Loss of tumor protein 53 protects against alcohol-induced facial malformations in mice and zebrafish

Abstract

Background:

Alcohol exposure during the gastrulation stage of development causes the craniofacial and brain malformations that define fetal alcohol syndrome. These malformations, such as a deficient philtrum, are exemplified by a loss of midline tissue and correspond, at least in part, to regionally selective cell death in the embryo. The tumor suppressor protein Tp53 is an important mechanism for cell death, but the role of Tp53 in the consequences of alcohol exposure during the gastrulation stage has yet to be examined. The current studies used mice and zebrafish to test whether genetic loss of Tp53 is a conserved mechanism to protect against the effects of early developmental stage alcohol exposure.

Methods:

Female mice, heterozygous for a mutation in the Tp53 gene, were mated with Tp53 heterozygous males, and the resulting embryos were exposed during gastrulation on gestational day 7 (GD 7) to alcohol (two maternal injections of 2.9 g/kg, i.p., 4 h apart) or a vehicle control. Zebrafish mutants or heterozygotes for the tp53^zdf1 M214K mutation and their wild-type controls were exposed to alcohol (1.5% or 2%) beginning 6 h postfertilization (hpf), the onset of gastrulation.

Results:

Examination of GD 17 mice revealed that eye defects were the most common phenotype among alcohol-exposed fetuses, occurring in nearly 75% of the alcohol-exposed wild-type fetuses. Tp53 gene deletion reduced the incidence of eye defects in both the heterozygous and mutant fetuses (to about 35% and 20% of fetuses, respectively) and completely protected against alcohol-induced facial malformations. Zebrafish (4 days postfertilization) also demonstrated alcohol-induced reductions of eye size and trabeculae length that were less common and less severe in tp53 mutants, indicating a protective effect of tp53 deletion.

Conclusions:

These results identify an evolutionarily conserved role of Tp53 as a pathogenic mechanism for alcohol-induced teratogenesis.

INTRODUCTION

The tumor suppressor protein Tp53 (tumor protein p53) coordinates adaptive cellular responses to a variety of cellular signals and chemical stimuli. Tp53 regulates cell death and differentiation, the cell cycle, cellular senescence, metabolic adaptation, and DNA repair by affecting gene expression and through direct protein interactions (for review, see Kruiswijk et al., 2015). These activities of Tp53 maintain the delicate balance of cell growth and cell death and defend against the replication of damaged cells. When Tp53 activity is compromised by mutations, there is decreased resistance to tumors and these mutations are linked to about 50% of all cancers (e.g., Aubrey et al., 2018; Vousden & Prives, 2009) Although the roles of Tp53 in cancer susceptibility and treatment are intensely studied, there is a growing appreciation of Tp53’s importance to embryonic and fetal development (Jain & Barton, 2018).

Early in life, Tp53 mRNA is expressed throughout the embryo and later becomes more restricted to regions undergoing differentiation (Schmid et al., 1991). The protein itself, however, is only active in response to cellular stressors and/or DNA damage, which inhibits MDM2 and induces posttranslational modifications that increase the stability of Tp53. Stable Tp53 is found throughout the cell, including the nucleus, where it affects gene transcription. It is also found in the cytoplasm and mitochondrial membrane, where it interacts with proteins in the Bcl2 family (Vaseva & Moll, 2009). Early studies in mice suggested that the absence of the Tp53 gene had a negligible effect on mouse embryonic development (Donehower et al., 1992), but later studies revealed that the absence of Tp53 can indeed affect neural tube, urogenital, skeletal, and cardiovascular development (Armstrong et al., 1995; Chen et al., 2008; Rinon et al., 2011). The developmental effects of knocking out Tp53 are not fully penetrant and vary across background strains (Armstrong et al., 1995), but are consistently more prevalent in female mice (Chen et al., 2008), which may contribute to the higher rates of fetal death in female Tp53 knockout mice. Nonetheless, the overall survival of the Tp53 knockout mouse is likely due to overlapping functions of Tp53 with its family members Tp63 and Tp73 (Van Nostrand et al., 2017). Significantly more damaging to early life than the absence of Tp53 is excessive Tp53 activation which can lead to early embryonic demise (Oca Luna et al., 1995) and a host of syndromes associated with congenital malformations (Bowen et al., 2019). Therefore, Tp53 activity must be precisely controlled to avoid abnormal development.

Tp53 is activated by a number of biochemical signals including oxidative stress, DNA damage, and glucose deprivation. Importantly, these signals can be elicited by exposure to many developmental insults, such as radiation (Komarov et al., 1999), hyperthermia (Hosako et al., 2007), nutrient deficiency (Li et al., 2018), hydroxyurea (El Husseini & Hales, 2018b), benzopyrenes (Nicol et al., 1995), and alcohol (Anthony et al., 2008; Flentke et al., 2019; Ignacio et al., 2014; Kuhn & Miller, 1998). Studies examining teratogenesis in Tp53 knockout mice reveal the complex roles of Tp53 in the embryo. There is increased sensitivity to benzo[a]pyrene (Nicol et al., 1995) and hydroxyurea (El Husseini & Hales, 2018a,b), in Tp53 knockout mice, indicating that Tp53 has a protective effect. Similarly, Tp53 gene deletion protects embryos from apoptosis following radiation, but also appears to impair the embryo’s ability to recover from radiation, as surviving Tp53^−/− fetuses have significantly more radiation-induced defects (Norimura et al., 1996). On the other hand, Tp53 deletion protected against embryonic cell death induced by 2-chloro-2′-deoxyadenosine and from the subsequent development of eye defects (Wubah et al., 1996). Similar protection has been reported for cyclophosphamide-induced birth defects (Pekar et al., 2007). The actions of Tp53, as a “lifeguard with a license to kill” (Kruiswijk et al., 2015), may therefore depend on which pathogenic mechanisms are induced by specific teratogens.

