Ketamine Modulates Zic5 Expression via the Notch Signaling Pathway in Neural Crest Induction

Ketamine is a potent dissociative anesthetic and the most commonly used illicit drug. Many addicts are women at childbearing age. Although ketamine has been extensively studied as a clinical anesthetic, its effects on embryonic development are poorly understood. Here, we applied the Xenopus model to study the effects of ketamine on development. We found that exposure to ketamine from pre-gastrulation (stage 7) to early neural plate (stage 13.5) resulted in disruption of neural crest (NC) derivatives. Ketamine exposure did not affect mesoderm development as indicated by the normal expression of Chordin, Xbra, Wnt8, and Fgf8. However, ketamine treatment significantly inhibited Zic5 and Slug expression at early neural plate stage. Overexpression of Zic5 rescued ketamine-induced Slug inhibition, suggesting the blockage of NC induction was mediated by Zic5. Furthermore, we found Notch signaling was altered by ketamine. Ketamine inhibited the expression of Notch targeted genes including Hes5.2a, Hes5.2b, and ESR1 and ketamine-treated embryos exhibited Notch-deficient somite phenotypes. A 15 bp core binding element upstream of Zic5 was induced by Notch signaling and caused transcriptional activation. These results demonstrated that Zic5 works as a downstream target gene of Notch signaling in Xenopus NC induction. Our study provides a novel teratogenic mechanism whereby ketamine disrupts NC induction via targeting a Notch-Zic5 signaling pathway.


INTRODUCTION
Neural crest (NC), the vertebrate-specific pluripotent cell population, derives from the border region between the epidermis and the neural plate. NC gives rise to numerous derivatives including cranial cartilages, neurons and glia of peripheral nervous system, melanocytes, and heart outflow tract . Abnormal development of NC derivatives results in neurocristopathies, leading to many human congenital disorders such as Hirschsprung disease, Treacher Collins syndrome, Waardenburg-Shah syndrome, DiGeorge syndrome, CHARGE syndrome, neuroblastoma, and melanoma (Bolande, 1997;Takahashi et al., 2013;Monsoro-Burq, 2015).
The development of NC undergoes induction, specification, migration, and differentiation, which is tightly regulated by a well-orchestrated gene regulatory network (GRN). After germ layer separation, a set of transcription factors including Pax3, Zic1, Zic5, Msx1, Hairy2, Ap2α, and c-Myc are expressed at the border region between the neural plate and the epidermis. Such expression is induced by morphogens like Wnt, BMP, FGF, Notch and retinoic acid in a gradient sensitive manner. These border determination transcription factors further trigger the expression of a group of genes including Slug, FoxD3, Snail, Sox9, and Twist in the emerging NC cells, thereby accomplishing NC induction (Steventon et al., 2005;Milet and Monsoro-Burq, 2012;Shyamala et al., 2015). After induction, the NC cells undergo epithelial to mesenchymal transition (EMT) and migrate to their destination in a contact inhibition of locomotion manner (Theveneau and Mayor, 2012). Terminal differentiation of NC cells is extensively cross-regulated by different networks.
The Notch signaling pathway plays vital roles in early embryonic development (Artavanis-Tsakonas et al., 1995. Following ligand binding, the Notch receptor is cleaved continuously by an ADAM (Brou et al., 2000;Mumm et al., 2000;Dyczynska et al., 2007;Bozkulak and Weinmaster, 2009) and γ-secretase complex (De Strooper et al., 1999;Song et al., 1999;Zhang et al., 2000) and releases the active N-terminal intracellular domain (NICD). The NICD then translocates to nucleus and interacts with CSL family DNA binding proteins to activate the transcription of downstream target genes (Schroeter et al., 1998;Struhl and Adachi, 1998). Activation of the Notch signaling pathway by binding to adjacent cells facilitates boundary determination. During NC induction, Notch and its ligands are expressed in the prospective anterior NC territory, where Notch coordinates with the BMP signal to determine the NC border region. In somitogenesis, the expressed protein amount of the NICD in nucleus forms a cyclical clock manner through degradation and thereby establish somite boundary formation (Dale et al., 2003).
Notch signaling pathway has been shown to play a role in NC development in different animal models. Disturbance of Notch signaling in mice leads to craniofacial structure abnormity, cardiac outflow deficiency, and decreased enteric neurons (Okamura and Saga, 2008;Mead and Yutzey, 2012). Notch signals refine the neural plate border region through negative regulation of PRDM1α in zebrafish (Hernandez-Lagunas et al., 2011). Studies using chick and Xenopus models demonstrate a similar BMP-regulating mechanism of Notch pathways in NC development (Endo et al., 2002). Hairy family genes required for NC induction are down-regulated by BMP-Smad signaling and positively regulated by the Notch/Delta-Su(h) pathway (Vega-Lopez et al., 2015). Another report suggests a role for Tsk in balancing Notch and BMP pathways via directly binding to BMP and the extracellular domain of Delta1 during neural plate border determination in Xenopus (Kuriyama et al., 2006).
Ketamine has been commonly used as anesthetic, analgesic, or sedative, and has recently been shown to be a promising acute anti-depressant (Ionescu et al., 2015;Kirby, 2015;Reardon, 2015). As a dissociative anesthetic, ketamine users can develop cravings for the drug and become addicted. Ketamine is now one of the most popular and most abused recreational illicit drugs in the world. A large number of the drug users are women at childbearing ages. The impact of illicit ketamine use on embryonic development during pregnancy could be detrimental. Previous studies using zebrafish, Xenopus and rat models showed its disruptive effects on motor-neuron development , primary germ cell layers specification (Akeju et al., 2014), cardiac morphogenesis (Guo et al., 2016), neural genesis and survival (Dong and Anand, 2013;Dong et al., 2014). More recently, a report suggested craniofacial and trunk phenotypes upon embryonic ketamine exposure in Zebrafish (Felix et al., 2017). Furthermore, ketamine has been shown to act as teratogen in a NMDA-independent ways (Felix et al., 2014). Although ketamine's roles as NMDA antagonist have been extensively studied, the effect on early embryonic development and its underlying mechanisms remain elusive.
In this study, we report that ketamine exposure affected NC development at a very early stage, equivalent to the first month of human pregnancy. Ketamine down-regulates Zic5 expression in the presumptive neural plate border region and thereby disrupting NC induction. Our data indicate that ketamine disturbs Notch signaling, and that Zic5 works downstream of Notch during Xenopus NC developmental stages. Our data suggest a completely new teratogenic mechanism whereby ketamine disrupts NC induction via targeting a NOTCH-Zic5 signaling pathway.

