Table 1.
Significant influences on facial shape at E11.5 (Procrustes ANOVA).
Table 2.
Significant influences on facial shape from E12.5-E14.5 (Procrustes ANOVA).
Fig 1.
Efnb1 mutant embryos have quantitative facial shape effects that mimic CFNS.
(A-F) Facial landmarks identified on representative Efnb1wt (A-B), Efnb1Δ/Y (C-D), and Efnb1+/Δ (E-F) E14.5 specimen surfaces. Scale bar, 1000 μm (G-H) Common facial shape effects of Efnb1Δ/Y (cyan) and Efnb1+/Δ (red) genotypes on facial landmark position, compared to Efnb1wt (black) from the (G) anterior and (H) lateral views. The lengths of these shape difference vectors are magnified three times to allow for easy comparison. Thin black lines are placed for anatomical reference. (I-L) Plots to illustrate facial shape variation of Efnb1Δ/Y (cyan) and Efnb1+/Δ (red) and Efnb1wt (black) genotypes across E12.5 (triangle), E13.5 (square), and E14.5 (circle). (I) Facial shape variation across E12.5–14.5 specimens is illustrated along the first two principal components. (J) A linear relationship exists between facial size and a multivariate summary score of facial shape, which indicates a strong allometric effect across this period of development. (K) The first two principal components of facial shape after accounting for this developmental allometry illustrate a common genotype effect across ages. (L) Facial shape variation of only E14.5 specimens, with 95% confidence intervals, illustrates similarities in the effect of both genotypes compared to control specimens. Number of embryos analyzed is presented in S1 Table.
Table 3.
Age-specific comparisons of the Procrustes distances between the mean shape of affected and control genotypes, after accounting for allometry.
Fig 2.
Post-migratory neural crest cells mosaic for EPHRIN-B1 expression undergo cell segregation in craniofacial primordia.
(A, A’) Immunostaining E11.5 frontal sections for EPHRIN-B1 (magenta) and GFP (green) reveals that Efnb1+XGFP/lox control embryos demonstrate a fine-grained mosaic pattern of XGFP expression, and EPHRIN-B1 expression is strong in the maxillary prominences and (B, B’) the lateral FNP. (C, C’) Efnb1+XGFP/Δ; Sox10-CreTg/0 embryos with EPHRIN-B1 mosaicism specifically in NCCs show dramatic cell segregation in the maxillary prominences and (D, D’) the lateral FNP, indicating that NCCs are capable of undergoing EPHRIN-B1-mediated segregation resulting in aberrant EPHRIN-B1 expression patterns in craniofacial mesenchyme. Scale bars, 200 μm. (E) Distribution of percentage of XGFP positive patches of various sizes over time in the MXP/secondary palate. Means of the size distributions across all sections measured for a given genotype are plotted, error bars represent S.E.M., **P<0.01; ****, P<0.0001 for comparison of each timepoint with the preceding timepoint (F) Distribution of percentage of XGFP positive patches of various sizes over time in the FNP. Means of the size distributions across all sections measured for a given genotype are plotted, error bars represent S.E.M., *, P<0.05; ****, P<0.0001 for comparison of each timepoint with the preceding timepoint. Number of embryos analyzed is presented in S1 Table.
Fig 3.
Craniofacial mesenchyme cell segregation correlates with local dysmorphology in the secondary palate.
