The phenotypes of Gpr161 KO embryos were varied yet included malformations of the pharyngeal arches and microcephaly, along with cranial and spinal NTDs (Mukhopadhyay et al., 2013; Kim et al., 2019). This pattern of altered development suggests that the craniofacial defects could occur in the later embryonic stages. We utilized Wnt1-Cre lines to investigate the role of Gpr161 on mouse neural crest-derived craniofacial development in the mouse. We initially characterized the Cre expression of Wnt1-Cre lines crossed with Rosa-LacZ reporter mice [Figure 1A; upper panel)]. Cre was expressed in the mesencephalon, the first and second pharyngeal arches, the trigeminal ganglia (V), and facial nerve ganglia (VII) at E9.5. By E11.5, the Cre expression was widely expanded into the neural tube, including the mid/hindbrain and most of the orofacial and pharyngeal arch regions (Figure 1A: lower panel). We observed the gross morphology of fetuses with Wnt1-Cre lineage-specific Gpr161 deletion, which resulted from Gpr161 ^f/f crossed with Gpr161 ^f/+ ; Wnt1-Cre/+ (Gpr161 ^f/f ; Wnt1-Cre/+, referred to as Gpr161 cKO from here on in this manuscript) (Figure 1B). The Gpr161 cKO fetuses survived until E18.5, although we failed to observe any liveborn pups. The Gpr161 cKO fetuses expressed midbrain protrusion, ano/microphthalmia, ano/microtia, and severe orofacial defects at E13.5, whereas Gpr161 ^f/+ ;Wnt1-Cre/+ (Cre control) or Gpr161 ^f/f (flox control) did not show any similar abnormal phenotypes (Figure 1B and Supplementary Table S1). Spinal edema was also apparent by E15.5 in Gpr161 cKO fetuses. In addition, we could observe widened mandible and maxilla, which are the representative phenotypes of increased Shh signaling in the face (Supplementary Figure S1A). We also observed encephalocele in ∼69% of the Gpr161 cKO fetuses at E17.5 and E18.5 (Figure 1B and Supplementary Table S2).

A histological analysis was performed to further confirm the gross phenotypic malformations of the Gpr161 cKO fetuses (Figure 2A). The tectum in Gpr161 cKO was extended and mesencephalic vesicle and fourth ventricle at E13.5 were enlarged. The dorsal midbrain was enlarged by E15.5 (Figure 2A: lower right) and the brain herniation along with protruded meninges, which is the representative phenotypes in encephaloceles (Naidich et al., 1992), was detected in the mesencephalic ventricles of Gpr161 cKO fetuses at E17.5 (Supplementary Figure S1). The maxillary bone was shortened and showed irregular shapes, and nasal septum and mandible were underdeveloped. In addition, hard palatal shelves were not fused along the midline, creating a mild cleft palate in Gpr161 cKO fetuses at E15.5 and E17.5 (Figure 2A: lower right and Supplementary Figure S1B). These morphological findings were consistent with the phenotypes of the Gpr161 cKO that were grossly examined in terms of structural malformations of the craniofacies and the mesencephalon.

As one of the characteristic phenotypes of Gpr161 cKO fetuses was a midbrain protrusion initially observed at E13.5 and at E15.5, we sought to investigate the underlying cellular defects responsible for this abnormal phenotype. We were interested in clarifying whether these midbrain protrusive phenotypes are primarily due to the ectopic Wnt1 overexpression or not, as was previously reported (Lewis et al., 2013). We observed that Cre control fetuses failed to express any similar midbrain protrusive phenotypes as were seen in the Gpr161 cKO fetuses (Figure 1B). In addition, Gpr161 ^f/f ;Nestin-Cre fetuses showed similar protrusive tectal defects at E13.5 (Supplementary Figure S2 and Supplementary Table S3) although the phenotypes are less severe than that of Gpr161 cKO. Both supported that protrusive tectal phenotypes in Gpr161 cKO resulted from Gpr161 deletion.

