlk2 (b) embryos. Histograms represent means ± standard deviation (X ± SD) of 2−ΔCT values of examined BMP target genes (Id1, Snai1, Snai2, and Runx2) in individual branchial arches (BAs) of C57BL/6J WT (a; n = 6), C57BL/6J WT/caAlk2 (black bars in b; n = 3), Cre⁻/caAlk2 littermate control (blue bars in b; n = 3) and Wnt1-Cre/caAlk2 (red bars in b; n = 3) mice. *p ≤ .05 (a) one-way ANOVA followed by Fisher’s least significant difference (LSD) analyses. *p ≤ .05, **p ≤ .01 (b) compared to corresponding BAs of Cre⁻/caAlk2 littermate control using two-way ANOVA followed by Fisher’s LSD analyses

Of note, differences of BMP target gene expression in E10.5 BAs were detected between C57BL/6J WT and Cre⁻/caAlk2 littermate controls. To understand whether the discrepancies were caused by the transgene insertion, C57BL/6J WT were bred with caAlk2 to generate caAlk2 heterozygotes (C57BL/6J WT/caAlk2) for analyses of the gene expression in BAs. Of all the BMP target genes tested, the expression patterns in C57BL/6J WT/caAlk2 BAs (Figure 6b) resembled those in Cre⁻/caAlk2 littermates (Figure 6b) and were different from those in C57BL/6J WT controls (Figure 6a).

To understand the relevance of cell proliferation and apoptosis with the posteriorly truncated microglossia in Wnt1-Cre/caAlk2 mutants, transcripts of proliferation relevant gene Cyclin-D1 and key apoptosis gene p53 in individual BAs were analyzed in E10.5 Wnt1-Cre/caAlk2 mutants and Cre⁻/caAlk2 littermate controls (Figure 7a). Increased gene expression of Cyclin-D1 was detected in BAs 1 (p < .05; Figure 7a) and 2 (p < .01; Figure 7a), whereas p53 expression levels were higher in BAs 1 (p < .01) and 4 (p < .05) of Wnt1-Cre/caAlk2 mutants compared to the corresponding Cre⁻/caAlk2 littermate control BAs (Figure 7a). No significant changes in Cyclin-D1 expression were observed in BA 3–4 in E10.5 Wnt1-Cre/caAlk2 mutants (p > .05; Figure 7a) compared to Cre⁻/caAlk2 littermate controls (Figure 7a). Similarly, no significant changes in p53 expression were observed in BA 2–3 in E10.5 Wnt1-Cre/caAlk2 mutants (p > .05; Figure 7a) compared to Cre⁻/caAlk2 littermate controls (Figure 7a). Ki67+ proliferating cells were more densely distributed per unit area (mm2) in BA 1 and 2 of E10.5 Wnt1-Cre/caAlk2 mutants (Figure 7b₂,b₄, p < .05 in d) than that of Cre⁻/caAlk2 littermate controls (Figure 7b₁,b₃,d). In contrast, the density of Ki67+ proliferating cells was decreased in BA 4 of Wnt1-Cre/caAlk2 mutants (Figure 7b₆, p < .05 in d) compared to the Cre⁻/caAlk2 littermate control BA 4 (Figure 7b₅,d). No obvious changes in Ki67+ cells were detected in the BA 3 of Wnt1-Cre/caAlk2 mutants (Figure 7b₆, p > .05 in d) compared to those of Cre⁻/caAlk2 littermate controls (Figure 7b₅,d). Further, mesenchymal cells marked by immunosignals of apoptotic cell marker cleaved (c) Caspase-3 were significantly higher in E10.5 BA 1, 3, and 4 of Wnt1-Cre/caAlk2 mutants (Figure 7c₂,c₆, p < .05 in e) than corresponding BAs of Cre⁻/caAlk2 littermate controls (Figure 7c₁,c₅,e). However, no significant changes in cCaspase-3+ cells was observed in Wnt1-Cre/caAlk2 mutant BA 2 (Figure 7c₄, p > .05 in e) compared to the Cre⁻/caAlk2 littermate control BA 2 (Figure 7c₃,e).

Figure 7.

