Examining craniofacial variation among crispant and mutant zebrafish models of human skeletal diseases

Abstract

Genetic diseases affecting the skeletal system present with a wide range of symptoms that make diagnosis and treatment difficult. Genome-wide association and sequencing studies have identified genes linked to human skeletal diseases. Gene editing of zebrafish models allows researchers to further examine the link between genotype and phenotype, with the long-term goal of improving diagnosis and treatment. While current automated tools enable rapid and in-depth phenotyping of the axial skeleton, characterizing the effects of mutations on the craniofacial skeleton has been more challenging. The objective of this study was to evaluate a semi-automated screening tool can be used to quantify craniofacial variations in zebrafish models using four genes that have been associated with human skeletal diseases (meox1, plod2, sost, and wnt16) as test cases. We used traditional landmarks to ground truth our dataset and pseudolandmarks to quantify variation across the 3D cranial skeleton between the groups (somatic crispant, germline mutant, and control fish). The proposed pipeline identified variation between the crispant or mutant fish and control fish for four genes. Variation in phenotypes parallel human craniofacial symptoms for two of the four genes tested. This study demonstrates the potential as well as the limitations of our pipeline as a screening tool to examine multi-dimensional phenotypes associated with the zebrafish craniofacial skeleton.

Keywords: computational anatomy; cranial morphology; geometric morphometrics; klippel-feil syndrome; osteogenesis imperfecta; osteoporosis; sclerosteosis.

FIGURE 1

Groupwise difference plots among datasets. (a) Digitizing accuracy analysis showing PC3 and PC5 from combined manual landmarks of the plod2 dataset, with colors indicating different individuals who landmarked datasets in this study and numbers indicating individual fish. (b–d) PC1 and PC2 from groupwise analyses using different landmarking methods including (b) 21 manually placed landmarks, (c) 308 pseudolandmarks, and (d) 21 MALPACA placed landmarks. Point color indicates gene grouping; shape indicates type of fish. Heat maps show dorsal (top) and lateral (bottom) cranial skeletons, with colors representing Procrustes distances between mean shape and the extreme shapes across each PC. All axis labels show the percentage of variation represented by each PC.

FIGURE 2

Principal component plots from analyses that show separation of plod2 crispants and their control siblings. (a) PC2 and PC3 from pseudolandmarks, and (b) PC1 and PC2 from symmetrical components of shape variation. Point color indicates groupings. Heatmaps represent lateral (left) and dorsal (right) cranial skeletons and show Procrustes distances between mean shape and the extreme shapes across each PC. Axis labels show the percentage of variation represented by each PC.

FIGURE 3

Principal component plots from analyses that show separation of meox1 crispants and their control siblings. (a) PC1 and PC3 from pseudolandmarks, and (b) PC1 and PC3 from symmetrical components of shape variation. Point color indicates groupings. Heatmaps represent lateral (left) and dorsal (right) cranial skeletons and show Procrustes distances between mean shape and the extreme shapes across each PC. Axis labels show the percentage of variation represented by each PC.

FIGURE 4

PC1 and PC2 from symmetrical components of shape variation for wnt16 crispants. Point color indicates groupings. Heatmaps represent lateral (left) and dorsal (right) cranial skeletons and show Procrustes distances between mean shape and the extreme shapes across each PC. Axis labels show the percentage of variation represented by each PC.

FIGURE 5

Principal component plots from manual landmark analyses that show separation of wnt16 mutants and their (a) control and (b) heterozygote siblings. Point color and shape indicate groupings. Heatmaps represent lateral (left) and dorsal (right) cranial skeletons and show Procrustes distances between mean shape and the extreme shapes across each PC. PC. Axis labels show the percentage of variation represented by each PC.

FIGURE 6

Principal component plots from manual landmark analyses that show separation of sost mutants, heterozygotes, and their control siblings. PC1 and PC2 from (a) pseudolandmarks and (b) symmetrical components of shape variation. Point color and shape indicates groupings. Heatmaps represent lateral (left) and dorsal (right) cranial skeletons and show Procrustes distances between mean shape and the extreme shapes across each PC. Axis labels show the percentage of variation represented by each PC.