Integration of Brain and Skull in Prenatal Mouse Models of Apert and Crouzon Syndromes

Abstract

The brain and skull represent a complex arrangement of integrated anatomical structures composed of various cell and tissue types that maintain structural and functional association throughout development. Morphological integration, a concept developed in vertebrate morphology and evolutionary biology, describes the coordinated variation of functionally and developmentally related traits of organisms. Syndromic craniosynostosis is characterized by distinctive changes in skull morphology and perceptible, though less well studied, changes in brain structure and morphology. Using mouse models for craniosynostosis conditions, our group has precisely defined how unique craniosynostosis causing mutations in fibroblast growth factor receptors affect brain and skull morphology and dysgenesis involving coordinated tissue-specific effects of these mutations. Here we examine integration of brain and skull in two mouse models for craniosynostosis: one carrying the FGFR2c C342Y mutation associated with Pfeiffer and Crouzon syndromes and a mouse model carrying the FGFR2 S252W mutation, one of two mutations responsible for two-thirds of Apert syndrome cases. Using linear distances estimated from three-dimensional coordinates of landmarks acquired from dual modality imaging of skull (high resolution micro-computed tomography and magnetic resonance microscopy) of mice at embryonic day 17.5, we confirm variation in brain and skull morphology in Fgfr2c^C342Y/+ mice, Fgfr2^+/S252W mice, and their unaffected littermates. Mutation-specific variation in neural and cranial tissue notwithstanding, patterns of integration of brain and skull differed only subtly between mice carrying either the FGFR2c C342Y or the FGFR2 S252W mutation and their unaffected littermates. However, statistically significant and substantial differences in morphological integration of brain and skull were revealed between the two mutant mouse models, each maintained on a different strain. Relative to the effects of disease-associated mutations, our results reveal a stronger influence of the background genome on patterns of brain-skull integration and suggest robust genetic, developmental, and evolutionary relationships between neural and skeletal tissues of the head.

Keywords: Apert syndrome; Crouzon syndrome; brain; craniofacial; craniosynostosis; development; morphological integration; skull.

Figure 1

3D reconstructions of high resolution HRμCT images of skull and HRMRM images of brain of E17.5 embryos superimposed to reveal structural associations. (A) Fgfr2^+/+ unaffected littermate of the Apert syndrome mouse model; (B) Fgfr2^+/S252W Apert syndrome mouse model; (C) Fgfr2c^+/+ unaffected littermate of the Crouzon syndrome mouse model; (D) Fgfr2c^C342Y/+ Crouzon syndrome mouse model. Scale bar = 1 mm. For details of image acquisition see Section Materials and Methods.

Figure 2

Relative position of brain landmarks (blue dots) and skull landmarks (red dots) used in analysis as positioned on a superimposition of brain and skull (A, B lateral view with face to the left). Ten skull landmarks shown on cranial HRμCT isosurface revealing landmarks located on the cranial base (S1, S2, S6, S10) that are visible due to large non-mineralized areas between developing cranial vault bones at E17.5 (C, lateral view; D, superioinferior view). Ten brain landmarks shown on a HRMRM 3D image reconstruction. Subcortical landmarks (B1, B2, B3) are shown but ghosted (E, lateral view; F, superioinferior view). Scale bar = 1 mm.

Figure 3

Results of PCA analyses of form based linear distances estimated among landmarks for skull and brain. (A,B) Scatter plots of individual scores based on PCA of skull form (shape + size). (A) Distribution of Fgfr2c^C342Y/+ mutant mice and unaffected littermates (Fgfr2c^+/+) along first and second Principal Components axes (PC1 and PC2) estimated using all unique linear distances among 10 cranial landmarks of each observation, scaled by the observation's geometric mean. (B) Distribution of Fgfr2^+/S252W Apert syndrome mice and unaffected littermates (Fgfr^+/+) along first and second Principal Components axes (PC1 and PC2) estimated using all unique linear distances among 10 cranial landmarks of each observation, scaled by the observation's geometric mean. (C,D) Scatter plots of individual scores based on PCA of brain form (shape + size). (C) Distribution of Fgfr2c^C342Y/+ mutant mice and unaffected littermates (Fgfr2c^+/+) along first and second Principal Components axes (PC1 and PC2) estimated using all unique linear distances among 10 brain landmarks of each observation, scaled by the observation's geometric mean. (D) Distribution of Fgfr2^+/S252W Apert syndrome mice and unaffected littermates (Fgfr^+/+) along first and second Principal Components axes (PC1 and PC2) estimated using all unique linear distances among 10 brain landmarks of each observation, scaled by the observation's geometric mean.

Figure 4

Brain (in blue) and skull (in red) linear distances whose association was statistically significantly different between Fgfr2c^C342Y/+ Crouzon syndrome mice and unaffected littermates at E17.5 pictured on HRMRM and HRμCT reconstruction of a Fgfr2c^+/+ unaffected littermate (A, lateral view; B, superoinferior view). The two brain metrics (BR7 and BR29; see Supplementary Table 2) were in included in 17 of 61 (~28%) of the correlations that were significantly different between Fgfr2c^C342Y/+ Crouzon syndrome mice and unaffected littermates. Scale bar = 1 mm.

Figure 5

Brain (in blue) and skull (in red) linear distances whose association was statistically significantly different between Fgfr2^+/S252W Apert syndrome mice and Fgfr2^+/+ unaffected littermates at E17.5 pictured on HRMRM and HRμCT reconstruction of a Fgfr2^+/+ unaffected littermate (A, lateral view; B, superoinferior view). The two brain metrics (BR29 and BR34; see Supplementary Table 2) were in included in 71 of 139 (~51%) of the correlations that were significantly different between Fgfr2^+/S252W Apert syndrome mice and Fgfr2^+/S252W Apert syndrome mice and Fgfr2^+/+ unaffected littermates. Scale bar = 1 mm.

Figure 6

Brain (in blue) and skull (in red) linear distances whose associations were statistically significantly different between Fgfr2c^C342Y/+ Crouzon syndrome and Fgfr2^+/S252W Apert syndrome mice at E17.5 pictured on HRMRM and HRμCT reconstruction of a Fgfr2c^C342Y/+ Crouzon syndrome mouse (A, lateral view; B, superoinferior view). Pictured are two brain metrics (blue lines, BR15, and BR34) whose correlation with the skull metrics (red lines) were included in in ~30% of significantly different correlations in the two mutant mouse models. Scale bar = 1 mm.

Figure 7

Brain (in blue) and skull (in red) linear distances whose associations were statistically significantly different between Fgfr2c^+/+ and Fgfr2^+/+ mice at E17.5 pictured on HRMRM and HRμCT reconstruction of a Fgfr2c^+/+ mouse (A, lateral view; B, superoinferior view). (A,B) Two brain metrics [blue lines, BR15 (lpol&midcb) and BR34 (rpol&midcb) whose correlation with the skull metrics (red lines) were included in in ~14% of significantly different correlations in the two mutant mouse models]. (C,D) Three brain linear distances representing metrics associated with the corpus callosum BR39 (spcc&gcc), BR18 (obnp&gcc), BR1 (aptc&ac) are shown in blue. These brain metrics were involved in 10% of the total significant differences in correlations with specific skull measures (shown in red). (E,F) Two brain linear distances, BR30 (rpol&aptc) and BR8 (lpol&aptc) (in blue) were involved in an additional 5.60% of brain-skull correlations that were significantly different between CD1 and B6 mice. Skull linear distances are shown in red. The linesets represented in (A–F) represent nearly 30% of the correlations between brain and skull that were significantly different between CD1 and B6 mice. Scale bar = 1 mm.