Craniofacial divergence by distinct prenatal growth patterns in Fgfr2 mutant mice

Abstract

Background: Differences in cranial morphology arise due to changes in fundamental cell processes like migration, proliferation, differentiation and cell death driven by genetic programs. Signaling between fibroblast growth factors (FGFs) and their receptors (FGFRs) affect these processes during head development and mutations in FGFRs result in congenital diseases including FGFR-related craniosynostosis syndromes. Current research in model organisms focuses primarily on how these mutations change cell function local to sutures under the hypothesis that prematurely closing cranial sutures contribute to skull dysmorphogenesis. Though these studies have provided fundamentally important information contributing to the understanding of craniosynostosis conditions, knowledge of changes in cell function local to the sutures leave change in overall three-dimensional cranial morphology largely unexplained. Here we investigate growth of the skull in two inbred mouse models each carrying one of two gain-of-function mutations in FGFR2 on neighboring amino acids (S252W and P253R) that in humans cause Apert syndrome, one of the most severe FGFR-related craniosynostosis syndromes. We examine late embryonic skull development and suture patency in Fgfr2 Apert syndrome mice between embryonic day 17.5 and birth and quantify the effects of these mutations on 3D skull morphology, suture patency and growth.

Results: We show in mice what studies in humans can only infer: specific cranial growth deviations occur prenatally and worsen with time in organisms carrying these FGFR2 mutations. We demonstrate that: 1) distinct skull morphologies of each mutation group are established by E17.5; 2) cranial suture patency patterns differ between mice carrying these mutations and their unaffected littermates; 3) the prenatal skull grows differently in each mutation group; and 4) unique Fgfr2-related cranial morphologies are exacerbated by late embryonic growth patterns.

Conclusions: Our analysis of mutation-driven changes in cranial growth provides a previously missing piece of knowledge necessary for explaining variation in emergent cranial morphologies and may ultimately be helpful in managing human cases carrying these same mutations. This information is critical to the understanding of craniofacial development, disease and evolution and may contribute to the evaluation of incipient therapeutic strategies.

Figure 1

Landmark and suture placement. Placement of anatomical landmarks (A-D) and sutures (E,F) on E17.5 mouse skull. Views are left lateral (A), superior (B), inferior (C), endocranial (D), left lateral (E), and inferior (F). Landmarks (A-D) are color coded by region: Face (red); Base (green); Vault (blue); Palate (orange). Additional landmarks that were used only in analysis of the global skull are shown in black. Sutures (E, F) are color-coded by region [Face (red); Vault (blue); Palate (orange)] and indicated by number: 1,2) Left, right coronal ; 3, 4) Left, right zygomatic-maxillary; 5,6) Left, right premaxillary-maxillary; 7) Intermaxillary; 8) Interpalatine; 9) Inter Premaxillary; 10, 11) Left, right maxillary-palatine; 12, 13) Left, right fronto-maxillary. Sutures 1-6 and 12-13 are bilateral; only left sutures are shown on the left lateral view (E). An additional table provides more complete definitions of the landmarks, as well as identification of the skull region in which each landmark is located [see Additional file 1]. More information on landmark identification and location can be found at: http://getahead.psu.edu/landmarks_new.html.

Figure 2

Skull morphology of Fgfr2^+/P253R and Fgfr2^+/S252W Apert syndrome mice and unaffected littermates at E17.5 and P0. Left lateral views of 3D HRμCT reconstructions of representative mice from our study samples: Fgfr2^+/P253R mutant at E17.5 (A) and P0 (B): unaffected littermate of the P253R model at E17.5 (C) and P0 (D); Fgfr2^+/S252W mutant at E17.5 (E) and P0 (F); unaffected littermate of the S252W model at E17.5 (G) and P0 (H). By P0, midfacial retrusion is more severe and fusion of multiple sutures is apparent.

Figure 3

Principal Components Analysis of form and shape at E17.5. A) Placement of all E17.5 mouse crania on PC1 and PC2 in the skull form space (as estimated by principal component analysis of all possible linear distances among cranial 33 landmarks on all E17.5 mice). B) Placement of all E17.5 mouse crania on PC1 and PC2 in the skull shape space (as estimated by principal component analysis of all possible linear distances of each observation, scaled by the observation’s geometric mean).

Figure 4

Morphological variation at E17.5. Differences in morphology between Fgfr2^+/S252W mice and unaffected littermates at E17.5 (A, lateral view; B, superior view) and P0 (C, lateral view; D, superior view) and between Fgfr2^+/P253R mice and unaffected littermates at E17.5 (E, F) and P0 (G, H). For all views rostral is left, caudal is right. The linear distances pictured are limited to those that differed significantly by ≥ 5% between mice carrying one of the two Fgfr2 mutations and their respective unaffected littermates (using α = 0.10 confidence limits). The magnitude of these differences varies across the skull. Lines represent distances among landmarks that are significantly larger (blue) and significantly smaller (fuchsia) in mutant mice. Thin lines indicate linear distances that are increased/decreased by 5-10% in mice carrying one of the two Fgfr2 mutations while thick lines indicate linear distances that differ by >10% between unaffected and mutant mice. Bone segmented from HRμCT images is shown as partially transparent to better visualize the differences. See supplementary videos for a full 360-degree rotation view of differences in cranial morphology between Fgfr2^+/S252W mice and unaffected littermates at E17.5 [Additional file 3] and P0 [Additional file 5], and differences in cranial morphology between Fgfr2^+/P253R and unaffected littermates at E17.5 [Additional file 4] and P0 [Additional file 6].

