To investigate the mechanism involving ALX1 in frontonasal development, we generated a new mutant mouse line carrying a deletion of exon-2 of the Alx1 gene (Alx1 ^del ) using CRISPR/Cas9-mediated genome editing in the C57BL/6N inbred mice (Figures 1A–C). Sequencing of RT-PCR products from Alx1 ^del/+ and Alx1 ^del/del embryos confirmed that the Alx1 ^del allele produced mutant mRNAs from splicing exon-1 to exon-3, which led to a frame-shift and is predicted to produce a truncated protein product containing only the N-terminal region of the ALX1 protein lacking the homeodomain and the C-terminal Aristaless domain (Brouwer et al., 2003). Indeed, western blot analysis confirmed that the Alx1 ^del/del embryos lacked full-length ALX1 protein and only produced a truncated product that was expressed at low levels but still detectable using the polyclonal anti-ALX1 antibody (Figure 1D).

Alx1 ^del/+ mice appeared indistinguishable from wildtype littermates. Examination of pups and fetuses from Alx1 ^del/+ mice intercrosses revealed that Alx1 ^del/del pups exhibited shortened snout, flattened nasal bridge with increased distance between the nostrils, short philtrum, and midline notching of the upper lip (Figures 2A–D, n = 6 for each genotype; and Supplementary Figure S1). Alx1 ^del/del pups were born alive but died soon after, most likely due to the cleft palate defect (n = 4 for each genotype examined by serial frontal sections at E16.5). In addition, 35 out of 50 E18.5 or newborn Alx1 ^del/del pups examined showed open eyelid unilaterally (Figures 2B,D). Analysis of skeletal preparations revealed that Alx1 ^del/del embryos had hypoplastic premaxilla and presphenoid bone, and malformed palatal processes (Figures 2E,F) (n = 5 for each genotype). Histological analyses of E16.5 Alx1 ^del/del embryos and control littermates showed that, in addition to cleft palate defect, the Alx1 ^del/del embryos had disruptions of nasal cartilages (Figures 2G,H) (n = 4 for each genotype). No neural tube defect has been detected in over 200 Alx1 ^del/del embryos analyzed in the C57BL/6N background.

We outcrossed the Alx1 ^del/+ mice to 129/S6 inbred mice, a subline derived from the original 129/SvEv inbred strain (Simpson et al., 1997) that was used to analyze the Alx1 ^tm1Crm allele (Zhao et al., 1996), for two generations and then intercrossed the N2 Alx1 ^del/+ male and female mice to analyze Alx1 ^del/del pups in the 129 X C57BL/6 hybrid background. Only 5 of 48 (∼10%) Alx1 ^del/del embryos harvested from E16.5 to E18.5 in this hybrid background displayed an exencephaly phenotype (Supplementary Figures S2A–C). While the frequency of Alx1 ^del/del embryos exhibiting anterior neural tube defect in this study is low, these results confirm that loss of Alx1 function causes a genetic background-dependent defect in anterior neural tube closure in mice. The Alx1 ^del/del embryos with exencephaly showed more severe ocular defect but fused secondary palate (n = 5), whereas cleft palate was observed in all E16.5 or older Alx1 ^del/del embryos that did not have exencephaly and subjected to examination of palate morphology (n = 16) (Supplementary Figures S2E,F). All Alx1 ^del/del pups in the 129 X C57BL/6 hybrid background exhibited similar frontonasal defects as in the C57BL/6N inbred background, including flattened nasal bridge, midline notching of the upper lip, and disruption of nasal cartilages (Supplementary Figures S2A–F). These midfacial defects in Alx1 ^del/del mice closely recapitulate the midfacial developmental defects in ALX-related patients.