(A, A’) Immunostaining E13.5 frontal sections for EPHRIN-B1 (magenta) and GFP (green) reveals that EPHRIN-B1 protein is strongly expressed in the anterior-middle palatal shelves. Evenly distributed and intermixed XGFP expressing cells are apparent in control Efnb1+XGFP/lox embryos. (B, B’) Cell segregation is visible in the palatal shelves of Efnb1+XGFP/Δ embryos as large patches of EPHRIN-B1 and GFP expression in these structures. The palatal shelves are also smaller and dysmorphic, with changes in shape occurring at boundaries between EPHRIN-B1 expressing and non-expressing domains (white arrow). (C, C’) Generation of EPHRIN-B1 mosaicism specifically in neural crest cells using Sox10-Cre results in dramatic cell segregation in Efnb1+XGFP/lox; Sox10-CreTg/0 palatal shelves, which are smaller and dysmorphic, with regions of dysmorphogenesis correlating with EPHRIN-B1 expression boundaries (yellow arrow). (D, D’) EPHRIN-B1 mosaicism in Shox2IresCre-expressing cells results in cell segregation in Efnb1+XGFP/lox; Shox2IresCre/+ palatal shelves. Areas of dysmorphogenesis are visible at the interface between EPHRIN-B1 expression and non-expression domains (blue arrow). (E) Distribution of percentage of XGFP-positive patches of various sizes. Column height represents means of the distributions across all sections measured for a given genotype, error bars represent S.E.M., *, P<0.05.; **, P<0.01; ****, P<.0001. (F) Patch sizes represented as scatterplots. Horizontal bars represent means, and error bars represent S.E.M. ****, P<.0001. Number of embryos analyzed is presented in S1 Table.
Fig 4.
Craniofacial mesenchyme cell segregation correlates with local dysmorphology in the FNP.
(A, A’) Immunostaining of frontal sections of control Efnb1+XGFP/lox embryos at E13.5 for EPHRIN-B1 (magenta) demonstrates strong expression in the LNP lateral to the nasal concha of the anterior frontonasal process (FNP). XGFP (green)-expressing cells are evenly distributed and intermixed with GFP non-expressing cells. (B, B’) In Efnb1+XGFP/Δ embryos with ubiquitous mosaicism for EPHRIN-B1 expression, cell segregation is evident throughout the anterior FNP, and bifurcation of the nasal concha occurs at an aberrant EPHRIN-B1 expression boundary (white arrowhead). (C, C’) Generation of EPHRIN-B1 mosaicism specifically in neural crest cells in Efnb1+XGFP/lox; Sox10-CreTg/0 embryos results in cell segregation visible throughout the anterior FNP and bifurcation of the nasal concha visible at EPHRIN-B1 expression boundaries (yellow arrowhead). (D, D’) Restriction of EPHRIN-B1 mosaicism to post-migratory neural crest cells using Shox2IresCre does not cause cell segregation or dysmorphology in the nasal conchae of the anterior FNP, as Shox2 is not expressed in this region. Scale bars, 200 μm. (E) Distribution of percentage of XGFP-positive patches of various sizes. Column height represents means of the distributions across all sections measured for a given genotype, error bars represent S.E.M., *, P<0.05.; **, P<0.01. (F) Patch sizes represented as scatterplots. Horizontal bars represent means, and error bars represent S.E.M. **, P<0.01; ****, P<.0001. Number of embryos analyzed is presented in S1 Table.
Fig 5.
EPHRIN-B1 mosaicism in neural progenitors produces cell segregation in the brain.
(A, A’) Immunostaining of E13.5 coronal sections for EPHRIN-B1 (magenta) and GFP (green) shows high EPHRIN-B1 expression, with an absence of cell segregation as shown by the fine-grained mosaic pattern of XGFP expression. (B, B’) In Efnb1+XGFP/Δ embryos with ubiquitous mosaicism for EPHRIN-B1 expression, cell segregation is evident throughout the brain as large patches of EPHRIN-B1 and GFP expression. (C, C’) Generation of EPHRIN-B1 mosaicism specifically in neural progenitor cells using Sox1Cre results in dramatic segregation throughout the brain of E13.5 Efnb1+XGFP/Δ; Sox1Cre/+ embryos, visible as large patches of EPHRIN-B1 and GFP expression. (D) Distribution of percentage of XGFP-positive patches of various sizes. Column height represents means of the distributions across all sections measured for a given genotype, error bars represent S.E.M., *, P<0.05.; **P<0.01; ***P<.005 (E) Patch sizes represented as scatterplots. Horizontal bars represent means, and error bars represent S.E.M. ****, P<.0001. Number of embryos analyzed is presented in S1 Table.
Table 4.
