Novel Ethanol‐Sensitive Mutants Identified in an F3 Forward Genetic Screen

Background Fetal alcohol spectrum disorders (FASD) collectively refer to all deleterious outcomes due to prenatal alcohol exposures. Alterations to the face are common phenotypes in FASD. While alcohol exposure is the underlying cause of FASD, many variables modify the outcomes of such exposures. Genetic risk is one such variable, yet we still have a limited understanding of the nature of the genetic loci mediating susceptibility to FASD. Methods We employed ENU‐based random mutagenesis in zebrafish to identify mutations that enhanced the teratogenicity of ethanol (EtOH). F3 embryos obtained from 126 inbred F2 families were exposed to 1% EtOH in the medium (approximately 41 mM tissue levels). Zebrafish stained with Alcian Blue and Alizarin Red were screened for qualitative alterations to the craniofacial skeleton between 4 and 7 days postfertilization (dpf). Results In all, we recovered 6 EtOH‐sensitive mutants, 5 from the genetic screen itself and one as a background mutation in one of our wild‐type lines. Each mutant has a unique EtOH‐induced phenotype relative to the other mutant lines. All but 1 mutation appears to be recessive in nature, and only 1 mutant, au29, has apparent craniofacial defects in the absence of EtOH. To validate the genetic screen, we genetically mapped au29 and found that it carries a mutation in a previously uncharacterized gene, si:dkey‐88l16.3. Conclusions The phenotypes of these EtOH‐sensitive mutants differ from those in previous characterizations of gene–EtOH interactions. Thus, each mutant is likely to provide novel insights into EtOH teratogenesis. Given that most of these mutants only have craniofacial defects in the presence of EtOH and our mapping of au29, it is also likely that many of the mutants will be previously uncharacterized. Collectively, our findings point to the importance of unbiased genetic screens in the identification, and eventual characterization, of risk alleles for FASD.

P RENATAL ETHANOL (ETOH) exposure can result in fetal alcohol spectrum disorders (FASD). Phenotypes present in individuals with FASD are highly variable and frequently include alterations to the nervous system and/or face. Numerous variables contribute to the effects of prenatal EtOH exposures, including EtOH dosage, timing of the exposure, and individual genetics (Riley et al., 2011;Streissguth and Dehaene, 1993).
We are gaining an understanding of the genetic risk to FASD. Allelic variants of genes involved in EtOH metabolism have been shown to associate with risk for FASD in humans (Warren and Li, 2005). However, most of our knowledge of the genetic risk for FASD comes from candidate gene studies in animal models. Candidate genes such as members of the Sonic hedgehog, retinoic acid, and nitric oxide pathways have been studied due to the similarities in phenotypes between these pathway mutants and individuals with FAS (Eberhart and Parnell, 2016).
We have previously used available zebrafish craniofacial mutants as candidates to identify and characterize EtOHsensitive mutants (McCarthy et al., 2013;Swartz et al., 2014). Our results have identified EtOH-sensitive mutations that would not have been readily predicted based on phenotypes of individuals with FASD. While we were able to assess the EtOH sensitivity of many mutants, our genetic screens were still biased to genes with known functions in craniofacial development. To fully appreciate the genetics underlying FASD, it is essential to assay gene-EtOH interactions in a more unbiased fashion.
Due to their high fecundity, zebrafish have been used extensively in forward genetic screens to identify genetic pathways mediating development and behavior in an unbiased fashion . Due to the high level of genetic conservation between zebrafish and humans (Postlethwait, 2006;Woods et al., 2005), these screens have provided important insights into human birth defects (Eberhart et al., 2008;Ingham, 2009;Neuhauss et al., 1996;Piotrowski et al., 1996). Importantly, the genetic pathways regulating craniofacial development are exquisitely conserved across vertebrates (Knight and Schilling, 2006). External fertilization makes zebrafish even more useful for screens to identify risk loci for environmental exposures as potential teratogenic agents can be added directly to the fishes' water.
Here, we demonstrate that zebrafish is a powerful model organism for the unbiased identification of EtOH-sensitive mutants. Through a small forward genetic screen, we recovered 5 EtOH-sensitive mutants and an additional background mutant from our wild-type stocks. The phenotypes of the mutants vary widely, and the facial skeleton of the majority of the mutants develops normally in the absence of EtOH. We demonstrate that at least one of these mutants is in a previously uncharacterized gene. Collectively, our genetic screen will provide new insights into the genetic inputs involved in the variability within FASD.