In contrast to the developmental abnormalities observed in some Tp53-deficient mice, tp53^Zdf1 fish develop normally and are indistinguishable from wild-type siblings in terms of fertility and viability (Berghmans et al., 2005). Homozygous tp53 ^Zdf1 mutants demonstrate suppressed irradiation-induced apoptosis, impaired cell cycle checkpoint response, and reduced transcription of downstream target genes mdm2, p21, bax, and ccng1 (Lee et al., 2008). Interestingly, the most common tumors observed in tp53 ^zdf1 fish are malignant peripheral neural sheath tumors, which originate from neural crestderived Schwann cells and express the neural crest marker sox10 (Lee et al., 2016).

Fetal alcohol exposure, the leading preventable cause of birth defects worldwide, induces apoptosis in key regions of the embryo depending on the timing of the alcohol exposure. While apoptosis is but one of the pathogenic mechanisms of fetal alcohol exposure, its regionally selective expression correlates with distinct alcohol-induced birth defects (Dunty et al., 2001). Alcohol exposure induces neural crest cell death in the first and second pharyngeal arches during early development in genetically sensitized pdgfra mutant and heterozygous zebrafish (McCarthy et al., 2013). During gastrulation, alcohol exposure causes excessive apoptosis that is restricted to the neuroectoderm, particularly in the region that normally gives rise to the cortex, eyes, and parts of the face (Dunty et al., 2001). Previous studies have demonstrated that prenatal alcohol exposure (PAE) can induce Tp53 expression (Kuhn & Miller, 1998) and that shortly after gastrulation, neural crest cell apoptosis depends on Tp53 (Chen et al., 2015; Flentke et al., 2019; Rinon et al., 2011; Yuan et al., 2017), but studies have yet to elucidate the role of Tp53 in gastrulation stage apoptosis.

The objective of the current study was to examine whether Tp53 is an important mediator of birth defects caused by early developmental stage alcohol exposure, testing the specific hypothesis that loss of Tp53 is an evolutionarily conserved mechanism to protect against alcohol-induced dysmorphologies. The alcohol exposure periods examined correspond to the first 3 to 4 weeks of human gestation, when most pregnancies are unrecognized. In the mouse, alcohol exposure occurred during gastrulation, while in the zebrafish the exposure included gastrulation as well as early neurogenesis. The results in both species confirm the hypothesis that Tp53 mediates at least some of the teratogenic effects of PAE.

METHODS

Animals

Female mice heterozygous for the Tp53 mutation in exon 2 to 6 (gift from Mohanish Desmukh, UNC Chapel Hill, originally obtained from the Jackson Lab, B6.129S2-Trp53^tm1Tyj/J) were mated with heterozygous males for 1 to 2 h. Upon detecting a copulatory plug, the start of the mating session was defined as gestational day 0 (GD 0) and females were moved to breeding cages and disturbed only for injection. A separate set of heterozygous Tp53 females (n = 6) were used for the measurement of blood alcohol levels (BALs). All mice were housed in covered, ventilated polycarbonate cages with Purina Isopro RMH 3000 (Purina) freely available through the wire food hopper and water freely available through a drinking spout. The cages were lined with cob bedding and contained both nesting material and a polycarbonate hut. The vivarium was 21 ± 1°C and 30 to 40% humidity and had a 12:12 light:dark cycle (lights on at 7:00). All mouse experiments followed the NIH guidelines using methods approved by the IACUC of University of North Carolina at Chapel Hill.

Zebrafish husbandry and embryo collection were conducted in accordance with the recommendations in The Zebrafish Book, fifth edition (Westerfield, 2007), and the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the University of Texas at Austin Institutional Animal Care and Use Committee (protocol number AUP-2018–00002). All zebrafish were housed at the University of Texas at Austin under IACUC-approved conditions. Developmental staging of embryos was determined by the identification of well-characterized morphological features (Kimmel et al., 1995). We used zebrafish with the tp53^zdf1 M214K mutation in exon 7 within the DNA-binding domain (Berghmans et al., 2005), purchased from the Zebrafish International Resource Center.

Procedures

During gastrulation (GD 7), dams received 2.9 g/kg ethanol (EtOH; two intraperitoneal injections of 25% volume: volume, 4 h apart, n = 17 dams) or a lactated Ringer’s control (n = 14 dams) in a volume of 1.5 ml/100 g. In C57BL/6J dams, this alcohol treatment causes peak BALs 30 min after the second injection to be ~440 mg/dl (Godin et al., 2010). On GD 17, the dams were euthanized with CO₂ inhalation followed by cervical dislocation; the uteri were removed and placed into ice-cold phosphate-buffered saline (PBS). The position of fetuses and the number of resorptions were noted before the fetuses were removed from the uterus and placed in ice-cold PBS. Each fetus was examined under a Nikon SMZ-U stereoscopic zoom dissecting microscope and photographed with a Micropublisher 5.0 digital camera and QCapture Suite software (QImaging). Gross malformations of the body and/or face were noted and scored as unaffected or dysmorphic. Eye morphology was initially scored with a 7-point rating scale from unaffected (1) to complete anophthalmia (7; Gilbert et al., 2016). Representative eyes are shown in Figure 1A-G, with scores >1 being considered dysmorphic due to reduced globe size and/or a pupil shape that deviates from the typically round pupil observed in the GD 17 fetus. Pupil shape was later quantified from the photographed images using the ImageJ software shape analysis. The critical shape parameter was the roundness of the pupil, for which a score of 1.0 represented a perfectly round pupil. All morphology ratings were made by experimenters who were blind to both maternal treatment and fetal genotyping. Tails were removed for genotyping, the fetuses were weighed, measured for crown–rump length, and placed in Bouin’s fixative. The heads of select fetuses were later cleared of Bouin’s using 70% EtOH, embedded in paraffin, and sliced coronally at 10 µm, and serial sections were mounted on glass sides to be stained with hematoxylin and eosin (H&E). Genotyping was performed for Tp53 DNA using the primers: WT forward (CCCGAGTATCTGGAAGACAG); WT reverse (ATAGGTCGGCGGTTCAT); a common exon 7 (TATACTCAGCCGACCT); and a mutant forward for the Neocassette in exon 2 to 6 (TCCTCGCTTTACGGTATC). BALs were measured enzymatically by an Analox alcohol analyzer (Model AM1, Analox Instruments) using plasma samples separated from 30 µl of tail blood that had been collected 30 min after the second injection of alcohol (2.9 g/kg).