Effects of elevated ALK2-BMP signaling on cell proliferation and apoptosis in E10.5 branchial arches (BAs). (a) Histograms (X ± SD; n = 3) of 2−ΔCT values of cell proliferation associated gene Cyclin-D1 and apoptosis associated gene p53 in individual BAs of Wnt1-Cre/caAlk2 and Cre⁻/caAlk2 littermate control. *p ≤ .05, **p ≤ .01 compared to the corresponding BA of littermate control using two-way ANOVA followed by Fisher’s least significant difference (LSD) analyses. (b, c) Single-plane laser scanning confocal images of transverse BA sections immunostained with cell proliferation marker Ki67 (b, magenta) or apoptosis marker cleaved (c) Caspase-3 (c, green) in littermate control (b₁, b₃, b₅/c₁, c₃, c₅) and Wnt1-Cre/caAlk2 (b₂, b₄, b₆/c₂, c₄, c₆) mice. Scale bars: 50 μm. (d, e) Histograms (X ± SD; n = 3) of the total number of Ki67+ cells per mm2 (d) or cCaspase-3+ cells per mm2 (e) in individual BAs of littermate controls and Wnt1-Cre/caAlk2 mutants. *p ≤ .05, **p ≤ .01 compared to the corresponding BA of littermate control using two-way ANOVA followed by Fisher’s LSD analyses

3 |. DISCUSSION

Our present study demonstrated that elevated ALK2-BMP signaling in neural crest and derived mesenchyme leads to posteriorly truncated microglossia and disorganized lingual tissues. The more severe tongue deformity along the anteroposterior axis and distinct alterations of gene expression and cell behavior in BAs 1–4 of Wnt1-Cre/caAlk2 mutants indicate that neural crest-derived mesenchymal cells in BAs 1–4 respond to elevated ALK2-BMP signaling in a region-specific manner. Overall, our results indicated that appropriate levels of ALK2-BMP signaling activity in neural crest-derived mesenchyme plays essential roles in coordinating various tissue and cell types for the proper formation of the tongue organ and its epithelial appendages (e.g., taste papillae and taste buds). The phenotype is severe and different from the mild defect resulted from the elevated ALK3-BMP signaling activity (Li et al., 2013), which indicates distinct roles of specific receptor-mediated BMP signaling in the tongue organogenesis.

3.1 |. Appropriate levels of ALK2-BMP signaling activity in neural crest and the derived lingual mesenchyme is required for proper tongue organogenesis

The neural crest is a multipotent cell population derived from the lateral ridges of the neural plate in early vertebrate embryos (Leikola, 1976; Trainor, 2015). In mice, cranial neural crest cells populate directly under the epithelium of the tongue primordium (branchial arch 1) as early as 1–2 somite stages (E8.0–8.5; Serbedzija, Bronner-Fraser, & Fraser, 1992). Neural crest-derived cells are extensively distributed in the mesenchyme/connective tissue of the tongue (Boggs et al., 2016; Liu, Komatsu, et al., 2012; Thirumangalathu et al., 2009) and are regarded as scaffolds of the tongue organ (Chai & Maxson, 2006; Cordero et al., 2011). Structural and molecular disruption of neural crest and neural crest-derived cells causes severe defects in tongue formation. For instance, lacking primary cilia in neural crest cells causes aglossia (Millington et al., 2017). Deficient molecular signaling of Wnt (Liu et al., 2009; Liu, Grosse, et al., 2012; Zhong, Zhao, Mayo, & Chai, 2015; Zhu et al., 2017) and TGF-β (Iwata et al., 2013) result in microglossia due to defects in cells migrating into and surviving in the tongue anlage. Our study in demonstration of the importance of ALK2-BMP signaling in neural crest and derived mesenchymal cells in the tongue formation suggest that organogenesis of the tongue involves a complex network of molecular signaling pathways. Detection of an increased p-Smad1/5/8 and unchanged p-Smad2/3 level in the E10.5 Wnt1-Cre/caAlk2 mutant BAs confirmed that the intracellular cascades of ALK2-BMP signaling were through p-Smad1/5/8 that is separate from TGF-β signaling that utilizes p-Smad2/3 (Grönroos et al., 2012). Moreover, our observations of an uneven severity of tongue development in Wnt1-Cre/caAlk2 mutants suggest that tongue formation is finely tuned in a region-specific manner.

