Genetic regulation of enteric nervous system development in zebrafish

The peripheral nervous system (PNS) encompasses all neurons and glia that reside outside of the brain and spinal cord and largely regulates unconscious events — such as the fight or flight response and digestive functions [1]. The largest branch of the PNS is the enteric nervous system (ENS). Also known as the ‘gut brain,’ the ENS is an autonomous network of neurons and glia located within the walls of the entire gut. The ENS mediates peristalsis, local digestive functions, and water balance in the gut, among many other functions [2]. However, although the human ENS contains more than 600 million neurons — rivaling the numbers of those found within the spinal cord — far less is known about its formation [2].

The vertebrate ENS primarily originates from neural crest cells (NCCs) [3–6]. NCCs are multipotent embryonic stem cells that migrate throughout the embryo, eventually contributing to numerous cell types in the vertebrate body [7,8]. NCCs are born along the anterior–posterior extent of the neural tube, and are subdivided according to their spatial level of origin — namely the ‘cranial’ from anterior, followed by the ‘vagal’, ‘trunk’, and ‘sacral’ populations in the most posterior [9]. Following emigration from the neural tube, vagal NCCs that approach the foregut are hereafter referred to as enteric NCCs (ENCCs). ENCCs are competent to become either enteric neurons or enteric glia [3,4]. ENCCs then migrate into and along the outer walls of the incipient gut tube and progress caudally along its length until reaching their final destinations, where they spatially position along the gut circumference. During their journey, ENCCs mainly migrate together in a ‘follow the leader’ chain pattern, such that chains of cells navigate the landscape of the growing gut tube [10–12]. Incredibly, during their migration ENCCs robustly expand in numbers sufficient to keep up with the growing gut tissue [13] and give rise to various enteric neuronal subtypes, as well as glia, which can be classified by molecular means [2,14].

Defects in ENS development and function manifest as various congenital and adult-onset gastrointestinal diseases. Such diseases include; Hirschsprung disease (HSCR), Esophageal Achalasia, Chronic Constipation, and Gastroesophageal Reflux Disease [15], among many others. In particular, HSCR is a birth defect characterized by the absence of ENS along variable lengths of the infant's gut, with colonic aganglionosis being the most common form, occurring every 1 in 5000 births [16,17]. The principal treatment for HSCR is invasive surgical resection of the aganglionic gut segment, highlighting the need for alternative therapies that may perhaps be based upon a more comprehensive understanding of ENS genesis in vivo.

To date, the field has identified evolutionarily conserved genetic and cellular signaling factors required for ENS development. Largely in the mouse model, progress has been made in identifying transcription factors required for ENS formation; these include Sox10, Phox2b, and Pax3 [18–20]. Among the most studied cellular signaling pathways during ENCC development, it is known that receptors, Rearranged during Transfection proto-oncogene (Ret) and Gdnf family member receptor alpha1 (Gfra1), and their ligand, glial-derived neurotrophic factor (GDNF), are instrumental for ENCC infiltration, proliferation, and survival in the gut [21–23]. Indeed, it has been appreciated that the loss of combinations of various intrinsic and extrinsic factors results in severe ENS defects such as HSCR, suggesting that complex phenomena underlie the formation of ENS [17].

For over 20 years now, the zebrafish, Danio rerio, has been leveraged as a relevant model to understand vertebrate ENS development, form, and function [24,25]. Zebrafish ENS development is under the control of genetic circuits, such as Sox10 [26], Phox2b [27], and the Ret-GDNF pathway [23,28–30], demonstrating the conservation of genetic pathways giving rise to the ENS, and thus the utility of this model for studying ENS development. In a nutshell, zebrafish offer many outstanding characteristics to study ENS development, including: 1. Large clutches of transparent, externally developing embryos and larvae; 2. Homologous organs and genes with high conservation to mammals; 3. A large number of fluorescently tagged transgenic lines for live cell and whole-gut imaging; 4. High amenability to transgenic, drug, and genetic approaches [24,25].

In this review, I summarize knowledge regarding the genetic and signaling level regulation of enteric neuronal development in zebrafish, and offer some brief perspectives about possible future directions in the field.

Following emigration from the neural tube, zebrafish vagal NCCs localize in ventrolateral domains posterior to the otic vesicles (Figure 1A). By 36 hours post fertilization (hpf), ENCCs then begin to migrate as two single chains medially toward and along the foregut entrance [28,31,32] (Figure 1B). Zebrafish ENCCs express a defining combination of marker genes that encode for various transcription factors and receptors; including sox10, foxd3, hoxb5b, phox2bb, ret, and gfra1a/b [23,27,28,33–37]. Once resident along the gut, ENCC chains migrate caudally within gut mesenchyme until they reach their final destinations, with the wavefront (a.k.a. vanguard cells) reaching the hindgut's distal region ∼66–72 hpf [30,31,38]. Along the gut tube, mesenchymal tissues express chemoattractant-encoding genes, such as gdnfa [28], while intestinal tissues express additional signaling factor-encoding genes, such as shha [39].