FIGURE 1.

Effects of Tp53 gene deletion on the incidence and severity of eye defects in mice. The percent of GD 17 fetuses with a defect in at least 1 eye in severity categories ranging from 1 (unaffected) to 7 (complete anophthalmia) is portrayed for Tp53 wild-type (+/+), heterozygous (+/−), and homozygous mutant (−/−) mice. Representative eyes from each category are shown in (A–G). A score of 2 (B) is considered a minor defect and is assigned if either the size of the globe is reduced or the shape of the pupil is altered. Scores of 3 or greater are considered severe defects and are assigned to small eyes with altered pupil shapes. The effects of a gastrulation stage treatment with either lactated Ringer’s vehicle control solution or alcohol are shown in (H). The percent of mice showing an FAS-like facial dysmorphology is indicated by inset text

Zebrafish embryos were dosed with media containing 1.5% or 2% EtOH at the onset of gastrulation (6 h postfertilization (hpf)) until either 48 hpf, covering the early neurogenesis/segmentation stages of embryogenesis, or 4 days postfertilization (dpf), when the zebrafish are freely swimming larvae. It has been demonstrated that the tissue levels in embryonic zebrafish are only a quarter to a third of the EtOH in the media (Lovely et al., 2016). As such, these are high, binge-like doses. At 4 dpf, all zebrafish embryos were stained with Alcian Blue and Alizarin Red for cartilage and bone, respectively (Walker & Kimmel, 2007). Heads were scored for craniofacial malformations and photographed for linear measures of eye size; then, tails were removed for genotyping. Genotyping for tp53^zdf1 was accomplished via Sanger sequencing of PCR products (University of Texas at Austin DNA Sequencing Facility) and evaluating the sequence using SnapGene Viewer software. Primers used were as follows: tp53 forward-standard PCR: ACATGAAATTGCCAGAGTATGTGTC and tp53-reverse-standard PCR: TCGGATAGCCTAGTGCGAGC. Images of zebrafish heads were taken with a Zeiss Axio Imager-A1 microscope, edited using Adobe Photoshop, and measured using ImageJ. Eye width was measured as the widest distance between 2 points for each eye, from a ventral view (Figure S1B). This overall measurement does not determine which specific ocular cell types were affected by alcohol, as previously reported (e.g., Muralidharan et al., 2015, 2018). Graphs were produced with RStudio and Microsoft PowerPoint. Full body zebrafish images were taken with an Olympus szx7 stereomicroscope (Figure S2).

Zebrafish embryos were fixed and prepared for terminal deoxynucleotidyl transferase dUTP nick end-labeling (TUNEL) staining following the procedure outlined previously (Wilfinger et al., 2013). The staining was performed using the in situ cell death detection kit, TMR red (Roche, Cat. No. 12156792910) according to the manufacturer’s instructions. TUNEL-stained embryos were imaged by Zeiss LSM 710 using Zen software. Images were analyzed using ImageJ.

Statistical analysis

The incidence of defects (proportion affected) was analyzed using chi-square tests comparing the wild-type, heterozygous, and homozygous knockout mice. Mouse fetal body weight, crown–rump length, as well as eye size and shape measurements were analyzed using two-way analysis of variance with genotype and treatment as between-subjects variables. Number of resorptions and live fetuses were not normally distributed and were analyzed by Mann–Whitney rank-sum tests. Alpha was set to p < 0.05. Zebrafish trabeculae length, eye width, and cell death were analyzed for each treatment using one-way analysis of variance followed by Tukey’s honestly significant difference post hoc between wild-type, heterozygous, and homozygous mutants. Alpha was set to p < 0.05. Outliers for zebrafish trabeculae length and eye width were detected using Tukey’s 1.5IQR rule and removed from the analysis.

RESULTS

Loss of Tp53 protects against EtOH teratogenesis in mouse

The distribution of Tp53 wild-type, heterozygous, and homozygous mutant fetuses approximated expected mendelian ratios in both the vehicle and alcohol treatments (Table 1). Among the 39 -homozygous mice, a Tp53 genotype effect, independent of alcohol treatment, occurred on the incidence of exencephaly (n = 5), spina bifida (n = 4), forelimb syndactyly (n = 1), hindlimb polydactyly (n = 1), and genital hyperplasia (n = 9). Consistent with previously published reports (Chen et al., 2008), these defects were observed only in female mice. More than one defect was observed in 7/12 of the affected fetuses, with genital hyperplasia being present in six of these seven cases and exencephaly and spina bifida present in the other. The incidence of spontaneously occurring eye defects was similar between the Tp53 wild-type and Tp53 homozygous mutant mice (~17% vs. ~20%) and was consistent with previous studies using the C57BL/6J strain (Parnell et al., 2006). When the data from each genotype were partitioned into the mean for each litter, there was a main effect of Tp53 genotype on body weight, F (2, 68) = 6.2; p = 0.003, and crown–rump length, F (2, 68) = 6.4; p = 0.003, such that regardless of alcohol treatment, the homozygous mutant mice weighed significantly less and were shorter than the wild-type mice (Table 2).

TABLE 1.

Litter measures for vehicle and alcohol exposures of Tp53 heterozygous mouse dams

Exposure Group	Litter (n)	Live Fetuses	Resorptions	n by GT (+/+, +/−, −/−)	% GT (+/+, +/−, −/−)
Vehicle	14	6.9±0.5	1.1±0.3	29, 52, 15	30, 54, 16
Alcohol	17	5.6±0.7	1.9±0.5	35, 39, 24	36, 40, 24

Note: Data for live mouse fetuses and resorptions are expressed as the mean litter average (±1 SEM).

