The transcription factor Hypermethylated in Cancer 1 (Hic1) regulates neural crest migration via interaction with Wnt signaling

Abstract

The transcription factor Hypermethylated in Cancer 1 (HIC1) is associated with both tumorigenesis and the complex human developmental disorder Miller-Dieker Syndrome. While many studies have characterized HIC1 as a tumor suppressor, HIC1 function in development is less understood. Loss-of-function mouse alleles show embryonic lethality accompanied with developmental defects, including craniofacial abnormalities that are reminiscent of human Miller-Dieker Syndrome patients. However, the tissue origin of the defects has not been reported. In this study, we use the power of the Xenopus laevis model system to explore Hic1 function in early development. We show that hic1 mRNA is expressed throughout early Xenopus development and has a spatial distribution within the neural plate border and in migrating neural crest cells in branchial arches. Targeted manipulation of hic1 levels in the dorsal ectoderm that gives rise to neural and neural crest tissues reveals that both overexpression and knockdown of hic1 result in craniofacial defects with malformations of the craniofacial cartilages. Neural crest specification is not affected by altered hic1 levels, but migration of the cranial neural crest is impaired both in vivo and in tissue explants. Mechanistically, we find that Hic1 regulates cadherin expression profiles and canonical Wnt signaling. Taken together, these results identify Hic1 as a novel regulator of the canonical Wnt pathway during neural crest migration.

Introduction

Developmental processes and cancer formation and progression share many common regulators. Growth factors that control early cell fate determination and cell differentiation are often dysregulated in cancer, and genes that govern cell movements and tissue morphogenesis are frequently reactivated during cancer metastasis to promote invasive cell behaviors. Because of reiterative usage of genes and pathways in embryogenesis and tumorigenesis, discoveries made about animal development often help to shed light on cancer machineries. Similarly, cancer research regularly provides fresh insight about developmental mechanisms. Parallel studies of genes involved in both processes can deepen our understanding on gene functions and their underlying mechanisms.

Hypermethylated in Cancer 1 (HIC1) is a tumor suppressor with its genomic locus on human chromosome 17 telomeric to p53 within a region that frequently is lost in tumor cells. Single allele loss of the locus is frequently accompanied by silencing of the second HIC1 allele via promoter hypermethylation (Makos et al., 1993; Wales et al., 1995). HIC1 expression levels are decreased in multiple tumor types compared with their neighboring healthy tissues, including lung, colon and brain. Exogenous expression of HIC1 in the SW40 colon cancer cell line, which exhibits HIC1 hypermethylation and low basal expression, decreased cell proliferation and migration, and cells regained a more epithelial-like morphology (Wales et al., 1995). Additional studies in other tumor types support the finding that HIC1 expression is inversely related to cancer progression (Cheng et al., 2013; Fujii et al., 1998; Kanai et al., 1999; Markowski et al., 2015). However, detailed mechanisms underlying the tumor suppression activity of HIC1 are not completely understood.

HIC1 is a transcription factor (TF) classified as a member of the BTB/POZ (Broad complex, Tramtrack, Bric à brac or poxvirus and zinc finger) zinc finger family. These TFs are characterized by the presence of an N-terminal POZ domain involved in protein-protein interactions and a C-terminal zinc-finger binding domain for direct DNA interaction. Founder proteins within this class include the Drosphila proteins Broad complex (BR-C), Tramtrack (Ttk) and Bric à brac (Bab) for whom the family is named. All of these proteins are involved in fly embryonic development through their roles as transcriptional repressors (Kelly and Daniel, 2006). HIC1 is shown to recruit CtBP and proteins of the Polycomb Repressive Complex 2 to suppress transcriptional activities (Boulay et al., 2012; Deltour et al., 2002). However, a recent report reveals that HIC1 can act as both a transcriptional repressor and an activator during induction of human regulatory T cells (Ullah et al., 2018). A number of individual genes have been identified as direct targets of HIC1 in mammalian cell culture, including Sirtuin 1 (SIRT1), Fibroblast growth factor binding protein 1 (FGF-BP1), Cyclin D1 (CCND1), ephrinA1 and EphA2 (Briones et al., 2006; Chen et al., 2005; Foveau et al., 2011; Van Rechem et al., 2010; Y. Wang et al., 2018; W. Zhang et al., 2010). These identified genes are important regulators of cellular processes, such as cell proliferation, angiogenesis, migration and invasion, that play crucial roles in both cancer and development. The results thus imply that HIC1 may control key regulatory factors to influence both tumorigenesis and embryogenesis.

The evidence that supports a function of hic1 in early vertebrate development comes from hic1 loss of function studies in mouse (Carter et al., 2000; Pospichalova et al., 2011). Several independent hic1 deletion alleles have been generated, and the resulting mice with complete removal of hic1 show embryonic lethality accompanied with a number of developmental defects, including limb abnormalities, omphalocele, and craniofacial malformation. Interestingly, the spectra of the developmental abnormalities in hic1 null embryos are reminiscent of the phenotypes in human patients with Miller-Dieker Syndrome, the complex human developmental disorder associated with deletion of a large genomic region containing HIC1 (Carter et al., 2000). The shared defects in neural and craniofacial structures between hic1 gene knockout mice and human patients suggest that hic1 may regulate embryonic tissues that give rise to these structures, but little work has been done to understand how hic1 may function during embryonic development.

The vertebrate craniofacial structures are largely derived from the multi-potent progenitor cells called the neural crest cells (NCC). This transient cell population, often referred to as the fourth germ layer, arises early during development within the embryonic ectoderm at the border between the neural plate and the non-neural ectoderm (neural plate border, NPB) (Hall, 2001), (Alfandari et al., 2010; Szabó and Mayor, 2018). Upon cell specification and determination, NCC undergo an epithelial-to-mesenchymal transition (EMT) and delaminate from the surrounding cells to migrate along distinct routes before differentiation into a variety of cell types depending on their positions. NCC that arise from the anterior of the embryo, referred to as cranial neural crest cells (CNCC), migrate into the head region and contribute to formation of cranial ganglia and craniofacial skeleton. A complex gene regulatory network governing sequential NCC induction, specification, EMT, migration and differentiation has been discovered from collective and extensive investigation from many groups (Sauka-Spengler and Bronner-Fraser, 2008). Intriguingly, many of the transcription factors identified from such studies on NCC development are shown to be activated frequently during progression of epithelial cancers (Micalizzi and Ford, 2009). For example, the genes that activate the EMT process and initiate cell migration, such as the TFs Snail and Twist, are often associated with higher-grade metastatic tumors (Moody et al., 2005). The data emphasize again the fluid transition between developmental regulators and tumor factors.

Based on the targeted hic1 gene deletion studies in mouse and the similarity of craniofacial defects to the phenotypes of Miller-Dieker Syndrome patients, we speculate that hic1 may regulate neural crest development during vertebrate embryogenesis. To test this hypothesis, we use the Xenopus model here because of the available research tools and the rich literature of using this model to explore gene function and mechanisms in regulating neural crest development. We show in this study that hic1 regulates migration, but not specification, of the cranial neural crest cells, and it does so by influencing cadherin expression and canonical Wnt signaling.

