Craniofacial Defects in Embryos with Homozygous Deletion of Eftud2 in Their Neural Crest Cells Are Not Rescued by Trp53 Deletion

Embryos with homozygous mutation of Eftud2 in their neural crest cells (Eftud2ncc−/−) have brain and craniofacial malformations, hyperactivation of the P53-pathway and die before birth. Treatment of Eftud2ncc−/− embryos with pifithrin-α, a P53-inhibitor, partly improved brain and craniofacial development. To uncover if craniofacial malformations and death were indeed due to P53 hyperactivation we generated embryos with homozygous loss of function mutations in both Eftud2 and Trp53 in the neural crest cells. We evaluated the molecular mechanism underlying craniofacial development in pifithrin-α-treated embryos and in Eftud2; Trp53 double homozygous (Eftud2ncc−/−; Trp53ncc−/−) mutant embryos. Eftud2ncc−/− embryos that were treated with pifithrin-α or homozygous mutant for Trp53 in their neural crest cells showed reduced apoptosis in their neural tube and reduced P53-target activity. Furthermore, although the number of SOX10 positive cranial neural crest cells was increased in embryonic day (E) 9.0 Eftud2ncc−/−; Trp53ncc−/− embryos compared to Eftud2ncc−/− mutants, brain and craniofacial development, and survival were not improved in double mutant embryos. Furthermore, mis-splicing of both P53-regulated transcripts, Mdm2 and Foxm1, and a P53-independent transcript, Synj2bp, was increased in the head of Eftud2ncc−/−; Trp53ncc−/− embryos. While levels of Zmat3, a P53- regulated splicing factor, was similar to those of wild-type. Altogether, our data indicate that both P53-regulated and P53-independent pathways contribute to craniofacial malformations and death of Eftud2ncc−/− embryos.


Introduction
Pathogenic variants in EFTUD2, which encodes a GTPase and a core component of the U5 subunit of the spliceosome [1], are responsible for the congenital syndrome Mandibulofacial Dysostosis with Microcephaly (MFDM) (OMIM#610536). MFDM patients have a range of craniofacial abnormalities that include micrognathia, small dysplastic pinnae, malar hypoplasia, hearing loss, microcephaly and developmental delay [2]. Mutant mice carrying a heterozygous loss-of-function mutation in Eftud2 are viable and fertile, and do not model MFDM [3]. On the other hand, embryos that are homozygous for the same loss of function mutation fail to implant and arrest at the blastocyst stage [3]. However, Eftud2 ncc−/− embryos with Wnt1-Cre2 mediated deletion of exon 2 of Eftud2 have severe brain and craniofacial malformations, in the same precursor that is affected in MFDM patients, thereby modelling those head and face defects that are found in patients [4].
RNAseq analysis using heads of Eftud2 ncc−/− embryos, prior to the onset of malformations, revealed increased skipping of exon 3 of Mdm2, a master regulator of Trp53 (P53). Additionally, the levels of nuclear P53 and P53-regulated genes, including Mdm2, were increased in Eftud2 ncc−/− embryos and in O9-1 neural crest cells after siRNA mediated knockdown of Eftud2. Furthermore, levels of P53 target genes were reduced when fulllength Mdm2 was overexpressed in O9-1 cells with Eftud2 knockdown, and treatment of pregnant females with pifithrin-α improved craniofacial development in Eftud2 ncc−/− embryos [4]. Therefore, we proposed that craniofacial malformations that are found in Eftud2 ncc−/− embryos was a consequence of the mis-splicing of Mdm2 leading to hyperactivation of the P53 pathway.
Herein, we show that P53 level and activity are reduced in Eftud2 ncc−/− embryos that were treated with pifithrin-α. These embryos also had reduced apoptosis in their neural tube. We generated Eftud2 ncc−/− ; Trp53 ncc−/− double mutant embryos and showed that they had reduced nuclear P53 and P53-activity, no changes in mitosis, and reduced apoptosis in their neural tube. Furthermore, although Eftud2 ncc−/− ; Trp53 ncc−/− mutants have more SOX10 positive cranial neural crest cells, when compared to Eftud2 ncc−/− (Eftud2 ncc−/− ; Trp53 +/+ ) embryos, they are morphologically abnormal from embryonic day (E)9.5 to E14.5 and they die before birth, similar to the Eftud2 ncc−/− mutants. Additionally, Eftud2 ncc−/− ; Trp53 ncc−/− embryos have a significant increase in exon-skipping in P53-regulated transcripts, including Mdm2 and FoxM1, and a non-significant increase in levels of the P53independent transcript Synj2bp, when compared to controls or Eftud2 ncc−/− mutants. Our data indicate that P53 hyperactivation contributes to apoptosis in the neural tube and suggests that P53 attenuates mis-splicing in Eftud2 mutant cells. Notably our study shows that craniofacial malformations and death of Eftud2 ncc−/− mutant embryos are independent of P53-activation.