TABLE 2.

Effects of Tp53 genotype and gastrulation stage alcohol exposure on the severity of eye defects, body weight, and body length in mouse fetuses

Exposure Group	Tp53 Genotype	Eye Defect Severity Score							Body Weight	Body Length
		1	2	3	4	5	6	7
Vehicle	(+/+)	24	4	1	0	0	0	0	0.79 ± 0.01	17.1 ± 0.01
	(+/−)	46	4	1	1	0	0	0	0.80 ± 0.01	17.2 ± 0.02
	(−/−)	12	0	2	1	0	0	0	0.66 ± 0.03*	16.5 ± 0.03*
Alcohol	(+/+)	9	11	8	1	2	4	0	0.72 ± 0.01	16.7 ± 0.02
	(+/−)	25	10	4	0	0	0	0	0.78 ± 0.02	17.2 ± 0.02
	(−/−)	19	3	1	0	1	0	0	0.73 ± 0.03*	16.6 ± 0.02*

Note: Data for eye defect severity are expressed as number of mouse fetuses per eye score. Body weight and crown–rump length are expressed as the mean litter average (±1 SEM). Emboldened data with asterisks indicate significant genotype main effects (*p < 0.05 vs. WT, following a 2-way ANOVA).

BALs in Tp53 heterozygous dams were 434 ± 20 mg/dl, which is similar to previous measurements from the C57BL/6J strain (Godin et al., 2010). Alcohol treatment did not significantly affect the number of live fetuses or the number of resorptions (Table 1), although the litter sizes are somewhat smaller than is typical for C57BL/6J mice. Confirming prior studies (Dou et al., 2013; Godin et al., 2010; Parnell et al., 2006), alcohol increased the incidence and severity of fetal dysmorphology, as measured by eye defects (χ² = 18.4; p < 0.001) and fetal alcohol syndrome (FAS)-like craniofacial features (Figure 1). The effect of alcohol on eye dysmorphology depended on the Tp53 genotype; Tp53 heterozygous and homozygous mutant mice had significantly fewer eye defects than did the Tp53 wild-type mice (χ² = 9.5 and 15.1 vs. wild-type, for heterozygous and homozygous mutant mice, respectively, p ≤ 0.002). Tp53 deletion also protected mice from fetal alcohol syndrome (FAS)-like facial features; only the wild-type mice manifested the severe dysmorphology, which included one case of micrognathia, and two philtrum deficiencies, one of which had a single nostril, extensive cleft palate, and severe holoprosencephaly (Figure 2). Moderate edema occurred in one heterozygous mouse treated with alcohol, and mild edema was noted in one vehicle-treated wild-type mouse.

FIGURE 2.

Representative photographs of GD 17 mouse faces and brains following alcohol or vehicle treatments. The left panel shows a typical unaffected Tp53 wild-type (+/+) fetus with hematoxylin- and eosin-stained coronal sections taken through the face at the level of the vomeronasal organ and medial septum of the brain. The center and right panels show alcohol treated Tp53 wild-type (+/+) and homozygous mutant (−/−) fetuses. The black arrows indicate areas of severe dysmorphology in the wild-type fetus, including decreased intraocular distance, flat philtrum, small vomeronasal organ, cleft palate, and holoprosencephaly. These FAS-like defects were not observed in the Tp53 homozygous mutant, or heterozygous (+/−) fetuses

A supplemental analysis using shape and size measurements taken from the facial photographs revealed similar findings to those derived from the dysmorphology scale. When the data from each genotype were partitioned into means for each litter and subjected to a two-way ANOVA, there was a significant interaction between alcohol treatment and Tp53 genotype on the roundness of both the right and left pupils, F (2, 68) = 5.1; p = 0.009 and 6.7; p = 0.002, for the right and left eyes, respectively (Figure S3). Post hoc Bonferroni comparisons revealed that there was no effect of genotype following vehicle treatment, but that the wild-type pupils were significantly less round after alcohol treatment than were the pupils of the heterozygous and homozygous mutant mice. Alcohol significantly reduced pupil roundness in the wild-type mice only.

The incidence of alcohol-induced eye defects in the wild-type mice was much higher than anticipated based on previous studies with C57BL/6J mice which typically find about 50% of gastrulation stage alcohol-exposed fetuses have an eye defect. In the current study, about 75% of wild-type mice had a defect. Although the Tp53 exon deletion was initially created on a mixed C57/129 background (Boettcher et al., 2019) and partially backcrossed onto the C57BL/6J strain, it is possible that this elevated penetrance reflects a contribution from the 129 background.

An evolutionarily conserved function for tp53 in alcohol teratogenesis

Because apoptosis is a common mechanism for alcohol teratogenesis across species, we tested if loss of tp53 would protect against alcohol teratogenesis in zebrafish. We used zebrafish tp-53^Zdf1 mutants which have been shown to have impairments in apoptosis, to evaluate attenuation of alcohol-induced phenotypes. We compared the effects of 1.5% EtOH from 6 hpf to 4 dpf, 2% alcohol from 6 hpf to 4 dpf, and 2% alcohol from 6 to 48 hpf relative to unexposed zebrafish (See Figure S2, for full body images). Zebrafish treated with 1.5% alcohol from 6 to 48 hpf had no apparent defects. Therefore, we did not pursue this treatment group further. In each of the analyses presented here, nearly all zebrafish survived (n = 85/87). The two deaths observed were in the highest exposure group (2% alcohol, 6 hpf-4 dpf). Thus, the doses used do not result in a high mortality, at least prior to 4 dpf. We next used these dosing windows to determine the effect of Tp53 on alcohol-induced craniofacial and ocular defects.

