Mutation in Eftud2 causes craniofacial defects in mice via mis-splicing of Mdm2 and increased P53

Abstract

EFTUD2 is mutated in patients with mandibulofacial dysostosis with microcephaly (MFDM). We generated a mutant mouse line with conditional mutation in Eftud2 and used Wnt1-Cre2 to delete it in neural crest cells. Homozygous deletion of Eftud2 causes brain and craniofacial malformations, affecting the same precursors as in MFDM patients. RNAseq analysis of embryonic heads revealed a significant increase in exon skipping and increased levels of an alternatively spliced Mdm2 transcript lacking exon 3. Exon skipping in Mdm2 was also increased in O9-1 mouse neural crest cells after siRNA knock-down of Eftud2 and in MFDM patient cells. Moreover, we found increased nuclear P53, higher expression of P53-target genes and increased cell death. Finally, overactivation of the P53 pathway in Eftud2 knockdown cells was attenuated by overexpression of non-spliced Mdm2, and craniofacial development was improved when Eftud2-mutant embryos were treated with Pifithrin-α, an inhibitor of P53. Thus, our work indicates that the P53-pathway can be targeted to prevent craniofacial abnormalities and shows a previously unknown role for alternative splicing of Mdm2 in the etiology of MFDM.

Figure 1

Homozygous mutation of Eftud2 in neural crest cells causes craniofacial malformations. (A) Eftud2^loxP/−; Wnt1-Cre^tg/+ mutant embryos are not distinguishable from controls at E9.0. (B) At E9.5, (C) E10.5 and (D) E11.5, Eftud2^loxP/−; Wnt1-Cre2^tg/+ mutant embryos exhibit hypoplasia of the midbrain (mb) and pharyngeal arches (pa). (E) The star indicates the open neural tube in an E11.5 mutant embryo. (F) Embryos found alive at E14.5 had exencephaly and an absence of snout and head structures. A frontal view of an E14.5 mutant embryo showing a protruding tongue from the oropharyngeal region, a cleft maxillary process and the fused mandibular process. (G) Images of live wild type (left) and dead mutant (right) embryos at E17.5. Mutant image shows a hypoplastic lower jaw, cleft maxillary process, missing head structures and absence of eyelid closure. nt: neural tube, pa: pharyngeal arch, hb: hindbrain, mb: midbrain, fb: forebrain, ov: otic vesicle, mx: maxillary, md: mandibular prominence, pn: pinnea, t: tongue, e: eye. Scale bar = 500 μm.

Figure 2

Head and face cartilages are missing or hypolastic in E14.5 Eftud2^loxP/−; Wnt1-Cre2^tg/+ mutant embryos. (A) side view of a control (Eftud2^loxP/+) and Eftud2^loxP/−; Wnt1-Cre^tg/+ mutant embryo. Cartilages did not differentiate in the head or face of this mutant embryo. (B) Ventral view of a control and another Eftud2^loxP/−; Wnt1-Cre2^tg/+ mutant embryo showing a hypoplastic Meckel’s cartilage (M) and hypoplasia of the chondrocranium. Structures missing in the mutant are labeled with a star. A: alas temporalis cartilage, B: basitrabecular process, O: orbital cartilage, T: trabecular cartilage, PN: paranasal cartilage, OA: occipital arch cartilage, P: parachordal cartilage, Y: hypochiasmatic cartilage, CO: cochlear part of the auditory capsule, CA: canalicular part of auditory capsule. (C) Frontal and (D) sagittal view of H&E stained sections of control and mutant embryos. v: vertebrea, h: heart, l: liver, b: brain. (E) Higher magnification of serial frontal sections showing a single outflow tract in mutant embryo compared with control (open versus full arrowhead) ve: ventricle, a: aorta, p: pulmonary trunk. Scale bar = 500 μm.

Figure 3

Abnormal trigeminal cranial ganglia formation and reduced neural crest cells in the craniofacial region of Eftud2^loxP/−; Wnt1-Cre2^tg/+ or Eftud2^loxP/loxP; Wnt1-Cre2^tg/+ mutant embryos. Cranial nerves V, VII, IX and X can be seen in representative images of control (Eftud2^loxP/+) and mutant E10.5 embryos (A) after wholemount in-situ hybridization to detect Sox10 expression (purple), and (B) wholemount IHC staining with 2H3 neurofilament antibody (brown). Representative images of control and mutant embryos showing neural crest cells in the developing craniofacial region (blue) (C) occupy a similar region in E9.0 embryos. (D) At E9.5 and (E) E10.5, the region occupied by neural crest cells in the pharyngeal region of Eftud2^loxP/−; Wnt1-Cre^tg/+ mutant embryos is reduced. Post-otic streams of neural crest cells destined for pharyngeal arches 3 and 4 (black arrowhead), were absent in Eftud2^loxP/−; Wnt1-Cre^tg/+mutant embryos (white arrowheads). (F–H) Frontal sections of wholemount stained embryos at (F) E9.0 (G) E9.5 and (H) E10.5. ne: neuroepithelium, hm: head mesenchyme. PA: pharyngeal arch, ov: otic vesicle, h: heart. Scale bar = 500 μm.