To understand the progressively increased severity of tongue deformity along the anteroposterior axis, we detected expression levels of genes in individual BAs including BMP target genes associated with cell growth (Id1, Snai1, Snai2, and Runx2), cell proliferation regulating gene (Cyclin-D1) and apoptosis control gene (p53) in Wnt1-Cre/caAlk2 mutant, Cre⁻/caAlk2 littermate control and C57BL/6J wild type (WT) mice. Individual BAs had significantly different expression levels of BMP target genes Snai2 and Runx2 in E10.5 C57BL/6J WT embryos suggesting distinct levels of BMP signaling activity. Moreover, alterations of gene expression levels in each BA of Wnt1-Cre/caAlk2 mutants were different suggesting that individual BAs respond to elevated ALK2-BMP signaling distinctly. BA 4 was most responsive such that all tested BMP target genes were altered. Ultimately, a significantly lower cell proliferation in BA 4 and higher apoptosis in BA 3 and 4 (Pietsch, Sykes, McMahon, & Murphy, 2008) might be the cause of the decreased cell number in BA 4 mesenchyme and absence of pharyngeal tongue since BAs 3–4 are the main contributor to the formation of pharyngeal tongue (Cobourne et al., 2018; Parada, Han, & Chai, 2012). Differences of BMP target gene expression pattern in E10.5 BAs were detected between C57BL/6J WT and Cre⁻/caAlk2 littermate controls. This was confirmed by the data from an independent experiment in caAlk2 hemizygous embryos through crossing with C57BL/6J WT breeders. Although we do not have evidence-based explanations, this discrepancy of gene expressions might be caused by (a) the genetic background difference—since caAlk2 mice are maintained mixture of 129S6 and C57BL/6J, and (b) alterations in gene expressions by transgene insertion, which is not uncommon. However, these Cre⁻/caAlk2 littermate controls did not exhibit phenotypic alterations demonstrating that the phenotypes we observed in Wnt1-Cre/caAlk2 mutants were due to the activation of caAlk2.

We observed a significantly higher level of p53 transcripts and more apoptotic cells in BA 1 of Wnt1-Cre/caAlk2 mutants, however, the cell proliferation regulating gene Cyclin-D1 (Yang, Hitomi, & Stacey, 2006) as well as cell proliferation were increased in BA 1–2. A promoted cell proliferation in the mutant BA 1 and 2 may overwrite the cell death and be responsible for the formation of the small oral tongue to which BA 1 and a portion of BA 2 contribute (Cobourne et al., 2018; Parada et al., 2012). This is supported by the increased cell count in the mesenchyme of mutant BA1.

Interestingly, we noticed that even though Wnt1-Cre labeled most, if not all, of the cells in the mesenchyme of BAs 1–4, only a population of mesenchymal cells had promoted cell proliferation or apoptosis. The data indicate that the mesenchymal cells in BAs responded to elevated ALK2-BMP signaling activity distinctly.

In summary, our data showed that expression levels of BMP target genes are distinct among individual BAs and that neural crest and the derived mesenchymal cells in individual BAs respond differentially in gene expression and cell behaviors to the constitutive activation of ALK2-BMP signaling. The results suggest that distinct levels of ALK2-BMP signaling activity in individual BAs are required for the cell behavior and functions during normal tongue development.

3.2 |. ALK2-BMP signaling regulates the interactions of neural crest-derived mesenchyme with other lingual tissues for their proper development and organization

As aforementioned, the tongue is comprised of multiple tissues that are connected by the mesenchyme/connective tissue arising primarily from neural crest and serving as a scaffold of the organ (Rinon et al., 2007). The neural crest-derived tongue mesenchyme compartmentalizes lingual muscles and allows nerves and blood vessels to traverse through. In the present study, we provide evidence for an important role of mesenchymal ALK2-BMP signaling in regulating the development of multiple tissues in the tongue. Constitutive activation of ALK2-BMP signaling in tongue mesenchyme led to altered orientation of muscle fibers, reduced branching and growth of nerves, thinner epithelium and arrest of further development of fungiform papillae and taste buds at late embryonic stages.

During the initial formation of the tongue organ, neural crest and neural crest-derived cells migrate into the tongue primordium before myogenic progenitors (Parada et al., 2012), and it has been speculated that neural crest-derived cells in the tongue mesenchyme acts as a niche that release molecular instructions to direct survival, proliferation, and differentiation of myogenic progenitors as well as the patterning of the muscle (Parada et al., 2012; Parada & Chai, 2015). In the present study, we observed missing of muscle fibers at certain orientations (e.g., longitudinal and transverse muscle fibers) in Wnt1-Cre/caAlk2 mutant mice suggesting that ALK2-BMP signaling regulate how mesenchymal cells interact and guide myogenic processes and patterning.