Mutagenized Zebrafish Lines
Zebrafish were mutagenized with ENU according to established procedures for an F3 forward genetic screen (Solnica-Krezel et al., 1994). Eye phenotypes obtained from this screen have previously been described (Lee et al., 2012). The same mutagenized stocks were screened for EtOH-induced facial phenotypes (see Fig. 1 for wildtype morphology). In all, we screened the offspring from 126 inbred F2 families. For both EtOH-treated and control conditions, we screened a minimum clutch size of 50 embryos for our initial assessment. Parents of clutches with apparent EtOH-induced mutations were mated a second time, a minimum of 50 embryos were screened to validate the initial assessment, and lines were established for these validated mutants. All zebrafish were housed and cared for with IACUC approval at The University of Texas at Austin.

Histological Staining and Imaging
We stained cartilage and bone in 4 to 7 days postfertilization (dpf) zebrafish using a previously described Alcian Blue/Alizarin Red staining protocol (Walker and Kimmel, 2007). Fluorescent Alizarin Red images were performed as described (Eames et al., 2013). Whole mount and flat mount images of histological stains were collected on a Zeiss Axio Imager. Fluorescent images were collected on a Zeiss 710 confocal microscope.

Genetic Mapping of au29 and Generation of au113 Mutants
Carriers for au29, derived in an AB genetic background, were crossed to the WIK genetic background to generate a hybrid stock for genetic mapping. Carriers from this hybrid stock were identified, and their mutant offspring were used for SNP mapping. The DNA from 25 mutant offspring was pooled, and whole genome sequencing was performed by the GSAF at The University of Texas at Austin. Sequence reads were assembled and mapped to the zebrafish genome (ZV9) using MegaMapper (Obholzer et al., 2012). MegaMapper calculated SNP frequency and mapping. To validate the au29 allele, we generated a second mutation in si:dkey-88l16.3 via CRISPR/Cas9, using the Alt-R CRISPR system (IDT). A gRNA targeting exon 18 (GAGAGAAGCCAGAGCTGCGC) was complexed with Cas9 and injected into 1-cell stage embryos. P0 fish were backcrossed to wild-type AB fish and screened for germline transmission of deleterious mutations via RFLP using FspI, which would be disrupted in indel-containing fish. The nature of the mutations was determined via Sanger sequencing. We selected a 5-bp deletion for our analyses due to its predicted frame shift.

The au26 Allele Is an EtOH-Sensitive Background Mutation That Disrupts Palatal Development
To initiate our forward genetic screen, we first sought to ensure our wild-type stocks did not harbor background sensitivity to EtOH. We tested both wild-type AB and Tubingen strains. We found that our AB stock is resistant to facial alterations caused by exposure to 1% EtOH in the medium (McCarthy et al., 2013;Swartz et al., 2014). Interestingly, we found a portion of clutches obtained from pair-wise matings of Tubingen fish appeared to have an EtOH-sensitive background mutation (n = 2/11). A mating colony was established from these fish, and through subsequent analyses, we determined that the fish carry a simple recessive mutation causing EtOH sensitivity. We designated this mutant allele au26.
We observe no defects in untreated embryos from incrosses of au26 carriers (not shown). In EtOH-exposed mutants, the overall size of the craniofacial skeleton at 4 dpf appears slightly smaller (Fig. 2). However, the most striking phenotype in au26 mutants is a profound defect to the palatal skeleton. The trabeculae are reduced resulting in the ethmoid plate being partially or completely separated from the posterior neurocranium (Fig. 2, arrow). The percentage of mutant phenotypes across clutches varies from 15 to 25%, suggesting that the au26 mutation is strongly, but perhaps incompletely, penetrant. Once the causative mutation is identified, precise calculations of the penetrance of au26 can be determined.