Microinjection and in Situ Hybridization
This study was carried out in accordance with the recommendations of the animal study guideline of the Animal Experimentation Ethical Committee of Chongqing Medical University. The protocol was approved by the Animal Experimentation Ethical Committee of Chongqing Medical University in Chongqing, China. Embryo in vitro fertilization, culture, whole mount in situ hybridization, mRNA preparation and microinjection were carried out as previously described (Shi et al., 2009. The probes for whole-mount in situ hybridization including Chordin, Xbra, Wnt8, Fgf8, Msx1,Pax3, Zic1, Zic5, Slug, Six1, MyoD were applied as described (Smith et al., 1991;Sasai et al., 1994;Nakata et al., 2000;Li et al., 2006;Guemar et al., 2007;Zhang et al., 2014). For enteric neuron labeling, intestines were first manually dissected out from stage 40 embryos which were freshly fixed with 4% paraformaldehyde. Then the enteric neurons were stained through in situ hybridization with N-tubulin probe (Oschwald et al., 1991).
To measure the reporter activity in embryos, 12.5 pg constructed firefly plasmid, 2.5 pg renilla construct and 500 pg NICD mRNA or 500 pg LacZ mRNA (for control) were co-injected into both side of the dorsal blastomeres at the 4-cell stage. The embryos were then harvested at the neural plate stage (stage 15), divided into three groups (>10 embryos each group), lysed and analyzed using Dual-Luciferase Reporter Assay System (Promega).

Gene
Primers (forward and reverse) Cycles Annealing temperatures ( • C) Extension time (s) Reference Frontiers in Molecular Neuroscience | www.frontiersin.org using an Odyssey system (LI-COR Biosciences). Unlabeled wildtype and mutant oligonucleotides at X100 molar excess were applied for the competition assay.