Pifithrin-α Reduces Levels of Nuclear P53, P53-Target Genes and Apoptosis in the Neural Tubes of Eftud2 ncc−/− Embryos
We previously showed that pifithrin-α treatment from E6.5 -E8.5 partially improves brain and craniofacial development in E9.5 Eftud2 ncc−/− mutants [4]. To determine if reduced P53 is responsible for this rescue, we examined P53 accumulation and activity in these previously generated E9.5 embryos. The levels of the three P53-regulated genes Ccng1, Trp53inp1, and Phlda3 which were significantly upregulated in heads of E9.0 Eftud2 ncc−/− embryos [4], were also increased in the E9.5 mutant embryos from vehicle-treated females when compared to their control littermates ( Figure 1A-C), although this difference was only significant for Ccng1 and Trp53inp1 ( Figure 1A,B). In contrast, the levels of these P53target genes were comparable in the heads of pifithrin-α-treated controls and Eftud2 ncc−/− embryos ( Figure 1A-C). We examined nuclear P53 accumulation in the pifithrin-α-treated embryos and found a significant increase of P53 positive nuclei in the neural tube, anterior to rhombomere 4, of the vehicle-treated Eftud2 ncc−/− embryos when compared to the controls ( Figure 1D,E), but not in their first pharyngeal arch ( Figure 1F,G). In contrast, the number of P53 positive cells in the neural tube or pharyngeal arches of pifithrin-α treated Eftud2 ncc−/− embryos was comparable to those of the controls ( Figure 1D-G). These findings show that pifithrin-α treatment abolished increased P53-accumulation and activity in the neural tube of Eftud2 ncc−/− embryos.
The number of dying cells found in the head of E9.5 Eftud2 ncc−/− embryos was significantly increased when compared to the controls [4]. To identify the cell population undergoing apoptosis and to determine if pifithrin-α treatment modulates this increase, we quantified the number of cleaved Caspase-3-positive (apoptotic cells) in the neural tube, anterior to rhombomere 4, and the first pharyngeal arch of vehicle and pifithrin-α treated E9.5 embryos ( Figure 1H-K). At this stage, Eftud2 ncc−/− embryos from vehicle treated females exhibited an increase in cleaved Caspase-3 positive cells in their neural tube and first pharyngeal arch, when compared to controls ( Figure 1H,J). In contrast, although the proportion of apoptotic cells in both structures was increased in Eftud2 ncc−/− embryos that were treated with pifithrin-α the difference was only significant in the pharyngeal arches Figure 1. Treatment of Eftud2 ncc−/− embryos with pifithrin-α from E6.5-E8.5 reduces P53-activity, nuclear P53 accumulation, and apoptosis. (A-C) RT-qPCR analysis revealed significant increases in the levels of (A) Ccng1 and (B) Trp53inp1, and a non-significant increase in the level of (C) Phlda3 in the heads of Eftud2 ncc−/− embryos (mut, n = 3) treated with vehicle (veh), when compared to the controls (ctl, n = 3). In pifithrin-α (pif)-treated mutants (mut. n = 6), levels of these genes are similar to the controls (ctl, n = 4). The Y-axis indicates fold change over the control, the error bars represent SEM, * p < 0.05 by t-test. (D,F) Quantification of nuclear P53-positive cells in the neural tube (D) and first pharyngeal arch (F) of the vehicle and the pifithrin-α treated control (n = 7 veh, n = 5, pif) and Eftud2 ncc−/− embryos (n = 6 veh, n = 4, pif). The percentage of P53-positive nuclei was significantly increased in the neural tube of the vehicle treated mutant embryos when compared to the controls. Each dot represents the average percentage of positive cells in a single embryo (* p < 0.01 by t-test). (E,G) Representative images of P53 nuclear staining in vehicle (left) or pifithrin-α (right)-treated embryos. Sections were counterstained with nuclear fast red (red), P53-staining is in brown. The arrows indicate P53 positive cells. (H) Quantification of cleaved caspase-3-positive cells showing significant increase in the percentage of apoptotic cells in the neural tube of vehicle-treated Eftud2 ncc−/− embryos (n = 7) and (J) in the first pharyngeal arch of pifithrin-α-treated Eftud2 ncc−/− embryos (n = 4), when compared to the control (n = 5, veh, n = 5, pif). Each dot represents the average percentage of positive cells in a single embryo (* p < 0.05, ** p < 0.01 by t-test). (I,K) Representative images of cleaved Caspase-3 staining of vehicle or pifithrin-α treated embryos. Arrows indicate cleaved Caspase-3 positive cells. (Genotypes of the control embryos included Eftud2 loxp/− or Eftud2 loxp/+ ). Scale bar = 25 µm. nt = neural tube, pa = pharyngeal arch.