As H&E staining provided evidence of the increased cell proliferation in the midbrain regions in Gpr161 cKO (Figure 2A), we performed immunohistochemistry (IHC) with Ki67 and pHH3 markers to affirm that the proliferation was regulated in Gpr161 cKO at E13.5. The Ki67 positive cells in the midbrain regions of Gpr161 cKO fetuses were significantly increased compared to those of the controls (Gpr161 ^f/f ) (Figures 2B,C), whereas pHH3 positive cells trended towards being increased in the midbrain regions of Gpr161 cKO fetuses but were not statistically significant. Interestingly, the Ki67 positive cells were widely spread in the ventricular zone (VZ). The pHH3 positive cells tend to be located in the mitotic zone of VZ (Arimura et al., 2019) in WT whereas they were more widely spread out in VZ of the dorsal midbrain in Gpr161 cKO fetuses (Figure 2B). In addition, we observed the increased Gli1 expression in the dorsal midbrain of Gpr161 cKO fetuses (Supplementary Figure S3).

Sonic hedgehog signaling as well as Wnt signaling are known to regulate midbrain patterning and proliferation (Brault et al., 2001; Bayly et al., 2007; Blaess et al., 2008). Gpr161 is an established negative regulator of sonic hedgehog signaling in multiple developmental contexts (Mukhopadhyay et al., 2013; Hwang et al., 2018; Kim et al., 2019), and is additionally involved in regulating Wnt signaling as well (Li et al., 2015; Kim et al., 2019). To determine if Shh and Wnt signaling are involved in the increased cell proliferation in the protrusive midbrain of Gpr161 cKO fetuses, we measured Shh and Wnt signaling activities within dissected midbrain tissues from floxed/Cre controls and Gpr161 cKO fetuses (Figure 3). The protein levels of Gli1 and the RNA levels of Gli1, Ptch1, and Fgf15, Shh target genes, were increased (Figures 3A,B). The repressor form of Gli3 was decreased, consistent with increased Gli1 levels in Gpr161 cKO. These results revealed increased Shh signaling activities in the midbrain tissues of Gpr161 cKO fetuses. Intriguingly, the protein levels of the Wnt signaling molecules, p-LRP6, Dvl2 (both significantly upregulated) and β-catenin (tends to increase, but not statistically significant), were increased, and one of Wnt target genes, CyclinD1, was increased (Figures 3A,B). However, classical target gene, Axin2, was not significantly changed in Gpr161 cKO midbrain tissues (Figure 3B). These results suggested that Gpr161 possibly regulated Wnt signaling in multiple ways.

To further unbiasedly identify the molecular basis of protrusive tectum phenotypes in Gpr161 cKO fetuses, we performed RNA sequencing analysis using midbrain tissues of Gpr161 ^f/f (floxed control), Gpr161 ^f/+ ;Cre/+ (Cre control), and Gpr161 cKO (Gpr161 ^f/f ;Cre/+) fetuses at E13.5 (Figure 4). The heatmap showed that the floxed control and the Cre control had a similar gene expression pattern, except Wnt1 as the Cre control had a higher Wnt1 expression compared to floxed control (Figure 4A). The top 10 differentially expressed genes (DEGs) in the midbrain tissues of Gpr161 cKO fetuses included: Gli1, Hhip, Ptch2, Nkx6-2, Fgf15 (Shh target genes-upregulated), Fgf8, Spry1 (Fgf signaling related genes-upregulated), Hes3 (Notch signaling related gene-upregulated), Wnt1 (agonist in Wnt signaling-upregulated) and Draxin (antagonist in Wnt signaling-downregulated) (Figure 4B). These results demonstrated that there was increased Shh, Wnt, as well as Fgf and Notch signaling in the midbrain regions of Gpr161 cKO fetuses. The Gene Ontology (GO) analysis further demonstrated that DEGs were highly enriched in the processes involved with neurogenesis, neuronal differentiation, neuronal morphogenesis, and mitotic cell cycle (Figure 4C). Taken together, the increased Shh and Wnt signaling are associated with the etiology of protrusive tectum phenotypes found in Gpr161 cKO.