To understand the cellular mechanism of ALX1 function in frontonasal and ocular development, we analyzed the patterns of Alx1 mRNA expression during early CNCC development. Previous lineage tracing and single-cell RNA sequencing (scRNA-seq) studies have shown that CNCCs first initiate delamination from the anterior neural plate border at about 4-somite stage (SS4) in mouse embryos and migrate to the facial prominences by SS8–SS10 (Yoshida et al., 2008; Zalc et al., 2021). At SS10, the Wnt1-Cre;Rosa26 ^mTmG/+ mouse embryos, in which all CNCCs as well as Wnt1-expressing dorsal mid- and hind-brain neuroepithelial cells were genetically labeled by the expression of green fluorescent protein (GFP), clearly showed CNCCs populating the periocular region and the first branchial arches (Figure 3A). In situ hybridization analysis also showed strong expression of the Sox10 mRNAs, a marker of pluripotent neural crest cells, in the CNCCs populating the periocular region and first branchial arches in SS10 wildtype mouse embryos (Figure 3B). By SS12, in addition to expression in the CNCCs, Sox10 mRNAs were detected in the trunk neural crest cells migrating ventrally from the dorsal neural tube region (Figure 3C). In contrast to the patterns of robust Sox10 mRNA expression in migrating neural crest cells, the earliest Alx1 mRNA expression in the cranial region of the mouse embryos was detected in the periocular region from SS9 to SS11 (Figures 3D,E). Alx1 mRNA expression was not detected in migrating CNCCs lateral to the mid- and hindbrain tissues at these stages, in contrast to the GFP expression pattern in the Wnt1-Cre;Rosa26 ^mTmG/+ embryos and Sox10 mRNA expression pattern in the wildtype embryos (compare Figures 3D,E with Figures 3A,B). At SS12, Alx1 mRNA expression in neural crest cells was still restricted to the periocular region and was not detected in Sox10-expressing migrating neural crest cells in either the cranial or trunk regions by in situ hybridization analysis (Figure 3F, compared with Figure 3C).

To further clarify the patterns of Alx1 expression during early CNCC development, we analyzed the scRNA-seq data of early CNCCs harvested from SS4 - SS10 mouse embryos that were recently reported by Zalc et al. (2021). This dataset provides an extremely deep whole transcriptome sequencing of the individual CNCCs around the time of active delamination from the anterior neural plate border and migration towards the early facial primordia, with the median number of over 6,500 detected genes per cell (Zalc et al., 2021). As shown in Supplementary Figure S3A, unsupervised clustering of this early CNCC scRNA-seq dataset clearly clustered the cells into six major groups, identified as the neuroepithelial and pre-delamination CNCC “precursors,” the “migrating CNCCs,” the “post-migratory CNCCs,” cranial “placode” cells, “cranial mesoderm” cells, and “endothelial cells,” respectively, according to their marker gene expression profiles (Supplementary Figures S3A–L; Supplementary Table S1). Whereas high levels of Sox10 mRNA expression was detected in all migrating and post-migratory CNCCs, while Foxd3 was highly expressed in the migrating CNCCs but down-regulated in post-migratory CNCCs in these samples (Supplementary Figures S3D,E), Alx1 mRNA expression was mainly detected in post-migratory CNCCs harvested from SS8 and SS10 embryos, with low level of Alx1 mRNAs detected in a subset of migrating CNCCs harvested from SS6 embryos (Supplementary Figure S3G). We also analyzed expression of Alx3 and Alx4 in this scRNA-seq dataset and found that expression of both Alx3 and Alx4 mRNAs was detected in a subset of post-migratory CNCCs harvested from SS10 embryos (Supplementary Figures S3H,I), indicating that expression of both Alx3 and Alx4 in the CNCC lineage was activated later than that of Alx1. Altogether, while the high sensitivity of the scRNA-seq analysis detected Alx1 mRNA expression in a subset of migrating CNCCs that was not detected in our in situ hybridization analysis, these data consistently demonstrate that Alx1 expression was activated after the onset of CNCC migration and that Alx1 exhibited a more restricted pattern of expression than that of Sox10 in CNCCs during early craniofacial development.