Removing Two Alleles of Trp53 Decreases P53-Activity and Apoptosis in E10.5 Mutant Embryos
We next used a previously described mutant mouse line carrying Trp53 conditional mutation [5] and Wnt1-Cre2 to remove Trp53 in the neural crest cells and confirm that its loss improves craniofacial development in Eftud2 ncc−/− embryos. We used two different mating schemes to generate embryos with heterozygous or homozygous loss of Trp53 (Eftud2 ncc−/− ; Trp53 ncc+/− or Eftud2 ncc−/− ; Trp53 ncc−/− (see Section 4) (Table S1)) and compared them to Eftud2 ncc−/− embryos that were generated in our previous study [4].
We postulated that craniofacial development would be improved in E10.5 Eftud2 ncc−/− ; Trp53 ncc+/− or Eftud2 ncc−/− ; Trp53 ncc−/− embryos with reduced P53-levels and activity. Therefore, we first examined the levels of the P53-target genes Ccng1, Trp53inp1, and Phlda3 which were significantly increased in E9.0 and E9.5 Eftud2 ncc−/− embryos. At E10.5, the levels of these genes were reduced in Eftud2 ncc−/− ; Trp53 ncc−/− when compared to Eftud2 ncc−/− , and comparable to those of the controls, as expected ( Figure 2A). Next, we evaluated if nuclear P53 accumulation was reduced in the neural tube, anterior to rhombomere 4, and the first pharyngeal arch of E10.5 embryos with genetic deletion of Trp53 ( Figure 2B,C). We found a non-significant increase in P53 accumulation in the neural tubes of Eftud2 ncc−/− and Eftud2 ncc−/− ; Trp53 ncc+/− embryos that was reduced in Eftud2 ncc−/− ; Trp53 ncc−/− mutants ( Figure 2B). On the other hand, nuclear P53 accumulation in the first pharyngeal arch of Eftud2 ncc−/− embryos was not affected by the Trp53 genotype ( Figure 2C). We next tested if the loss of P53 impacts apoptosis or proliferation in the head of double mutant embryos. Apoptosis, as detected by nuclear cleaved-Caspase 3, was increased in the neural tubes of Eftud2 ncc−/− and Eftud2 ncc−/− ; Trp53 ncc+/− mutant embryos, although the difference was only significant in Eftud2 ncc−/− ; Trp53 ncc+/− , when compared to controls ( Figure 2D). However, this increase was no longer observed in Eftud2 ncc−/− ; Trp53 ncc−/− embryos ( Figure 2D). At E10.5, apoptosis was not significantly increased in the first pharyngeal arch of embryos with any of the three mutant genotypes ( Figure 2E). Additionally, no significant changes were observed in proliferation in the head of E10.5 mutant embryos with mutations in both Eftud2 and Trp53 ( Figure S2). Taken together, these data indicate that deletion of a single allele of Trp53 was not sufficient to reduce P53 accumulation or apoptosis in the neural tube. However, homozygous deletion of both Eftud2 and Trp53 did abolish the increase of nuclear P53 and the increase in apoptosis found in the neural tube of Eftud2 ncc−/− embryos. Thus, we conclude that increased P53 activity mediates apoptosis in the neural tube of Eftud2 ncc−/− embryos. a non-significant increase in P53 accumulation in the neural tubes of Eftud2 ncc−/− and Ef-tud2 ncc−/− ; Trp53 ncc+/− embryos that was reduced in Eftud2 ncc−/− ; Trp53 ncc−/− mutants ( Figure 2B). On the other hand, nuclear P53 accumulation in the first pharyngeal arch of Eftud2 ncc−/− embryos was not affected by the Trp53 genotype ( Figure 2C). To determine if reduction of P53 activity and apoptosis in the neural tube results in an increase in the number of post-migratory cranial neural crest cells, we examined expression of SOX10, which is expressed in these cells, in E9.0 embryos. As shown in Figure 3A-C, SOX10-expression was reduced in the frontonasal mass and first pharyngeal arch of Eftud2 ncc−/− mutant embryos (n = 4) ( Figure 3B) when compared to controls (n = 4) ( Figure 3A). However, SOX10 immunoreactivity in Eftud2 ncc−/− ; Trp53 ncc−/− mutants (n = 4) ( Figure 3C) was comparable to that of controls. Thus, we concluded that Eftud2 ncc−/− ; Trp53 ncc−/− mutant embryos have more SOX10-positive post-migratory neural crest cells than Eftud2 ncc−/− embryos.

