Mycn deficiency underlies the development of orofacial clefts in mice and humans

Abstract

Non-syndromic cleft lip with or without cleft palate (NSCL/P) is the most common subphenotype of non-syndromic orofacial clefts arising from genetic and/or environmental perturbations during embryonic development. We previously identified 2p24.2 as a risk locus associated with NSCL/P in the Chinese Han population, and MYCN is a candidate risk gene in this region. To understand the potential function of MYCN in craniofacial development, we generated Wnt1-Cre;Mycn^flox/flox mice that exhibited cleft palate, microglossia and micrognathia, resembling the Pierre Robin sequence (PRS) in humans. Further analyses indicated that the cleft palate was secondary to the delayed elevation of palatal shelves caused by micrognathia. The micrognathia resulted from impaired chondrogenic differentiation in Merkel’s cartilage, which limited tongue development, leading to microglossia. In terms of mechanism, Mycn deficiency in cranial neural crest cells (CNCCs) downregulated Sox9 expression by inhibiting Wnt5a in a CNCC-derived chondrogenic lineage in Merkel’s cartilage. To investigate whether MYCN deficiency contributed to NSCL/P, we performed direct sequencing targeting all exons and exon–intron boundaries of MYCN in 104 multiplex families with Mendelian NSCL/P and identified a novel pathogenic variant in MYCN. Taken together, our data indicate that ablation of Mycn in mouse CNCCs could resemble PRS by suppressing the Wnt5a-Sox9 signaling pathway in Merkel’s cartilage and that mutations in MYCN may be novel potential causes of NSCL/P.

Introduction

Orofacial clefts (OFCs) are multifactorial birth defects affecting the integrity of the face and oral cavity, usually classified as cleft lip with or without cleft palate (CL/P) or as cleft palate only (CPO). Although orofacial clefting occurs in more than 300 syndromes (1), the vast majority of OFCs are non-syndromic (NS) (2), and caused by the interactions of multiple genetic and environmental risk factors (3). In explorations of the genetic architecture of non-syndromic orofacial clefts (NSOFCs) in the past decade, genome-wide association studies (GWAS) and whole-exome sequencing (WES) have identified considerable common noncoding variants and some causative mutations in candidate genes associated with NSOFCs; however, the underlying genetic mechanisms remain largely unclear (4–13). Further functional annotation of the candidate genes and identification of functional variants are warranted (5).

In our previous study, 2p24.2 was identified as a NSCL/P risk locus in a Han Chinese population (7). The tagged single nucleotide polymorphisms (SNPs) in this region lie between MYCN and FAM49A, but conventional Fam49a knockout mice do not show craniofacial developmental defects (our unpublished data). The transcription factor MYCN belongs to the MYC proto-oncogene family along with two other paralogs, MYC (C-MYC) and MYCL. For each of these three proteins, a basic-helix-loop-helix-zipper (bHLHZ) domain forms an individual heterodimer with the bHLHZ domain of either MYC-associated protein X (MAX) or MAX-like protein X (MLX) and confers the functions of integrating extracellular and intracellular signals and modulating global gene expression (14–16).

MYC is pivotal in both tumorigenesis and embryonic development (17). Ablation of Myc in mouse cranial neural crest cells (CNCCs) impairs craniofacial development, leading to defects in the skull and in hearing (18) In addition, a noncoding region at chromosome 8q24 has been identified as containing multiple SNPs with significant associations with NSCL/P in a White population. This noncoding interval lies between MYC and FAM49B, and the distal enhancers in this interval control Myc expression during facial morphogenesis (19). MYCN is involved in multiple processes that are important for carcinogenesis and embryonic development (20), including cell growth, proliferation and differentiation (21,22), but the role of MYCN in the pathogenesis of OFCs remains unknown. To elucidate the specific function of Mycn during craniofacial development, we ablated Mycn in CNCCs by crossing Wnt1-Cre mice and Mycn^fl/fl mice. Compared with the control mice, Wnt1-Cre;Mycn^fl/fl (CKO) mice exhibit cleft palate, microglossia and micrognathia, resembling the Pierre Robin sequence (PRS) in humans that is caused by suppression of the Wnt5a–Sox9 signaling pathway in Merkel’s cartilage. To further investigate whether the mutations in MYCN contribute to the NSCL/P, we performed sanger sequencing in 104 NSCL/P multiplex families with Chinese Han ancestry and identified a novel missense mutation. Together, our data indicate the critical role of Mycn in craniofacial development and the mutation in MYCN may be a novel potential causes of NSCL/P.

Results

Mycn deletion in the CNCC lineage leads to cleft palates

To study the potential function of Mycn in craniofacial development, we first examined its localization in the palate, tongue and mandible from E12.5 to E15.5 by immunohistochemistry. Our results showed Mycn expression in the palatal shelves, the developing tongue septum region and mandibles at E12.5 and E13.5, with a decrease at E14.5 and total disappearance at E15.5 (Supplementary Material, Fig. S1).