Results

Hic1 is expressed throughout early Xenopus development

Previous work has shown that Hic1 is expressed during early development in both mouse and zebrafish (Bertrand et al., 2004; Pospichalova et al., 2011). Additionally, RNA-sequencing identified hic1 expression in embryos of both Xenopus laevis (Session et al., 2016 and Xenbase.org, our unpublished data) and tropicalis (Owens et al., 2016). To confirm hic1 expression during early Xenopus development in our hands, we performed RT-PCR using cDNAs obtained from whole Xenopus laevis embryos at various stages of early embryogenesis (Fig. 1A). We found that hic1 is expressed in the 2-cell stage embryos and its expression remains until at least stage 25, indicating that hic1 mRNA may be both maternally transferred and expressed by the zygotic genome. While previous work showed hic1 mRNA expression patterns at two stages of X.l. development (Z. Zhang et al., 2017), we next sought to determine the spatial distribution of hic1 mRNA across a broader range of developmental timepoints by performing in situ hybridization (ISH) using an antisense probe generated against hic1 (Fig. 1 B). hic1 is seen to have a weak, diffuse distribution in both animal and marginal regions in blastula and gastrula stage embryos (i-iv). Beginning at neurulation, hic1 expression is apparent in the developing neural/neural crest regions (v-viii, arrows) with strongest expression seen around the optic cup and along the trunk. At tadpole stages, hic1 is expressed highly throughout the craniofacial region (ix). Parallel ISH using a hic1 sense probe did not show obvious signals (i’ to ix’). Additional experiments show that hic1 mRNA is expressed in similar regions in Xenopus tropicalis embryos (Fig. 1 C), including within the branchial arches and around the optic cup (iii, arrows). The expression pattern of hic1 hence implies a possible role of the gene in early ectodermal development that is conserved between X. laevis and tropicalis species.

Figure 1. Expression of hic1 RNA throughout early Xenopus laevis development.

Representative image of RT-PCR using cDNA generated from whole embryos at indicated stages (Nieuwkoop-Faber, NF) across early developmental timepoints (A). In situ hybridization with hic1 anti-sense and sense control probes identifies mRNA expression pattern over early developmental timepoints (B). (i) Blastula NF stage 9 animal view and (ii) vegetal view, (iii) gastrula NF stage 10.5 animal view and (iv) vegetal view, (v) neurula NF stage 19 anterior view and (vi) dorsal view, (vii) tailbud NF stage 24 side view, (viii) tailbud NF stage 32 side view, (ix) tadpole NF stage 40 side view. Arrows indicate hic1 expression within the placodal ectoderm (v), neural plate border (vi) and branchial arches (vii, viii). In situ hybridization with hic1 anti-sense probe identifies mRNA expression pattern in Xenopus tropicalis embryos (C). (i) neurula NF stage 17 anterior view and (ii) dorsal view (iii) tailbud NF stage 27 side view and (iv) tadpole NF stage 35 side view. Arrows indicate hic1 expression within the neural plate border (i, ii) and branchial arches (iii).

Hic1 regulates craniofacial development

To investigate directly a role for Hic1 during ectodermal development, we sought to manipulate hic1 expression levels specifically within the presumptive neural and neural crest domains via targeted injection of molecules into the dorsal animal region of early Xenopus embryos. Injection of either hic1 mRNA (total 500 pg, overexpression) or a translation-blocking anti-sense morpholino oligonucleotide directed against hic1 (total 50 ng hic1 MO, knockdown) divided between both dorsal animal (DA) cells of 8-cell stage embryos results in developmental defects predominately restricted to the cranial region (Fig. 2 A, C). Defects are most apparent in the developing eyes, which are frequently smaller and underdeveloped, and altered overall shape of the head. Additionally, manipulated embryos sometimes display differences in density and localization of melanocytes compared to control embryos. Overall, hic1 knockdown results in a more severe phenotype than hic1 overexpression. To examine whether the phenotype of hic1 morphant is specific, we co-injected hic1 MO along with a sub-phenotypic dose of hic1 mRNA that lacks the MO target sequence (50 pg hic1*). The expression of hic1* mRNA results in rescue of the hic1 knockdown phenotype (Fig. 2 B, C), indicating that the morphant phenotype is due to specific knockdown of hic1 mRNA. To directly test the ability of hic1 MO to block translation of hic1 mRNA, an immunofluorescence assay was performed (Fig. 2 D). Both dorsal animal cells of 8-cell-stage embryos were injected with mRNA encoding hic1 containing an HA epitope tag at the C-terminal end (hic1-HA, 250 pg each cell) or hic1-HA mRNA that additionally contained the sequence upstream of the translation start site that could be recognized by the MO (hic1-5’UTR-HA). In one of the DA cells mRNA was co-injected with hic1 MO (25 ng) plus fluorescein dextran as a tracer. Embryos were allowed to develop until early neurula stages at which time transverse sections were stained using an anti-HA antibody. In embryos injected with hic1-HA mRNA, nuclear HA staining is seen on both sides of the midline including in the MO-co-injected region (as indicated by the green fluorescent signal and outlined by dashed lines). However, in embryos injected with hic1-5’UTR-HA mRNA, nuclear HA staining is substantially or completely lost in the co-injected region, indicating that hic1 MO efficiently blocks protein translation from hic1 mRNA containing the 5’ UTR target sequence.

Figure 2. Disruption of normal hic1 expression results in head defects.

Injection of either hic1 mRNA (500 pg total) or hic1 MO (50 ng total) in both dorsal animal cells of 8-cell stage embryos results in head defects including altered head shape and small eyes (A). Knockdown of hic1 expression by MO injection is rescued by co-injection with hic1 mRNA (50 pg) that lacks the MO binding site (B). Experiments were performed in biological and technical quadruplicate and phenotypes were quantified as percent of total embryos. Total number of embryos for each group is indicated above bars within the graphs (C). Immunofluorescence staining of transverse tissue sections shows Hic1 protein expression after injection of hic1-HA mRNA (250 pg) (D). Co-injection of hic1 MO (25 ng) in one side of embryo results in loss of Hic1 protein only when the MO binding site is present (hic1 5’UTR-HA, 250 pg) in injected hic1 mRNA. Dashed lines indicate region of co-injection as assessed by fluorescein dextran.