Figure 5

Principal Components Analysis of form and shape: E17.5 and P0 data combined. A) Placement of E17.5 and P0 mouse crania on PC1 and PC2 in the skull form space as estimated by principal component analysis of all linear distances among 33 cranial landmarks. B) Placement of the all E17.5 and P0 mouse crania on PC1 and PC2 in the skull shape space (as estimated by principal component analysis of all possible linear distances scaled by the age and genotype-specific geometric sample mean) among cranial 33 landmarks on all E17.5 and P0 mice.

Figure 6

Variation in growth of Apert syndrome models. Differences in growth between Fgfr2^+/S252W and unaffected littermates (A, lateral view; B, superior view) and Fgfr2^+/P253R and unaffected littermates (C, lateral; D, superior). For all views rostral is left, caudal is right. The linear distances pictured are limited to those whose growth from E17.5 to P0 differed significantly by ≥ 5% using α = 0.10 confidence limits. The magnitude of these differences in growth varies across the skull. Linear distances that grew significantly more (blue) and significantly less (white) in mutant mice are shown. Thin lines indicate linear distances whose growth was increased or decreased by 5-10% in mice carrying one of the two Fgfr2 mutations relative to their respective unaffected littermates; thick lines indicate linear distances whose growth differed by >10% between unaffected and mutant mice. Bone segmented from HRμCT images is shown as partially transparent to better visualize the growth differences. See supplementary videos for a full 360-degree rotation of differences in cranial growth between Fgfr2^+/S252W mice and unaffected littermates [Additional file 8] and between Fgfr2^+/P253R and unaffected littermates [Additional file 9].

Figure 7

Mouse cranial in the skull form space. Placement of all mouse crania on PC1 and PC2 in the skull form space estimated by PCA of all unique linear distances among 33 cranial landmarks (on all mice including two hypothetical forms) to show the estimated growth trajectory for each sample as vectors. GDMA revealed growth patterns between mutant mice and unaffected littermates to be statistically different (Figure 6; Additional file 7). Group-specific mean vectors begin at the location of the E17.5 group mean form in the skull form space defined by the PCA and end at the location of the P0 group-specific mean form. Hypothetical form A represents the average Fgfr2^+/+ (S252W) cranial morphology grown using the Fgfr2^+/S252W growth pattern. Hypothetical form B represents the average Fgfr2^+/+ (P253R) cranial morphology grown using the Fgfr2^+/P253R growth pattern. The growth trajectories for the hypothetical forms (ghosted) begin at the group-specific unaffected E17.5 mean and are equivalent in direction and magnitude to the growth trajectories of the mutant mice, ending at the position of the hypothetical forms (see text).

Figure 8

Craniofacial suture patency in Fgfr2^+/S252Wand Fgfr2^+/P253Rmice and unaffected littermates visualized by HRμCT. A) Patent sutures are green, partially patent (closing) sutures are yellow, and sutures that are no longer patent (closed) are blue. Suture identification is along the X-axis while the percent of individuals within a sample showing a patency state is given on the Y-axis. Numeric codes for the sutures: 1, 2 – Left, right coronal; 3, 4 – Left, right zygomatic-maxillary; 5,6 – Left, right premaxillary-maxillary; 7 - Intermaxillary; 8 - Interpalatine; 9 - Inter premaxillary; 10, 11 – Left, right maxillary-palatine; 12,13 – Left, right fronto-maxillary. Suture locations are shown in Figures 1 and 8B. Only coronal suture patency was recorded for the neurocranium as all other neurocranial sutures were patent in all animals at these ages. The internasal suture was patent in all mice. B) Timeline of cranial suture closure based on data from E17.5 and P0 mice. Colors used on trajectories between the observed time points are interpolated. Arrows at the end of trajectories indicate that the status observed at P0 is ongoing. Trajectories without an arrow suggest that future suture patency states could not be interpolated from available data. Suture patency patterns were similar across unaffected littermates so data were pooled.

Figure 9

Relationship of suture patency patterns and craniofacial shape as estimated by PCA. (A) Distribution of all individuals along PC1 and PC2 following Figure 5A. (see Figure 5A and Methods for details of computing the PCA). (B) Distribution of individuals along PC1 and PC2 coded by the average patency of all facial sutures (coronal suture patency is not included in this average); (C-H) Distribution of individuals along PC1 and PC2 coded for patency of: zygomatic-maxillary suture (C); frontomaxillary suture (D); premaxillary-maxillary suture (E); inter-premaxillary suture (F); inter-maxillary suture (G); inter-palatine suture (H).