Then, we investigated the role of Mycn during craniofacial development by knocking out Mycn in the CNCC-derived cells using the Wnt1-Cre transgenic allele and floxed Mycn mice. The efficiency of Mycn gene deletion by the Wnt1-Cre allele in the domain populated from CNCC was confirmed by quantitative (q)PCR analysis and western blot (Supplementary Material, Fig. S2). Histological examinations revealed that morphological structures of the palatal shelves along the anterior–posterior axis in the mutants were comparable with the controls up to E13.5. However, an obvious aberration in palate development was observed in the mutants at E14.5, by which time the palatal shelves in controls had elevated to above the tongue and begun to fuse at the midline. In mutant embryos, the palatal shelves failed to elevate and remained in the vertical position either on both sides or on one side of tongue (Fig. 1A). At E16.5, the palatal shelves had fused completely in the controls, whereas among the mutant palate shelves, 64% had unilateral deficiency in palate shelf elevation, 30% showed bilateral deficiency and 6% had fused palate shelves (Supplementary Material, Fig. S3, Table S2). All the mice with cleft palate died perinatally without any milk in stomach (Fig. 1B and C). These observations suggest an important role for Mycn in palatogenesis.

Figure 1

Conditional ablation of Mycn in the embryonic mouse CNCC-derived cells results in cleft palate. (A) Mutants exhibit palatal shelf elevation delay. Coronal H&E palate sections of control and mutant embryos from E13.5 to E16.5. From E14.5 to E16.5, mutant embryos exhibited elevation delay and unfused palate compared to littermate controls. Scale bars: 100 μm. P: palatal shelves; T: tongue; M: mandibles; * indicates the clefting. (B) Stereomicroscope images and SEM image of the cleft palate at the newborn stage. The arrow indicates the clefting. Scale bars: 200 μm. (C) Representative whole-mount images show no obvious change at the newborn stage, but the mutants died at newborn stage without any milk in stomach (the arrow indicates the milk in stomach of controls). (D) Stereomicroscope images of embryonic head from E13.5 to the newborn stage show micrognathia (red arrow) whereas controls do not. Scale bars: 100 μm.

Cleft palate in Wnt1-cre;Mycn^fl/fl mice is a consequence of tongue malposition and micrognathia

We carefully examined the mice at the newborn stage with MicroCT scans and stereomicroscopy and found that mutant mice exhibited micrognathia (Fig. 1D). The 3D reconstruction of MicroCT scans revealed a mild reduction in mandible volume in mutant mice compared with control mice. Of note, the mandibular body was present as a comparatively mild change, the unique morphological structure showed no significant change, and the mandibular morphological angles were the same as in the control groups (Fig. 2A). These results were confirmed by Alcian blue/Alizarin Red double staining (Fig. 2B). The MicroCT scans of tongues have shown that the CKO mice also exhibited a rather mild decrease in the tongue volume compared with the controls (Fig. 2C), along with paramorphic differences (Fig. 2D). Double staining of Sp7 (Osterix), a marker of osteogenic progenitors, and myosin heavy chain (MHC), found that the genioglossus showed abnormal patterning. The aberrant genioglossus pattern was detectable immediately at the time of palatal shelf elevation in E13.5 but manifested as a mild change. These phenotypes were present throughout palate morphogenesis from E13.5 to E15.5, indicating that tongue extrinsic muscle asymmetry is consistent with the timing of the failure of palatal shelf elevation (Fig. 2E and Supplementary Material, Fig. S4). H&E stains of sagittal sections of control and CKO tongues uncovered a tongue malposition (Fig. 2E). All these phenotypes found in mutants showed no sex bias.

Figure 2

Wnt1-Cre;Mycn^fl/fl Mice exhibit micrognathia and tongue abnormality. (A) 3D reconstruction of microCT scans reveal mild decreases in mandible volume and length of CKO mice compared with controls at newborn stage (broken lines indicate the length), but the unique mandibular morphological angles are the same as in the controls groups (red arrows indicate the angles). Samples were quantified in T-test. ^*P < 0.05, n = 4. (B) Alcian blue/Alizarin Red double staining at E15.5 and the newborn stage confirms that mutants exhibit micrognathia (red arrows and broken lines indicate the mandible tissue). Scale bars: 100 μm. (C) 3D reconstruction of microCT scans reveals a mild reduction in tongue volume of CKO mice compared with controls (the colored areas indicate the tongue tissue). Samples were quantified in T-test. ^*P < 0.05, n = 4. (D) Intraoral and SEM views of the tongues of CKO newborns show paramorphia (indicated by red arrows) compared with controls. (E) MHC (green) and Sp7 (red) double immunostaining of coronal and sagittal sections of E16.5 heads of control and Wnt1-Cre;Mycn^fl/fl mice showing an aberrant genioglossus pattern (red arrows). Sagittal sections of H&E stains and the height of tongue at posterior part show a mild tongue malposition. P: palate, M and broken red lines indicate mandible, yellow broken red lines indicate the posterior height of tongue we counted. * indicates the clefting, T-test. ^*P < 0.05, n = 3. Scale bars: 200 μm.

Because of the tongue malposition and micrognathia, we hypothesized that the palate clefts in mutant mice arise from malposition of the tongue and secondary to the mandibular defects. To test this hypothesis, we used both in vivo and in vitro approaches. First, we generated Osr2-Cre;Mycn^fl/fl mice because Osr2 is expressed in the palatal shelves from E12.5 to the newborn stage but not in the mandible at any developmental stage (23). Indeed, palate clefts in Osr2-Cre;Mycn^fl/fl mice would suggest that Mycn plays an intrinsic role in the elevation of the palatal shelves. However, Osr2-Cre;Mycn^fl/fl groups did not have cleft palate (Fig. 3A). MicroCT scans of the mandibles in Osr2-Cre;Mycn^fl/fl mice also showed no obvious changes even at newborn stage (Supplementary Material, Fig. S5). Second, we performed in vitro experiments using a rotational culture system. The mandible and tongue of Wnt1-Cre;Mycn^fl/fl mice were removed at E13.5, a time before palatal shelf elevation, and isolated in culture. After 2 days of culture in a rotational culture system, the palatal shelves of control mice and Wnt1-Cre;Mycn^fl/fl mice had both elevated and fused. H&E staining was performed for evaluation of the palatal integrity (Fig. 3B). Last, Keratin 17 (K17) and p63 double staining has defined the natural periderm in Wnt1-Cre;Mycn^fl/fl mice compared with controls, which suggested there were no abnormal oral epithelial adhesions in mutant mice (Supplementary Material, Fig. S6).