FoxM1, a P53-Target, Is Abnormally Spliced in Eftud2 ncc−/− Mutant Embryos
In our previous RNAseq analysis using the heads of morphologically normal E9.0 Eftud2 ncc−/− embryos, we showed that mis-splicing of Mdm2 contributed to upregulation of P53 [4]. We postulated that mis-splicing of a P53-independent or dependent target contributes to these malformations. Therefore, we further analyzed our previously published RNAseq data and found a significant increase in the skipping of exon 7 (p = 8.63 × 10 −6 , FDR = 0.005) of FoxM1, which is negatively regulated by P53 [7], in Eftud2 ncc−/− embryos. RT-PCR analysis revealed that a transcript with the predicted skipping of exon 7 was present and enriched in Eftud2 ncc−/− mutants, and not found in controls ( Figure S5A). Furthermore, at E9.5, mis-splicing of FoxM1 was significantly increased in Eftud2 ncc−/− ; Trp53 ncc−/− embryos when compared to controls or Eftud2 ncc−/− mutants with wild-type Trp53 ( Figure 6A). However, although mis-splicing and expression levels of FoxM1 was increased at E10.5, the expression of its target Cdc25b was non-significantly reduced in the heads of E10.5 Eftud2 ncc−/− ; Trp53 ncc−/− embryos ( Figure S5B,C).
Thus, we conclude that increased mis-splicing of FoxM1 was unlikely to be a major contributor to craniofacial defects in Eftud2 ncc−/− ; Trp53 ncc−/− embryos.