To further explore the consequence of altered hic1 levels on craniofacial development, we cultured embryos injected unilaterally with either 250 pg of hic1 mRNA or 25 ng of hic1 MO to stage 45 and examined the craniofacial cartilage structures upon staining with Alcian blue. Both overexpression and knockdown of hic1 results in a spectrum of malformations to the craniofacial cartilages (Fig. 3). Dorsal view of the un-stained embryos shows that while the uninjected side of the embryos appears normal, the injected sides are misshapen and overall the heads are asymmetrical (Fig 3 A). These defects are confirmed with Alcian blue staining and can be divided largely into two groups. Some embryos have minor loss of single cartilages or misshapen and unconnected cartilage across the injected to uninjected side of the embryos. We characterize these embryos as having mild defects (Fig. 3A, B, arrows). Other embryos have more overt loss of large regions of cartilages, including the absence of the entire Meckel’s, ceratohyal or branchial cartilages, or loss of multiple cartilages (Fig. 3 A, arrows). We characterize these embryos as having severe defects (Fig. 3A, B). Consistent with morphological phenotypes, the defects in craniofacial cartilages are more severe in hic1 morphant embryos than in hic1 overexpression embryos (Fig. 3B). Together, our results indicate that hic1 plays a crucial role in craniofacial development.

Figure 3. Disruption of normal hic1 expression levels results in defects of the craniofacial cartilages.

Unilateral injection of either hic1 mRNA (250 pg) or hic1 MO (25 ng) in one DA cell results in misshapen craniofacial regions and malformations of the craniofacial cartilages as shown by alcian blue staining (A). Cartilage phenotypes were categorized as mild or severe depending on the extent of cartilages involved. Both dorsal and ventral views of individual heads are shown. Arrows indicate regions of disrupted cartilage development. CH: ceratohyal cartilage M: Meckel’s cartilage BR: branchial cartilage. Experiments were performed in biological and technical quadruplicate. Phenotypes were quantified and shown as percent of total embryos, with total embryo numbers indicated above bars within the graph (B).

Hic1 affects expression patterns of genes involved in cranial neural crest migration

Our above data show that hic1 is expressed within the domains of developing cranial neural crest cells (CNCC) and its proper level is required for normal development of CNC derived tissues, such as craniofacial cartilages. This implies that hic1 regulates CNC development. We therefore addressed this issue directly by analyzing markers expressed at different stages of CNC development. CNC is specified and patterned at the neural plate border at neurula stages, undergoes epithelial-to- mesenchymal transition (EMT) and delamination from the neighboring ectoderm at the end of neurulation, and migrates away from the dorsal neural tube and into the cranial region and through the branchial arches. Known markers of CNC development at these various stages have been reported previously (Pegoraro and Monsoro-Burq, 2012; Sauka-Spengler and Bronner-Fraser, 2008), allowing us to use these markers to identify when and where hic1 may be involved in regulating CNC development. In these experiments, we co-injected a lineage tracer encoding nuclear beta-galactosidase together with hic1 mRNA or MO into one dorsal animal cell of 8-cell stage embryos. The injected embryos were collected from neurula to tailbud stages, stained with the red-Gal substrate, and subjected to ISH analysis. Results of marker expression in stage 14 embryos show that markers of neural (sox2), non-neural (xk70), neural plate border (ap2α, msx1), and early neural crest (snail, slug) are all expressed at normal levels in hic1 mRNA- or MO-injected side (marked by the red-gal staining) when compared with the un-injected side in the same embryos (Fig. 4 A). Similarly, the CNC marker slug was unchanged at stage 16, at a point when CNC cells are fully specified but are characterized as pre-migratory (Fig. 4 B). These results indicate that early patterning and specification of the neural crest is unaffected by either hic1 overexpression or knockdown.

Figure 4. in situ hybridization shows changes in CNC gene expression patterns upon hic1 manipulation.

(A-D) Embryos were unilaterally injected with hic1 mRNA (250 pg) or hic1 MO (25 ng) and β-galactosidase mRNA (200 pg) as a tracer (red color) in one DA cell and ISH was performed at various stages throughout CNC development. NF stage 14 embryos were assessed for markers of neural ectoderm (sox2), non-neural ectoderm (xk70), neural plate border (ap2α, msx1) and neural crest (snail, slug) (A). NF stage 16 embryos were probed for the neural crest-expressed gene slug (B). NF stage 19–20 embryos were assessed for the expression of the neural crest-expressed genes slug, snail and sox9 (C). NF stage 24 embryos were assessed for expression of the neural crest-expressed gene sox10 (D). (E) Embryos were unilaterally injected with hic1 MO (25ng), hic1 mRNA (50 pg) and β-galactosidase as a tracer. NF stage 19–20 embryos were assessed for the neural crest marker slug. All experiments were performed in at least biological and technical triplicate and number of phenotypic embryos observed out of the total number of embryos assayed is indicated.

To see whether hic1 modulates migration of the neural crest, we examined several markers at stages 19–20 when CNC has begun to migrate away from the dorsal neural tube. This is shown by the expression patterns of the CNC markers slug, snail and sox9, all of which have started to shift both anteriorly and laterally in the control embryos (Fig. 4 C). In contrast, when hic1 levels are elevated or knocked down, all three markers appear to remain close to the dorsal neural tube (Fig. 4 C). At stage 24, the neural crest cells expressing sox10 have migrated ventrally in distinct streams to populate domains in the branchial arches and more anteriorly around the developing optic vesicle (Fig. 4 D). In both hic1 overexpression and knockdown embryos, sox10 remains highly expressed; however, its expression domains are restricted to the most dorsal aspect of the embryos and do not reach the ventral branchial arch regions or surround the optic cups (Fig. 4D). Additionally, the altered expression pattern of the CNC marker slug observed in MO-injected embryos at stage 19–20 is recued upon coinjection with hic1* mRNA that lacks the MO-recognition sequence (Fig. 4 E). Taken together, our ISH data suggest that while hic1 is not required for initial neural crest specification and patterning, disruption of proper hic1 expression levels leads to defects in CNC migration.

hic1 expression is required for proper neural crest migration both in vivo and in vitro

To more directly assess a role for hic1 in CNC migration, we turned to a well-established system of in vivo CNC migration using labeled transplant from a donor embryo (Borchers et al., 2001b; Milet and Monsoro-Burq, 2014) and focused our studies on knockdown of the endogenous hic1 transcript. We injected embryos unilaterally in one dorsal animal cell at the 8-cell stage with 25 ng of either a control or hic1 MO along with 200 pg of mRNA encoding a membrane targeted GFP (GFP-CAAX). These embryos were allowed to develop until stage 16 when CNC are specified correctly but pre-migratory in both control and hic1 morphants (Fig. 3 B). At this point, the CNC can be identified easily and physically removed from the embryos. The GFP positive (GFP+) CNC population was then removed and transplanted into a wild-type un-injected sibling host. The resulting embryos were cultured to the late tailbud stages before being examined under the fluorescence microscope. The embryos containing transplanted CNC tissue from the control MO-injected embryos show that the GFP-expressing cells migrate normally as distinct streams into individual branchial arches (Fig. 5 A). However, the majority of the embryos harboring CNC transplants from the hic1 morphant displayed defects in CNC migration (Fig. 5 A, B). CNC either failed to migrate away from the transplant site all together (17% of the embryos without migration), or CNC cells were able to leave the site of transplant but were unable to migrate into the branchial arches and remained clumped together (61% of the embryos with aberrant migration). In some cases, individual cells or small groups of cells could be seen separated from the main CNC cluster, but were localized in areas not normally populated by CNC. These results suggest that hic1 may be required autonomously in the neural crest cells for their proper migration out of the dorsal neural tube.