Figure 3

Cleft palate in mutant mice is a secondary defect to micrognathia and tongue malposition. (A) H&E staining of coronal sections of E15.5 control and Osr2-cre;Mycn^fl/fl mice. (B) Intraoral views of heads from E13.5 control and CKO embryos from which the mandible and tongue were removed and cultured for 3 days. H&E staining of coronal sections of E13.5 cultured explants, where palatal shelves are fused and no epithelial cells remain (broken lines represent the intact midline of palate. Red arrowheads indicate the complete palatal fusion in both groups). P: palate, n = 3. (C) immunohistochemical staining and WB analysis of Scleraxis (Scx) in mandible at E15.5 in control and CKO mice (broken lines represent the mandible tissue. Red arrowheads indicate location of tendons). n > 3.

Then we examined myogenic differentiation by analyzing the expression of MHC, Myogenic differentiation 1 (Myod1) and Myogenesis (Myog), which label muscle fibers. We detected no significant differences in signal intensity in tongues between control and mutant animals at E14.5. Western blot analyses of Myog, Myod1 and MHC expression at E14.5 also confirmed that muscle differentiation was unaffected in the tongues of Wnt1-Cre;Mycn^fl/fl embryos compared with controls (Supplementary Material, Fig. S7).

Next, we checked whether differentiation of neural crest-derived mesenchyme was affected in the tongues of Wnt1-Cre;Mycn^fl/fl mice by analyzing the expression of scleraxis (Scx). Scx is a marker for tendons, which are derived from neural crest cells but cross with muscle and function as a bridge for bone and skeletal muscle. In CKO mice, Scx expression in tendons disappeared (Fig. 3C). Because of the absence of Scx in Wnt1-Cre;Mycn^fl/fl tongues, we speculated that the absence of Mycn may be a primary cause of aberrant tongue muscle development because of its effect on tendon morphogenesis.

To elucidate the potential cause of tongue and mandible dysplasia, we checked cell survival and proliferation in the tongue by using caspase-3 and BrdU staining. However, we found no changes in Wnt1-Cre;Mycn^fl/fl mouse tongues at E13.5 (Fig. 4A, Supplementary Material, Fig. S8). An indistinguishable fluorescence pattern of Wnt1-Cre;Mycn^fl/fl;ROSA^mT/mG mice and Wnt1-Cre;ROSA^mT/mG animals indicated the natural migration of post-migratory neural crest populating cells (Supplementary Material, Fig. S9).

Figure 4

Micrognathia in mutants is accompanied with partially compromised expression of Wnt5a and Sox9 in Meckel’s cartilage. (A) Meckel’s cartilage and tongue BrdU staining of control and CKO mice at E13.5 (dotted box and circle are selected areas for quantification of BrdU staining cells in tongue and mandible). Quantificative results of proliferating cells from (A) (n > 5). MC: Meckel’s cartilage. (B) Aniline Blue staining of control and CKO mice at E13.5 (red dotted circles stand for Meckel’s cartilage). immunofluorescence staining of Col2a1 in Meckel’s cartilage at E13.5 in control and CKO mice (dotted circles stand for Meckel’s cartilage, n > 3) (C) Sox9 and Wnt5a immunofluorescence staining of Meckel’s cartilage at E13.5 in control and CKO mice (dotted circles stand for Meckel’s cartilage, n > 3). (D) Western blot of Sox9, Col2a1 and Wnt5a in mandibles from control and CKO mice at E13.5. Samples were quantified in T-test. ^*P < 0.05, n = 5. (E) qPCR analysis of the relative expression of Sox9 and Wnt5a in mandible mRNA from control and CKO mice at E12.5 and E13.5. Samples were quantified in T-test. ^*P < 0.05, n = 6. Scale bars: 100 μm.

Taken together, these findings strongly suggested that the palatal shelf elevation defect in Mycn mutant mice is secondary to a primary malformation in the tongue and/or mandible.

Meckel’s cartilage differentiation is partially compromised in Mycn mutant mice by suppressed Wnt5a-Sox9 signaling

As the phenotype of CKO mice would highly resemble the PRS in humans, we hypothesized the observed glossoptosis was the consequence of micrognathia, which eventually impedes the elevation of the palatal shelves and results in cleft palate. To determine the cause of the mandibular defects at an early stage, we carefully examined correlative osteogenic factors such as Runx2, SP7 and type I collagen alpha 1 chain (Col1a1). These factors remained unchanged (Supplementary Material, Fig. S10). And cell proliferation in the mandibles of CKO mice were indistinguishable from controls at all time points (Supplementary Material, Fig. S11). Sirius Red staining also showed that the accumulation and formation of the mineralized matrix remained stable (Supplementary Material, Fig. S12). These data indicated no initial decrease in the pool of osteogenic progenitors, followed by a significant delay in the osteogenic process, resulting in a small mandible in CKO mice. The decrease in mandibular volume thus appears to arise from a defect in the development of Meckel’s cartilage at early stage. To verify this hypothesis, we checked the expression of serval key chondrogenesis-related factors including connective tissue growth factor (CTGF), SRY (sex determining region Y)-box 9 (Sox9) and type II collagen alpha 1 chain (Col2a1), and found Sox9 expression was partially compromised. Although the toluidine staining and immunostaining showed type II collagen remained stable, the western blot indicated a rather mild change of Col2a1 expression (Fig. 4B and C). The expression of CTGF was unchanged (Supplementary Material, Fig. S13). We concluded that Mycn was essential for the expression level of Sox9, which was a key regulator of Meckel’s cartilage development (24).