We next analyzed Alx1 expression during frontonasal development from E8.75 to E10.5 (Figures 4A–D). Frontal views of the embryos showed that Alx1 mRNAs were concentrated in the lateral regions of the FNP but absent from the anterior midline region overlying the developing forebrain at this developmental stage (Figures 4B,D). At E10.5, strong Alx1 mRNA expression was detected in the MNP and LNP, with moderate levels of expression also detected in the distal region of the MxP directly adjacent to the LNP (Figures 4E,F). Immunofluorescent staining of serial sections from the E10.5 embryos showed that ALX1 protein was strongly expressed in the CNCC-derived periocular mesenchyme as well as in the mesenchyme of the LNP and MNP, but no ALX1 expression was detected in any of the epithelial tissues, such as the neural epithelium of the brain, optic cup, optic stalk, and the facial and nasal epithelium (Figures 4G,H). These data indicate that ALX1 primarily acts in the CNCCs to regulate frontonasal development.

We next examined whether ALX1 is required for CNCC migration to the facial primordia by using the Wnt1-Cre:Rosa26 ^mTmG mediated genetic lineage tracing (Danielian et al., 1998; Muzumdar et al., 2007). As shown in Figure 5, similar patterns of GFP-labeled CNCCs in the frontonasal and periocular regions as well as in the branchial arches were observed in the control and Alx1 ^del/del embryos at E9.5 and E10.5 (Figures 5A–D). For embryos examined at E10.5, lateral views of the head consistently detected smaller eyes in the Alx1 ^del/del embryos but the contribution of GFP-labeled CNCCs in the nasal, maxillary, and mandibular processes appeared similar in the Alx1 ^del/del embryo and control littermates (Figures 5C,D). To further verify that the frontonasal neural crest cells migrated normally to the FNP in the Alx1 ^del/del embryos, we analyzed the expression of known frontonasal mesenchyme markers, Alx3 and Alx4, and found that both were similarly expressed in the frontonasal prominence in the control and Alx1 ^del/del littermates at E9.5 (Figures 5E–H). These results indicate that early migration of cranial neural crest cells to the FNP was not overtly affected in the Alx1 ^del/del mouse embryos, which is consistent with our finding that Alx1 expression was absent in early migrating CNCCs and was highly expressed in post-migratory periocular CNCCs and the frontonasal mesenchyme.

Since FND3 patients exhibited extreme microphthalmia and other ocular defects including eyelid coloboma and asymmetric optic nerves (Uz et al., 2010; Pini et al., 2020), we analyzed the ocular developmental defects in Alx1 ^del/del embryos. In addition to open eyelids, we found that the Alx1 ^del/del pups consistently exhibited smaller eyeballs but the inter-eye distance was not significantly different from wildtype littermates (Supplementary Figures S1D,E). Histological analysis of E16.5 embryos showed that the Alx1 ^del/del mutants had deformed optic cup and optic stalk (Figures 6A,B) (n = 4 for each genotype). During early eye development, invagination of the optic vesicle results in the formation of asymmetric optic cup with a ventral groove, called the optic fissure, at the ventral side of the optic cup and optic stalk (Tao and Zhang, 2014). From E10.5 to E12.5 in mouse embryogenesis, as the distance between the ventral diencephalon and optic cup increases, the optic stalk epithelium extends along both the medial-lateral and dorsal-ventral axes and changes from an initially 10–12-cell thick neuroblastic epithelial layer to a 1-2-cell thick bilayered epithelial structure around the proximally projected retinal ganglion axons in the optic groove (Evans and Gage, 2005). To better understand the ocular developmental defects in the Alx1 ^del/del embryos, we analyzed the patterns of molecular marker expression for the optic cup, optic stalk, and axonal neurofilament in serial sections through the developing optic stalk. At E12.5, while the optic cup and optic stalk neuroepithelia were marked by expression of PAX6 and PAX2, respectively, in both the control and Alx1 ^del/del littermates, the optic cup was abnormally extended along the medial-lateral axis in the Alx1 ^del/del embryos compared with the control embryos (Figures 6C,D) (n = 6 for each genotype). In addition, while the optic stalk epithelium in the E12.5 control embryos had thinned out to a 1–2-cell thick bilayered epithelial structure around the retinal ganglion axons (positive for 2H3 immunostaining) in the optic groove, with only the ventral/inner layer of the optic stalk epithelium expressing PAX2 (Figure 6E) (n = 3), the optic stalk epithelium in the Alx1 ^del/del littermates remained as an epithelial tube, with a central lumen surrounded by a 4-6-cell thick PAX2+ epithelium, connecting the ventral diencephalon to the optic cup (Figures 6D,F), and with the retinal ganglion axons lying outside of the optic stalk (Figure 6F) (n = 3). By E14.5, the PAX2+ cells of the optic stalk had delaminated and integrated with the retinal ganglion axons, forming the organized optic nerve bundle in the control embryos (Figure 6G) (n = 3). In the E14.5 Alx1 ^del/del embryos, however, the PAX2+ optic stalk cells and the retinal ganglion axons remained largely segregated, with the PAX2+ epithelium partly wrapping around the nerve fibers (Figure 6H) (n = 3). These results indicate that ALX1 function is required for optic stalk and optic nerve morphogenesis in mice.