Loss of tp53 reduces susceptibility to alcohol-induced malformations in the zebrafish palate

The zebrafish anterior neurocranium is analogous to the mammalian palate in function, gene expression, and cell origin (Swartz et al., 2011). It consists of the midline ethmoid plate and bilateral trabeculae that fuse to the posterior neurocranium (see Figure 3). We first analyzed craniofacial phenotypes by scoring individuals into six categories based on anterior neurocranium morphology: (1) normal morphology; (2) mild defects to the anterior neurocranium, in which the ethmoid plate was slightly smaller or malformed; (3) moderate anterior neurocranium, with malformed ethmoid plate and trabeculae; (4) unilateral trabeculae break, moderate defects plus a trabeculae that is shortened and fails to fuse to the posterior neurocranium; (5) bilateral trabeculae breaks, like 4 but with bilateral failure to fuse; (6) severe anterior neurocranium, in which the ethmoid plate is severely reduced or absent and trabeculae dramatically malformed.

FIGURE 3.

Loss of tp53 ameliorates craniofacial malformations in alcohol-exposed zebrafish. (A-F) are representative images of normal morphology, mild anterior neurocranium, moderate anterior neurocranium, unilateral trabecula break, bilateral trabeculae breaks, and severe anterior neurocranium, respectively. (G) shows the percentage of individuals in each category, by wild-type (+/+), heterozygous (+/−), and homozygous mutant (−/−) genotypes

Consistent with previous reports that zebrafish tp53 mutants develop normally (Berghmans et al., 2005), we found that loss of tp53 had no striking deleterious effect on the craniofacial skeleton. Across all genotypes, we find a very low proportion of zebrafish with mild anterior neurocranial defects in the absence of alcohol exposure, although there is a slightly higher proportion of mutants with defects (13% higher compared to wild-type zebrafish; Figure 3, see Table 3 for incidences of craniofacial phenotypes). Thus, tp53 genotype has little or no effect on craniofacial development in the absence of additional insults, such as an alcohol exposure.

TABLE 3.

Effects of alcohol and tp53 genotype on craniofacial morphology in zebrafish

Exposure Group	tp53 Genotype	Normal Anterior Neurocranium	Mild Anterior Neurocranium	Moderate Anterior Neurocranium	Unilateral Trabecula Break	Bilateral Trabeculae Break	Severe Anterior Neurocranium
No Treatment	(+/+)	13	1	0	0	0	0
	(+/−)	16	1	0	0	0	0
	(−/−)	8	2	0	0	0	0
1.5% Alc, 6 hpf to 4 dpf	(+/+)	2	1	0	2	2	0
	(+/−)	3	4	0	3	5	0
	(−/−)	3	3	0	2	0	0
2% Alc, 6 hpf to 48 hpf	(+/+)	0	1	1	2	4	1
	(+/−)	0	0	5	3	6	1
	(−/−)	0	2	2	2	0	0
2% Alc, 6 hpf−4 dpf	(+/+)	0	0	0	2	4	0
	(+/−)	0	0	2	5	4	3
	(−/−)	0	1	3	3	1	0

Note: Data for craniofacial morphology are the number of zebrafish within each severity category.

Our results indicate that tp53 genotype does modulate the severity and proportion of deleterious craniofacial phenotypes produced by embryonic alcohol exposure. Mutants exposed to 1.5% alcohol from 6 hpf to 4 dpf had higher rates of either normal phenotypes or mild defects to the anterior neurocranium than did wild-type or heterozygous fish (Table 3; Figure 3). None of the tp53 mutant fish had bilateral trabeculae breaks, but breaks were observed in wild-type and heterozygous fish (Table 3; Figure 3). Thus, while there does not appear to be an effect of gene dosage, with phenotypes in heterozygotes being similar to those in wild-type zebrafish, loss of tp53 protects against severe alcohol-induced malformations.

We observed a similar result in zebrafish exposed to 2% alcohol (Table 3; Figure 3). Mutant fish displayed the highest proportion of mild and moderate anterior neurocranium phenotypes, and none displayed the severe phenotype. Again, we did not observe a clear reduction in severity of palate defects in heterozygous fish.

We quantified craniofacial malformations by measuring the length of both trabeculae and the ethmoid plate of each fish. While we did not find any significant alterations to the length of the ethmoid plate (data not shown), there were substantial gene–environment effects on trabeculae length. Genotype did not have an effect on trabeculae size in the absence of alcohol (Figure 4; see Table 4 for descriptive statistics and Table 5 for ANOVA results); however, there was a statistically significant improvement in trabeculae length in tp53 mutants compared with their wild-type siblings in all three treatment groups (Figure 4; see Table 4 for descriptive statistics and Table 5 for ANOVA results). Mutant fish also have statistically significant increases in trabeculae length compared with their heterozygous siblings in the 1.5% alcohol 6 hpf-4 dpf and 2% alcohol 6 hpf-4 dpf exposure groups (Figure 4; see Table 4 for descriptive statistics and Table 5 for ANOVA results). We observed a trend for increased trabeculae length in the 2% alcohol 6 hpf-48 hpf group, but this effect was not statistically significant (Figure 4; see Table 4 for descriptive statistics and Table 5 for ANOVA results). Interestingly, the heterozygous fish have statistically significant increases in trabeculae length compared with their wild-type siblings in the 2% alcohol 6 hpf-48 hpf group. In addition to defects in the neurocranium, we observed malformations in the viscerocranium in both treatment groups exposed to 2% alcohol, including dysmorphology of the ceratohyal and Meckel’s cartilage (2% alcohol 6hpf-48hpf group +/+ n = 2/9, +/− n = 3/15, −/− n = 2/6 and 2% alcohol 6hpf-4dpf +/+ n = 2/6, +/− n = 8/14, −/− 2/8). These phenotypes were independent of tp53 gene dosage. Collectively, these data indicate that loss of tp53 is protective against some, but not all, alcohol-induced facial defects.

FIGURE 4.