EtOH-Sensitive Mutants Disrupting Specific Craniofacial Developmental Events
Similar to au26, half of the mutants isolated in our forward genetic screen disrupted development of specific craniofacial elements. In EtOH-treated au15 and au28 mutants, skeletal elements of the ventral first arch and dorsal second arch are disrupted, respectively. These effects are only seen in the presence of EtOH; unexposed mutants appear to have normal facial development.
In EtOH-exposed au15, there is a severe reduction or loss of the lower jaw. In 5 dpf EtOH-treated mutants, the dorsal first arch cartilage element, the palatoquadrate, is present albeit reduced ( Fig. 3A,B, arrows). However, the ventral cartilage element, Meckel's cartilage, is lost ( Fig. 3B) as are the dentary (d) and entopterygoid (e) bones (Fig. 3A). Within the second pharyngeal arch, the basihyal cartilage (Fig. 3A, asterisk) is lost in EtOH-treated mutants (Fig. 3B). The mutation appears to behave as a fully penetrant recessive mutation, resulting in 25% mutant phenotypes. We note, however, that in some backcrosses of au15, we have observed a reduction in this level of penetrance (from 25% to 10% of a clutch), suggesting that au15 is subject to genetic modifiers. These currently unknown background modifiers may give important insight into susceptibility to FASD. We observed no defects in clutches that were not exposed to EtOH, suggesting that the normal function of au15 buffers against EtOH-induced craniofacial defects.
In EtOH-exposed au28 mutants, there are subtle defects to the hyosymplectic cartilage. The hyosymplectic cartilage consists of the symplectic cartilage rod extending ventrally from the body of the hyomandibular. Based on mutant phenotypes, the hyomandibular can be subdivided into an anterior region, which is sensitive to signals from the adjacent first pharyngeal pouch, and a posterior region, which is insensitive to such disruption (Crump et al., 2004b). The anterior hyomandibular is specifically disrupted in EtOH-treated au28 mutants (Fig. 4, arrows). The posterior hyomandibular as well as the opercle bone that attaches to it is present (Fig. 4). Phenotypes range from complete loss to partial loss of the anterior hyomandibular in 1 to 20% of the clutch, respectively. Thus, the au28 allele is likely incompletely penetrant with variable expressivity.