Figure 4

Eftud2 mutant neural crest cells have increased levels of cell death. (A) Representative images of E9.0 control (Eftud2^loxP/+) and mutant (Eftud2^loxP/−; Wnt1-Cre^tg/+) embryos showing phospho-histone H3 positive (red) nuclei. Scale bar = 100 μm. (B) quantification of the % of PH3 stained cells/DAPI. (C) Representative images of E9.0 control and mutant embryos showing TUNEL positive cells (red) indicated by white arrows. Scale bar = 100 μm. (D) quantification of the % of TUNEL positive cells/DAPI. (E) Representative images of E9.5 control and mutant embryos showing TUNEL positive cells (red) nuclei, DAPI (blue) scale bar = 280 μm (F) quantification of the % of TUNEL positive cells/DAPI. ^*P < 0.05. (G) Boxplot of Annexin V fluorescence intensity, in arbitrary units, for control (n = 6) and knockdown (n = 4) 09-1 neural crest cells 24–30 h post-transfection with negative control or Eftud2 siRNAs, respectively. A Kruskal–Wallis rank sum test was used to determine the significance associated with the difference between the treatment groups (^*P ≤ 0.05).

Figure 5

RNAseq analysis at E9.0 reveals increased exon skipping and upregulation of the P53 pathway. The number of various alternative splicing events is represented and a sign test performed to compare the tendency for each type of event to occur preferentially in mutants versus controls (A) skipped exons (P < 0.0001) (B) intron retention (P < 0.05, more frequent in controls) (C) alternative 5′ splice site (D) alternative 3′ splice site (P < 0.05 when jointly testing 5′ and 3′ splice site preference for the usage of downstream site). (E) Venn Diagram showing genes with skipped exons in mutant embryos that were also upregulated (including Mdm2) or downregulated (including Eftud2) (F) The percentage of neural crest cells in heads of control and mutant embryos at E9.0 and E9.5 using the exclusion of exon 2 of Eftud2 from the RNAseq data as a surrogate for the neural crest cells population (see Materials and Methods). (G) GO terms analysis of genes that were upregulated or downregulated in mutant versus controls.

Figure 6

Alternatively spliced and overexpressed Mdm2 leads to increased p53 when Eftud2 is mutated. (A) RT-qPCR showing increased expression of P53 target genes in E9.0 Eftud2^loxP/−; Wnt1-cre^tg/+ mutant embryos (n = 3) compared with control embryos (n = 3). Samples were assayed in duplicates and are described in Supplementary Material, Table S6. ^*P < 0.05 ^**P < 0.01by t-test. (B) RT-PCR using primers flanking exon 3 of Mdm2 confirms a significant increase in transcripts with skipped Mdm2 exon 3 in E9.0 Eftud2^loxP/−; Wnt1-cre^tg/+ mutant embryos. Left, representative gel image and right, quantification of splicing of Mdm2 exon 3.^***P < 0.001 by t-test. FL: full-length transcript ∆3: transcript without exon 3. (C) Representative RT-PCR using the same primers for Mdm2 in O9-1 cells showing increase of the Mdm2 transcript with skipped exon 3 in cells transfected with Eftud2 siRNA (with empty pCAGIG or Mdm2-pCAGIG vector) compared with control siRNA with empty pCAGIG vector. (D) RT-qPCR showing fold change of Mdm2 transcripts with exons 1–4 in LCL lines isolated from MFDM patients when compared with controls. Schematic highlights primer flanking locations. P-values calculated from ΔΔCt levels (ANOVA). Mean values with SEM error bars are shown (n = 3). (E) Fold change in Eftud2, Ccng1, Trp53inp1, Phlda3 and Mdm2 levels in O9-1 neural crest cells transfected with Eftud2 siRNA and pCAGIG (black bars) or pCAGIG_Mdm2 (gray bars) expression vectors. Dashed horizontal lines denote expression in control cells transfected with pCAGIG. Data presented as average fold-change. (F) Representative image of sections of E9.0 control (Eftud2^loxP/+) and Eftud2^loxP/−; Wnt1-cre^tg/+ mutant embryos showing increased P53 (brown) positive cells after IHC. ne: neuroepithelium, hm: head mesenchyme. Scale bars, lower magnification = 75 μm, higher magnification = 25 μm.

Figure 7

Pifithrin-α partially rescues craniofacial abnormalities in Eftud2^loxP/−; Wnt1-cre^tg/+ mutant embryos. Pregnant females were treated daily with vehicle (veh:2% DMSO/PBS; n = 3) or 2.2 mg/kg pifithrin-α (pif) (n = 7) I.P. for 3 days from E6.5 to E8.5, and embryos were collected at E9.5. (A) Representative pictures of control embryos (Eftud2^loxP/+ and Eftud2^loxP/+; Wnt1-cre^tg/+) treated with vehicle (ctl + veh), neural crest mutant embryos (Eftud2^loxP/−; Wnt1-cre^tg/+) treated with vehicle (mut + veh), and neural crest mutant embryos treated with pifithrin-α (mut + pif). The latest were categorized by their head morphology (green area) as normal, n = 3/11 or abnormal, n = 8/11. (B) Perimeter of first left pharyngeal arches (black dotted lines). ^*P < 0.05 by t-test. Scale bar = 500 μm.