Next, we tried to figure out whether Mycn could regulate the Sox9 expression directly. Nevertheless, the ChIP-PCR assay showed Mycn could not interact with Sox9 promoter (Supplementary Material, Fig. S14). Then, RNA-seq of CON and CKO mandible tissues was carried out to examine the possible mechanisms and key transmitters between Mycn and Sox9 in mesenchyme that induce chondrogenesis associated with Meckel’s cartilage morphogenesis (Supplementary Material, Fig. S15). We selected several chondrogenesis-related signaling pathways from RNA-seq enriched data for further analyses. Informed by the previous studies, we finally emphasized on Wnt signaling pathway and validated the expression of several important chondrogenesis-related Wnt ligands, including Wnt4, Wnt5a, Wnt5b, Wnt9a, Wnt9b and Wnt10b (Supplementary Material, Fig. S16) (25–28). Expression analyses of Wnt5a showed a significant decrease in mutant mandibles at E13.5 compared with controls, corresponding with a similar trending of the expression in Sox9 (Fig. 4D and E). In addition to a well-known bidirectional factor in cartilage morphogenesis, Wnt5a was also approved as one of the regulators of Sox9 in chondrocytes (29,30). So, we hypothesized that the Mycn–Wnt5a–Sox9 signaling pathway was a possible mechanism involved in cartilage differentiation. Seven possible binding sites for the Mycn protein in the Wnt5a promoter domain were predicted by the JASPAR database. DNA sequences harboring potential E-box binding site (CACGTG) are relatively frequent in the genome. ChIP-PCR of mandible tissues by using anti-Mycn antibody and verified two functional sites (Site 1, 1377 basepairs from the TSS domain of Wnt5a, and Site 2, 741 basepairs away from the TSS domain of Wnt5a). Luciferase reporter assays of MC3T3 cells transfected with pCMV-MCS-Mycn, MCS-Wnt5a promoter-firefly and MCS-Wnt5a mutant promoter A/B/C-firefly luciferase confirmed the two functional binding sites (Fig. 5A and B). At last, we tried to rescue the Sox9 expression by adding exogenous Wnt5a protein during mandible segregated culture. After 2 days culturing, we harvested the tissue and found indeed the expression of Sox9 was obviously upregulated in mutant mandibles by WB analyses (Fig. 5C). All these data illustrated that the disrupted Mycn–Wnt5a–Sox9 signaling pathway might underlie cartilage differentiation defects.

Figure 5

Mycn modulated Sox9 expression by interacting with Wnt5a directly. (A) MC3T3 was co-transfected with the indicated expression vector and Wnt5a promoter reporter with different mutation sites, and luciferase reporter assay results were processed after 36 h of culture in medium. Luciferase activity was normalized to Renilla luciferase. Samples were quantified in T-test. ^*P < 0.05, n = 3. (B) ChIP analysis of Mycn protein binding with the Wnt5a promoter. ChIP assays were performed in mandibles of control mice at E13.5, using anti-Mycn antibodies or IgG. PCR was performed with primers targeting Wnt5a promoters as described in the methods. (C) Mandibles of control and CKO mice at E13.5 were cultured segregated for 2 days. Ectogenic recombinant protein Wnt5a was added for rescue of the phenotype. Western blot for Sox9 was carried out to examine Wnt5a-associated upregulation of Sox9 expression (n = 3).

Sanger sequencing reveals a role of MYCN in NSCL/P

To investigate whether MYCN variants contribute to NSCL/P, we performed direct sequencing of all exons and exon–intron boundaries of MYCN in 104 NSCL/P multiplex families. A missense variant in exon 2 of MYCN (MIM: 164840, NM_001293228, c.703G > C [p.A235P]) was found in family #21. This mutation was confirmed by reverse Sanger sequencing in the proband and his affected mother and excluded in the unaffected sister, showing co-segregation with phenotype in this family (Fig. 6A). It located at downstream of a low complexity region in exon 2, leaving the most important bLHL domain unaffected (Fig. 6B). This variant is either rare or absent in gnomAD (v.2.1). In addition, this variant was not found in 446 independent individuals of Han Chinese ancestry (data not shown). An analysis using ANTHEPROT (31) showed that this missense variant affects a conserved domain in the MYCN protein (Fig. 6C). The variant had partial altered hydrophilicity, antigenicity and secondary structure in ways predicted to be deleterious and indicating a metastable evolution of 3D protein structure (Supplementary Material, Fig. S17, S18 and S19).