Mutants Broadly Disrupting Craniofacial Development
The remaining 3 mutants isolated in our analyses have more broad effects on the craniofacial skeleton. Extensive loss of bone is observed in EtOH-treated au27, and cartilage is broadly disrupted in au32 and au29 mutants.
Similar to the other mutants obtained in this forward genetic screen, au27 mutants only display defects when exposed to EtOH. At 5 dpf, the entire craniofacial skeleton is The viscerocranium consists of 7 segments. The first and second segments have distinct ventral and dorsal skeletal elements at this age. Meckel's cartilage (mc) and the palatoquadrate (pq) reside ventrally and dorsally, respectively, in the first pharyngeal arch. Ventral and dorsal elements within the second pharyngeal arch are the ceratohyal cartilage (ch) with its associated bone the branchialstegal ray (bsr) and the hyosymplectic cartilage (hs) and opercle bone (op), respectively. The remaining 5 pharyngeal arches house ceratohyal cartilages 1 to 5 (numbered), with the fifth ceratohyal harboring the pharyngeal teeth. The mesoderm-derived cleithrum (cl) resides just posterior to the pharyngeal arches. Modified from Swartz and colleagues (2014). slightly smaller in EtOH-treated au27 mutants. More striking, however, is the severe reduction in the amount of bone (Fig. 5). There is loss of neural crest-derived bone (the opercle and the ossification of the fifth ceratobranchial along with its associated teeth) as well as mesoderm-derived bone (the cleithrum; Fig. 5B). To determine whether this reflected a loss of bone per se or a delay in bone differentiation, we analyzed mutants at 7 dpf with a more sensitive live Alizarin Red staining procedure. Confocal imaging of ventral bones demonstrates that the ossification of the ceratohyal and the premaxilla, dentary, opercle, retroarticular, and branchiostegal ray bones had formed appropriately in untreated and EtOH-treated wild-type fish (Fig. 6A). However, EtOH-treated au27 mutants lack nearly all Alizarin Red staining (Fig. 6B). These phenotypes appear in 25% of fish within a clutch. Thus, we conclude that au27 is a recessive mutation that sensitizes fish to bone loss following EtOH exposure.
Mutants for au32 also appear normal in the absence of EtOH. In EtOH-treated clutches of au32, just under 50% of fish appear normal in the presence of EtOH, strongly suggesting that au32 is a partially dominant EtOH-sensitive allele. The phenotypes of affected zebrafish are variable but fall within 3 major phenotypic classes. The largest phenotypic class (32%) consisted of zebrafish with microphthalmia, cardiac edema, and reduced cartilage (Fig. 7B).
Roughly 14% of EtOH-treated fish were dead or dying by 4 dpf. In addition to microphthalmia, edema, and reduced cartilage, the trabeculae and/or Meckel's cartilage were absent or greatly reduced in about 6% of fish (Fig. 7C, arrow). Based on these proportions, it is not clear that there is a straightforward genotype-phenotype correlation in EtOHtreated au32 fish, as we observed with EtOH-treated pdgfra mutants and heterozygotes (McCarthy et al., 2013). These more severely affected fish also had an overall reduction to the body axis and appeared very unhealthy (Fig. 8, bottom). Thus, it is possible that the early embryonic lethal/severe craniofacial defects associate with homozygous mutants and heterozygosity associates with microphthalmia, cardiac edema, and reduced cartilage. Identification of the causative mutation will be necessary to determine if any such correlation exists.
The last mutant from our forward genetic screen, au29, is unique among the mutants identified in this screen in that mutant phenotypes are apparent in the absence of EtOH. The size of the eye is notably reduced in untreated au29 mutants (Fig. 9A,B). Additionally, there are slight disruptions to the ethmoid plate (Fig. 9A, arrow), the angle of the ceratohyal is altered (Fig. 9B, arrow), and the number and size of the ceratobranchial cartilages are reduced (Fig. 9B). Following EtOH exposure, these mutants have profound  craniofacial defects. The ethmoid plate is severely reduced (Fig. 9C, arrow). EtOH-treated mutants lose most of the ceratobranchial cartilages (Fig. 9D). Additionally, the first and second arch cartilage elements are severely reduced (Fig. 9D, arrows). Mutant phenotypes are generated in 25% of fish within a clutch, suggesting that au29 is a fully penetrant recessive mutant. Given that the EtOH-induced phenotypes present in our mutants have not been previously described, our collective findings suggest that forward genetic screens are an efficient way to identify gene-EtOH interactions.