Embryonic Ketamine Exposure Results in Neurocristopathies
Previous data indicates that 2 mM ketamine applied in culture yields approximately 0.4% accumulation, i.e., 8 µM inside embryo , which is comparable to the anesthetic concentration in human plasma (Dong et al., 2014). 2 mM or similar culture concentration was commonly chosen to evaluate the embryonic ketamine exposure in most studies Felix et al., 2014;Lantz-McPeak et al., 2015;Robinson et al., 2015). To examine ketamine's effects on early embryonic development, we exposed Xenopus embryos to 2 mM ketamine-MBS culture media from pre-gastrulation (stage 7) to early neural plate (stage 13.5), and then cultured in MBS media until harvest time. Upon ketamine incubation, pigment cells in trunk diminished and eyes exhibited developmental abnormities ( Figure 1A). Ocular pericytes located in retina ganglionic cell layer (GCL) and inner nuclear layer (INL) are derived from the NC (Trost et al., 2013). Further analysis showed disorganization of GCL, INL, and outer nuclear layer (ONL) (Figure 1B DAPI stain, white arrow) and loss of β-tubulin III expression in cell layer boundary ( Figure 1B, white arrowhead) and lens (Figure 1B, asterisk) expression, indicating abnormal differentiation of lens and retina. At the tailbud stage, ketamine caused enlargement of intestinal system combining loss of enteric nervous cells (Figure 1C, black arrowhead), a mega-colon like (Hirschsrpung disease) phenotype ( Figure 1C) and marked reduction or even complete loss of all the cranial cartilage elements including Meckel's (Mk), ceratohyal (ch), basihyal (bh), and branchial (ba) cartilages as demonstrated by alcian blue staining ( Figure 1D). These data clearly indicate that ketamine impairs NC development, causing neurocristopathies.

Ketamine Blocks NC Induction Independent of Mesoderm Genesis
Induction is the first event in NC genesis. To investigate the effect of ketamine on NC induction, we examined the expression pattern of the NC specification gene Slug. The criteria for judging the expression pattern are based on two aspects, expression Frontiers in Molecular Neuroscience | www.frontiersin.org area and intensity. Compared to control, decreased expression area or signal intensity was determined as 'reduced expression level.' At the early neural plate stage (stage 14), when the NC has been induced, control embryos express a normal Slug expression pattern (in 32/32 embryos) (Figure 2A). However, ketamine at a 0.36 mM culture concentration showed reduced Slug expression (36/37). The inhibition of Slug expression increases in a dose dependent manner. At a 2 mM ketamine concentration, Slug expression was severely reduced (41/41) (Figure 2A). Morphogens from the mesoderm are critical for inducing the NC (Klymkowsky et al., 2010). Thus, we examined mesoderm development by in situ hybridization assay with mesoderm markers on the ketamine-incubated and control embryos prior to NC induction (stage 11.5) ( Figure 2B). There was no significant difference between ketamine and control groups in the expression pattern of four mesoderm maker genes including Chordin (100% normal expression for ketamine, n = 56; and 98.41% normal expression for control, n = 63), Xbra (98.48% normal expression for ketamine, n = 66; and 100% normal expression for control, n = 46), Wnt8 (95.92% normal expression for ketamine n = 49; and 98.08% normal expression for control, n = 52), and Fgf8 (100% normal expression for ketamine, n = 77; and 100% normal expression for control, n = 55) ( Figure 2C). These results suggest that ketamine does not affect mesoderm at the beginning of NC induction. These data indicate that ketamine blocks NC induction independent of mesoderm genesis.