Figure 5. hic1 knockdown leads to defects in migration of cranial neural crest transplants.

Embryos were injected unilaterally with hic1 MO (25 ng) or control MO (25 ng) plus GFP-CAAX mRNA (200pg) in one DA cell. At NF stage 16, GFP+ CNC were transplanted into a naïve stage-matched embryo and embryos were imaged at NF stage 29 for CNC migration by GFP fluorescence (A). Migration was categorized as normal (into the three branchial arches), aberrant (altered pattern of migration) or no migration (CNC did not leave the site of explant) and quantified as percent of total embryos. Total embryo numbers are indicated above bars within graph (B).

To further examine the function of hic1 in regulating neural crest migration, we next analyzed how hic1 knockdown would affect the ability of CNC cells to migrate on fibronectin-coated dishes. In this case, external cues are eliminated and we can more directly assess the innate ability of CNC to undergo EMT and initiate their migratory program. We injected and incubated embryos as above, isolated the CNC population, and allowed the explants to adhere to tissue culture dishes coated with fibronectin (10μg/ml). The explants were then incubated at 15°C overnight and imaged 14–16 hours later. By this time, the majority of the explants injected with control MO have completely spread out into a monolayer of cells that are moving outward from the center of the explants (63% with extensive migration), while a smaller proportion of the explants exhibit limited migration around the cluster of the cells remaining at the core of the explants (30% with limited migration) (Fig. 6 A, C). In the case of the explants from hic1 morphant, a much smaller proportion of the explants exhibit extensive migration (25%) while the majority of the explants show either limited (58%) or no (16%) migration (Fig. 6 A, C). The cells tend to remain highly clumped together without spreading out away from the center of the explants. Upon further inspection of the explants, we find that while all of the migrating regions from the explants of the control morphant contain GFP positive cells, migrating cells from the explants of the hic1 morphant are GFP negative with the GFP positive cells remaining clumped together in the center of the explants (Fig. 6 B). This result indicates that the hic1 MO-containing cells (GFP positive) have an impaired ability to initiate EMT and migration in a cell autonomous fashion. To further examine the effect of hic1 on NC migration, we next assessed cell migration of individual cells. In this experiment, explant tissues were first dissociated and plated on fibronectin-coated dishes as single cells. Cells were allowed to adhere for one hour followed by time-lapse microscopy with 3-minute intervals for 150 minutes. Individual cell tracks were generated to investigate the migratory patterns of single CNC over the entire time course (Fig. 6 D). We observed that while control MO-injected CNC actively migrate in random patterns over the entire time period, hic1 MO-injected CNC move very little. Indeed, a majority of these cells remain round in shape and have a complete lack of movement. Combining all collected cell tracks (N = 60) in rosette plots shows that there is a clear decrease in overall distance migrated in hic1 MO-injected CNC compared to control CNC (Fig. 6 E). This result further suggests that while loss of hic1 does not disrupt the ability of CNC cells to adhere to fibronectin, these cells lack the ability to spread and adopt a migratory morphology and activity. In combination with the results from the in vivo transplant experiments, we conclude that hic1 function is required within the CNC population for proper migration.

Figure 6. hic1 knockdown leads to defects in migration of cranial neural crest explants.

Embryos were injected unilaterally with hic1 MO (25 ng) or control MO (25 ng) plus GFP-CAAX mRNA (200pg) in one DA cell. (A-C) At NF stage 16, GFP+ CNC were removed and explanted onto a fibronectin-coated dish. Explants were imaged 14–16 hours later to assess for migration (A). Migrating cells from control MO embryos are GFP+, while migrating cells from hic1 MO embryos are GFP- (B). Migration was categorized as extensive (cells in a complete migrating monolayer), some (partial migration with some cells remaining clumped in the center of the explant) or no migration (all cells remain clumped in the explant) and quantified as percent of total explants. Total explant numbers are indicated above bars within graph (C). (D-E) At NF stage 16, GFP+ CNC were removed, dissociated, and plated onto a fibronectin coated dish as single cells. Cells were imaged for 150 minutes and individual cell movements were tracked. Still images from start (time = 0) and end (time = 150) of imaging show individual cell tracks (D). Cell tracks were normalized to a zero-point origin and plotted to show individual cell movements within the population. N= 65 control MO, 63 hic1 MO taken over three separate experiments.

hic1 knockdown disrupts canonical Wnt signaling dynamics in cranial neural crest

Neural crest cell EMT and subsequent migration are regulated by a complex network of signaling molecules and transcriptional regulators. Wnt signaling plays crucial roles throughout NCC development, including during migration (Pegoraro and Monsoro-Burq, 2012); but interestingly, Wnt signaling has to be transiently dampened in order for CNC to delaminate prior to the onset of migration (Maj et al., 2016; Rabadán et al., 2016). Sustained Wnt signaling during this time period results in migration phenotypes similar to what we observe upon hic1 knockdown. So, we next asked whether knockdown of hic1 disrupts Wnt signaling dynamics during CNC migration. Embryos were injected with 25 ng control or hic1 MO together with 200 pg GFP-CAAX, as performed previously, and CNC explants were dissected at stage 16 and placed on glass coverslips coated with fibronectin (100μg/ml). Explants were fixed either 30 minutes (pre-migratory) or 7 hours (migratory) after culture and assessed for β-catenin localization by immunostaining (Fig. 7 A, B). In both control and hic1 morphant explants, β-catenin is localized in the nucleus and at the cell membranes at the pre-migratory stage (Fig. 7 A, arrow/arrowhead).

Figure 7. hic1 knockdown disrupts Wnt signaling dynamics within migrating cranial neural crest cells.

Embryos were injected unilaterally with hic1 MO (25 ng) or control MO (25 ng) plus GFP-CAAX mRNA (200pg) in one DA cell. At NF stage 16, GFP+ CNC were explanted onto fibronectin-coated glass coverslips and allowed to develop for 30 minutes or 7 hours. Explants were fixed and immunostaining was performed for β-catenin and filamentous actin (phalloidin) (A). β-catenin was quantified as fluorescence pixel intensity from immunostained explants and plotted against position along cell diameter. Regression analysis was performed to identify the overall pattern of β-catenin localization (B). Total RNA was collected from explant cells and qRT-PCR was performed to assess level of axin2 gene expression. Fold change in gene expression was calculated relative to control explants at the pre-migration stage. * p < 0.05, unpaired t-test.

