Gene-ethanol interactions underlie craniofacial variability in a zebrafish model of FASD

Variation is common in human birth defects and this variability is influenced by genes and the environment. How genes and the environment interact in causing birth defect variability is a fundamental question in biology. I sought to investigate these interactions using a zebrafish model of fetal alcohol spectrum disorder. Fetal alcohol spectrum disorder (FASD) is an umbrella term that describes all ethanol-induced fetal defects. It is highly variable and also highly prevalent, with an estimated 1% of the population being affected. FASD can cause variable defects, including those affecting the craniofacial skeleton, a neural crest- and mesoderm-derived structure. Although both timing and dosage can influence FASD variability, genetics is an underlying factor. Little is known of the genes that cause susceptibility to FASD, and so I sought to uncover these genes and the mechanism of their interaction with ethanol. Using a novel zebrafish genetic screen to uncover gene-ethanol interactions, we found a synergistic interaction between platelet-derived growth factor receptor alpha (pdgfra) and ethanol. Pdgfra is a receptor tyrosine kinase involved in cell migration, proliferation, and survival. Untreated pdgfra mutants display cleft palate. In a percentage of ethanol in which wildtype zebrafish are unaffected, pdgfra mutants display exacerbated loss of the entire palatal skeleton. Furthermore, pdgfra heterozygotes display variable craniofacial defects, uncovering latent haploinsufficiency. Genetic analysis of a group of children with and without FASD suggests that this interaction is highly conserved. In zebrafish, further analysis of the mechanism of this pdgfra-ethanol interaction revealed a protective role of pdgfra to ethanol-induced neural crest cell death. Analysis of the Pdgfra downstream pathway PI3K/AKT/mTOR revealed an inhibitory effect of ethanol at the level of mTOR. Together, these data suggest that genes functioning in growth factor signaling could predispose to ethanol-induced defects. Analysis of another growth factor signaling gene, fgf8a, supported this hypothesis. Ethanol interacts with fgf8a to cause posterior mesoderm-derived craniofacial defects. This phenotype can be recapitulated by blocking both fgf8a and fgf3, suggesting ethanol broadly attenuates growth factor signaling. Analysis of the fgf8a;fgf3 phenotype suggests that Fgf signaling is required for proper specification, via hyaluronan synthetase 2, of the mesoderm-derived posterior craniofacial skeleton. To test whether ethanol may also broadly attenuate Pdgf signaling, we analyzed pdgfra;pdgfrb mutant phenotypes and saw a recapitulation of the pdgfra-ethanol interaction. The synergistic pdgfra;pdgfrb phenotype may be due to a similar increase in neural crest cell death observed in the pdgfra-ethanol interaction. Together, these data reveal genes involved in growth factor signaling may act to protect against ethanol-induced craniofacial defects.