Mis-Splicing Is Increased in Eftud2 ncc−/− ; Trp53 ncc−/− Embryos
Since we found an unexpected increase in mis-splicing of FoxM1 in Eftud2 ncc−/− ; Trp53 ncc−/− mutant embryos, we next evaluated exon-skipping in two transcripts which were mis-spliced in Eftud2 ncc−/− mutants wild-type for Trp53. RT-PCR analysis of the heads of mutant and control embryos revealed a significant increase in the proportion of Mdm2 transcripts missing exon 3 in Eftud2 ncc−/− ; Trp53 ncc+/− and in Eftud2 ncc−/− ; Trp53 ncc−/− embryos, when compared to controls or Eftud2 ncc−/− that are wild-type for Trp53 ( Figure 6B). To determine if P53-independent targets also showed increased mis-splicing, we next interrogated the list of genes that were alternatively spliced in our previous RNAseq analysis from both E9.0 and E9.5 head embryos [4]. Amongst the transcripts with a FDR < 0.01, an inclusion level difference >0.1 and with an average number of reads close to 100, we selected Synj2bp for further validation. We designed primers flanking the skipped exon and used RT-PCR to quantify the ratio of the Synj2bp full-length and short transcripts in our samples. Although not significant, we found a higher proportion of the shorter transcript without exon 2 of Synj2bp in Eftud2 ncc−/− ; Trp53 ncc+/− and in Eftud2 ncc−/− ; Trp53 ncc−/− mutant heads when compared to controls or Eftud2 ncc−/− embryos wild-type for Trp53 ( Figure 6C). Altogether these data indicate that the presence of wild-type Trp53 might attenuate mis-splicing in Eftud2 ncc−/− mutant cells. 6A). However, although mis-splicing and expression levels of FoxM1 was increased at E10.5, the expression of its target Cdc25b was non-significantly reduced in the heads of E10.5 Eftud2 ncc−/− ; Trp53 ncc−/− embryos ( Figure S5B,C). 2.9. Zmat3 Is Not Upregulated in Eftud2 ncc−/− ; Trp53 ncc−/− Embryos ZMAT3 is a P53-regulated RNA splicing factor that regulates exon skipping in a number of transcripts, including exon 3 of Mdm2 [8,9]. Therefore, we examined the expression of this gene to determine if it is responsible for the increased mis-splicing that was found in Eftud2 ncc−/− mutants with reduced P53 activity. At E9.5 and E10.5, Zmat3 expression was significantly increased in Eftud2 ncc−/− mutants that are wild-type for Trp53, when compared to controls ( Figure 6D,E). However, homozygous deletion of Trp53 attenuated this increase at E9.5 ( Figure 6D). Additionally, although levels of Zmat3 were increased in E10.5 Eftud2 ncc−/− ; Trp53 ncc−/− mutants ( Figure 6E), this difference was not significant when compared to controls. Thus, we conclude that P53 reduces mis-splicing in Eftud2 ncc−/− mutant cells in a ZMAT3-independent fashion.