The gross morphology (Figure 1B) and histological analysis (Figure 2A) suggested abnormal facial and cranial structures in Gpr161 cKO fetuses. We observed up to 69% of the Gpr161 cKO fetuses had encephalocele (Supplementary Table S2). The observed encephalocele could also be secondary to skull defects. To investigate craniofacial bone development in Gpr161 cKO fetuses, we performed skeletal staining with Alcian Blue (unmineralized cartilages) and Alizarin Red S (mineralized cartilages and bones). We observed a significant loss of mineralized skull and facial bones in Gpr161 cKO fetuses at E17.5 (Figure 5A). Specifically, the frontal, maxillary, and mandibular bones, which are derived from neural crest cell lineages, were significantly underdeveloped while the frontal bones failed to even form. The formation of parietal bones, which are derived from paraxial mesodermal cell lineages, was also severely reduced.

To further validate the skeletal staining results, we performed a bone segmentation study using 3D micro-CT imaging and further measured the volume of each bone based on micro-CT data within the craniofacial regions of floxed control, Cre control and Gpr161 cKO fetuses at E17.5 (Figures 5B,C). We failed to observe any overall head size differences between controls and Gpr161 cKO fetuses. Consistent with the findings of the skeletal staining studies, the neural crest lineage derived bones in the cranial vault and facial bones, including the maxilla, premaxilla, mandible, and frontal, were significantly underdeveloped or completely absent (Figure 5B) and their volumes were significantly reduced in Gpr161 cKO fetuses (Figure 5C). However, palatine bone formation was not changed, and the volume of nasal bones was increased in Gpr161 cKO fetuses. Additionally, segments and volume of the parietal bones were significantly reduced as shown in the skeletal staining of the fetuses, whereas the formation of other bones derived from paraxial mesoderm, such as the interparietal and occipital bones, were not affected in Gpr161 cKO fetuses. These results strongly suggest that Gpr161 has a significant role in the formation of the neural crest derived cranial vault and facial bones.

As previously reported (Lewis et al., 2013) and shown in Figure 1A, Wnt1-Cre driver-mediated recombination initially occurs in the midbrain dorsal neuroectoderm and the neural crest cells derived from mesencephalon forms craniofacial cartilages and bones (Santagati and Rijli, 2003). The brain hypertrophy, specifically protrusive tectal phenotypes in Gpr161 cKO fetuses with Wnt1-Cre, appears as if it is a phenocopy of Ptch1 cKO with Nestin-Cre (Martinez et al., 2013). We also observed identical phenotypes in Gpr161 cKO with Nestin-Cre fetuses (Supplementary Figure S2), indicating the critical role that Gpr161 serves during midbrain morphogenesis. The dorsal midbrain proliferation was increased in both Gpr161 cKO with Wnt1-Cre and Ptch1 cKO with Nestin-Cre, suggesting that Shh signaling is required for the tectal progenitor cell proliferation. Indeed, Shh target gene expression was increased (Figures 3, 4B), and Gli3 processing (Figure 3A) was inhibited in midbrain tissues of Gpr161 cKO with Wnt1-Cre, indicating the increased Shh signaling. In addition, five out of the top ten DEGs in the RNA seq data set were Shh target genes, including Gli1, Hhip, Ptch2, Fgf15, and Nkx6.2 (Figure 4B), which were upregulated in Gpr161 cKO fetuses, providing further supporting evidence of an increased Shh signaling in the Gpr161 cKOs. Together, these results explain that increased Shh signaling involved in the increased mesencephalon proliferation, thereby causing the protrusive tectum in the affected fetuses. The forebrain ventricular surface was also reported to be expanded during embryogenesis from radial glial over-proliferation upon Nestin-Cre mediated deletion of Gpr161 (Shimada et al., 2019). Cortical phenotypes that were observed included polymicrogyria in the medial cingulate cortex, increased proliferation of intermediate progenitors and basal radial glia, and altered neocortical cytoarchitectonic structure with increased upper layer and decreased deep layer neurons. Overall results support the role of Gpr161 in the cell proliferation during fore/mid brain morphogenesis. In addition, the protrusive mesencephalon in RhoA cKO with Wnt1-Cre fetuses results from the hyperproliferation of midbrain progenitor cells via increased Shh signaling, providing yet more evidence supporting the role of Shh signaling in dorsal midbrain progenitor cell proliferation (Katayama et al., 2011). As Ptch1 cKO with Nestin-Cre showed similar midbrain protrusive phentoypes as did Gpr161 cKO, it will be an interesting future study to identify the relation between Gpr161 and other Shh signaling molecules (e.g. Ptch1 or Smo) during embryonic midbrain morphogenesis.