In addition to the defects of muscle cells in Wnt1-Cre/caAlk2 mutants, there were also changes in the nerve outgrowth and branching. Although nerve fibers were not tracked to distinguish the gustatory (chorda tympani) from somatosensory (trigeminal) innervating fibers, we observed that Wnt1-Cre/caAlk2 mutant tongues lack βIII-tubulin⁺ nerve fibers in the vicinity of taste papillae and surrounding epithelium at E14.5, within early fungiform taste buds and surrounding epithelium at E18.5. The results indicate that both nerves were affected by the elevated ALK2-BMP signaling in tongue mesenchyme. Glial cells including the myelinating Schwann cells are derived from neural crest (Mayor & Theveneau, 2013). Our results support the idea that ALK2-BMP signaling is involved in regulating the interactions of neural crest-derived stromal and/or Schwann cells with axons for proper nerve outgrowth, pathfinding, and/or branching.

The development of epithelial appendages requires the interactions with the underlying mesenchyme (Thesleff, 2003). Both direct (Ribatti & Santoiemma, 2014) and indirect (Jussila & Thesleff, 2012) cell–cell interactions between mesenchymal and epithelial cells have been reported. Multiple signaling pathways are involved in the mesenchymal–epithelial interactions (Puthiyaveetil, Kota, Chakkarayan, Chakkarayan, & Thodiyil, 2016), among which BMP signaling plays important roles (Merrill, Eames, Weston, Heath, & Schneider, 2008). Profound effects of BMPs (2, 4, and 7) and antagonists (noggin and follistatin) in taste papilla formation and patterning have been reported (Beites et al., 2009; Zhou et al., 2006). Here, we provide evidence that ALK2 is a mediating receptor of BMP signaling in lingual mesenchyme for its interactions with other tissues.

Moreover, our results showed that Wnt1-Cre/caAlk2 mutants had well-formed fungiform papillae on the small tongue, however at E18.5 tongue epithelium became thinner and further development of fungiform papillae and taste buds were arrested. The data indicate that appropriate level of ALK2-BMP signaling is critical in mesenchymal interactions with the overlying epithelium for epithelial cell proliferation and differentiation, thus advancing fungiform papilla development. It is noteworthy that early taste buds were formed in the less advanced fungiform papillae without innervation, indicating that taste bud formation and cell differentiation are nerve-independent which is also supported by other reports (Castillo et al., 2014; Farbman & Mbiene, 1991; Fritzsch, Sarai, Barbacid, & Silos-Santiago, 1997; Ito, Nosrat, & Nosrat, 2010; Mbiene et al., 1997; Ren et al., 2017). However, lack of innervation might be one of the causes for the arrest of further development of fungiform taste buds observed in E18.5 mutant tongues. Further, our observation that the Wnt1-Cre/caAlk2 mutants had an enlarged Shh+ patch of tissue that was presumably a circumvallate papilla placode at E12.5 but lacked a circumvallate taste papilla at E14.5 indicates that the circumvallate is initially specified but eventually is not sustained. It is possible that the pharyngeal tongue region is important for sustaining the development of the circumvallate papilla.

Collectively, our data provide evidence that ALK2-BMP signaling in neural crest-derived mesenchyme regulates mesenchymal interactions with other lingual tissues including epithelium, muscles, and nerves for their development and organization. Since Type I BMP receptors are present in both tongue epithelium and mesenchyme (Beites et al., 2009; Jung et al., 1999; Kawasaki et al., 2012; Li et al., 2013; Suga et al., 2007; Zhou et al., 2006), further studies are needed to understand the specific roles of each receptor in distinct tissue compartments.