Figure 6

Sanger sequencing reveals a role of the MYCN in NSCL/P. (A) Pedigree #21 showing segregation of the MYCN variant. Segregation of indicated variants was performed by Sanger sequencing in all individuals for whom DNA was available. (B) MYCN protein structure. LCR, low-complexity region; HLH, bHLH binding domain; blue, pathogenic missense variant. (C) Analysis of evolutionary conservation of MYCN Ala235. (D) Western blot analysis of MYCN, Wnt5a and β-catenin from HEK293 cells transfected with empty vector (Mock) or vectors expressing WT or mutant MYCN. (E) Analysis of relative luciferase activity. Relative luciferase activity was measured following transfection with either TOPflash or FOPflash reporter plasmid and transfection with empty vector (Mock) or vectors expressing WT or mutant MYCN in HEK293 cells. Luciferase assay reporter results were processed after 36 h of culture in medium. Experiments were performed in triplicate, and the results were averaged. Relative luciferase activity was reported as the ratio of firefly/Renilla luciferase activity. Samples were quantified in T-test. ^*P < 0.05, n = 4.

Expression of MYCN was detected by western blot analysis of human embryonic kidney 293 cells (HEK293) transfected with vectors expressing wild-type (WT) or mutant MYCN. We found that compared with WT MYCN, expression of p.A235P was unchanged. To investigate whether mutant MYCN affects the activation level of the Wnt5a, we performed western blots, which showed that p.A235P was associated with decreased β-catenin and Wnt5a (Fig. 6D). In addition, the dual-luciferase assay revealed that compared with WT MYCN, the p.A235P mutant drove decreased levels of luciferase activity, which led to suppression of Wnt signaling pathway activation (Fig. 6E).

Discussion

Facial development begins in the fourth week of pregnancy in humans and on the 11th day of embryonic development in mice, when neural crest cells begin migrating to form the five facial primordia, with the appearance of the palatal primordia from the maxillary process. Two secondary palatal shelves are elevated along the two sides of the tongue, grow toward each other and make contact at the midline, and form a union of the primary palate and two secondary palatal shelves (30,31). Disruption of any of these steps could result in a cleft palate. The whole elevation process is not fully understood but may involve several complicated extracellular or intracellular mechanisms along the anterior–posterior axis (32). The anterior palatal shelves are elevated by rotating, whereas the middle and posterior shelves are elevated from a vertical to horizontal position via a remodeling mechanism (33). Palatal shelf elevation failure is hypothesized to arise from two possible mechanisms: the shelves cannot be elevated because of effects from extrinsic structures, and/or the shelves themselves are involved in the internal process. Both intrinsic and extrinsic factors that cause failed palatal shelf fusion or delayed palatal shelf elevation can ultimately lead to cleft palate formation. Our unpublished data of Mycn staining in developing medial/lateral nasal processes at E10.5, along with the previous studies regarding Mycn expression in whole-mount mouse embryos and single-cell sequencing of developing mouse ectodermal, mesenchymal and endothelial cell populations (34–36), suggested that Mycn may also play a potential role in lip development. However, in our study, the cleft palate in Wnt1-cre;Mycn^fl/fl mice is secondary to tongue malposition and micrognathia, but not caused by intrinsic factors. Mycn^+/− mice have reduced survival but no other observable phenotype (21,37–39). In the Mycn^−/− embryos, those tissues that typically show high expression of Mycn have structural deficiencies, reduced sizes and lower mitotic rates, and Mycn^−/− embryos died at E10.5, which impeded us to use Mycn^−/− in investigating the role of Mycn in craniofacial development. Further studies are warranted for investigating the role of Mycn in lip development.

In humans, mutations in several genes, including SATB2, SOX9 and BMP2 (40–44), have been implicated in PRS-like clefting. However, the developmental mechanisms of PRS and PRS-like phenotype are still not fully understood today (45–47). The function of Mycn in embryonic development has already been sectionally confirmed, in the nervous system, with a strict spatial and temporal expression pattern. In addition, its role in normal embryonic expansion is associated with cell proliferation and cell growth and is crucial for embryonic development (48). The Mycn-involved cell cycle is a strictly regulated process under the control of the CDK family of serine/threonine kinases. These proteins act as heterodimers with two functionally important subunits, a catalytic subunit, CDK, and a regulatory subunit, cyclin, which activates the kinase activity of the protein (49–52). Our findings here are the first indication that Mycn could alter cartilage morphogenesis without doing so through a cell proliferation mechanism. Of interest, a previous study of Myc (C-myc) mutant mice uncovered a similar defect of Meckel’s cartilage, with hearing loss and an abnormal malleus, which differentiates from the posterior part of Meckel’s cartilage (without cleft palate) (18). We highly speculate that all three paralogs of the MYC proto-oncogene family (Myc, Mycl, Mycn) might affect morphogenesis of Meckel’s cartilage in a spatial-dependent way.

Previous GWAS have implicated more than 40 chromosomal loci, each of small effect size, with common variants in only a few of the syndromic CL/P genes contributing to NSCL/P. Functional gene(s) and variants in each locus need further investigation for elucidating their specific mechanism in OFC pathogenesis. The genetic findings and analyses of chondrodysplasia, a subset of skeletal dysplasia, encompassing a group of genetic disorders of cartilage that affect its function to act as the template for bone growth have identified the loci 2q24.1-33.3, 4q32-qter, 11q21-23.1 and 17q21-24.3 were significantly associated with PRS in humans (53–55). But only a few of the hundreds of genes in these regions have been shown to be involved in causing cleft palate or PRS-like phenotype in transgenic mice, such as SATB2 and GAD67 in 2q31, PVRL1 in 11q23-q24 and SOX9 in 17q24.3-q25.1. Taking together with our data, there would be some sort of possible interactions going on between 2p24.2 (MYCN locus) and 2q24.1 or 2q32 (breakpoints for PRS on chr. 2).