The defect in optic stalk morphogenesis in the Alx1 ^del/del mutants appeared remarkably similar to the optic stalk morphogenesis defect previously reported in mice with neural crest lineage-specific inactivation of Pitx2 (Evans and Gage, 2005). While Pitx2 mRNAs were expressed in the periocular mesenchyme surrounding the developing eye in the wildtype embryos at E10.5, expression of Pitx2 mRNAs in the periocular mesenchyme was apparently reduced in the E10.5 Alx1 ^del/del embryos (Figures 7A,B). Immunodetection of PITX2 protein revealed a domain-specific loss of PITX2 protein expression in the periocular mesenchyme surrounding the optic cup (Figures 7C,D). Furthermore, we found that expression of Lmx1b, another key ocular developmental regulator (Pressman et al., 2000), was also reduced in the periocular mesenchyme in the E10.5 Alx1 ^del/del embryos in comparison with control littermates (Figures 7E,F). Since Alx1 is expressed in the periocular neural crest cells but not in the optic stalk or optic cup epithelium (Figures 4G,H), these data suggest that ALX1 acts upstream of Pitx2 and Lmx1b in the periocular neural crest cells to regulate eye development. In addition, we found that the Alx1 ^del/del embryos consistently exhibited increased apoptosis of the periocular mesenchyme cells located dorsally to the optic cup at E10.5, compared with control littermates (Figures 7G,H) (n = 4 for each genotype). This regional loss of periocular mesenchyme during early eye development likely also contributed to the ocular defects in Alx1 ^del/del mice.

To characterize the developmental mechanism underlying the frontonasal defects in Alx1 ^del/del embryos, we compared patterns of expression of several marker genes for distinct regions of the CNCC-derived frontonasal and jaw mesenchyme, respectively. First, we analyzed the expression of Pax7, which is critical for nasal cartilage formation (Mansouri et al., 1996). Pax7 is specifically expressed in the lateral nasal mesenchyme at E10.5 in wildtype embryos (Figure 8A). In the Alx1 ^del/del embryos, the Pax7 expression domain is reduced and is disrupted at the caudal third of the lateral nasal process adjacent to the maxillary process (Figure 8B). Disruption or loss of caudal lateral nasal mesenchyme gene expression was further validated by comparing the patterns of expression of Gsc and Rnf128 in the wildtype and Alx1 ^del/del littermates (Figures 8C–F). Next, we analyzed whether the loss of LNP marker gene expression in the Alx1 ^del/del embryos reflected a defect in patterning of the facial mesenchyme by comparing the expression of marker genes of the maxillary and mandibular mesenchyme in the Alx1 ^del/del embryos and their littermates. Whereas Lhx6 and Lhx8 were strongly expressed in the maxillary and mandibular mesenchyme but absent in the lateral nasal mesenchyme in E10.5 wildtype embryos (Figures 8G,I), both Lhx6 and Lhx8 mRNAs were found ectopically expressed in the caudal region of the lateral nasal processes in the Alx1 ^del/del embryos (Figures 8H,J). These data indicate that ALX1 plays a crucial role in patterning the lateral nasal mesenchyme.