Discussion
Herein, we investigated the contribution of the P53-pathway to craniofacial malformations in Eftud2 ncc−/− mutant embryos. We show that we can reduce P53 levels and activity in the neural tube by treating Eftud2 ncc−/− embryos with pifithrin-α or using Wnt1-Cre2 to delete both alleles of Trp53. Furthermore, although reducing P53-activity attenuated apoptosis in the neural tube of Eftud2 ncc−/− embryos, it did not rescue craniofacial development. We also show that mis-splicing is increased in Eftud2 ncc−/− ; Trp53 ncc−/− mutants and that this increase was not associated with a significant change in levels of the P53-regulated splicing factor, Zmat3. These data indicate that P53-independent pathways contribute to craniofacial defects in Eftud2 ncc−/− mutants and reveal a previously unappreciated role for P53 in regulating mis-splicing.
The increased accumulation and activity of P53 secondary to mutations in genes that are involved in a diverse array of cellular processes, including splicing, contribute to malformations in animal models of developmental syndromes [10,11]. In a few cases, increased P53 activity was also found in clinical samples from patients with developmental syndromes [11]. More specifically, mouse models of neurocristopathies have overactivation of the P53 pathway and reducing the levels of P53 improve and in some cases rescue malformations in these models [12,13]. A similar partial rescue was found when Eftud2 was mutated in zebrafish [14,15]. Based on our previous findings that pifithrin-α improved neural tube and craniofacial development in Eftud2 ncc−/− embryos we postulated that reducing P53 in mouse would lead to a similar outcome [4].
In the present study, we show that P53-activity was increased in the heads of E9.5 Eftud2 ncc−/− embryos and that this increase was attenuated by pifithrin-α. More specifically, we show that Eftud2 ncc−/− mutant embryos at this stage had increased nuclear P53 accumulation and apoptosis in the neural tube, but not the first pharyngeal arch. Therefore, and not surprisingly, treating these mutant embryos with pifithrin-α led to reduced apoptosis in the neural tube and not in the first pharyngeal arch. We previously showed that the perimeter of the first pharyngeal arch of pifithrin-α treated Eftud2 ncc−/− embryos is larger than those of vehicle-treated mutants [4]. Therefore, it is surprising that we found a significant increase in apoptosis in pharyngeal arches of embryos that were exposed to pifithrin-α from E6.5 to E8.5. Although this data suggest that P53-independent activity leads to the death of Eftud2 mutant cells in the pharyngeal arches, further work is needed to decipher how a loss of Eftud2 results in abnormal development of the first arch and its derivatives, including Meckel's cartilage. Although first identified as a P53 inhibitor, pifithrin-α can also have P53-independent activity [16]. In fact, its mechanism of action remains unclear and its specific effect on P53 may vary in different cell types. For instance, pifithrin-α was associated with cell-type specific phosphorylation of P53 and the differential expression of its target genes [16]. Moreover, it was reported to reduce intracellular reactive oxygen species (ROS) independently of P53 [16]. To rule out the contribution of a P53-independent mechanisms to the partial rescue of craniofacial development that were found in pifithrin-α-treated Eftud2 ncc−/− embryos, we used a genetic approach and deleted Trp53 in Eftud2 mutant cells.
When we used the Wnt1-Cre2 transgenic mouse line to remove both alleles of Eftud2 and one allele of Trp53 (Eftud2 ncc−/− ; Trp53 ncc+/− embryos), there was still an increase in nuclear P53 accumulation and in P53 activity. Additionally, most Eftud2 ncc−/− ; Trp53 nc+/− embryos resembled Eftud2 ncc−/− mutants, except for one morphologically normal embryo that was recovered at E14.5. Therefore, we cannot completely rule out the possibility that reducing the levels of P53 can prevent craniofacial defects in Eftud2 ncc−/− mutants. However, this is likely to be a very rare event. In fact, when all stages were combined, no additional morphologically normal Eftud2 ncc−/− ; Trp53 ncc+/− (n = 42) or Eftud2 ncc−/− ; Trp53 ncc−/− (n = 35) embryos were found. Unfortunately, molecular studies could not be performed on this embryo as the litter was collected for cartilage preparations.
Since Eftud2 and Trp53 both map to chromosome 11 and are only 33cM apart, it was difficult to generate both Eftud2 ncc−/− and Eftud2 ncc−/− ; Trp53 ncc−/− embryos from the same cross. Therefore, we generated Eftud2 ncc−/− ; Trp53 ncc+/− and Eftud2 ncc−/− ; Trp53 ncc−/− embryos in separate crosses and compared the morphological and the molecular defects that were found to those that we previously described in Eftud2 ncc−/− embryos. Our current studies show that P53 accumulation and activity were reduced in Eftud2 ncc−/− ; Trp53 ncc−/− mutant embryos, suggesting that the P53-pathway was efficiently inhibited. Furthermore, as was found for pifithrin-α treated Eftud2 ncc−/− mutants, apoptosis was significantly reduced in the neural tube of double homozygous mutant embryos. However, in double homozygous mutant embryos the onset of craniofacial malformations, and the size of the pharyngeal arches and the midbrain, were no different from Eftud2 ncc−/− embryos wild type for Trp53. Nor did a loss of P53 extend the life of Eftud2 ncc−/− embryos. Altogether, these data indicate that the malformations found in Eftud2 ncc−/− embryos are predominantly P53-independent. In the future, we will identify the pathways that are responsible for these malformations.
Nonetheless, our study suggests that the reduced apoptosis in the neural tube most likely leads to an increase in the number of neural crest cells in Eftud2 ncc−/− ; Trp53 ncc−/− embryos. Supporting this, we found reduced SOX10 expression in E9.0 Eftud2 ncc−/− mutants that were morphologically similar to controls. Moreover, removing both copies of Trp53 resulted in an increase of SOX10 expression in the head of Eftud2 ncc−/− ; Trp53 ncc−/− embryos. Using the ROSA26R reporter, we previously showed that the proportion of Cre-expressing cells was similar in control and E9.0 Eftud2 ncc−/− mutant embryos, suggesting that these embryos have a similar number of neural crest cells [4]. Our results from this study indicate that at this stage, Eftud2 ncc−/− mutant cranial neural crest cells have attenuated expression of SOX10, a marker of post-migratory neural crest cells. Thus, we propose that increased P53 activity in the cranial neural crest cells of E9.0 Eftud2 ncc−/− mutants as they exit the neural tube leads to the abnormal expression of proteins such as SOX10, important for their survival and patterning in the pharyngeal arches. Since removing P53 increases expression of SOX10 in Eftud2 ncc−/− mutant cells but does not rescue craniofacial development, we further postulate that SOX10 is not sufficient to protect Eftud2 ncc−/− mutant neural crest cells in the first pharyngeal arch from undergoing cell death at E9.5. However, although we assumed that the number of surviving neural crest cells is insufficient to rescue brain and craniofacial development, we cannot rule out the possibility that surviving neural crest cells cannot differentiate and form the cartilage and bones of the head and face. Furthermore, although we did not see increased splicing of Xbp1 suggesting that the IRE1α pathway was not activated, we cannot exclude that one of the remaining two-arms of the ER-stress pathway is activated and contributes to Caspase-3 independent cell death of Eftud2 ncc−/− mutant cells.
We next examined our RNAseq dataset for additional pathways/transcripts which may contribute to defects. We examined a role for FoxM1, a transcription factor which is essential for G2 progression into mitosis [17], that was mis-spliced in these mutants. FoxM1 is required for embryonic survival and normal craniofacial development [18], and is expressed in the head and the pharyngeal arches of E10.5 embryos [19]. However, its expression and that of its downstream target, Cdc25b, was not significantly changed in mutant embryos, ruling it out as a major contributor to craniofacial defects found in Eftud2 ncc−/− embryos that were wild-type or mutant for Trp53.
Surprisingly, the mis-splicing of at least three-transcripts: Foxm1, Mdm2 and Snynj2b was increased in Eftud2 ncc−/− ; Trp53 ncc−/− embryos, suggesting that wild-type P53 activity tampered mis-splicing. Therefore, we examined the expression of Zmat3 which encodes for an RNA binding protein that is regulated by P53 [9]. We hypothesized that the expression of this splicing factor would be further increased in Eftud2 ncc−/− ; Trp53 ncc−/− embryos, thus explaining the increased exon skipping that was found. However, Zmat3 overexpression was completely abrogated when Trp53 was deleted, consistent with the observation that it is regulated by P53 [9]. We propose that one, or more, yet to be identified splicing factors are negatively regulated by P53 and are over expressed in Eftud2 ncc−/− ; Trp53 ncc−/− embryos.
Identifying this gene(s) will be essential for determining how P53 attenuates splicing of Foxm1, Mdm2 and Snynj2b in Eftud2 mutant cells. Further RNAseq analysis using the heads of double mutant embryos will be needed to confirm a global increase in mis-splicing.