To date, variants in MYCN have been identified as the cause of bilateral nephroblastomatosis, Feingold syndrome, hearing impairment, digit abnormalities, multiple congenital anomalies with bilateral Wilms tumor, and neuroblastoma (56–59). Most of these conditions develop because of the effects of the gene variants in the nervous system. The exception is Feingold syndrome, which is characterized by variable combinations of esophageal and duodenal atresias, microcephaly, learning disability, syndactyly, cardiac defects and micrognathia (with occurrence rate at 30%), associated with heterozygosity for MYCN variants (60). The MYCN missense variant (MIM: 164840, NM_001293228, c.703G > C [p.A235P]) we characterized here originated from Chinese central Han population, with a incidence at 0.96% (1 in 104 multiplex pedigrees). However, we believe it can be a possible pathogenic candidate in other populations, especially in cases of PRS-like clefting. This result offers a new entry point for genetics analyses in these patient groups.

In summary, experimental evidence of mouse model, Wnt1-Cre;Mycn^fl/fl, supports the involvement of Mycn in a PRS-like phenotype. The sequencing results from NSCL/P multiplex pedigrees strongly suggest that mutations in MYCN may be involved in human NSOFCs. All our data have uncovered an important role of Mycn in craniofacial development and provides a promising candidate gene for genetic screening in patients with OFCs.

Materials and Methods

Conditional knockout mouse generation

Mycn floxed mice (Mycn^fl/fl [Stock No. 06933]), mT/mG mice (ROSA^mT/mG [Stock No. 07676]) and Osr2-IresCre (or Osr2IresCre [Stock No. 09388]) mice were brought from Jackson lab, and Wnt-1/GAL4/cre-11 transgenic mice was a gift from Shanghai Model Organism. Mating Wnt1-Cre;Mycn^fl/+ mice with Mycn^fl/fl or Mycn^fl/fl; ROSA^mT/mG mice to generate Wnt1-Cre;Mycn^fl/fl (CKO mice mentioned below) and Wnt1-Cre;Mycn^fl/fl;ROSA^mT/mG mice, respectively. Osr2-Cre;Mycn^fl/+ mice were crossed with Mycn^fl/fl mice to generate Osr2-Cre;Mycn^fl/fl mice. Animal usage was approved by the Institutional Animal Care and Use Committee (IACUC) at the State Key Laboratory Breeding Base of Basic Science of Stomatology, Wuhan University. All mice were genotyped using the relevant primers sets; embryonic and neonatal sex was determined by genotyping for Sry. All primers for PCR and q-PCR are available in Supplementary Material, Table S1.

Histological analysis of mouse tissues

Fresh mouse embryos were fixed in Bouin’s solution and 4% paraformaldehyde, respectively, for over 24 h. After a mild gradient dehydration with ethanol and stored in n-butyl alcohol for 4 h, all the samples were embedded in paraffin and sectioned at 5 μm intervals for the generation of tissue slides. After deparaffinization, slides were stained with hematoxylin and eosin or Sirius red via standard methods for visualization of structures (32).

Western blot

For cells, 48 h after transient transfection, cell layers were washed three times with PBS and then whole-cell lysates were prepared using RIPA Lysis Buffer and phenylmethanesulfonylfluoride (PMSF) (Beyotime, China). For tissues, the mandible of embryo at E13.5 were harvest in RIPA buffer + protease inhibitor. Those tissues were sonicated at 40% amplitude for 30 s, spun down for 10 min, and quantified using the BCA assay. Equal amounts of protein were loaded onto a 10% SDS–polyacrylamide gel and then transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore, Billerica, MA, USA). The blot was blocked with tris-buffered saline with 0.1% tween 20 (TBST) containing 5% skim milk for 1 h and then incubated with antibody Mycn (CST, 84406S, 1:2000), MHC (Abcam; ab234431; 1:2000), Sox9 (Abcam, ab158966; 1:2000), Myod1 (Abcam, ab203383; 1:1000), Myogenesis (Myog; Abcam, ab1835; 1:1000), Cola1a1 (Abcam; ab34710; 1:1000), Cola2a1 (Abcam; ab34712; 1:1000), SCXA (Abcam, ab58655; 1:400), SP7/osterix (Abcam, ab22552; 1:200), β-catenin antibody (Abcam, ab32572; 1:2000) and β-Tubulin antibody (Antgene, Wuhan, China, 1:5000). After overnight incubation at 4°C, the blot was washed with TBST and incubated in the secondary antibody (Antgene, Wuhan, China, 1:8000) for 1 h at room temperature. The blot was visualized using WesternBright ECL (Advansta, Menlo Park, CA, USA) and then exposed. Chemiluminescence was detected using SuperSignal West Pico Plus ECL substrate (ThermoFisher #34580) for 2 min and band intensities in imaged were quantified by Image Studio. All those experiments were repeated at least three times.