Quantification of β-catenin levels, as measured by pixel intensity across the cell diameter, shows that distribution is similar between the two groups (Fig. 7 B). Upon explant spreading and migration, nuclear-localized β-catenin is reduced in the control explants, a result consistent with a previous report (Maj et al., 2016). In hic1 morphant explants, however, nuclear localized β-catenin persists at the migratory stage (arrow), and its distribution is similar to what is seen at the pre-migratory stage (Fig. 7 A, B). Additionally, while migratory stage CNC from control embryos show an increase in F-actin (phalloidin staining), explants from hic1 morphant exhibit only low levels of F-actin at the migratory stage (Fig. 7 A). To determine whether the persistent β-catenin nuclear localization exhibited by hic1 morphant explants is functional, we next performed quantitative reverse transcription PCR (qRT-PCR) to assess the expression level of the known Wnt target gene axin2. In control explants, axin2 expression is significantly downregulated in migratory compared to pre-migratory cells (Fig. 7 C, 0.216 fold, p < 0.05) correlating with the loss of nuclear-localized β-catenin (Fig. 7 A, B). In hic1 morphant explants, however, axin2 expression in migratory cells remains at the level seen in pre-migratory cells (0.626 vs. 0.673 fold compared to control pre-migratory cells) indicating that the nuclear-localized β-catenin observed in hic1 morphant migratory explants results in sustained Wnt signaling. These results suggest that in the absence of Hic1, Wnt signaling remains active within CNC and this may contribute to the altered migration of these cells.

Wnt signaling inhibition rescues cranial neural crest spreading in hic1 knockdown explants

To examine whether dysregulated Wnt signaling may at least partially underlie the CNC migration defects in hic1 morphant, we sought to test whether inhibition of the Wnt pathway would rescue the CNC migration phenotype. In these experiments, we applied the Tankyrase inhibitor XAV939 to block Wnt signaling. The Tankyrase enzymes normally function to promote ubiquitination and proteosomal degradation of Axin1 and Axin2, key components of the β-catenin destruction complex. It has been shown that blocking Tankyrase function leads to stabilization of the β-catenin destruction complex, effectively reducing canonical Wnt signal transduction (Huang et al., 2009). Embryos were injected unilaterally as before with control or hic1 MO plus GFP-CAAX and allowed to develop until stage 16. GFP positive CNC tissues were dissected and plated on fibronectin coated dishes. Immediately after the explants adhered to the fibronectin dishes, they were imaged to determine the starting size of each explant tissue. The explants were then incubated at 15°C in the presence of Tankyrase inhibitor (XAV939, 5μm concentration) or DMSO as a control. CNC explants were again imaged after 15 hours of incubation as done in the previous experiments. The area occupied by the CNC explants at both pre- and post-migration stages was measured using ImageJ software, and the ratio of post- to pre-migration areas was calculated as fold change in area. (Fig. 8 A, B). Consistent with our above results (Fig. 6), the spread of the CNC explants from hic1 morphant was decreased significantly relative to that from control morphant in the DMSO treatment conditions. In the presence of the Tankyrase inhibitor, migration of control explants was greatly inhibited. This would be expected given that Wnt signaling needs to be reactivated after a transient down-regulation during CNC delamination in order for efficient migration (Rabadán et al., 2016). In contrast, migration of the explants from hic1 morphant was restored to that of normal explants upon Tankyrase inhibition so that the extent of explant spreading on fibronectin was indistinguishable from that of DMSO-treated control group (Fig. 8 B). This result indicates that the elevated canonical Wnt signaling that we observed in CNC of hic1 morphant is at least partially responsible for the impairment in CNC migration, and migration can be restored through dampening the Wnt signal.

Figure 8. Wnt inhibition in hic1 morphant explants rescues neural crest spreading and cadherin expression.

Embryos were injected unilaterally with hic1 MO (25 ng) or control MO (25 ng) plus GFP-CAAX mRNA (200pg) in one DA cell. At NF stage 16, GFP+ CNC were explanted onto fibronectin-coated dishes and imaged after 30 minutes. Explants were treated with 5μM concentration of Tankyrase inhibitor XAV939, or DMSO control, and imaged after 15 hours (A). Area of neural crest spreading was determined and presented as fold change in area as compared to pre-migratory explant (B). Experiment was performed in biological and technical triplicate. Statistical analysis was performed using the unpaired t test. Total RNA was collected from explant cells and qRT-PCR was performed to assess cadherin gene expression levels (C). Experiment was performed in biological and technical quintuplicate. Gene expression was compared within each individual experiment relative to control pre-migratory explant cells and statistical analysis was performed using the paired t test. * p < 0.05, ** p < 0.01, *** p < 0.001

Altered Wnt signaling dynamics in hic1 morphants correlates with changes in cadherin expression profile

Previous work has shown that sustained Wnt signaling inhibits CNC delamination (Rabadán et al., 2016). One of the hallmarks of delamination is a change in adhesive properties of neural crest cells and indeed the failure of neural crest migration in hic1 morphant transplants and explants implies that cell adhesion may be altered by the loss of hic1 gene product. Several cadherin family of adhesion proteins have been shown to regulate neural crest migration, and changes in cadherin protein repertoire and physical (e.g. proteolytic cleavage) or functional (e.g. adhesive properties) characteristics are reported to associate with neural crest EMT and/or delamination (McKeown et al., 2012). Pre-migratory CNC expresses high levels of the tightly adhesive E-cadherin (cdh1), whereas as CNC delaminates and initiates migration, a different cadherin expression profile emerges, disruption of which can lead to failure in CNC migration (Alfandari et al., 2010; Shellard and Mayor, 2019). In Xenopus, moderate increase in N-cadherin (cdh2) and strong increase in cadherin 11 (cdh11) accompany neural crest migration (Borchers et al., 2001a; Theveneau et al., 2010) To address whether hic1 regulates expression of adhesion molecules, and how adhesion might be affected by Wnt signal modulation, we performed qRT-PCR analysis of cadherin levels in control and hic1 morphant CNC explants. A total of 25 ng of hic1 or control MO plus GFP-CAAX was injected unilaterally as before and pre-migratory CNC were dissected and plated on fibronectin coated dishes. Explants were then treated with Tankyrase inhibitor or DMSO and incubated at 15°C for 15 hours to allow for CNC migration. Total explant RNA was collected at pre- and post-migratory stages for gene expression analysis (Fig. 8 C). In control explants, post-migratory cells exhibit expression levels of E-cadherin similar to pre-migratory stage (0.77 fold, no significance.) while expressing significantly lower levels of N-cadherin (0.42 fold, p < 0.01) and increased levels of Cadherin-11 (9.38 fold, p < 0.05). The low level of N-cadherin expression in these cells may reflect the switch from collective cell migration to individual cell migration as it has been shown previously in chick that N-cadherin is down-regulated in later stages of neural crest migration (Cheung et al., 2005). In contrast, hic1 morphant explants show a significant increase in E-cadherin expression upon migration (3.17 fold, p < 0.05). In addition, while hic1 morphant cells express significantly less N-cadherin at the pre-migratory stage compared to controls (0.57 fold, p <0.001), there is no difference in expression levels between pre- and post-migratory hic1 morphant cells. Similar to control explants, hic1 morphants show great increase of cadherin11 (23.53 fold, p < 0.05). Upon Wnt inhibition with the Tankyrase inhibitor XAV939, hic1 morphant explants show a specific decrease in E-cadherin expression (0.92 fold of control pre-migratory cells, p < 0.05) that may contribute to the observed restoration of migration (Fig. 8 B). Together, these results suggest that in the absence of Hic1, CNC cells have altered expression of cadherin and impaired organization of F-actin, and these defects are coupled with persistent Wnt signaling that may synergize with and enhance adhesive defects to prevent efficient CNC migration.