4 |. MATERIALS AND METHODS

4.1 |. Mouse lines and tissue collection

The animal use was approved by the institutional animal care and use protocols of the University of Georgia and in accordance with National Institute of Health guidelines for proper care and use of animals for research. Mice carrying constitutively activated form of Type I BMP receptor Alk2 (CAG-Z-EGFP-caAlk2) transgene (hereafter caAlk2) were provided by Dr. Yuji Mishina, University of Michigan (Fukuda et al., 2006). Heterozygous Wnt1-Cre mice [B6.Cg-Tg (Wnt1-Cre) 11Rth Tg(Wnt1-GAL4) 11Rth/J, Jackson Laboratory, Stock# 009107] were bred with homozygous caAlk2 breeders to generate the Wnt1-Cre/caAlk2 mice. Homozygous RFP mice [B6.Cg-Gt (ROSA)26 Sortm14(CAG-tdTomato)Hze/J, The Jackson Laboratory, Stock# 007914] were used as the reporter to bred with Wnt1-Cre breeders to generate Wnt1-Cre/RFP mice. Wild type (C57BL/6J, Jackson Laboratory, Stock# 000664) mice were used to cross with homozygous caAlk2 breeders to generate the C57BL/6J WT/caAlk2 mice.

Littermates that were negative for Wnt1-Cre were used as controls (Cre⁻/caAlk2). Genotype of mice was confirmed using specific primers. PCR product of Cre (200 bp) were identified using primers CreF (5′-CTC GTG ATC TGC AAC TCC AGTC-3′) and CreR (5′-GAG ACT AGT GAG ACG TGC TACT-3′). PCR products of caAlk2 (580 bp) were identified using primers TF41 (5′-GTG CTG GTT ATT GTG CTG TCTC-3′) and TF61 (5′-TGT GAG CGA GTA GTA ACA ACC-3′). PCR product of RFP (200 bp) were identified using primers oIMR9103 (5′-GGC ATT AAA GCA GCG TAT CC-3′) and oIMR9105 (5′-CTG TTC CTG TAC GGC ATG G-3′).

Timed pregnant mice were euthanized using CO2 followed by cervical dislocation for embryonic tissue collection. Embryos were harvested at E10.5, E12.5, E14.5, E16.5, and E18.5. Noon of the day of vaginal plug detection was designated as embryonic day 0.5 (E0.5). All embryo dissections were performed between 12.00 p.m. and 3.00 p.m. for consistency across litters. The uterus with embryos was removed and embryos were retrieved under a microscope. Stages of the embryos were confirmed by the number of somite pairs and development of multiple organs. Dissected tissues were further processed for different analyses as described below.

4.2 |. Whole mount immunohistochemistry

Immunohistochemistry against Sonic Hedgehog (Shh), a developing papilla marker, was performed on whole tongue on the mandible (E12.5, E14.5) using goat polyclonal, affinity-purified antibody against mouse Shh N-terminal peptide [amino acids 25–198] (AF464, R&D Systems, Minneapolis, MN) as previously described (Liu et al., 2004; Mistretta, Liu, Gaffield, & MacCallum, 2003; Zhou et al., 2006). E12.5 and E14.5 tongues with mandibles were fixed in 4% paraformaldehyde (PFA) in 0.1 M PBS, at 4°C for 2 hr, then transferred to 100% methanol for storage at −20°C until use.

Tongues were treated with 6% H₂O₂ in 100% methanol at room temperature for 5 hr to block endogenous peroxidase activity followed by rehydration through a descending methanol series (50, 25, and 0% in 0.1 M PBS) at room temperature for 30 min each. Antigen retrieval was performed by heating the tissues at 92–95°C for 5 min in Universal Antigen Retrieval Agent (CTS015; R&D Systems, Minneapolis, MN). PBS/MT (0.1 M PBS with 2% skim milk powder and 0.1% Triton X-100) was used to block nonspecific staining. Tongues were then incubated with primary antibody at a 1:300 dilution in 10% normal donkey serum (NDS; D9663; Sigma Aldrich, St Louis, MO) in PBS/MT at 4°C for 48 hr. After rinses with PBS/MT (five times, 1 hr each on ice) tissues were incubated at 4°C for 24 hr with biotin-conjugated rabbit anti-goat secondary antibody (1:500, BA-5000; Vector Laboratories, Burlingame, CA) in 1% NDS in PBS/MT. Following the rinses with PBS/MT (five times, 1 hr each on ice), tongues were incubated at 4°C for 24 hr with peroxidase-conjugated streptavidin in blocking solution (1:500, PK6200; Vector Laboratories, Burlingame, CA). After five rinses in PBS/MT and two rinses in PBT (0.1 M PBS with 0.1% Triton X-100 and 0.2% bovine serum albumin) for 1 hr each on ice, tongues were preincubated in nickel-intensified DAB solution (SK4100; Vector Laboratories, Burlingame, CA) without H₂O₂ at room temperature for 30 min followed by incubation with DAB solution containing 0.0003% H₂O₂. The reaction was stopped by PBS rinses twice for 1 hr each. Tongues were then transferred into 4% PFA in 0.1 M PBS at 4°C for further fixation and photographed in PBS.