Once reaching their destinations along the gut, which can extend from the foregut to the hindgut, ENCCs spatially pattern along the gut circumference, and begin to differentiate into enteric neurons [31,35] (Figure 1C), or glia [40,41], between 54 hpf and 7 days post fertilization (dpf) during the larval stage. Differentiating neurons express enteric neuronal subtype markers, such as vipb, nos1, chata, and pbx3b [36,41–43], while cells with differentiating glial identity express the markers her4, cx43, and s100b [40,41]. The zebrafish ENS continues neurogenesis after 7 dpf [40,43]; however, the ENS already contains an identifiable neuropil structure by 7 dpf [44]. The mature zebrafish ENS exists in one diffuse layer of neurons and glia, called the myenteric plexus, which is sandwiched between circular and longitudinal muscle layers of the gut [45]. While the zebrafish gut [46] and ENS represent simplified versions of the amniote counterparts — where zebrafish ENS is not arranged in ganglia and solely contains a myenteric plexus yet lack a submucosal plexus, the larger functions of the zebrafish ENS are conserved [25].

Spectacularly, research in the zebrafish ENS field has been especially fruitful for not only confirming the conservation of gene function across vertebrates — it has also functionally brought to light novel candidates that are required for ENS development. To date, we know of at least 65 genes (Table 1) [47] that are required for, and/or are functionally important for, zebrafish ENS development. While I do not discuss all the factors listed in Table 1, below I touch upon key gene regulatory network (GRN) takeaways from the data compiled and discuss what the results suggest regarding zebrafish enteric neuronal development.

Not surprisingly, many genes encoding for transcription factors that are expressed in and required for the specification of NCC populations are functionally important for zebrafish ENS development. Specifically, genes within the GRN of zebrafish NCC [48] are crucial for the subsequent development of ENCCs. For example, zebrafish with reduced pax3a levels [49], or harboring mutations in tfap2a [50,51], sox10 [26,27,52,53], foxd3 [50,51,54], and zeb2a [55] (formerly known as sip1a), all present with near complete ENCCs loss and total aganglionosis in larvae.

Across vertebrates, a core enteric GRN has been constructed from data largely using the mouse and chicken, which centers — the genes Sox10, Pax3/7, Phox2b, GDNF, Ret, Hand2, and Ascl1a [9]. Similarly, the execution of zebrafish enteric neuronal development relies upon an enteric neuron GRN (Figure 2). The zebrafish enteric neuron GRN represents transcription factors, chromatin-associated proteins, and signaling molecules, which together are expressed in vagal NCCs, ENCCs, microenvironmental tissues along ENCC pre-gut entry migratory paths, and/or in the developing gut tissues.

In zebrafish, morpholino-knockdown of pax3a [49], or loss of function mutation in foxd3 [54], decreases the expression of sox10 in migratory vagal NCC, later resulting in the near complete loss of ENCCs, and therefore, total aganglionosis in larvae. As evident in data from a zebrafish sox10 mutant line (formerly known as colorless) [27], sox10 is required for the expression of phox2bb and ret in ENCCs en route to and along the gut. While a paltry number of ENCCs reach the foregut in sox10 mutants, the ENCCs do not further colonize the gut, and total aganglionosis occurs. In foxd3 mutants [54], hand2 expression along the gut is abolished, highlighting foxd3’s upstream role within the GRN. Furthermore, mutation and/or morpholino-knockdown of ascl1a [56], hand2 [39], or phox2bb [27], results in compromised ENS development; although ENCCs are initially present, there is a reduction in enteric neuron numbers by the larval stage, likely due to defects in proliferation, migration, survival and/or differentiation of ENCCs.

The Ret signaling pathway, a key part of the enteric GRN, is necessary for zebrafish ENS development. Reduction in ret expression and/or function, via morpholino-knockdown or mutations, elicits enteric neuron loss, manifesting as colonic, and/or total aganglionosis [23,29,30,33,57–59]. In particular, dose-dependent disruption of ret in zebrafish, via heterozygous and homozygous mutant analyses [29,30], leads to colonic aganglionosis or total aganglionosis, respectively, demonstrating that even partial disruption of ret causes severe ENS defects. mapk10, encoding for the intracellular kinase Mapk10, has been shown to genetically interact with ret. Heterozygous disruption of ret combined with homozygous mutations in mapk10 increases the severity of gut aganglionosis [29], therefore, putting forth Mapk10 as a timely player that may function to influence ENS development synergistically with Ret. Moreover, similarly to mapk10, various other genes [58,60] have been discovered to genetically interact with ret, such as sema3c/d [58], expanding the Ret-associated network. Finally, morpholino-knockdown of the Ret pathway-associated genes, gdnfa [28], encoding for ligand GDNF, and gfra1a/b [23], encoding for co-receptors Gfra1a/b, causes drastic reductions in enteric neuron numbers along the larval gut.