Ketamine Disrupts NC Induction by Inhibiting the Expression of Zic5
Neural crest induction requires high Wnt and intermediate BMP signaling input (Steventon et al., 2009). At the onset of gastrulation, animal cap, a pluripotent cell group located in at the animal pole of Xenopus embryos, can develop into almost any cell type given the proper signaling stimulation in vitro. In Xenopus animal caps, co-injection of Wnt8 and BMP4 truncated receptor (tBR) mimics endogenous NC stimulation and results in the expression of NC genes (Zhang et al., 2014). To determine the targeting molecules mediating ketamine's effect on NC induction, we first examined gene expression in the wnt8-tBR co-injected animal cap system. The default of animal cap cells is an epidermis fate, characterized by a high BMP signaling niche Montagner et al., 2016). BMP target genes Msx1 (Tribulo et al., 2003), Dlx3, Dlx5 (Luo et al., 2001;McLarren et al., 2003) and AP2α (Luo et al., 2002(Luo et al., , 2003 are involved in neural plate border determination and are normally expressed in control animal caps. Co-injection of 500 pg Wnt8 mRNA and 500 pg tBR mRNA successfully triggers animal cap cells to express the secreted genes Fgf8, Wnt8, and neural border genes including Zic1, Pax3, Zic5 as well as NC specific genes Slug ( Figure 3A).
In the presence of ketamine, Wnt8-tBR induced animal caps markedly reduces the expression of Zic5 and Slug ( Figure 3A). Zic5 is expressed in the NC area and plays an indispensable role in the induction process (Nakata et al., 2000). Since embryos undergo NC induction in the late gastrula (approximate at stage 12.5), we examined the NC markers at this critical time with a 1-2 mM ketamine exposure or 0.8 nmol direct injection. At stage 12 the pre-induction state, neither control nor ketaminetreated embryos expressed Slug. At stage 13, all of the NC regulatory network genes were expressed in control embryos, whereas Zic5 and Slug transcripts were reduced in embryos cultured in the ketamine-containing media or directly injected with ketamine. However, the Zic5 and Slug expression levels from whole embryos returned to normal at the beginning of NC migration (stage 16). The recovery NC gene expression may be due to the feedback of complicated GRNs ( Figure 3B). Consistent with RT-PCR data from animal caps and whole embryos, in situ hybridization results showed that ketamine treatment did not affect most NC induction-related genes including Fgf8 (97.96% normal expression, n = 49), Wnt8 (100% normal expression, n = 77), Msx1 (100% normal expression, n = 66), Pax3 (97.56% normal expression, n = 41), and Zic1 (96.15% normal expression, n = 52) (Figures 3C,F). However, the expression of Zic5 and the induction marker gene Slug were markedly inhibited by ketamine, with 97% (n = 70,) and 100% (n = 43) of embryos having inhibited expression respectively (Figures 3D,F). Finally, exogenous injection of Zic5 mRNA at two-cell stage significantly rescued the ketamine-induced Slug inhibition (63.64% of normal expression, n = 44) (Figures 3E,F). Our data suggest that ketamine disrupts NC induction by inhibiting Zic5 expression.

Targeting Zic5 by Notch Signaling in NC Induction
To investigate the upstream signaling pathways mediating ketamine-induced Zic5 inhibition during NC induction, we screened the most common NC induction morphogens including Wnt, Fgf, and Notch. Previous work showed that prospective paraxial mesoderm or Fgf8 induces Zic5 expression (Monsoro-Burq et al., 2003). Our results show that Fgf8 and its downstream NC inducing gene Msx1 (Monsoro-Burq et al., 2005) were normally expressed and not affected by ketamine treatment (Figure 3). However, ketamine exposure markedly reduced the expression of Notch target genes Hes5.2a, Hes5.2b, and ESR1 ( Figure 4A). Somite labeling with MyoD and Six1 transcripts showed slightly reduced signal ( Figure 4B) and a shortened embryonic body axis ( Figure 1A) that mirrors phenotypes seen in Notch-deficient embryos (Jen et al., 1997). To analyze ketamine's effect on Notch signaling, the human T-cell leukemia line Jurkat was used. 2 mM Ketamine exposure for 10 h increased the amount of ubiquitinated Notch ( Figure 4C) and promoted Notch protein degradation (Figure 4D). To ask whether Zic5 functions as a downstream target gene of Notch signaling, we assayed Zic5's response to high and low Notch environment during NC induction in Xenopus embryos. Blocking Notch pathway with microinjection of Delta-stu mRNA (Revinski et al., 2010) at the 4-cell stage markedly inhibited Zic5 expression ( Figure 4E). Activation of Notch signaling by injection of NICD (Fryer et al., 2002) induced Zic5 ectopic expression in whole neural plate and NC area ( Figure 4E).