4.3 |. Immunohistochemistry on sections

Embryos were fixed with 4% PFA in 0.1 M PBS at 4°C for 2 hr, cryoprotected in 30% sucrose at 4°C for 48 hr, then embedded in OCT (#23730571; Fisher Scientific, Waltham, MA) and rapidly frozen. Serial sections were cut at 10 μm in thickness. Sections were air-dried, rehydrated, and blocked with 10% NDS in PBS-X (0.1 M PBS, 0.3% Triton X-100) at room temperature for 1 hr. Primary antibodies listed in Table 1 were diluted in PBS-X containing 1% NDS (carrier solution). Sections incubated with carrier solution without primary antibody were used as negative controls.

TABLE 1.

Primary antibodies used for immunohistochemistry

Primary antibody	Source (catalog number, company)	Dilution
Goat anti Shh	AF464, R&D systems, Minneapolis, MN	1:300
Rabbit anti p-Smad1/5/8	#13820, Cell signaling, Danvers, MA	1:1,000
Rabbit anti p-Smad2/3	#3101, Cell signaling, Danvers, MA	1:1,000
Mouse anti GAPDH	G8795, Sigma Aldrich, St Louis, MO	1:10,000
Rabbit anti βIII-tubulin	ab18207, Abcam, Cambridge, United Kingdom	1:1,000
Chicken anti Vimentin	AB5733, EMD millipore, Burlington, MA	1:1,000
Rabbit anti Desmin	ab15200, Abcam, Cambridge, United Kingdom	1:1,000
Goat anti E-cadherin	AF748, R&D systems, Minneapolis, MN	1:1,000
Rat anti keratin 8	TROMA-I, Developmental Studies Hybrydoma Bank, Iowa city, IA	1:1,000
Sheep anti Ki67	AF7649, R&D systems, Minneapolis, MN	1:500
Rabbit anti cCaspase-3	#9661, Cell signaling, Danvers, MA	1:500

After incubation with primary antibodies at 4°C for 24 hr, sections were washed three times with 0.1 M PBS, and then incubated with secondary antibodies conjugated with Alexa Fluor 488, or 647 (1:500, Jackson immune research, West Grove, PA) in carrier solution at room temperature for 1 hr. After rinsing in PBS, sections were counterstained with DAPI solution (200 ng/ml in PBS, D1306; Life Technologies, Carlsbad, CA) at room temperature for 10 min. After thorough rinsing in PBS, the sections were air-dried and mounted with ProLong® Diamond antifade medium (P36970; Fisher Scientific, Waltham, MA).

4.4 |. Early taste bud labeling with keratin 8 (K8) immunosignals on epithelial sheets

Peeling of epithelial sheets from E18.5 tongues of Wnt1-Cre/caAlk2 mutants and Cre⁻/caAlk2 littermate controls was performed as previously described (Venkatesan, Boggs, & Liu, 2016). A mixture of collagenase A (1 mg/ml, #10103578001; Roche Diagnostics, Basal, Switzerland) and dispase II (2.5 mg/ml; # 10374300; Roche Diagnostics, Basal, Switzerland) in 0.1 M PBS was injected into the subepithelial space of E18.5 Wnt1-Cre/caAlk2 mutant and littermate control tongues. After the injection, tongues were incubated at 37° C for 30 min followed by fixation in 4% PFA in 0.1 M PBS at room temperature for 1 hr. The epithelial sheet of the tongue was separated from the lamina propria and then rinsed three times in 0.1 M PBS at room temperature for 30 min each.