Wnt signaling is well known to be associated with mesencephalon development (Brault et al., 2001; Panhuysen et al., 2004; Chilov et al., 2010), and our results further demonstrated that the activated Wnt signaling is also involved in the mesencephalic cell proliferation. The activities of upstream signaling molecules in Wnt signaling, phosphorylated LRP6 and Dvl2, were increased and one of target genes, CyclinD1, was increased (Figure 3). At the same time, the expression of Draxin, an inhibitor of canonical Wnt signaling (Miyake et al., 2009), was decreased in Gpr161 cKO fetuses (Figure 4B), providing evidence that the increased Wnt signaling secondary to Gpr161 depletion contributes to the observed mesencephalon cell proliferation. However, we cannot rule out the possibility that Wnt and Shh signaling regulate cell proliferation in the midbrain tissues of Gpr161 cKO in parallel. Additionally, the level of downstream Wnt signaling molecules, such as β-catenin and Axin2, was unchanged (Figure 3), maintaining the complexity of the interactions between Shh and Wnt signaling. Nonetheless, the interplay between Shh and Wnt signaling in mesencephalon progenitor cell proliferation and differentiation in Gpr161 cKO fetuses remains an inadequately resolved question (Tang et al., 2010). Clearly, the molecular interactions between Gpr161, Fgf, Notch, and Wnt signaling requires future investigation.

Gpr161 depletion with Wnt1-Cre resulted in severe craniofacial skeletal defects (Figure 5). It is notable that the portions of the craniofacial skeleton derived from neural crest cells were completely absent, while those portions derived from mesenchyme were only partially reduced. The partial reduction of parietal bones, which derived from mesodermal lineages, in Gpr161 cKO could be explained by two possibilities. Jiang et al. (2002) suggests the possibility that the neural crest derived meninges are required for the mesodermal derived bone ossification. The other possibility is the possible Cre expression in non-neural crest cells due to leakage or due to insertion side effect in Wnt1-Cre line (Doro et al., 2019).

The Gpr161 depletion in mesenchymal lineages was reported to include phenotypes with the posterior cranial vault defects caused by the lack of intramembranous ossification (Hwang et al., 2018), supporting the role of Gpr161 during cranial vault skeletogenesis. One possible explanation regarding craniofacial skeletal defects in Gpr161 cKOs is that the increased neural crest cell populations due to increased Shh signaling resulting from the Gpr161 depletion might cause craniofacial skeletal defects, as was previously observed in Fuz knockout mice (Tabler et al., 2016). Another possible explanation is the involvement of Wnt/β-catenin signaling as it plays a critical role in the intramembranous bone formation (Day and Yang, 2008). This possibility is supported by the previous report that β-catenin cKO embryos with Wnt1-Cre was phenocopied in Gpr161 cKO with Wnt1-Cre in terms of their complete lack of neural crest cell derived cranial vault and facial bones (Brault et al., 2001). However, the underlying cellular and molecular processes of craniofacial skeletal defects in Gpr161 cKO need to be further explored. It will be an additional future study to investigate the relation between Gpr161 and other Shh signaling molecules, such as Sufu or Smo, during embryonic craniofacial skeletogenesis.