Identification of a 15 bp cis-Acting
Notch-Response Element in the Zic5
Electrophoretic mobility shift assay (EMSA) was performed to further examine whether Notch signaling activation promotes the binding of the transcriptional factors to the 15 bp NRE. Nuclear proteins with or without NICD-transfected were extracted from SHSY5Y cells. NICD-transfection significantly increased the amount of the protein-NRE DNA binding complex ( Figure 5E). The competitor probe containing the same non-fluorescently labeled 15 bp sequence reversed the fluorescent binding signaling to background level, and a mutant probe had little effect to compete with the binding (Figure 5E). These results unequivocally demonstrate that Notch activates Zic5 expression via the newly identified 15 bp NRE.

DISCUSSION
Ketamine is a dissociative anesthetic and the most popular club drug abused by youths for recreational purposes (Lee et al., 2015). Lately ketamine has attracted extensive attention for its quick and significant alleviation of depression (DiazGranados et al., 2010;Li et al., 2011). Ketamine usage grows steadily. Young people, including those at childbearing age, are a major addict-prone population. In human, NC induction happens within the first month of pregnancy. At this early point in the pregnancy, most women are not aware that they are pregnant until the next regular menstrual cycle. While ketamine exposure has been shown to affect cell survival (Brambrink et al., 2012;Bai et al., 2013), germ layer specification (Akeju et al., 2014), neurogenesis Kanungo et al., 2013;Dong et al., 2014), cardiac morphogenesis (Guo et al., 2016), and craniofacial and trunk genesis (Felix et al., 2017), the effect of ketamine on early embryonic development is poorly understood.
In this study, we discovered that early embryonic ketamine exposure specifically blocks NC induction via targeting Zic5. This effect on the NC is not due to the disruption of the mesoderm. Previous studies suggested the vital roles of Zic5 in NC genesis in mouse and Xenopus (Nakata et al., 2000;Inoue et al., 2004). Zic5 also works downstream of Wnt pathways and regulates tectum cell proliferation in zebrafish (Nyholm et al., 2007). However, the upstream signals regulating Zic5 expression in NC induction were not yet known. Our work revealed that the expression of Zic5 is controlled by Notch signaling during NC induction. Ketamine inhibited Notch targeted gene expression in Xenopus embryos. Xenopus embryos exposed to ketamine also exhibited shortened somite-segmentation, a typical Notch signaling deficiency phenotype. The Notch signaling pathway has long been known to participate in NC induction (Glavic et al., 2004;Kuriyama et al., 2006). In the neural-epidermal boundary, Notch balances BMP via activation of Hairy2 expression (Glavic et al., 2004). However other reports showed that Hairy2 is not regulated by Notch, but activates Notch via Stat3 (Nichane et al., 2008a,b). In this study, we demonstrated that Zic5 is a Notch targeted border gene involved in Xenopus NC induction. We also identified a 15 bp cis-acting element in the 5 regulatory region of Xenopus Zic5 gene that binds to NICD to mediate Notch's transcriptional activation of Zic5 gene expression. Our study further showed that ketamine enhanced ubiquitination of Notch proteins, and reduced Notch-1 protein level. These data suggest that enhanced degradation of Notch protein by ketamine may partly account for the phenotypes we observed in Xenopus, consistent with the similar phenotypes in Zebrafish (Felix et al., 2017). These results suggest a conserved mechanism of ketamine among different spices.
In summary, we discovered that Zic5 works as a downstream target gene of Notch signaling in Xenopus NC induction. Our study provides a novel teratogenic mechanism for ketamine whereby it disrupts NC induction by targeting a Notch-Zic5 signaling pathway.

AUTHOR CONTRIBUTIONS
YS, JL, and WS conceived and designed the experiments. YS, JL, CC, YXi, YL, and YXu performed the experiments. YS, JL, YXi, YXu, TL, and WS analyzed and contributed reagents/materials/analysis tools. YS, JL, and WS wrote the paper. All authors reviewed the manuscript.