Following the incubation with 10% NDS in PBS-X at room temperature for 1 hr to block nonspecific staining, the epithelial sheets were incubated with primary antibody against K8 in carrier solution at 4°C for 48 hr. After three rinses in 0.1 M PBS at room temperature for 30 min each, epithelial sheets were incubated at 4°C for 24 hr with Alexa Fluor 647-conjugated donkey anti-rat secondary antibody (1:500, #712-605-150; Jackson ImmuneResearch, West Grove, PA) in carrier solution. The epithelial sheets were rinsed with 0.1 M PBS and transformed into 4% PFA in 0.1 M PBS at 4°C for further fixation and photographed in PBS.

4.5 |. RNA extraction and quantitative reverse transcriptase polymerase chain reaction (RT-PCR)

Individual branchial arches (BAs) (1–4) from Wnt1-Cre/caAlk2 mutants and Cre⁻/caAlk2 littermate controls were dissected at E10.5 and immersed in Trizol (#15596018; Life technologies, Carlsbad, CA) solution for RNA extraction. A total of nine tissues (pooled three tissues × three replicates) were used from each BA for RNA extraction. Tissues were homogenized with a PorGen 700D tissue homogenizer (Fisher Scientific, Waltham, MA) and RNA was extracted using the RNeasy Plus kit (#74136; Qiagen, Hilden, Germany). RNA concentrations were determined using Nanodrop 8000 spectrophotometer (Nanodrop, Thermo Scientific, Waltham, MA). Complementary DNA was synthesized from the RNA extracts using SuperScript™ First-Strand Synthesis System (#11902018; Fisher scientific, Waltham, MA). The cDNA products were used to analyze the expression of selected genes using the primers in Table 2.

TABLE 2.

Primer sequences used for the quantitative reverse transcriptase polymerase chain reaction (qRT-PCR)

Gene	Primer sequence
Id1	Forward 5’ GAG TCT GAA GTC GGG ACC AC 3’ Reverse 5’ CCT CAG CGA CAC AAG ATGC 3’
Snai1	Forward 5’ CTT GTG TCT GCA CGA CCT GT 3’ Reverse 5’ CTT CAC ATC CGA GTG GGT TT 3’
Snai2	Forward 5’ GGC TGC TTC AAG GAC ACA TT 3’ Reverse 5’ TTG GAG CAG TTT TTG CAC TG 3’
Runx2	Forward 5’ AAG AAG AGC CAG GCA GGT GC 3’ Reverse 5’ CAT ACC GAG GGA CAT GCC TGAG 3’
Cyclin D1	Forward 5’ GAC GGC GTC AAA TAT GTC CT 3’ Reverse 5’ CTG GAG AGT GAC AGC ATG GA 3’
p53	Forward 5’ CGG GTG GAA GGA AAT TTG TA 3’ Reverse 5’ TAG CAC TCA GGA GGG TGA GG 3’

4.6 |. Western blotting

Proteins were extracted from the mesenchyme using Radio-immunoprecipitation assay buffer/RIPA buffer (1% NP-40, 150 mmol/L NaCl, 50 mmol/L Tris-HCI, 0.5% sodium deoxycholate, 0.1% SDS, 1 mmol/L EDTA, pH 7.4). The epithelium was removed from E10.5 BAs of Wnt1-Cre/caAlk2 mutant (n = 3) and Cre⁻/caAlk2 littermate control (n = 3) mouse embryos as previously described (Liu, Henson, Zhou, D’Silva, & Mistretta, 2008). Protein concentrations were determined using Pierce™ BCA protein assay kit (#23225; Thermo Fisher Scientific, Waltham, MA) and Synergy™ 4 microplate reader (#7161000; BioTek Instruments, Winooski, VT). Protein extracts were then resolved using sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. Membranes were blocked with 3% bovine serum albumin (A-420–100; Gold Biotechnology, St Louis, MO) in tris-buffered saline with Tween 20 buffer/TBST buffer (20 mM Tris pH 7.5, 150 mM NaCl, 0.1% Tween 20) at room temperature for 1 hr. The membranes were incubated with primary antibodies (Rabbit anti p-Smad1/5/8, rabbit p-Smad2/3, and GAPDH; Table 1) in blocking buffer (3% BSA in TBST) at 4°C for 24 hr. After three rinses in TBST (10 min each), membranes were incubated with Alexa Fluor 647-conjugated goat anti-rabbit (1:10,000, #711-605-152, Jackson Immunoresearch, West Grove, PA) and Alexa Fluor 647-conjugated donkey anti-mouse (1:10,000, #715-605-150, Jackson Immunoresearch, West Grove, PA) secondary antibodies in blocking buffer at room temperature for 2 hr. Band detection was performed using chemiluminescence (ChemiDocMP Imaging System, Bio-Rad ChemiDoc, Hercules, CA).

4.7 |. Scanning electron microscopy

E10.5 BAs and E12.5-E18.5 tongues from Wnt1-Cre/caAlk2 mutants and Cre⁻/caAlk2 littermate controls were fixed in 2.5% glutaraldehyde (#75520; Electron Microscopy Science, Hatfield, PA) and 4% PFA in 0.1 M PBS at room temperature for 24 hr. Tongues were then rinsed in 0.1 M PBS and subsequently postfixed in a sequence of aqueous 1% OsO₄ (#19150; Electron Microscopy Science, Hatfield, PA), 1% tannic acid (#16201; Sigma Aldrich, St. Louis, MO), 1% OsO₄, at 4°C for 1 hr each. Tissues were then dehydrated through an ascending series of ethanol (35, 50, 70, 90, and 100%, three changes at each concentration); and ethanol was displaced by three changes of hexamethyldisilazane (HMDS, #440191; Sigma Aldrich, St. Louis, MO) for 45 min each. Residual HMDS was evaporated slowly in a fume hood until the tissues were completely dry. Tissues were mounted onto specimen stubs, lightly sputter-coated with gold/palladium (Leica Gold/Carbon coater; Georgia Electron Microscope Core Facility, University of Georgia) and imaged using a scanning electron microscope (FEI Teneo FE-SEM; Georgia Electron Microscope Core Facility, University of Georgia).

4.8 |. Photomicroscopy and data analyses

Immunostained whole mount tissues were imaged under a stereomicroscope (Olympus SZX16). Sections were examined thoroughly under a fluorescent light microscope (EVOS FL, Life Technologies). Colocalization of immunosignals was confirmed and photographed using a laser-scanning confocal microscope (Zeiss LSM 710, Biomedical Microscopy Core at the University of Georgia). Adobe Photoshop (2015) was used for Figure assembling and image editing was minimal to improve clarity.

Quantitative analyses were made to obtain the number of p-Smad1/5/8+, cleaved (c) Caspase-3+, Ki67+, and DAPI+ total mesenchymal cells per unit area (mm2) on BA sections. Immunoreacted serial transverse sections of Cre⁻/caAlk2 littermate control (n = 3) and Wnt1-Cre/caAlk2 (n = 3) BAs were thoroughly examined under a fluorescent light microscope (EVOS FL, Life Technologies). Single-plane laser scanning confocal photomicrographs were obtained from every other sections to include the representative and corresponding regions of BAs of littermate control and Wnt1-Cre/caAlk2 mutants. Ki67+ and DAPI+ mesenchymal cells were counted in a selected area of BAs, while p-Smad1/5/8+ and cCaspase-3+ cells were quantified in the entire BAs 1–4 on every other sections.

Measurements of the length and width of the tongues were performed using the images of Shh immunostained E12.5 littermate control (n = 3) and Wnt1-Cre/caAlk2 mutant (n = 3) tongues. The distance between the anterior most point and posterior end of Shh+ circumvallate papilla placode was considered as the length. The widest region in the anterior 2/3 of the oral tongue was taken as the width. NIH-ImageJ software was used to measure the length and width of E12.5 tongues, western blot band intensities and the area of individual BAs. For the quantitative RT-PCR, changes of gene expression levels in individual BAs of Wnt1-Cre/caAlk2 and littermate control groups were presented as means ± standard deviation (X ± SD; n = 3) of 2−ΔCT values.

Student’s t test was used for the statistical analysis of the tongue length and width, and intensities of western blot bands of p-Smads between E12.5 Wnt1-Cre/caAlk2 mutants and littermate controls. One-way analysis of variance (ANOVA) followed by Fisher’s least significant difference (LSD) analyses was performed to test the statistical significance of gene expression levels among individual BAs of E10.5 C57BL/6J WT and C57BL/6J WT/caAlk2 embryos. Two-way ANOVA followed by Fisher’s LSD analyses was used to compare the gene expression levels in individual BAs between Wnt1-Cre/caAlk2 and Cre⁻/caAlk2 littermate controls. A p value <.05 was taken as a statistically significant difference.