Whereas the Alx1 ^del/del embryos exhibited evident frontonasal defects including shortened snout, flattened nasal bridge, upper lip notching and premaxillary hypoplasia, Alx4 ^−/− mouse embryos showed failure of eyelid closure but relatively normal frontonasal development (Supplementary Figures S4A,D), as previously reported (Qu et al., 1997; Curtain et al., 2015). The Alx4 ^- allele (official allele name Alx4 ^lst−2J ) used in this study originally arose in the C57BL/6J inbred background and carries a spontaneous 33.4 kb deletion of the 5′ and exon1 region of the Alx4 gene (Curtain et al., 2015). In our mouse colony the Alx4 ^+/− mice had been previously outcrossed to the wildtype CD1 mice. During this study of Alx1/Alx4 compound mutants, we examined over 30 Alx1 ^del/del embryos generated from Alx1 ^del/+ ;Alx4 ^+/− intercrosses and did not detect any phenotypic difference from the Alx1 ^del/del embryos analyzed in the C56BL/6 inbred background. However, deleting one allele of Alx4 in the Alx1 ^del/del embryos exacerbated the midfacial phenotype with widened and depressed nasal bridge, shortened philtrum, and further malformed premaxilla in the Alx1 ^del/del Alx4 ^+/- embryos (Supplementary Figures S4B,E). Furthermore, Alx1 ^del/del Alx4 ^−/− double homozygous mutants exhibited a wide-open midline facial cleft (Supplementary Figures S4C,F). These data indicate that, while ALX1 function is essential for frontonasal development, ALX4 partly complements ALX1 function in frontonasal morphogenesis.

We next investigated whether ALX4 plays a complementary role to ALX1 in the patterning of the LNP mesenchyme. Compared with the patchy reduction of Pax7 expression in the caudal region of the LNP in the E10.5 Alx1 ^del/del embryos (Figures 8A,B), E10.5 Alx1 ^del/del Alx4 ^+/- embryos exhibited a clear loss of Pax7 expression from the caudal region of the LNP and Alx1 ^del/del Alx4 ^−/− embryos exhibited a much-reduced domain of Pax7 expression in the LNP (Figures 9A–F). In a mirror image pattern to the loss of Pax7 mRNA expression in the caudal region of the LNP, both Lhx6 and Lhx8 exhibited enhanced ectopic expression in the caudal region of the LNP in the E10.5 Alx1 ^del/del Alx4 ^+/- embryos and further expanded domain of ectopic expression in the LNP in the Alx1 ^del/del Alx4 ^−/− embryos (Figures 9G–R) compared with the patterns of expression in the control and Alx1 ^del/del embryos (Figures 8G–J). Furthermore, the Alx1 ^del/del Alx4 ^−/− embryos exhibited ectopic activation of Lhx6 and Lhx8 expression in the MNP as well (Figures 9L,R). These data indicate that ALX1 plays a critical, predominant role in patterning the lateral nasal mesenchyme with ALX4 partly complementing its function for determining the frontonasal mesenchyme identity.