A Novel Member of the LRP Family Is Mutated in au29
Given the uniqueness of the phenotypes in our mutants, it is likely that they harbor mutations in genes not previously characterized in EtOH teratogenesis. To test this, Fig. 5. Bone loss in EtOH-treated au27 mutants. Flat mounted viscerocrania from wild-type (A) and au27 mutant (B) zebrafish. There is an overall reduction in the size of craniofacial cartilages in au27 mutants. More striking is the near complete loss of craniofacial bone. Only a tiny remnant of the pharyngeal teeth (arrows) remain in mutants. Op, opercle; cb5, ceratobranchial #5; cl, cleithrum. Anterior to the left. Fig. 6. Bone loss in au27 mutants is not due to a developmental delay. Relative to EtOH-exposed wild-type fish (A), ossification in au27 mutants is greatly diminished at 7 dpf. Pm, premaxilla; d, dentary; ra, retroarticular; ch, ceratohyal; op, opercle; bsr, branchiostegal ray. Ventral view, anterior to the left. we selected 1 allele, au29, on which to perform genetic mapping.
We used MegaMapper for genome assembly as well as variant calling and mapping in au29 relative to the reference genome. This positioned au29 on chromosome 10. SNPs predicted to be deleterious in skor2, spp1, and si:dkey-88l16.3 (an annotated, but uncharacterized gene) were identified within 3 Mb of the peak of homozygosity. Each of these SNPs was predicted to disrupt splice sites. We found that the SNPs in skor2 and spp1 were not in splice sites and likely represent miscalls by MegaMapper. However, we verified a T to C donor splice site mutation in intron 18 to 19 of si:dkey-88l16.3 (Fig. 9E). Using PCR and Sanger sequencing on a separate group of au29 mutants, we determined that the mutation was homozygous in au29 mutants but not wildtype zebrafish. To verify that au29 is a mutation in si:dkey-88l16.3, we generated a CRISPR/Cas9 allele (au113, Fig. 9E). The phenotype of au113 mutants phenocopies those observed in au29. Furthermore, we crossed an au29 carrier and an au113 carrier and screened a clutch of 200 embryos. We found that approximately 25% of the embryos had phenotypes identical to au29 demonstrating that au29 and au113 fail to complement. Thus, we conclude that au29 is a mutation in a previously uncharacterized gene. The predicted protein structure of si:dkey-88l16.3 would place it within the LDL receptor-related protein (LRP) family, important modulators of cell signaling. Mutant or morpholino-induced loss-of-function phenotypes have been described for 6 other members of the LRP family in zebrafish: lrp1aa, lrp1ab, lrp2a, lrp4, lrp5, and lrp6 Pi et al., 2012;Saint-Amant et al., 2008;Veth et al., 2011;Willems et al., 2015). None of these other Lrp family members have been tested for EtOH sensitivity. Only lrp5 morphants have been characterized to have facial phenotypes potentially due to altered Wnt signaling (Willems et al., 2015) but the predicted structure of si:dkey-88l16.3 is distinct from Lrp5. We are currently characterizing the identity and function of this gene (Kuka, et al., in preparation). It will be of great interest to determine whether the Lrp family is broadly sensitive to EtOH exposure and how au29 alters signaling within the developing embryos.

DISCUSSION
Our results are the first vertebrate forward genetic screen used to identify mutants with enhanced susceptibility to EtOH teratogenesis. Most human birth defects are thought to arise from complex interactions between genetics and the environment. Most of the mutants that we have recovered show no obvious deleterious phenotypes in the absence of EtOH. Consistent with this, our recovery rate of approximately 1 mutant/25 mutagenized genomes (4%) is higher than that of a large-scale zebrafish screen in which 109 mutations disrupting craniofacial development were identified in 3,857 mutagenized genomes (2.8%; Piotrowski et al., 1996;Schilling et al., 1996). Given that the mutagenesis procedure used results in approximately 1.1 deleterious mutations per mutagenized genome , it is also likely that a minority of mutations will confer EtOH sensitivity, although our screen was relatively small and may not reflect the true number. Thus, it is likely that many sensitizing genetic loci have yet to be identified through candidate approaches. The success of this forward screen in identifying mutations that enhance the teratogenicity of EtOH demonstrates the usefulness of such screens and should be useful in identifying sensitizing alleles for other teratogens.
Each mutant that we obtained has a unique EtOH-induced defect, suggesting that they are mutations in different genes; however, genetic mapping of the individual mutants will be required to directly determine the identity of the genes disrupted in these mutants. Furthermore, the majority of these mutants have no apparent deleterious phenotypes in the absence of EtOH, suggesting that many of the mutants may be in novel genes. Our genetic mapping of au29 demonstrates that our screen has already identified previously uncharacterized mutants. Given that all the other mutants appear normal in the absence of EtOH, it is likely that many of these mutants will also not have been identified in previous genetic screens. For those genes that have been characterized, the EtOH-induced phenotypes are likely to be unknown.
The mechanism of how EtOH disrupts development in this set of mutants is wholly unknown at this point. It remains possible that some mutations alter the uptake or elimination of EtOH, particularly those mutant more broadly disrupting development. However, the wide range of phenotypes that we observe would suggest that EtOH is disrupting genetic pathways regulating specific developmental events. Therefore, genetic mapping and characterizing these mutants will provide important insights into the genetic pathways that mediate susceptibility to EtOH teratogenesis. Fig. 8. Whole body defects in au32 mutants. Those au32 mutants with more profound craniofacial defects also display marked reductions in the overall body axis (au32, bottom) while less severe au32 mutants (au32, top) have a more normal body axis.