BrdU incorporation and immunohistochemistry

Cell proliferation within the cranial region was monitored by intraperitoneal 5′-bromo-2′-deoxyuridine (BrdU, Sigma) injection (100 μg/g body weight) at E13.5. Two hours after injection, mice were harvest and fixed in 4% PFA and processed for immunohistochemistry. Detection of BrdU-labeled cells in E13.5 embryos was carried out with an antibody to BrdU(Abcam) followed by incubation with a fluorescent antibody. Other antibodies used for immunohistochemistry included Mycn (CST, 84406S, 1:100), CTGF(Abcam, ab5097; 1:200), MHC (Abcam; ab234431; 1:100), caspase 3 (Abcam, ab2302; 1:100), K17 (CST, 12590S; 1:200), p63 (Abcam, ab735; 1:200), BrdU (Abcam, ab6326; 1:200), Ki67 (Abcam, ab15580; 1:100), Sox9 (Abcam, ab158966; 1:100), Myod1 (Abcam, ab203383; 1:50), Myogenesis (Myog; Abcam, ab1835; 1:100), Cola1a1 (Abcam; ab34710; 1:100), Cola2a1 (Abcam; ab34712; 1:100), SCXA (Abcam, ab58655; 1:200) and SP7/osterix (Abcam, ab22552; 1:100). Fluorescent secondary antibodies were Alexa Fluor 488 and 594 (Invitrogen Life Technologies; 1:200). Sections were counterstained with DAPI and imaged by fluorescence microscopy, followed by quantification of measurement intensities using ImageJ.

Organs culture in vitro

Timed-pregnant mice were sacrificed at E13.5, and embryos were placed in a Petri dish in PBS under a dissecting microscope. Small dissecting scissors and tweezers were used to separate the head from the body. Next, two incisions were made on both sides of the lip line, with the mouse cranium carefully held with forceps above eye level (avoiding the maxillary process). The mandible and tongue were removed, and the maxillary process region was isolated by making a cut at eye level and removing the brain and eyes. Palatal shelves were cultured in the organ culture dish using BGJP (BGJP 80%, serum 20%; Gibco) culture media, and then placed in the Rotary Cell Culture System TM (RCCS™—4SQ, Synthecon) with a rotary standard 15 rpm for 2 days (Medium changed every 24 h). Afterward, the samples were harvested and fixed in 4% PFA as preparation for histology and immunostaining. For the mandible culture, embryos were microdissected at E13.5, and mandibles were placed on a Whatman Nucleopore Track-Etch Membrane and cultured in BGJP. For both control and mutant embryos from timed-pregnant mice sacrificed at the same time, ectogenic recombinant protein Wnt5a (RD system 645-WN-010100 ng/ml) was added to the BGJP medium, followed by 48-h culture and harvest for western blot.

MicroCT scan

Control and Wnt1-Cre;Mycn^fl/fl newborn mice were sacrificed, and the heads were fixed in 4% PFA. The skulls were imaged using a microCT system (Scanco Medical_V1.2a) with energy/intensity 45 kVp,133 μA and 6 W. Visualization and 3D reconstruction of the skull were performed using Mimics Research 19.0. For tongue scanning, the tissue must be scored in I₂/KI solution for a month before microCT scan.

Scanning electron microscopy evaluation

For the scanning electron microscopy (SEM) assay, the prepared organs were immersed in 2.5% glutaraldehyde for 4 h at 4°C, rinsed in 0.1 mol/L phosphate buffer for 30 min, and post-fixed in 1% osmic acid for 1 h at 4°C. After being rinsed thoroughly in the same buffer for 30 min, and then progressively dehydrated in graded ethanols and isoamyl acetate gradient and dried under a CO₂ critical-point dryer (Eiko, Hitachi). Afterward, the organs were mounted on aluminum stubs, sputter coated by an ionic sprayer meter (Eiko, Hitachi), and analyzed by SEM (Stereoscan 260) under an accelerating voltage of 20 kV.

RNA-seq and analyses

Total RNA of mandibles at E13.5 was extracted using TRIzol Reagent (Invitrogen, Life Technologies, USA) following the manufacturer’s instructions, then treated with DNase I (Invitrogen, Life Technologies, USA). We checked the purity of the samples using a NanoDrop (NanoDrop products, Implen, Westlake Village, CA, USA), and the concentration was assessed in a Qubit 2.0 Fluorometer (Life Technologies, USA). Four RNA-seq libraries were prepared from four control mandibles and another four libraries were established from four mutant mandibles at E13.5. The eight libraries were sequenced in three lanes on the Illumina HiSeq 2500 platform (Illumina, USA) based on sequencing by synthesis with 100-bp paired-end reads (BGI Technologies, Shanghai). Approximately 44 879 835 clean reads were mapped to the reference genome Mus_musculus (GCF_000001635.26_GRCm38.p6, NCBI) using HISAT2 tools. After data were mapped and normalized, fragments per kilobase per million mapped reads (FPKM) was calculated using RESM software. A false discovery rate < 0.01 and absolute value of log2 ratio ≥ 2 was used to identify differentially expressed.

Quantitative (q)PCR analysis

Total RNA was isolated from dissected tissues using the RNeasy Mini Kit (TianGen, China). The QuantiTect Reverse Transcription Kit (Vazyme, China) was used for cDNA synthesis. qPCR was carried out on an iCycler (Bio-Rad) with gene-specific primers and SYBR Green (Bio-Rad). Values were normalized to Gapdh. All the primers were listed in Supplementary Material, Table S1.