Discussion

HIC1 was first identified due to its frequent silencing in epithelial cancers (Wales et al., 1995). As such, a multitude of studies have been performed in the context of cancers of various tissue origins to identify how HIC1 loss contributes to cancer progression. These studies have identified a number of direct and indirect targets of HIC1 and revealed p53-dependent and -independent regulation of cell proliferation and tumor invasion (Reviewed in (Rood and Leprince, 2013)). In addition to its role as a tumor suppressor gene, a function of HIC1 in developmental processes has begun to be uncovered. Human HIC1 gene locus resides in a chromosomal region that is frequently deleted in the complex human developmental disorder Miller-Dieker Syndrome (MDS), and a loss-of-function mouse model of Hic1 inactivation displays embryonic phenotypes reminiscent of those in human MDS patients (Carter et al., 2000). Spatial expression of mouse hic1 is shown only from mid-gestation onward, which reveals specific patterns including in craniofacial mesenchymes (Grimm et al., 1999). In zebrafish, the HIC1 ortholog, ZfHIC1, is observed to have broad expression during gastrulation and early somitogenesis, with subsequent expressions in the head mesenchyme, branchial arches and pronephric duct (Bertrand et al., 2004). However, no functional assays have been reported using zebrafish. Data mining of recent single cell RNA sequencing studies in zebrafish and Xenopus tropicalis reveals that hic1 is expressed in multiple cell lineages during early development in both species (Briggs et al., 2018; Wagner et al., 2018), implying that hic1 may potentially modulate developmental processes in these species. To explore the role of hic1 in embryogenesis in more detail, we set out to investigate the expression and function of hic1 in Xenopus laevis in the current study, taking advantage of this well-established model system.

Through both targeted overexpression and knockdown of hic1 in the tissues that give rise to neural and neural crest cells, we find that proper levels of hic1 are required for development of the craniofacial cartilages. The initial specification of the neural crest population is not affected by altered expression of hic1, but migration of the neural crest out of the dorsal neural tube is impaired. This suggests that hic1 acts at a step subsequent to neural crest fate determination. This activity is distinct from several other TFs that regulate both specification and migration of the neural crest, such as snail, slug/snail2, foxd3, and sox10 (Aybar et al., 2002; Honoré et al., 2003; LaBonne and Bronner-Fraser, 2000; Sasai et al., 2001). Although like the other TFs, hic1 is expressed during early neural plate stages, the level of hic1 expression is much lower than that of the other TFs. The lack of an early function of hic1 in neural crest specification may thus reflect this fact and indicate a requirement for a threshold level of hic1 to function in regulation of downstream targets. The other interesting feature of hic1 is that both loss- and gain-of-function of hic1 affect localization of markers of neural crest migration and the ultimate craniofacial cartilage phenotypes in similar ways. This is expected for many regulators of cell movements. Genes controlling cell motility often do so by regulating cell adhesion and/or cytoskeleton organization and contraction. Both strengthening and weakening such adhesive and cytoskeletal properties would result in defects in cell migration, leading to analogous phenotypic defects. One most prominent example is the control of tissue morphogenesis by planar cell polarity pathway components, where gain- and loss-of-function mutants often generate comparable defects. It will be important in the future to determine whether hic1 target genes are regulated in opposite fashions by overexpression and knockdown of hic1 to give similar neural crest phenotypes.

Previously identified HIC1 targets include genes involved in cell cycle control and cell signaling, such as cyclin D1, ephrinA1 and EphA2 (Foveau et al., 2011; Van Rechem et al., 2010; W. Zhang et al., 2010). Although the ephrin/Eph pathway is shown to regulate neural crest migration, the signaling mainly participates in selection of migratory routes (Smith et al., 1997). During neural crest EMT and delamination, a major regulatory node is the expression of the cell adhesion molecules in the cadherin family. The profile of cadherin expression is often altered during neural crest EMT, so that N-cadherin (cdh2) and cadherin 11 (cdh11) expression is elevated relative to E-cadherin (cdh1) in migratory neural crest cells in Xenopus (McKeown et al., 2012). This enhanced expression of cdh2 and cdh11 seems to be crucial in promoting NC migration (Borchers et al., 2001a; Theveneau et al., 2010). In hic1 knockdown CNC explants, the cadherin profile is altered such that cdh1 expression is elevated relative to control cells. The impaired switch in cadherin expression profile may partially underly the failure in EMT in NC from the hic1 morphant embryos. Indeed, when hic1 morphant explant migration is rescued upon Wnt inhibition, cdh1 expression returns to that of control explants supporting a link between increased adhesion and sustained Wnt signaling in failure of CNC migration. It is currently unclear whether Hic1 can directly bind to cadherin promoters to regulate their expression, or if the effect on cadherin expression is indirect via other Hic1 targets. Future assays on association of Hic1 with regulatory elements of the cadherin genes will help resolve this issue.

One interesting discovery from our study is the functional link between Hic1 and canonical Wnt signaling. Previous work has shown that Wnt signaling needs to be temporally regulated in order for CNC to delaminate properly and initiate migration. High levels of Wnt signaling required for neural crest specification need to be dampened temporarily during NC EMT and delamination, but the signaling levels have to be restored for subsequent migration. Disruption of the dynamic Wnt signaling patterns through either forced overexpression or downregulation results in a similar phenotype of failed CNC migration (Maj et al., 2016; Rabadán et al., 2016). In hic1 morphant embryos, Wnt signaling, as indicated by nuclear β-catenin levels and axin2 gene expression, remains high during NC EMT. This persistent Wnt signaling likely interferes with NC EMT and spreading in the explants, as inhibition of Wnt signaling rescues NC migration defects. The mechanism underlying hic1-dependent temporal control of Wnt signaling is currently unknown. One possibility is that Hic1 directly and competitively regulates Wnt target genes, such as cyclin D1, a gene shown to be a direct target of both Wnt and HIC1 (Van Rechem et al., 2010; Ziegler et al., 2005). A second possibility is that Hic1 interacts with specific components of the Wnt signaling pathway to dampen the response specifically during this time window. Indeed, a study in mammalian cells demonstrates that HIC1 could bind and sequester TCF4 protein into nuclear bodies termed “Hic1 bodies” to dampen Wnt signaling responses (Valenta et al., 2006). In Xenopus, ectopic expression of a Hic1-GFP fusion protein also shows that Hic1 is present in a punctate pattern both in the nucleus and the cytoplasm, consistent with the formation of Hic1 bodies in Xenopus (data not shown). Thirdly, Hic1 may regulate Wnt signaling indirectly through modulation of genes that themselves influence the Wnt pathway, such as the Yes-associated protein (YAP). YAP has been shown to regulate neural crest development in mouse, zebrafish, Xenopus and human in vitro-derived NCC (Gee et al., 2011; Hindley et al., 2016; Jiang et al., 2009; J. Wang et al., 2015). While the mechanisms of this regulation have yet to be fully elucidated, YAP is shown to positively interact with both the BMP and Wnt pathways in avian NCC development (Kumar et al., 2019). Hence, Hic1 may use different strategies to modulate Wnt signaling during NC development.