Insights Into Potential EtOH-Sensitive Pathways: Palatal Development
The phenotype of au26 mutants closely resembles that of gata3 mutants (Sheehan-Rooney et al., 2013) with a specific disruption to the trabeculae of the posterior palate. This striking phenotypic similarity may suggest that au26 is: (i) an allele of gata3, (ii) a member of the same genetic pathway as gata3, or (iii) a member of a parallel pathway that converges on the development of the trabeculae.
Interestingly, gata3 was in our original candidate-based screen for EtOH-sensitive mutants (McCarthy et al., 2013). From this, and subsequent analyses, we know that, unlike au26, loss of gata3 does not sensitize embryos to EtOH teratogenesis. Furthermore, we have performed complementation analyses of au26 and gata3 (data not shown). We find that all fish from these crosses appear normal. Collectively, these findings strongly suggest that au26 is not an allele of gata3.
Currently, little is known about the pathways specifically regulating the development of the trabeculae. Disruption of Bmp signaling, via expression of a dominant negative Bmp receptor, can result in specific loss of the trabeculae (Alexander et al., 2011). In mouse, Gata3 is a target of Bmp signaling (Bonilla-Claudio et al., 2012). EtOH exposure causes ectopic Bmp signaling in the zebrafish heart (Sarmah et al., 2016) and H9c2 cardiomyoblast cells (Shi et al., 2017). Similarly, EtOH exposure elevates the expression of Bmp4 in cultured rat cranial neural crest cells (Wentzel and Eriksson, 2009). It is possible that the overall level of Bmp signaling is critical for activation of gata3 and that this is perturbed in EtOH-exposed au26 mutants. We are currently characterizing the expression of gata3 in au26 and Bmp pathway mutants to test this possibility.
No other signaling pathway has been specifically implicated in trabeculae development. However, both the Sonic Hedgehog and Wnt/PCP pathways are important for palate development (Bush and Jiang, 2012;Eberhart et al., 2006;Fig. 9. Craniofacial defects in au29 mutants. In untreated au29 mutants (A, B), there are reductions to the ethmoid plate in the neurocranium (A, arrows). Within the viscerocranium, the ceratohyal cartilages (arrow) fail to project anteriorly and the number of ceratobranchial cartilages (numbered) is reduced (B). EtOH exposure dramatically exacerbates these phenotypes (C, D). The ethmoid plate is greatly reduced in EtOH-treated au29 mutants (C, arrow; the eyes have been dissected away to more clearly visualize the cartilage). (E) Schematic representation of the au29 and au113 alleles. Schematic of exons 14 through 21, of 24 predicted exons, is shown with the location of au29 and au113 noted. The splice donor of exon 15, underlined in wt sequence, is mutated in au29. A 5-bp indel is present in au113. The PAM sequence is underlined in the wt sequence, and the FspI site used for genotyping is shown in red. Kamel et al., 2013;Rochard et al., 2016;Wada et al., 2005) and have pathway members that modulate the risk to EtOH teratogenesis (Eberhart and Parnell, 2016). Once the molecular nature of the au26 allele is discovered, we will be able to more easily determine which genetic pathway(s) is disrupted in this mutant.