Mouse Lines
All the procedures and experiments were performed according to the guidelines of the Canadian Council on Animal Care and approved by the Animal Care Committee of the Montreal Children's Hospital. Wild-type CD1 mice (strain code 022) were purchased from Charles Rivers (Laval, QC, Canada) and wild-type C57Bl/6 mice (stock #000664) were purchased from Jackson Laboratories (Augusta, ME, USA). Wnt1-Cre2 mice on the 129S4 genetic background were purchased from Jackson's laboratory (stock# 022137). These Wnt1-Cre2 transgenic mice express Cre recombinase under the control of the mouse Wnt1, wingless-related MMTV integration site 1, promoter and enhancer [20]. The Trp53 tm1brn mouse line on the C57BL/6 genetic background with loxP sites flanking exons 2-10 of the Trp53 gene was purchased from Jackson's laboratory (Trp53 loxP/+ ) (stock# 008462) [5]. The generation and the characterization of the conditional Eftud2 em2Lajm (Eftud2 loxp/+ ) line on the inbred C57BL/6 and mixed CD1 genetic backgrounds, and the Eftud2 em1Lajm (Eftud2 +/− ) exon 2 deletion line on a mixed CD1 genetic background were described previously [3,4].
Cre recombinase activity was reported in the male germline of Wnt1-Cre2 mice on the 129S4 background [21]. Since the embryos carrying homozygous mutation of Eftud2 in all cells arrest pre-implantation, [3] we did not recover Eftud2 ncc−/− embryos with germline homozygous deletion of Eftud2. All embryos that were analyzed in this study were on a mixed genetic background following the multiple crosses that were needed to generate double homozygous mutants.