The sgRNA target sites were selected according to the on- and off-target scores from the CRISPR design web tool (http://CRISPOR.org) (Haeussler et al., 2016). The selected sgRNAs, targeting specific sequences in intron-1 (GTA​AGA​TGT​GGG​TGG​TAC​T) and intron 2 (TTA​CTA​AGT​ATA​GGG​ACA​GG) regions, respectively of the Alx1 gene, were transcribed in vitro using the MEGAshorscript T7 kit (ThermoFisher), purified by using the MEGAclear Kit (ThermoFisher) and stored at −80°C. Individual sgRNA was incubated with CAS9 protein (ThermoFisher) at 37°C for 5 min to form the ribonucleoprotein complex and validated in a small batch of mouse zygotes following electroporation, in vitro culture to the blastocyst stage, and genotyping. The mixture of sgRNAs (50 ng/μL each) and CAS9 protein (150 ng/μL) was injected into the cytoplasm of fertilized eggs of the C57BL/6N inbred mice using a piezo-driven microinjection technique (Scott and Hu, 2019). Injected eggs were transferred on the same day into the oviductal ampulla of pseudopregnant CD-1 female mice at approximately 25 eggs per recipient. Pups were born and genotyped by PCR to identify founder mice and Sanger sequencing to verify the deletion of exon-2 and flanking sequences of the Alx1 locus.

All animal work procedures were performed following the recommendations in the Guide for Care and Use of Laboratory Animals by the National Institutes of Health and approved by Institutional Animal Care and Use Committee (IACUC) at Cincinnati Children’s Hospital Medical Center.

Heterozygous Alx1 ^del/+ mice were maintained in the C57BL/6N background. Embryos from the Alx1 ^del/+ intercrosses were collected at respective stages with the noon of the vaginal plug observed as E0.5. Alx1 ^del/+ intercross embryos were genotyped using the following primers, Alx1 WT3F: GAA​GCA​TTC​TCA​GCT​AAG​ACT​TG, Alx1 WT3R: GCA​GTA​TTA​CGT​GCT​GAA​GTG​GT, Alx1 3F: GGA​TTC​ATA​CCT​CAT​TGC​AGT​C. For analysis in the 129 x C57BL/6 hybrid background, Alx1 ^del/+ male mice were bred with 129/S6 inbred females and the female progeny were back-crossed with 129/S6 males to generate N2 males and females, which were intercrossed and embryos were analyzed for skeletal preparations at E18.5 and histology at E16.5.

The Alx4 ^lst−2J/+ (referred here as Alx4 ^+/− ; Jax Stock #000221) (Curtain et al., 2015) mice were obtained from the Jackson Laboratory. Alx4 ^+/− mice had been outcrossed to wildtype CD1 mice in the past and had been maintained by crossing to C57BL/6N mice. For the generation and analysis of Alx1/Alx4 compound mutants, Alx1 ^del/+ Alx4 ^+/− mice were intercrossed, and the embryos dissected at predetermined developmental stages for skeletal preparations or in situ hybridization analyses.

PCR primers, forward 5′-GGA​GAC​GCT​GGA​CAA​TGA​GT-3′ and reverse 5′-AGG​CGA​GTG​AGA​GTA​AGG​TG-3′, were used to amplify a 673 bp fragment from exon1 to exon4 of Alx1 (NM_172553.4). RT-PCR products were gel-purified and sequence-verified at the DNA Core Facility in Cincinnati Children’s Hospital Medical Center.

Western blot analyses were carried out as reported previously (Iyyanar and Nazarali, 2017; Okello et al., 2017). In brief, frontonasal and periocular tissues of E11.5 wild-type and Alx1 ^del/del embryos, respectively, were lysed in RIPA buffer containing protease inhibitor cocktail. Proteins were quantified using a BCA assay kit (ThermoFisher) and 20 µg proteins were separated on a 4–20% mini-PROTEAN gradient gel (Bio-Rad) and blotted onto a PVDF membrane. Primary antibodies used were anti-ALX1 rabbit polyclonal (1:2000; Proteintech 16372-1-AP) and anti-β-tubulin mouse monoclonal (1:1000; DSHB E7).