Evidence for a possible intersection between Hic1, YAP, and Wnt signaling come from cancer studies. In bladder cancer cell lines, HIC1 expression levels negatively correlate with proliferation, invasion and migration, and loss of HIC1 leads to increased levels of Yes-associated protein (YAP) and increased YAP-activated transcriptional responses (Zhou et al., 2018). YAP is a binding partner of β-catenin and TBX5 in a protein complex that activates transcription in a TCF-independent manner in cancer cell lines of the GI tract (Rosenbluh et al., 2012). Taken together, the data suggest a possible mechanism whereby Hic1 regulates the levels of Wnt signaling through a negative regulation of YAP. Additional evidence for Hic1-regulated Wnt signaling comes from studies of colon cancer. In this best-studied Wnt-driven cancer, the loss of normal function of APC, a negative regulator of Wnt, through genetic mutations is often associated with familial inheritance of colorectal cancer. In mice, a single allele loss-of-function mutant of APC results in spontaneous tumors throughout the GI tract. Interestingly, one study shows that additional loss of a single allele of Hic1 in an APC deficient mouse model (Hic1^+/−, Apc^+/Δ716) leads to increased polyp number and increased markers of active Wnt signaling, including nuclear-localized β-catenin and elevated Sox9 levels (Mohammad et al., 2011). These results corroborate with those of our own to indicate that Hic1 negatively regulates Wnt signaling in cancer formation, and highlights the conservation of mechanisms that regulate both development and cancer. Hence, further studies of Hic1/Wnt interaction will have implications for both neural crest development and our understanding of Wnt-driven cancers.

In summary, we identify in this study an important function of the tumor suppressor gene hic1 in regulation of neural crest migration. We also show that hic1 does so by modulating cadherin expression and Wnt signaling levels. Many questions remain. For example, might Hic1 regulate Wnt signaling in vivo through both canonical transcriptional regulation as well as through protein-protein interactions? Does Hic1 act as a transcriptional activator or repressor in this context? Our ongoing and future studies are aimed at addressing these questions to identify a comprehensive list of genes that are differentially expressed in response to hic1 manipulation and to identify direct targets of Hic1 regulation during NCC development, including those genes that would contribute to Wnt regulation.

Materials and methods

Fertilization and embryo collection

Xenopus laevis and X. tropicalis frogs were maintained at the University of Alabama Birmingham animal housing facility in accordance with the UAB Institutional Animal Care and Use Committee. Female X. laevis frogs were injected with 800 units of human chorionic gonadotropin (HCG, Sigma) 16–20 hours prior to collection of mature eggs. In vitro fertilization was performed and embryos were maintained at 15°C until they reached the proper developmental stage for experiments.

Microinjection

Embryos were allowed to develop in 0.1X MMR media (Sive et al., 2000) until the 4-cell stage at which time they were transferred to 2% Ficoll in 0.5X MMR for injections. Embryos were micro-injected at the 8-cell stage with a 10 nl volume of RNA and/or morpholino and allowed to recover before transferring back into 0.1X MMR prior to the start of gastrulation.

Phenotype Characterization and Alcian Blue Staining

Embryos were injected at the 8-cell stage in both dorsal animal cells with either 50 pg or 500 pg total of hic1 mRNA (in vitro transcribed with mMessage mMachine SP6 enzyme kit, Invitrogen, cat # AM1340), 50 ng total of a morpholino directed against the translational start site of hic1 (Hic1 ATG MO sequence 5’-CCATCCCCAGCTGTGCGAGCGCGTC-3’, GeneTools), or co-injected with 50 pg hic1 mRNA and 50 ng hic1 MO. Embryos were imaged between stages 37–40 and phenotypes were compared to un-injected clutch-mate controls. Some embryos were injected unilaterally with either 250 pg hic1 mRNA or 25 ng hic1 MO and allowed to develop until stage 45 to assess craniofacial cartilage development. These embryos were fixed in 2% paraformaldehyde (PFA) in phosphate buffered saline (PBS) for one hour at room temperature (RT), rinsed in 50% ethanol in water for 10 minutes, and stained with Alcian blue (0.04% Alcian blue, 10 mM MgCl₂, 80% EtOH) for 5 days. Embryos were washed in 80% EtOH/10 mM MgCl₂ overnight, rehydrated in graded EtOH/water, bleached in 3% H2O2/0.5% KOH for 10 minutes at RT, and brought into 50% glycerol/0.1% KOH for clearing. Heads were isolated and brain tissue removed by manual dissection for better visualization of the cartilages. Images were obtained using a Nikon AZ100 stereo light microscope with either a 0.5X or 2X objective.

In situ hybridization

Embryos were injected at the 8-cell stage unilaterally in the left dorsal animal cell with 250 pg of hic1 mRNA, 25 ng hic1 MO, or 25 ng hic1 MO in combination with 50 pg hic1 mRNA lacking the MO recognition sequence (hic1*) along with 200 pg β-galactosidase mRNA as a tracer and allowed to develop until the indicated timepoint. Embryos were fixed in MEMFA (1X MEM salts, 3.7% formaldehyde) for 30 minutes at RT, incubated in Proteinase K solution (10 μg/mL in PBS) for 20 minutes if still contained within the vitelline membrane, followed by incubation in Red-gal staining solution until color was sufficiently developed. Following additional MEMFA fixation (1–2 hours at RT), embryos were dehydrated in two washes each of EtOH and MeOH (20 minutes at RT) and stored in MeOH at −20°C. In situ hybridization (ISH) was carried out according to previously published methods (Harland, 1991; Sinigaglia et al., 2017). ISH probes were generated by amplification of gene sequences out of embryo cDNA and subsequent cloning into the pBSKS vector.