Loss of tp53 protects against alcohol-induced reduction in trabeculae length in zebrafish. Shaded boxes represent the interquartile range, while horizontal lines represent the median values for tp53 wild-type (+/+, dark gray), heterozygous (+/−, gray), and homozygous mutant (−/−, light gray) genotypes, following no treatment (left set of bars), or different alcohol concentrations and durations of exposure. Each circle represents one trabecula. Statistical comparisons are indicated as inset text

TABLE 4.

Effects of alcohol and tp53 genotype on trabeculae length and eye width in zebrafish

Exposure Group	tp53 Genotype	Mean Trabecula Length (µm)	Standard Deviation	Sample Size	Mean Eye Width (µm)	Standard Deviation	Sample Size
No Treatment	(+/+)	159.8	19.7	28	313.4	11.4	26
	(+/−)	163.4	16.5	33	315.6	4.6	32
	(−/−)	163.6	25.4	20	308.2	22.5	20
1.5% Alc, 6 hpf−4 dpf	(+/+)	113.5	31.1	14	277.6	43.1	14
	(+/−)	115.1	36.4	30	267.3	43.3	30
	(−/−)	145.0*	25.6	16	293.4	45.5	16
2% Alc, 6hpf−48hpf	(+/+)	86.4	12.7	16	225.7	10.9	16
	(+/−)	102.3*	17.6	26	232.9	8.2	24
	(−/−)	111.8*	20.2	11	261.1*	37.8	12
2% Alc, 6 hpf−4 dpf	(+/+)	89.7	15.9	12	228.7	11.6	11
	(+/−)	101.0	16.6	27	224.0	8.9	26
	(−/−)	120.4*	20.9	16	237.9*	6.0	12

Note: Emboldened data with asterisks indicate significance versus (+/+). Emboldened, italicized data with asterisks indicate significance versus (+/+) and (+/−).

^*p < 0.05, following a one-way ANOVA.

TABLE 5.

Analysis of variance (ANOVA) of zebrafish eyes and trabeculae

Exposure Group	tp53 Genotype Comparison	F-values	Adjusted p-value Eye Width	F-values	Adjusted p-value Trabeculae Length
No Treatment	(+/+): (+/−)		N.D.		N.D.
	(+/+): (−/−)	F (2, 73) = 1.9; p = 0.16	N.D.	F (2, 76) = 0.31; p = 0.73	N.D.
	(+/−): (−/−)		N.D		N.D.
1.5% Alc, 6 hpf−4 dpf	(+/+): (+/−)		N.D.		0.99
	(+/+): (−/−)	F (2, 55) = 1.8; p = 0.18	N.D.	F (2, 55) = 4.9; p = 0.01	0.03*
	(+/−): (−/−)		N.D.		0.02*
2% Alc, 6 hpf−48 hpf	(+/+): (+/−)		0.50		0.28
	(+/+): (−/−)	F (2, 47) = 12.1; p < 0.001	<0.001*	F (2, 48) = 8.1; p < 0.001	0.001*
	(+/−): (−/−)		<0.001*		0.01*
2% Alc, 6hpf−4dpf	(+/+): (+/−)		0.33		0.20
	(+/+): (−/−)	F (2, 44) = 9.9; p = <0.001	0.04*	F (2, 49) = 11.1; p < 0.001	<0.001*
	(+/−): (−/−)		<0.001*		0.003*

Note: Emboldened data with asterisks indicate statistically significant differences (p < 0.05). N.D., not determined because of a nonsignificant ANOVA.

Loss of tp53 protects against eye defects in alcohol-exposed zebrafish

As in mice, alcohol exposure can result in eye defects in zebrafish and consistent with our findings in mice, we find tp53-alcohol interactions in these eye defects. As with the craniofacial skeleton, loss of tp53 does not reduce the overall width of the eye (Figure 5; see Table 4 for descriptive statistics and Table 5 for ANOVA results). We found highly variable effects on eye size in zebrafish exposed to 1.5% alcohol from gastrulation to 4 dpf. Likely due to this variability, we detect no statistically significant differences in this treatment group (Figure 5; see Table 4 for descriptive statistics and Table 5 for ANOVA results). Wild-type fish treated with 2% alcohol from either 6 to 48 hpf or 6 hpf-4 dpf had severe reductions in eye width (Figure 5; see Table 4 for descriptive statistics and Table 5 for ANOVA results).

FIGURE 5.

Loss of tp53 protects against alcohol-induced reduction in eye width in zebrafish. Shaded boxes represent the interquartile range, while horizontal lines represent the median values for tp53 wild-type (+/+, dark gray), heterozygous (+/−, gray), and homozygous mutant (−/−, light gray) genotypes, following no treatment (left set of bars), or different alcohol concentrations and durations of exposure. Each circle represents one eye. Statistical comparisons are indicated as inset text

Heterozygosity for tp53 did not modify the effect of alcohol on eye width in fish treated with 2% alcohol from either 6 to 48 hpf or 6 hpf-4 dpf (Figure 5; see Table 4 for descriptive statistics and Table 5 for ANOVA results). However, mutants from these same two treatment groups were significantly protected from eye defects (Figure 5; see Table 4 for descriptive statistics and Table 5 for ANOVA results). We found that independent of genotype, there was a strong correlation between anterior neurocranium malformations and eye width (Figure S4), in that zebrafish with more severe neurocranium malformations also had smaller eyes. We also observed a high correlation between eye width and trabeculae length (adjusted R² = 0.69, Figure S5). This corroborates what is known in human (reviewed in Abdelrahman & Conn, 2009) and mouse (Parnell et al., 2006) and may provide a complementary approach for quantifying alcohol teratogenesis in zebrafish.