ChIP assay

One gram of mandibular tissue was collected (refractory or low chromatin content of general 2–3 g) and placed in precooled PBS mixed solution with 0.4% PMSF. A scissor was used to cut the sample into small pieces, which were filtered by cell sieve. A formaldehyde cross-linking solution (formaldehyde final concentration 1%) was mixed with the sample and cross-linked on ice for 15 min. The crosslinking was terminated with 0.2 M glycine solution for 10 min. After addition of 400 μL of animal nuclear lysis buffer, the samples were placed on ice for 30 min for cracking. A Bioruptor non-contact ultrasonic breaker was used for ultrasonic breakdown, and the broken test was carried out to verify the efficiency of the breakdown. ChIP dilution buffer and protein beads was added to the samples and rested on the magnetic rack for 2 min. ChIP dilution (containing 20 μL Mycn antibody, CST, 84406S) buffer was added at 4°C and the samples placed on a magnetic shelf for 5 min. The magnetic beads were cleaned with ChIP dilution buffer and mixed at 4°C for 7–12 h. Then treated samples (containing the antibody–protein–DNA complex) were placed on a magnetic shelf for 5 min, discarded. The elution buffer then was added and shaken for 20 min to yield the DNA sample, purified by the EasyPure® PCR Purification Kit. The concentration and purity were determined using the NanoDrop 2.0.5 and stored at −80°C. For PCR analysis of the target gene, primers were designed to detect the input products, IP products and IgG products.

Cell culture and transfection analysis

Human embryonic kidney 293 cells (HEK293) and Mouse 3 T3 cells (MC3T3) were cultured in α-MEM media with high glucose, supplemented with 10% fetal bovine serum and penicillin/streptomycin. For expression constructs and site-directed mutagenesis, the expression vectors pCMV-MCS-MYCN, pCMV-MCS-Mycn and MCS-Wnt5a promoter-firefly Luciferase were purchased from Genechem (Shanghai, China). To generate pCMV-MCS-A235P and MCS-Wnt5a promoter mutant A/B/C-firefly luciferase, we used in vitro site-directed mutagenesis, using the Beyotime QuickMutation™ Plus Site-Directed Mutagenesis Kit (Beyotime, D0208S, Shanghai, China), according to the manufacturer’s protocol. All vectors were confirmed by sequencing the whole length.

Dual-luciferase reporter assay

HEK293 cells were co-transfected with TOP flash (a TCF reporter plasmid, Beyotime, D2503), FOP flash (which contains mutant TCF-binding sites; Beyotime, D2501) and vectors expressing either WT or mutant MYCN. After 72 h, firefly and Renilla luciferase activities were measured using a Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA), according to the manufacturer’s protocol. Relative luciferase activity was reported as the ratio of firefly/Renilla luciferase activity. Experiments were carried out in triplicate and repeated at least three times.

Recruitment and samples collection

All the patients are recruited from the Chinese population through collaboration with multiple hospitals in central China. The 104 families were selected for apparent segregation consistent with multigenerational autosomal-dominant inheritance. All cases were interviewed and clinically assessed by experienced clinicians. A detailed questionnaire was completed to identify any other anomalies that would suggest an underlying syndrome. We collected clinical information from the participants through a complete clinical checkup, and additional demographic information was obtained through a structured questionnaire that included four main parts (basic information, clinical features, family situation and lifestyle during the first trimester of pregnancy) and detailed clinical histories. Peripheral blood samples were collected from affected probands, their parents and from other available affected and unaffected family members. Peripheral blood samples were collected after the informed consent has been signed. Paternity test was carried out to define family consanguinity. DNA samples of 446 unrelated healthy controls were selected from our established sample database. DNA was extracted using standard protocols, and sample quality-control steps were taken including coefficient of inbreeding and XY genotyping. The research was approved by the institutional ethics committee of Hospital of Stomatology Wuhan University (2014-20; 2016-14; 2017-09).

Sanger sequencing

Sanger sequencing was used to validate identified variants in affected and unaffected relatives. The accession number for the genes described in this study was GenBank: MYCN (MIM: 164840, NM_001293228). Primers around the genomic region of the variants of interest were designed using Primer Premier 5. Primer sequences and conditions are available upon request. PCR products were sent for sequencing using an ABI 3730XL (Tingske Biosciences) and viewed with SnapGene program (v. 3.2.1).

Acknowledgements

We thank all of the patients and their families for participating in this study. Also, we would like to express our gratitude to Prof. Yiping Chen from Tulane University for his valuable suggestion for this study. The authors would also like to acknowledge the Genome Aggregation Database (gnomAD) and its contributors (https://gnomad. broadinstitute.org/about). Conflict of Interest statement. The authors declare no conflicts of interest with respect to the authorship and/or publication of this article.

Funding

National Natural Science Foundation of China (Grant Nos 81970904 to M.H., 81970923 to Z.B.); the National Key Research and Development Program (Grant No. 2016YFC1000505 to Z.B.) and Major Project of Technological Innovation in Hubei Province (Grant No. 2017ACA181 to Z.B.)

Author Contributions

Z.B. and M.H. conceived this study. Z.B. and M.H. provided financial support. Z.B. and M.H. designed the study. H.T.Y., H.L.W. and F.G. conducted sample selection and clinical data management. R.H.Y., Z.H. and Y.N.Z. performed genotyping and sequencing. R.H.Y. and R.Y.L. undertook data analyses, statistical and bioinformatics analyses, and data checking. R.H.Y. and F.W. performed the functional annotation. R.H.Y., Z.H., H.L.W. and Y.N.Z. conducted animal experiments. R.H.Y. and M.H. contributed to writing of the manuscript. Z.B. and M.H. helped to revise the manuscript.