For histological analyses, embryos were dissected at desired stages from timed pregnant mice, fixed in 4% paraformaldehyde, dehydrated through an ethanol series, embedded in paraffin, sectioned at 7 μm thickness, and stained with Alcian blue followed by hematoxylin and eosin. Skeletal preparations of E18.5 embryos were processed and stained with Alizarin red and Alcian blue as previously described (Ovchinnikov, 2009).

Snout length, philtrum length, diameter of the eyeballs, inter-nostril and inter-eye distances were measured using ImageJ software and lateral or frontal view pictures of three pairs of E16.5 control and Alx1 ^del/del littermates. Data are represented as mean ± SEM. Statistical analysis was performed by unpaired t-test using the GraphPad Prism software. p < 0.05 was considered as statistically significant difference.

Immunofluorescent staining was performed as previously described (Xu et al., 2016). The following primary antibodies were used: rabbit polyclonal anti-ALX1 (1:200; Proteintech 16372-1-AP), mouse monoclonal anti E-cadherin (BD Biosciences; 610,182), rabbit monoclonal anti-PAX2 (1:400; Abcam Ab79389), mouse monoclonal anti-PAX6 (1:25; DSHB AB_528,427), rabbit monoclonal anti-PITX2 (1:200; Abcam Ab221142), and mouse 2H3 monoclonal anti-neurofilament antibody (1:600; Developmental Studies Hybridoma Bank). Cell death was determined using a TUNEL assay kit in paraffin sections as per manufacturer’s protocol (Promega).

Whole mount in situ hybridization was carried out as previously described (Baek et al., 2011). Embryos were staged by counting somite numbers. For each probe analyzed, a minimum of three embryos of each genotype were analyzed and only probes that detected consistent patterns of expression in all samples were considered as valid results. The plasmid templates of the Alx1, Gsc, and Rnf128 probes were amplified by PCR and cloned into pBSKII vector using the following primers, Alx1 F: TAT​ACG​GGG​TTT​TCG​AAC​CA, Alx1 R: CAC​TCT​GTT​GCA​GCC​TCA​AG; Gsc F: CTG​TCC​GAG​TCC​AAA​TCG​CT, Gsc R: AGC​ATC​GAC​AAC​ATC​CTG​G; Pax7 F: GGG​TAG​GGG​GCA​CAG​AGG​CA, Pax7 R: CCG​GGC​CAG​CAG​GTG​GTT​TC; Pitx2 F: ACA​TAC​TCA​TAG​ATG​AGA​TG, Pitx2 R: GAA​ATC​AAA​AAG​GTC​GAG​TT; Rnf128 F: CAA​CAG​GAC​TGC​CAA​TCA​GG, Rnf128 R: TGC​ACC​GTA​ACC​AGT​TAC​CAA; Sox10 F: CGA​AGC​TTC​CAT​CTC​ACG​ACC​CCA​GTT​T, Sox10 R: CCG​GAT​CCA​GGC​GAG​AAG​AAG​GCT​AGG​T. The Lhx6 and Lhx8 (Grigoriou et al., 1998), and Lmx1b (Liu and Johnson, 2010) probes were received from published sources.

The scRNA-seq data for the SS4 - SS10 mouse embryonic CNCCs (Zalc et al., 2021) were obtained from the NCBI GEO database (accession number GSE162035). Data analysis was performed using the Seurat package (version 4.0.4, R version 4.0.3) (Hao et al., 2021). Data normalization was performed using the function SCTransform and the heterogeneity associated with ribosomal content and ERCC spike-in content were regressed out. Principal component analysis (PCA) was performed using the function RunPCA. Non-linear dimension reduction was carried out using the function RunUMAP utilizing the first 30 principal components (PCs). The function FindNeighbors was used to construct the Shared Nearest Neighbor (SNN) graph using the first 30 PCs and cell clustering was performed using the function FindClusters with resolution set to 0.2. The cell identities were determined based on their marker genes, which were identified using the FindAllMarkers function.