Insights Into Potential EtOH-Sensitive Pathways: Endoderm Development
While au15 and au28 have distinct phenotypes from one another, disrupting development of the ventral first arch versus dorsal second arch (respectively), both of these phenotypes are observed in mutants disrupting endoderm development. Endoderm is necessary for proper craniofacial development (Crump et al., 2004a;David et al., 2002;Lovely et al., 2016) and provides positional patterning information to the neural crest (Couly et al., 2002). Much work on the effects of EtOH on facial development has focused on the neural crest (Smith, 1997). These mutants may suggest an underappreciated role of EtOH on endoderm development.
The loss of Meckel's cartilage in au15 mutants phenocopies sphingosine-1-phospate (S1P) type 2 receptor (s1pr2) mutants. In s1pr2 mutants, the appropriate migration of anterior-most pharyngeal endoderm fails and signaling centers required for lower jaw development are not formed appropriately (Balczerski et al., 2012). We have found that s1pr2 mutants are not EtOH-sensitive (Lovely, et al, unpublished) and s1pr2 and au15 mutants complement (data not shown). Thus, like au26 and gata3, au15 and s1pr2 are EtOH-sensitive and insensitive mutations, respectively, that generate remarkably similar and highly specific phenotypes. Further characterizations will determine whether au15 disrupts the development of the anterior endoderm, facial signaling centers, or the neural crest cell response to these signaling centers.
The specific loss of the anterior hyomandibular cartilage in au28 may be due to alteration of the first pharyngeal pouch or the signals the pouch sends to the neural crest. In integrin alpha 5 (itga5) mutants, an endoderm autonomous requirement of itga5 results in the subsequent loss of the anterior hyomandibular (Crump et al., 2004b). We have yet to determine whether itga5 is EtOH-sensitive, but alcohol exposure has been shown to alter integrin levels in the rat brain, trophoblast cells, and human neurospheres (Rout, 2006;Rout and Dhossche, 2010;Vangipuram et al., 2008). There is no clear association of down-regulated integrin expression in these studies, rather some integrins are upregulated and some are down-regulated. It is also unclear whether these alterations reflect direct effects of EtOH on transcription or an indirect effect, such as alterations to the relative proportions of cell subtypes expressing different integrins. Regardless of mechanism, however, the phenotype of au28 mutants strongly suggests an involvement of endoderm or the neural crest response to endodermal signals.

Insights Into Potential EtOH-Sensitive Pathways: Bone and Cartilage Development
There are broad defects to bone development in au27 mutants and cartilage development in au29 and au32 mutants. Well-established genetic pathways regulate the development and differentiation of these tissues. Given that au27 mutants have defects to bone derived from both neural crest cells and mesoderm, it likely disrupts differentiation of osteocytes in general. Numerous markers of osteocytes and their progenitors, osteoblasts, exist (Hojo et al., 2017). Analyses of transcription factors, such as sp7 and runx2, as well as extracellular matrix molecules, such as col10a1a, will provide insight into which steps of bone development are disrupted in au27. Given that the au27 mutation is lethal in EtOH, our current analyses cannot fully rule out a developmental delay nor can we determine whether later forming axial skeletal elements are disrupted. These phenotypic characterizations will provide a deeper understanding of the nature of the au27 defect. Craniofacial cartilages are largely of neural crest origin. Thus, analyses of genes within the well described neural crest gene regulatory network (Martik and Bronner, 2017), such as foxd3, snail, sox10, and Dlx family genes, will provide insight into au29 and au32 mutants. Additionally, analysis of the genes critical for cartilage development, such as sox9a and cartilage-specific collagens, will provide insights into the genesis of the phenotypes in these mutants.
Collectively, these mutants will provide insights into the risk for FASD. As the mutations are identified, we will be able to thoroughly quantify the phenotypes in these mutants. Our characterizations of the genesis of these phenotypes will elucidate the effects of EtOH on the cell behaviors mediating craniofacial development.