Immunfluorescence

Embryos were injected in both dorsal animal cells at the 8-cell stage and allowed to develop until late neurula stages. In one cell was injected either 250 pg of hic1 mRNA fused to an HA epitope tag at the C-terminal end (Hic1-HA), or 250 pg of hic1-HA mRNA that additionally contained the 5’ UTR sequence recognized by the hic1 MO (Hic1–5’UTR-HA). In the second cell was injected the same dose of appropriate mRNA along with 25 ng of hic1 MO and fluorescein dextran as a tracer. At mid-neurula stages, the vitelline membrane was removed, embryos were fixed in 1X MEMFA for one hour at room temperature and subsequently dehydrated in EtOH and MeOH washes as above. Following overnight incubation in MeOH at −20°C, embryos were rehydrated in graded MeOH:PBS washes, washed three times in PBS and incubated for 48 hours in sucrose in PBS (20% w/v) at 4°C. Embryos were then equilibrated into OCT mounting medium, frozen in OCT blocks, and 12 μm sections were cut using a cryostat and affixed to glass slides. After drying, slides were washed with PBS (3 X 5 min.), permeabilized with 0.1% Triton-X in PBS for 10 minutes, blocked with 10% normal goat serum in 0.1% Triton-X in PBS for one hour at RT, and incubated in primary antibody (mouse anti-HA, 1:200 dilution, Covance cat #MMS-101P) in blocking buffer overnight at 4°C. Slides were then washed in PBS (3 X 5 min.) and incubated in secondary antibody (goat anti-mouse IgG AF 555, 1:1000, Life technologies cat #A21425) in 0.1% Triton-X in PBS for 2 hours at RT. Following final PBS washes (3 X 15 min.), slides were mounted with Fluormount G with DAPI (eBioscience, cat # 00–4959-52). Slides were imaged using an Olympus FV1000 laser scanning confocal microscope with a 20X objective, and images were processed using ImageJ to present maximum intensity projections of multiple z-planes.

Cranial neural crest transplant and explant culture

Embryos were injected at the 8-cell stage in the left dorsal animal cell with 25 ng of either hic1 MO or control MO plus 200 pg of EGFP-CAAX mRNA and allowed to develop until stage 16. Embryos were screened for proper GFP expression in the targeted CNC region and prepared for further experiments via removal of the vitelline membrane. For transplant experiments, injected and un-injected clutch-mate embryos were placed in 2/3 MMR media in a clay-bottomed dish. The GFP+ CNC population was removed from injected embryos and transplanted into un-injected embryos per previously published methods (Borchers et al., 2001b; Milet and Monsoro-Burq, 2014). Transplanted embryos were transferred to 0.1 X MMR in a standard petri dish and incubated at 15°C. Embryos were collected at stage 27–28 and imaged for GFP expression using a Nikon AZ100 stereo light microscope with 2X objective. For explant experiments, the GFP+ CNC population was removed and placed in 1X Danilchik’s for Amy (DFA) medium (Sater et al., 1993) on a fibronectin coated dish (10 μg/1 mL in PBS) to assess cell migration, or fibronectin coated cover glass (100 μg/1 mL in PBS) for immunostaining. Explants were cultured at 15°C until the indicated times for imaging. Some explants were imaged 30 minutes after explant procedure (pre-migration) followed by addition of Tankyrase inhibitor (XAV939, 5um concentration, Selleckchem) to decrease Wnt signaling. Explants were reimaged 15 hours later (post migration) to assess effect of Wnt inhibition on CNC migration. For each explant, the area of migration was determined using ImageJ software and compared to the area of the initial explant to calculate fold change in migration. Statistical analysis was performed using the unpaired t-test. To assess the motility of individual CNC cells, GFP+ CNC were removed and placed in calcium- and magnesium-free media for several minutes to dissociate cells prior to plating on a fibronectin-coated dish as previously described (Theveneau et al., 2010). Cells were allowed to adhere for one hour at room temperature prior to imaging. Areas of GFP+ single cells were imaged every three minutes for 150 minutes and individual cells were followed using the Manual Tracking plugin for ImageJ. Individual cell tracks were combined and plotted as rosette plots using the Chemotaxis and Migration Tool (Ibidi). For immunostaining, explants were fixed in MEMFA for 20 min at RT, washed 3 times with 1X PBS and permeabilized with 0.5% Triton-X in PBS 10 min each, blocked with Image-iT FX signal enhancer (life technologies, cat # I36933) 30 minutes in the dark, and washed 2 times with PBS. Primary antibody was added (rabbit anti-human β-catenin, Sigma cat # C2206) at 1:150 dilution in 0.5% Triton-X, 1% normal goat serum in PBS and incubated overnight at 4°C. Explants were then washed 3 times in PBS and secondary antibody was added (donkey anti-rabbit AF 647, Invitrogen, cat # A-31573, 500 μg/1 mL) along with Rhodamine-conjugated phalloidin (Invitrogen, cat # R415, 200U/1 mL) at 1:125 dilutions each in 0.5% Triton-X in PBS for two hours in the dark. Following final washes in PBS (3 X 10 min.), explants were mounted on a glass slide with Fluoromount-G containing DAPI and allowed to dry. Imaging was performed on an Olympus FV1000 laser scanning confocal microscope with a 60X objective, and images were processed using ImageJ to present maximum projections of multiple z-planes. To quantify β-catenin distribution from immunofluorescence images, ImageJ was used to draw a line through the center of individual cells and collect pixel intensity across the cell diameter. Measurements were taken from 27 total cells across three biological and technical repeats. Measurements of cell diameter were normalized to percent length across the cell and were plotted against fluorescent intensity. A regression curve was generated for each group using a fifth order polynomial equation.

RT-PCR

To assess presence of hic1 mRNA expression throughout early developmental time, whole embryos were collected, total RNA was isolated and cDNA was prepared. RT-PCR was performed using the following primer pairs: Hic1_F 5’-TACCCGGGAGACAAGACAGT-3’, Hic1_R 5’-CTCAATGCTGGGGTACCTGG-3’, H4_F 5’-ATAACATCCAGGGCATCACC-3’, H4_R 5’-ACATCCATAGCGGTGACGGT-3’. To assess changes in gene expression in CNC explants, total RNA was isolated using Trizol at the indicated timepoints. cDNA was prepared and qRT-PCR was performed using standard methods for the SYBR-green platform using the Step One Plus Real-time PCR System (Applied Biosystems) with the following primer pairs: Axin2_F 5’- CCCAAAGTTCCTCAGGGTATG-3’, Axin2_R 5’- CTGGCTTTCTGGAGCTTCTT-3’, Cdh1_F 5’- GACAAGGAACAGAACCAGAAGA -3’, Cdh1_R 5’- TTTCCCGTGACAATCCCATTA -3’, Cdh2_F 5’- TCCTGAGTTCACAGCAATGAC -3’, Cdh2_R 5’- GTGTGTGAGGTTGGTCCTTATC -3’, Cdh11_F 5’- GGACTCTCAGGGACAACTAAAG-3’, Cdh11_R 5’- AGCTTCTGACACAGACATGG-3’, EF1a_F 5’- TTGGAGCCCTCTCCCAATA-3’, EF1a_R 5’- GATGGTGTCCAAAGCCTCAA-3’. Data were collected from five biological and technical replicates and gene expression levels were determined using the ΔΔCT method with EF1a as the endogenous control and calculating each experiment independently relative to the control pre-migration explant. Statistics were performed between experimental groups using a paired t test.

• The transcription factor Hic1 is expressed during early Xenopus laevis development

• Hic1 disruption leads to craniofacial cartilage malformations

• Hic1 is necessary for neural crest migration in vivo and in vitro

• Hic1 loss results in sustained Wnt signaling and altered cadherin expression