The pathogenesis and the etiology of the structural malformations known as encephaloceles have not been comprehensively studied even though it is often classified as a type of neural tube defects. A recent publication (Rolo et al., 2019) showed that encephaloceles result from the defective surface ectoderm in a post-neurulation manner along with severe calvarial bone defects at a later embryonic stage in the surface ectoderm specific deletion of Rac1 mouse model. As Gpr161 cKO did not show any significant phenotypic malformations at E9.5 or 10.5, the encephaloceles in Gpr161 cKO clearly occur in post-neurulation stage embryos. In addition, encephaloceles are often observed to be associated with other human birth defects, including other NTDs, cleft palate, craniosynostosis (Naidich et al., 1992; Caplan et al., 2002; Ganapathy et al., 2014), which are similar phenotypes to those observed in mouse models with Gpr161 cKO or Rac1 cKO (Rolo et al., 2019). Therefore, our study provides a potential mouse model of encephaloceles, which enables us to further study the molecular pathogenicity of this much-understudied congenital malformation.

In this study, we attempted to unravel the role of Gpr161 in embryonic midbrain development and craniofacial bone formation in mouse models with Wnt1 lineage specific deletion of Gpr161, demonstrating the distinct role of Gpr161 in mesencephalon proliferation and neural crest derived craniofacial skeleton morphogenesis. Our data suggests that Gpr161 cKO can serve as a mouse model for enhancing our understanding of the basic developmental biology of encephaloceles. The results from this study also suggest a possible genetic association of GPR161 with such craniofacial defects as cleft palate, as well as encephaloceles in humans. Based on this study, we will further delineate the role of Gpr161 in neural crest cell differentiation and will also study the genetic association of GPR161 and associated Sonic Hedgehog genes and craniofacial birth defects in human patient samples.

All mice were housed and handled according to the guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of The University of Texas at Austin. Gpr161 conditional knock out mice (Gpr161 flox) were generated and graciously provided by Dr. Saikat Mukhopadhyay (UT Southwestern, Dallas) and the detailed information was previously reported (Hwang et al., 2018). The transgenic mice, Wnt1-Cre (#009107), Nestin-Cre (#003771), and Rosa26-lsl-LacZ (#003474), were purchased from Jackson laboratory. The genotypes of the mice and embryos/fetuses were determined by PCR-based genotyping.

Embryos were collected at E9.5 and E11.5 from timed mated breeding pairs between Wnt1-Cre and Rosa26-LacZ. The harvested embryos were fixed, and β-gal staining was performed according to the manufacturer’s instruction (Millipore Sigma). The images were captured by Leica stereomicroscope with a Nikon digital camera.

Fetuses were harvested at either E13.5 or E15.5 from timed mating between Gpr161 ^f/f and Gpr161 ^f/+ ;Wnt1-Cre/+. Collected fetuses were fixed, paraffin embedded, and sectioned with 4 um-thickness. The paraffin sections were deparaffinized, dehydrated, antigen-retrieved, blocked (blocking solution: Thermo Fisher scientific), and incubated with primary antibodies (Ki67, pHH3, and Gli1) diluted with Lab Vision™ Antibody Diluent Quanto (Thermo Fischer scientific) overnight at 4°C. After washing, sectioned were incubated with Horseradish peroxidase (HRP) polymer conjugate (UnltraVision™ LP detection system, Thermo Fisher scientific) and DAB (Boster Bio). The sections were counterstained with hematoxylin (Thermo Fisher scientific). Images were captured with All-In-One Fluorescent (Keyence) microscope using a ×2 and ×20 objective. To assess the positive cells for each proliferation marker Ki-67 and pHH3 in the dorsal midbrain regions of each embryo, the images were digitized and analyzed with Fiji (NIH), with image analysis. For quantification, a total of 4–6 fields per dorsal midbrain regions per each embryo were captured at ×20 magnification. Each field was divided into 100 equal squares and subjected to color deconvolution. For each of the images, the Shanbhag threshold (Shanbhag, 1994) was applied to DAB-only images. The stained fraction was measured in the threshold fields (color deconvolution also gives 4′6′-diamidino-2-phenylindole (DAPI), and helps identify each threshold cell), then percentage of DAB positive cells from each image was calculated and assessed. The mean of 10 values obtained across from the six fields was calculated for each of the markers and each of the embryos, which then were subjected for the statistical analysis.