Significant influence of facial size but not Efnb1; Sox1-Cre genotype on facial shape at E14.5 (Procrustes ANOVA).
Fig 6.
Disruption of Efnb1 in NCCs results in face shape changes but disruption in brain does not.
(A-D) Genotype-specific facial shape effects are plotted between predicted E14.5 facial shape landmark positions for Efnb1wt (grey points) and Efnb1+/lox; Sox1Cre/+ (orange points) from the (A) anterior and (C) lateral views and between Efnb1wt (grey points), Efnb1+/lox; Sox10-CreTg/0 (orange points), and Efnb1lox/Y; Sox10-CreTg/0 (blue points) from the (B) anterior and (D) lateral views. The lengths of these shape difference vectors are magnified three times to allow for easy comparison of shape effects. Thin black lines are placed for anatomical reference. (E-F) Facial shape variation of indicated genotypes is projected along the first two principal components from Fig 1L for direct comparison of Sox1Cre and Sox10-Cre mediated Efnb1 tissue specific disruption effects with full Efnb1 genotype effects. The large ovals are the 95% confidence intervals from Fig 1L. Number of embryos analyzed is presented in S1 Table.
Table 5.
Significant influences of facial size and Efnb1; Sox10-Cre genotype on facial shape at E14.5 (Procrustes ANOVA).
Table 6.
Procrustes distancesa of E14.5 facial shapes of Efnb1 mutant genotypes using tissue-specific Cre alleles.
Table 7.
Significant influences of facial size and Ephb receptor genotype on facial shape at E14.5.
Fig 7.
Distinct EPHB receptors exhibit additive non-equal quantitative effects on face shape.
A sample of all possible Ephb1, Ephb2, and Ephb3 null allele genotype combinations displays wide facial variation across the first two principal component axes representing allele facial shape variation (95% CIs from Fig 1L) defined by Efnb1wt (black ellipses), Efnb1Δ/Y (cyan ellipses) and Efnb1+/Δ mutant (red ellipses). (A) Ephb null series specimens are colored by total number of null alleles. A subset of these specimens that are homozygous null for only one Ephb gene (B) or two Ephb genes (C) are plotted alongside EphB wt controls and “all null” specimens that are triple Ephb1-/-; Ephb2-/-; Ephb3-/- homozygous mutants. In (B, C), unlisted Ephb genotypes include both +/+ and +/-, but not -/- genotypes. Comparisons of specific genotypes illustrate the influence of homozygous and heterozygous genotypes across Ephb1 (D), Ephb3 (E), and Ephb1; b3 homozygous null specimens (F). Number of embryos analyzed is presented in S1 Table.
Fig 8.
EPHB2 and EPHB3 receptors mediate cell segregation in secondary palatal shelves.
Secondary palatal shelves of E13.5 embryos harboring compound loss of Ephb1-3 receptors in combination with Efnb1+/Δ heterozygosity with specific genotype combinations shown. Immunostaining for EPHRIN-B1 expression (white) and DAPI (blue) is highlighted with a yellow dashed line at high magnification to demarcate cell segregated patches. (A-F) Compound loss of some EphB receptors does not reduce apparent EPHRIN-B1 driven cell segregation, with a relatively small number of large patches of cells observed. (G, G’) Compound loss of EphB2 and EphB3 receptor resulted in smaller patches, with greater intermingling of EPHRIN-B1 positive and negative cells. (H, H’) Loss of all known EPHRIN-B1 receptors (EphB1, EphB2, EphB3) also resulted in loss of cell segregation, but with the persistence of small patches of EPHRIN-B1 negative cells. Scale bars, 100 μm. (I) Distribution of percentage of EPHRIN-B1 negative patches of various sizes. Column height represents means of the distributions across all sections measured for a given genotype, error bars represent S.E.M., **, P<0.01; ****, P<.0001. (J) Patch sizes represented as scatterplots. Horizontal bars represent means, and error bars represent S.E.M. ****, P<.0001. Number of embryos analyzed are presented in S1 Table.
Fig 9.
Model of cell segregation and craniofacial dysmorphology in Efnb1+/- mutant embryos.