Collection of Embryos
For embryo collection, the day that a vaginal plug was seen was considered embryonic day 0.5 (E0.5). On the day of dissection, the embryos were removed from their extraembryonic membranes; for stages E8.5 to E10.5, the number of somites was counted under light microscope (Leica MZ6 Infinity1 stereomicroscope). The embryos were fixed in 1 or 4% paraformaldehyde at 4 • C overnight (unless otherwise stated), washed in PBS and kept at 4 • C. The yolk sacs were collected and used for genomic DNA extraction for genotyping.

Cartilage Preparation
To evaluate cartilage formation, embryos were stained with Alcian Blue as previously described [4]. BABB-cleared embryos were visualized under a light microscope (Leica MZ6 Infinity1 stereomicroscope).

Preparation of Embryos for Embedding and Histology
The dissected embryos were fixed in 1% paraformaldehyde overnight. For cryo-embedding, the fixed embryos were first cryoprotected in 30% sucrose overnight, embedded in cryomatrix and sectioned at 10 µm thickness for immunohistochemistry and immunofluorescence.

Immunohistochemistry (IHC) and Immunofluorescence (IF)
Cleaved Caspase-3 (1:250, cat#9661T, Cell Signaling, NEB, Whitby, Ontario, Canada), P53 (1:250, cat#2524, Cell Signaling, NEB, Whitby, ON, Canada), and Phosphohistone H3 (Ser10) (1:200, cat#06-570, Sigma-Aldrich, Saint-Louis, MO, USA) primary antibodies were used. For IHC, an avidin/biotin-peroxidase based system was used (VECTASTAIN ® Elite ABC HRP Kit: PK-620000, Vector Laboratories, Newark, CA, USA) containing biotinylated universal (anti-mouse/rabbit IgG) secondary antibody and visualized with DAB (Vector Laboratories). After rinsing with water, the slides were counterstained with Nuclear Fast Red before mounting with an aqueous mounting medium. For quantification of colorimetric signal, particle analysis on Image J was used. 2 to 3 sections per embryo were imaged. To count the number of total cells, the sections after the ones that were used for IHC was mounted with DAPI and particle analysis on Image J was used to determine the total number of cells in the head. The percentage of positive cells were determined as follows: the number of P53 or cleaved Caspase-3-positive cells/number of DAPI-positive cells in the head X100, and plotted using GraphPad (Prism).

RNA Isolation for RT-qPCR
RNA extraction and RT-qPCR analysis was performed as previously published [4]. The primers used are listed in Table 1.

Primers Used for Splicing Analysis
For Mdm2, FoxM1 and Synj2bp splicing analysis, cDNA was amplified with a RT-PCR program that included a hot start at 95 • C for 5 min, followed by 35 cycles of a denaturation step at 95 • C for 10 s, an annealing step at 55 • C for 30 s, an extension step at 72 • C for 45 s with a final extension step at 72 • C for 10 min. The products were visualized on a 2% agarose gel. The primers that were used are as follow: Mdm2 Forward (exon2-6/7): GATCAC-CGCGCTTCTCCTGC, Mdm2 Reverse (exon2-6/7): GATGTGCCAGAGTCTTGCTG [22]; FoxM1 forward: CTTTAAGCACATTGCCAAGC.
The intensity of the bands was quantified using Image J and the ratios of spliced vs. non-spliced were plotted by GraphPad Prism.

ER Stress
To determine if the unfolded protein response pathway is activated in Eftud2 ncc−/− ; Trp53 ncc+/− or Eftud2 ncc−/− ; Trp53 ncc−/− embryos, we used primers that were previously designed to amplify a 350 bp portion of Xbp1 (Forward primer: GATCCTGACGAG-GTTCCAGA and Reverse primer: GGTCCCCACTGACAGAGAAA), followed by digestion of the amplicon with Pst1 restriction enzyme, as previously described [23]. Since Pst1 does not cut the amplicon in the absence of ER stress, the presence of ER stress is indicated by the presence of a 240 bp and a 110 bp bands.

Statistical Analysis
Two-tailed non-parametric Mann-Whitney t-test analysis was performed using Excel and Prism Software. ANOVA test analysis followed by Tukey's post-test to compare all pairs of columns and a Chi-square test were performed using Prism. Significant p-values are represented as * p < 0.05, ** p < 0.01 and *** p < 0.001. All unique/stable reagents that were generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement.