Cleft Palate in a Mouse Model of SOX2 Haploinsufficiency

Abstract

Objective

While SEX-determining region Y-Box 2 (SOX2) mutations are typically recognized as yielding ocular and central nervous system abnormalities, they have also been associated with other craniofacial defects. To elucidate the genesis of the latter, Sox2 hypomorphic (Sox2^HYP) mice were examined, with particular attention to secondary palatal development.

Results

Clefts of the secondary palate were found to be highly penetrant in Sox2^HYP mice. The palatal clefting occurred in the absence of mandibular hypoplasia and resulted from delayed or failed shelf elevation.

Conclusions

Sox2 hypomorphism can result in clefting of the secondary palate, an effect that appears to be independent of mandibular hypoplasia and is thus expected to result from an abnormality that is inherent to the palatal shelves and/or their progenitor tissues. Further clinical attention relative to SOX2 mutations as a basis for secondary palatal clefts appears warranted.

Sex-determining region Y-Box 2 (SOX2) is an high mobility group box domain-containing transcription factor that is widely expressed in the developing nervous system and which is involved in a wide array of developmental processes (Kamachi et al., 1998; Schneider et al., 2008; Sarkar and Hochedlinger, 2013). Humans with SOX2 mutations/haploinsufficiency commonly exhibit severe ocular and central nervous system defects, hormone deficiencies, as well as craniofacial abnormalities, with retrognathia and facial asymmetry being reported (Fantes et al., 2003; Kelberman et al., 2006; Zenteno et al., 2006; Schneider et al., 2009). Notably, cleft palate has also been reported, but in only one case (Male et al., 2002). Sox2 hypomorphic (Sox2^HYP) mice were generated to model human SOX2 haploinsufficiency (Taranova et al., 2006). These mice express from 20% to 40% of wild type (WT) SOX2 protein levels and have been previously reported to develop neural tube and ocular defects that are consistent with those observed in SOX2 haploinsufficient humans (Taranova et al., 2006; Langer et al., 2012). The present study characterizes the craniofacial defects in this mouse model.

Materials and Methods

Animals

The generation of Sox2^IR and Sox2^EGFP alleles was previously described (Ellis et al., 2004; Taranova et al., 2006). For the current study, Sox2^IR/+ females were bred to Sox2^EGFP/+ males, generating Sox2^EGFP/IR (Sox2^HYP), Sox2^+/IR, and Sox2^+/+ embryos and fetuses. Mice with the latter two genotypes are phenotypically indistinguishable and were considered controls (Sox2^CONT) (Taranova et al., 2006). Genotyping was performed as described in Langer et al. (2012). All analyses were performed on a CD1 background. The morning of vaginal plug detection was considered embryonic day (E) 0.5, and embryo/fetus staging was based on limb morphology (Kaufman, 1992). Embryos and fetuses were harvested on E13.5 to E16.5 following maternal sacrifice. The described animal work was performed with Institutional Animal Care and Use Committee approval and in accordance with the University of North Carolina at Chapel Hill Division of Laboratory Animal Medicine and the American Psychological Association animal care guidelines.

Tissue Preparation for Whole Mount Palate Analyses

Immediately following collection, the E13.5 to E16.5 embryos/fetuses were fixed in 2.5% glutaraldehyde at 4°C. The mandibles were removed from the heads to permit visualization of the palatal primordium/palate. The remaining portion of the head was incubated for 5 minutes in a solution of 1:750 ethidium bromide and imaged under green fluorescent light.

Analysis of Relative Mandible Length

E16.5 Sox2^CONT and Sox2^HYP fetuses were dissected in chilled phosphate-buffered saline (PBS), cut below the forelimbs, and fixed in room-temperature 10% phosphate-buffered formalin for 2 weeks. The fixed fetuses were placed in a premade mold, enabling the consistent positioning of the sample. The vertical midpoint of the ear was established, and the nasomaxillary and mandibular lengths were measured from this point. To control for slight deviations in the orientation of the fetus and within-litter size variations, the mandibular to nasomaxillary length ratio, rather than absolute measurements, of the two groups were compared. Nine measurements for each metric were averaged, and these averages were used to determine the mandibular to nasomaxillary length ratios. The measurements were performed with ImageJ v. 1.43u software.

Tissue Preparation for In Situ Hybridization

E13.5 to E16.5 embryos and fetuses were fixed at 4°C in a solution of PBS and 4% paraformaldehyde. Following three PBS washes, the embryos were cryoprotected in a sucrose gradient and mounted in optimum cutting temperature mounting medium (O.C.T., Tissue-Tek). For in situ analyses, 20-μm frontal sections were incubated with digoxigenin-labeled probes and visualized using enzymatic detection, following the manufacturer’s protocol (Roche). A probe against Sox2 (a kind gift from Dr. Lovell-Badge) was used to show Sox2 expression, and a probe against Tbx2 was used to mark the palatal shelves for the morphologic analyses (Pontecorvi et al., 2008).

Statistical Analyses

The proportions of Sox2^HYP embryos that exhibited defects of either the left or the right palate were analyzed using a two-tailed chi-square test. The ratios of the mandibular to nasomaxillary lengths of E16.5 Sox2^HYP and Sox2^CONT fetuses were compared using Student’s t test. The data are given as the mean ± SEM. Significance was defined as P < .05.

Results

SOX2 Expression in the Palatal Primordium

In situ staining for Sox2 in the secondary palatal primordium of E13.5 to E16.5 Sox2^CONT mice reveals its expression throughout the secondary palatal epithelium, with stronger expression laterally prior to elevation and ventrally following elevation (arrows in Fig. 1A, B, and ventral palate in C).

FIGURE 1.

SOX2 expression in the palatal epithelium and the gross secondary palatal morphology of Sox2^CONT and Sox2^HYP embryos. A–C: Sox2 expression in the palatal epithelium of E13.5 (A), E14.75 (B), and E16.5 (C) Sox2^CONT embryos. The sections were taken at the mid optic level. SOX2 is expressed throughout the palatal epithelium at all of the examined stages (black arrows). D–F: Morphology of the Sox2^CONT palate at E13.5 (D), E14.75 (E), and E16.5 (F). G–I: Morphology of the Sox2^HYP palate at E13.5 (G), E14.75 (H), and E16.5 (I). At E13.5 (D,G), no clear difference can be observed between Sox2^HYP and Sox2^CONT secondary palates. By E14.75 (H), the palate is cleft and there is a clear defect in the anterior third of the Sox2^HYP palate (red arrowhead). At E16.5 (I), the cleft is broad and asymmetry of the palatal shelves can be observed in a subset of Sox2^HYP embryos (red arrow).

Gross Morphologic Analysis of the Secondary Palate in Sox2^HYP Mice

As expected, the examined embryos exhibited ocular defects that grossly manifested as variably severe microphthalmia. As was observed in whole mounts, the developing secondary palatal shelves of Sox2^HYP embryos were indistinguishable from those of the controls at E13.5 (Fig. 1D versus G). At E14.75, however, when Sox2^CONT palatal shelves have elevated and initiated fusion, the shelves of Sox2^HYP mice are abnormally separated, with consistently unilateral defects in the anterior third of the palatal shelf (Fig. 1E versus H). By E16.5, when fusion is complete in Sox2^CONT fetuses, a broad cleft is apparent in Sox2^HYP embryos (Fig. 1F versus I). In newborns, 67% (8/12) of Sox2^HYP mice exhibit secondary palatal clefting.

Histologic Analysis of the Sox2^HYP Secondary Palate

Frontal (coronal) histologic sections of Sox2^HYP mice also illustrate that at E13.5, the Sox2^HYP palatal shelves are similar to those of Sox2^CONT embryos (Fig. 2A,C versus B,D). At E14.75, all of the mutants that exhibited clefting (5/5) also exhibited a unilateral failure of shelf elevation, either specifically in the anterior region (4/5) or along the length of the secondary palate (1/5), such that the shelf was oriented vertically (Fig. 2E versus F). At E16.5, most of the affected Sox2^HYP embryos exhibited bilaterally elevated palatal shelves (Fig. 2I,K versus J,L). No significant bias was observed with respect to the laterality of the elevation defect (three left, eight right, χ² = 2.3, P = .13).

FIGURE 2.

Histologic analysis of palatal morphology of Sox2^CONT and Sox2^HYP embryos. A–D: Anterior and posterior frontal sections of E13.5 Sox2^CONT embryos (A,C) and Sox2^HYP embryos (B,D). No clear differences were observed between Sox2^HYP and Sox2^CONT palates at this stage. E–H: Anterior and posterior frontal sections of E14.5 Sox2^CONT embryos (E,G) and Sox2^HYP embryos (F,H). In affected Sox2^HYP embryos, one of the palatal shelves is consistently unelevated at this stage (arrowhead in F), and the shelves fail to extend in more posterior regions (arrow in H). I–L: Anterior and posterior frontal sections of E16.5 Sox2^CONT embryos (I,K) and Sox2^HYP embryos (J,L). Both palatal shelves have elevated in Sox2^HYP embryos, but fusion has not occurred. All of the sections were stained with an in situ probe against Tbx2.

Analysis of the Relative Mandibular Length in Sox2^HYP Embryos

SOX2 haploinsufficient humans exhibit retrognathia (Zenteno et al., 2006). Because this defect has been associated with delayed palatal shelf elevation, an analysis was performed of the relative length of the mandible to the nasomaxillary complex, an important metric in the context of the mandibular influence on palatal development (Latham, 1966; Diewert, 1979). The mandibular/maxillary length ratios of Sox2^HYP embryos were observed to be higher than those of Sox2^CONT embryos, indicating that the mandible is not shortened in Sox2^HYP embryos (Sox2^CONT: 0.913 ± 0.004; Sox2^HYP: 0.927 ± 0.004; P = .03) (Fig. 3A versus B). In fact, the data indicate that the mandible is relatively longer in Sox2^HYP embryos, indicating a shortened maxilla or a lengthened mandible. Analysis of the raw mandibular and maxillary length measurements did not reveal significant differences between the Sox2^HYP and the Sox2^CONT embryos with respect to the absolute lengths of these facial structures (data not shown); it can therefore not be concluded whether a shortened maxilla, an elongated mandible, or both, underlies this effect. However, these data indicate that the cleft palate that is observed in Sox2^HYP embryos is unlikely to be due to retrognathia.

FIGURE 3.

Analysis of relative mandible lengths in Sox2^HYP embryos. A,B: E16.5 Sox2^CONT (A) and Sox2^HYP (B) embryos with lines indicating the lengths of the maxilla (top line) and the mandible (bottom line), from which relative measures of mandibular length were calculated. No retrognathia is observed in the Sox2^HYP fetuses.

Discussion

Mutations in the gene for the transcription factor SOX2 result in abnormal development of the brain, eye, gut, and certain craniofacial structures (Kelberman et al., 2006; Zenteno et al., 2006; Schneider et al., 2009). The results of this study indicate that Sox2^HYP mice develop cleft palate at a high penetrance. Sox2^HYP mice have been previously demonstrated to variably exhibit ocular and hypothalamic defects that are consistent with those observed in SOX2 haploinsufficient humans, indicating the faithfulness of these lines as models of the human disorder (Taranova et al., 2006; Langer et al., 2012). Moreover, the human and mouse SOX2 amino acid sequences are highly similar, being identical in the DNA-binding domain (Gubbay et al., 1990; Collignon, 1993; Stevanovic et al., 1994). Although cleft palate has only rarely been reported in association with SOX2 mutations in humans, dental anomalies, including both widely spaced and supernumerary teeth, have also been reported, indicating this gene’s involvement in multiple aspects of normal oral development (Male et al., 2002; Ragge et al., 2005; Numakura et al., 2010).

That few human cases have been observed to exhibit cleft palate may be due to a combination of two factors: (1) the high variability and incomplete penetrance of the phenotypes that are associated with SOX2 mutations, and (2) that cleft palate is not classically associated with this condition (Zenteno et al., 2006; Zhou et al., 2008). SOX2 haploinsufficient humans who exhibit cleft palate in the absence of classically recognized phenotypes would therefore likely remain unidentified as mutation carriers.

Additional support for the role of SOX2 in human palatal development is provided by the presence of overlapping phenotypes associated with SOX2 haploinsufficiency and the CHARGE association, the latter of which includes cleft palate and is most commonly caused by mutations in the gene that encodes CHD7 (Kelberman et al., 2006; Schneider et al., 2009; Zentner et al., 2010). Notably, CHD7 physically interacts with SOX2 to regulate the expression of other genes that are mutated in several human syndromes (Engelen et al., 2011).

For these reasons, the present results support the premise that direct or indirect interference with SOX2 activity should be considered important relative to the genesis of craniofacial malformations, including clefting. Importantly, the addition of palatal clefting to the list of phenotypes that are suggestive of SOX2 mutation may lead to the identification of neural or ocular defects or hormonal deficiencies that may otherwise be overlooked in such patients and which would provide important information with regard to the treatment and/or genetic counseling of these individuals.

Although it is clear from this study that Sox2 hypomorphism-induced palatal clefting can occur in the absence of micro/retrognathia (i.e., is not a result of tongue obstruction as is considered likely in the Pierre Robin anomaly), the developmental basis for the clefting remains unknown. However, considering that Sox2 is expressed in the epithelium of the palatal primordia, failure of the secondary palatal shelves to unite appears likely to be an inherent abnormality of the palatal shelves. Additional experiments to explore this premise are warranted.