The midbrain tissues were dissected from E13.5 mouse fetuses as described (Weinert et al., 2015). The dissected midbrain tissues were lysed with Radioimmunoprecipitation assay (RIPA) lysis buffer. The denatured protein samples were immunoblotted using anti-Gli3 (Santa Cruz Biotechnology), Gli1, p-LRP6, Dvl2, β-actin (Cell signaling), β-catenin (BD bioscience) and then with 1RDye^® 800CW goat anti-rabbit IgG and 1RDye^® 680CW goat anti-mouse IgG secondary antibodies (LI-COR). The images were captured by Odyssey^® (LI-COR).

The dissected midbrain tissues from E13.5 mouse fetuses were lysed with Trizol and total RNA was purified with Direct-zol RNA kit (Zymo research). For quantitative RT-PCR, 500 ng-1 µg of RNA was used to synthesize cDNA using iScript reverse transcription Supermix kit (Bio-Rad). The quantitative RT-PCR was performed using SsoAdvanced™ Universal SYB R Green Supermix (Bio-Rad) according to the manufacturer’s instruction. The primers for qRT-PCR are as below.

The tissues were harvested as described in Midbrain dissection and total RNAs were extracted with Direct-zol RNA kit (Zymo Research). The quantity and integrity of RNAs were analyzed by Nanodrop (Thermo Fisher) and Bioanalyzer (Agilent Technologies). The library was prepared with NEBNext Ultra RNA with Poly-A selection and (NEB) was sequenced on an Illumina Hi-Seq 4000 (Admera Health LLC). The differential gene expression was determined with fold change >1.5 and p < 0.05 genes with <1 count per million (cpm). Any gene with a p-value greater than FDR, after Benjamini-Hochberg correction for multi-testing, was deemed significantly differentially expressed under the test condition as compared to the control. The dataset was analyzed by the Gene Ontology (GO) enrichment analysis.

Skeletal staining was performed using a modified Alcian Blue/Alizarin Red staining procedure (Kessel et al., 1990). Briefly, the E17.5 fetuses were eviscerated and fixed with 95% ethanol and then acetone. Fixed fetuses were incubated with staining solution (0.005% Alizarin red S 0.015% Alcian Blue GS in 5% acetic acid, 5% H₂O and 90% ethanol) for 3 days at 37°C. After washing, samples were kept in 1% KOH for 48 h. For long term storage, specimens were transferred into 20, 50 and 80% glycerol solutions and were ultimately maintained in 100% glycerol. The images were captured by a Leica stereomicroscope with a Nikon digital camera.

The E17.5 fetuses were fixed with 10% formalin followed by 70% ethanol. Specimens were scanned at the University of Texas High-Resolution X-ray CT Facility using the flat panel detector on a Zeiss Xradia 620 Versa. The X-ray source was set to 70 kV and 8.5 W with no filter. A total of 2001 0.1s projections were acquired over ±180 degrees of rotation with no frame averaging. A source-object distance of 18.0 mm and a detector-object distance of 251.7 mm resulted in 9.98-micron resolution. The resulting data were segmented in Avizo software v.2020.2.

The experiments were done in triplicate Unless specifically stated otherwise and the data was analyzed by the Standard Deviation (SD) with student t-test for comparing groups. For IHC data, the statistical analysis was performed with 2-way ANOVA, followed by Tukey’s test for multiple comparison (GraphPad Prism9).