Loss of tp53 protects against cell death in the cranial neural crest and eye of alcohol-exposed zebrafish

In order to determine whether the partial rescue we observe in tp53 mutants results from reduced cell death, we evaluated the number of apoptotic cell nuclei by TUNEL stain. By TUNEL staining zebrafish in the neural crest labeling Tg(sox10:eGFP) transgenic line, we were able to quantify neural crest cell death in the first pharyngeal arch and frontonasal stream at 24 hpf, during neural crest cell migration. We observed a clear reduction in TUNEL-positive nuclei in the cranial neural crest of 1.5% alcohol-exposed tp53 mutants compared with heterozygotes and wild types, as well as a reduction in heterozygotes compared with wild-type embryos (Figure 6D). This corroborates the changes we observe in trabeculae length (Figure 4). We also observed a reduction in apoptosis in the eyes of tp53 mutants compared with their heterozygous and wild-type siblings, matching the general trend we identified in eye width.

FIGURE 6.

Loss of tp53 protects against alcohol-induced cell death in zebrafish embryos. Representative images of 24 hpf TUNEL stained Tg(sox10:eGFP) transgenic embryos are shown for: (A) tp53 wild-type (+/+); (B) heterozygous (+/−, gray); and (C) homozygous mutant (−/−) genotypes. (D) Shaded boxes represent the interquartile range, while horizontal lines represent the median number of TUNEL-positive cells in the cranial neural crest (left bars) and eye (right bars) of tp53 wild-type (+/+, dark gray), heterozygous (+/−, gray), and homozygous mutant (−/−, light gray) genotypes, following 1.5% alcohol from 6 to 24 hpf. Each circle represents one individual. Statistical comparisons are indicated as inset text

DISCUSSION

The current study using the Tp53 knockout mouse and a tp53 mutant zebrafish highlights the role of this protein in the craniofacial and brain defects caused by gastrulation stage alcohol exposure. The reduced sensitivity of both Tp53 heterozygous and homozygous mutant mice to alcohol-induced eye and craniofacial defects supports the hypothesis that Tp53 is involved in the cascade of cellular events that follow PAE. These data are consistent with previous teratogenicity studies in Tp53 knockout mice following cyclophosphamide or 2-chloro-2′deoxyadenosine exposure (Pekar et al., 2007; Wubah et al., 1996). However, Tp53 deletion does not protect against all teratogens because Tp53 knockout mice were previously shown to be more sensitive to radiation, benzo[a]pyrene, and hydroxyurea (El Husseini & Hales, 2018a; Nicol et al., 1995; Norimura et al., 1996). These teratogen-specific effects of Tp53 deletion suggest that Tp53 may facilitate some teratogenic mechanisms, while attenuating others. The fact that the zebrafish tp53 mutant largely confirmed the findings in the mouse strongly supports Tp53 as a conserved mechanism of alcohol teratogenesis.

Prenatal alcohol-induced eye defects are consistently observed in animal models of early gestational alcohol exposure (e.g., Brennan & Giles, 2013; Dlugos & Rabin, 2007; Muralidharan et al., 2018; Parnell et al., 2006; Zhang et al., 2011) and have a high degree of external validity to human dysmorphisms, since reduced palpebral fissure length (PFL) is a hallmark facial feature of FAS (Del Campo & Jones, 2017). This PFL is reflective of alcohol-induced microphthalmia, but optic nerve hypoplasia and reduced visual acuity are also common following PAE (Stromland, 1987). Eye defects caused by gastrulation or early neurulation alcohol exposure in the mouse are thought to reflect tissue loss in the developing forebrain/frontonasal prominence, prior to the emergence of the optic sulci and specification of the optic primordium (Dunty et al., 2001; Parnell et al., 2006). The longer duration of alcohol exposure in the zebrafish study makes it difficult to ascertain a critical window for alcohol’s acute pathogenic effects on the developing eye, so it is likely that alcohol alters several processes of eye development, including gastrulation and subsequent formation of the neural ectoderm and the neural crest cells, both of which are precursors to the major structural components of the eye. Indeed, apoptotic cell death was observed in both the eye and the neural crest, replicating previous findings following various alcohol exposure durations and concentrations (Eason et al., 2017; Kashyap et al., 2007; Muralidharan et al., 2015). Tp53 regulated cell death in the eye and neural crest, since the full and partial Tp53 gene deletion attenuated the effects of alcohol, which can help explain its protective effect on the eye, the neurocranium, and the trabeculae. In the mouse, midline tissue apoptosis is the major morphological consequence of gastrulation stage PAE (Kotch & Sulik, 1992); the current results suggest that this requires Tp53, as it does in zebrafish. A recent RNA-sequencing study demonstrated that gastrulation stage alcohol exposure upregulates Tp53 pathways 12 h later (Boschen et al., 2021), a time that correlates with the largest amount of excessive apoptotic cell death (Kotch & Sulik, 1992). The protective effect of the Tp53 gene deletion suggests that the activation of Tp53 during this time is related to the pro-apoptotic effect of Tp53, rather than to its many other effects, such as cell cycle arrest and DNA repair, which are thought to reduce long-lasting genetic damage caused by repeated exposure to cellular stressors. Thus, we hypothesize that during gastrulation, Tp53 gene deletion protects the ectodermal cells that will give rise to the neural crest cells and that later in development, Tp53 can protect the neural crest cells themselves. This dual protection sustains normal organization and size of the craniofacies, even in the presence of PAE.

While apoptosis may be the critical final event that determines craniofacial teratogenesis, our prior research indicates several simultaneous events are also necessary, including the inhibition of growth factor pathways. We have previously shown that alcohol teratogenesis in zebrafish involves the inhibition of the phosphoinositide 3 kinase (PI3K)-mechanistic target of rapamycin (mTOR) pathway, downstream of platelet-derived growth factor receptor a (Pdgfra) (McCarthy et al., 2013). Zebrafish, mutant or heterozygous for pdgfra, had severe craniofacial deficits when exposed to a normally subteratogenic alcohol concentration, indicating that Pdgfra and the stimulation of this growth factor pathway protect against the effects of alcohol. Alcohol can decrease signaling molecules immediately downstream of Pdgfra, including PI3K, AKT, and mTOR (Hong-Brown et al., 2010; Vary et al., 2008; Xu et al., 2003). Of these molecules, AKT acts as a hub to stimulate mTOR signaling, restrict pro-apopotic proteins (BCL-2 and BAX), and, most relevant to the present study, regulate Tp53 signaling by activating MDM2. Thus, when alcohol inhibits the PI3K-mTOR pathway, there is an overall shift toward reduced growth and proliferation and increased apoptosis, mediated in part by Tp53.

One limitation of these experiments is the use of a constitutive Tp53 mutant, rather than a conditional gene deletion that is restricted to the critical time window for alcohol’s effects on apoptotic pathways. Thus, we cannot rule out possible compensatory effects, such as overall upregulation in growth pathways, or identify precisely when the protective effect of Tp53 deletion occurs. The presence of anterior and posterior neural tube defects (i.e., exencephaly and spina bifida) as well as genital hyperplasia, in a small percentage of Tp53 knockout mice supports the idea that growth and proliferation pathways may indeed be upregulated. However, these defects were observed exclusively in homozygous females and were neither exacerbated nor ameliorated by alcohol exposure, indicating that Tp53 interacts with sexually dimorphic genes (Chen et al., 2008), in a manner that appears unrelated to the effects of alcohol. Moreover, the protective effect of Tp53 gene deletion was observed even in the heterozygous mice, a genotype with no obvious morphological consequences. The heterozygous mice were about half as sensitive as the wild-type mice overall and showed no severe manifestations of alcohol exposure. A similar protection the heterozygous was found in zebrafish, with a higher percent of mild facial phenotypes and a lower percent of failed fusions and trabeculae breaks compared to the wild-type zebrafish following alcohol exposure. This gene dosage effect provides strong evidence that only a minor decrease in Tp53 is sufficient to diminish alcohol teratogenesis.

Both the zebrafish and mouse dysmorphologies are due to alcohol exposure beginning in gastrulation and it is possible that Tp53 may have different roles after alcohol exposure at other developmental stages which produce different dysmorphologies. First of all, the response of Tp53 itself may depend on the timing of alcohol exposure. While Tp53 levels increased following gastrulation alcohol exposure, Tp53 mRNA was downregulated following a mid-neurulation (GD 9) exposure (Boschen et al., 2020; Leitzinger et al., 2020). Similarly, maternal alcohol drinking from GD 6 to GD 16, increased Tp53 expression in fetal cortices (Kuhn & Miller, 1998); but lower levels were found following maternal alcohol drinking from GD 10 to GD 13 (Hicks & Miller, 2019). Studies on the neural crest, which develops just after gastrulation, find that alcohol increases protein levels of Tp53 (Flentke et al., 2019) and Tp53-mediated apoptosis (Chen et al., 2015) and that blocking the transcriptional pathway induced by Tp53 can prevent apoptosis in specific regions (Flentke et al., 2019). Likewise, in a neuronal stem cell model, Tp53 gene knockdown has also been shown to protect against apoptosis (Miller, 2019). However, in a postnatal alcohol exposure procedure modeling alcohol in the third trimester, Tp53 gene deletion did not have a protective effect on alcohol-induced apoptosis in the hippocampus, cortical regions, or cerebellar internal granule cell cultures (Camargo Moreno et al., 2017; Ghosh et al., 2009; Nowoslawski et al., 2005). Instead, Tp53 gene deletion modified the expression of genes related to the cell cycle and DNA damage, rather than apoptosis directly (Camargo Moreno et al., 2017). The above examples of how Tp53’s role in alcohol teratogenesis may depend on when the alcohol exposure occurs touch upon a common theme in Tp53 research, namely that specific Tp53 responses depend on a number of specific circumstances (Jain & Barton, 2018). In the gastrulation stage embryo exposed to alcohol, Tp53 drives regions undergoing developmentally appropriate, low levels of apoptosis toward excessive apoptosis which eventually leads to specific midline facial and eye defects.

The mechanisms of Tp53-induced apoptosis involve both transcriptional changes and direct protein interactions. Tp53 acts upon the intracellular (intrinsic) cell death pathway by binding to the anti-apoptotic Bcl-2 family proteins, Bcl-2, Bcl-xL, and Mcl-1, thereby derepressing the pro-apoptotic proteins Bax and Bak and facilitating mitochondrial membrane permeabilization. Transcriptionally, Tp53 also upregulates the expression of BAX, and its regulators NOXA, and PUMA, as well as death receptors involved in the extracellular cell death pathway and other modulators of nonapoptotic cell death (Yu & Zhang, 2005). Our results implicate the subsequent activity of these other cell death pathway members in alcohol teratogenesis. However, we do not know how alcohol stimulates Tp53 activity and future experiments can address whether compounds, such as N-acetylcysteine (Parnell et al., 2010), that prevent dysmorphologies when given concurrently with alcohol, also attenuate Tp53 activity.

Overall, our results indicate that Tp53 has a critical role in alcohol-induced teratogenesis in two animal models of gastrulation stage exposure, the primary defects of which were craniofacial, eye, and brain dysmorphologies. In light of the heterogeneity of prior studies on Tp53 and other models of PAE, the contribution of Tp53 to fetal alcohol spectrum disorders likely depends on the timing of the alcohol exposure. One might speculate that Tp53 mutations could protect individuals from some fetal alcohol effects, but render them more sensitive to others, the particular response being determined by the predominant actions of Tp53 at the time of the alcohol exposure (e.g., apoptosis vs. DNA repair). Regarding alcohol exposure that occurs in the early stages of pregnancy, mutations which decrease Tp53 signaling could cause exposed individuals to present without the hallmark features of FAS, but still have other symptoms consistent with alcohol exposure. The conserved actions of Tp53 in the mouse and zebrafish suggest that Tp53 also plays a role in the responses to fetal alcohol in humans and should prompt further study of Tp53 and/or its pathway in clinical populations.