Further building the zebrafish enteric GRN, various other studies have been able to link signaling level inputs to intracellular signaling effectors that drive ENS development. These recent studies have begun to pave the way for more thoroughly fleshing out the signaling networks underlying the complexity of ENS development.

One example includes the linkage of the bone morphogenetic protein pathway to downstream ENS regulators. bmp2b is expressed along the developing gut path of ENCCs during ENS development in zebrafish [61]. Morpholino-knockdown of bmp2b leads to total aganglionosis due to ENCC loss, as well as disruptions in enteric GRN factor expression. Specifically, bmp2b reduction leads to the abolishment of gdnfa expression in the gut tube, therefore, placing bmp2b within the zebrafish enteric GRN.

Another example is illustrated by the association between the Sonic Hedgehog (Shh) pathway and zebrafish ENS. shha, encoding for the ligand Shha, is expressed in gut endoderm during pre-gut entry phases ∼30 hpf, and along the intestine during colonization phases between 48 and 60 hpf, while ptch2 (formerly known as ptc1), encoding for the Shh receptor and repressor Patched 2, is expressed within gut mesenchyme and in a subset of ENCCs between 36 and 42 hpf [39]. A zebrafish shha mutant line (formerly known as sonic you), presents with total aganglionosis due to failure of ENCC colonization of the gut [39]. In addition, treatment with the Shh pathway inhibitor Cyclopamine phenocopies the shha mutant ENCC colonization phenotype, and depletes gdnfa expression along the gut, indicating that intestinal Shh signaling is upstream of gdnfa expression during zebrafish ENS development.

Further connecting signaling and intracellular inputs, Meis3 has been identified as an important factor during early zebrafish ENS development. meis3, which encodes for the co-transcriptional regulator Meis3, is expressed within ENCCs, and along ENCC migratory paths, during pre-gut entry and early colonization phases of ENS development [32,36]. Reduction in Meis3 via morpholino stalls ENCC migration along the gut and causes colonic aganglionosis. Meis3 knockdown also results in a near complete loss of intestinal shha expression, while also eliciting an expansion in ptch2 expression in the adjacent gut mesenchyme — suggesting that Meis3 either directly or indirectly up-regulates shha, while it represses ptch2 in the gut [32]. On the other hand, depletion of retinoic acid (RA) production via pharmacological inhibition of Aldh1a2 restricts meis3 expression, resulting in ENCC failure to colonize the gut, and total aganglionosis [62].

Recently, a connection between Ret signaling and enteric neuronal subtype expression has been brought to light. Zebrafish larvae heterozygous for a ret mutation [30] display colonic aganglionosis, due to defects in proliferation and migration along the gut. However, the ret mutant enteric neurons prematurely differentiate, whereby they express aberrantly elevated levels of pbx3b, encoding for Pbx3b, a transcription factor associated with excitatory neuron differentiation in mammalian ENS [63]. In congruence, immunohistochemical detection against choline acetyltransferase, an indicator of acetylcholine presence, was elevated within ret mutant enteric neurons, when compared with sibling controls [30]. Thus, while it is not yet clear whether Ret indirectly or directly regulates the expression of pbx3b, the results suggest pbx3b is downstream of the Ret pathway.

Epigenetics involves understanding how cell states and/or gene expression are altered without changes in the DNA sequence of a cell. While epigenetic regulation of gene expression is known to occur in NCC populations [64,65], largely through changes in chromatin state, we know relatively little about this in terms of the zebrafish ENS. Nonetheless, several studies have begun to shed light on how epigenetic factors regulate zebrafish ENS development and enteric GRN factor expression.

One epigenetic modification is DNA methylation. DNA methylation is associated with the suppression of gene expression and is mediated by DNA methyltransferases (Dnmt). De novo DNA methylation is created by Dnmt3a/b, while maintenance of DNA methylation marks is enabled by Dnmt1 [66]. The Ubiquitin-like protein containing PHD and RING finger domains 1 (Uhrf1) is responsible for recruiting Dnmts to unmethylated DNA [67].

While zebrafish carrying homozygous loss-of-function mutations in uhrf1 or dnmt1 display no apparent change in ENCC production during pre-gut entry and early gut migratory stages, later during larval stages mutants present with strong, yet variable intestinal hypoganglionosis and colonic aganglionosis phenotypes [68]. uhrf1 and dnmt1 mutant ENS phenotypes are likely due to diminished ENCC proliferation, migration, or survival; however, it is not yet clear what downstream ENS-related gene expression is altered in these zebrafish mutants. Interestingly, ENS phenotypes detected within double mutants for uhrf1 and dnmt1 are not more severe than those of the single mutants [68], suggesting Uhrf1 and Dnmt1 co-operate, and that DNA methylation as a whole is required for proper ENS development.

Besides DNA methylation, other epigenetic influences are essential for zebrafish ENS formation. The enzyme Chromodomain helicase DNA-binding protein (CHD) 7 is a chromatin remodeler [69]. CHD7 can also act as a co-transcriptional regulator at promoters and enhancer regions, either as a context-dependent repressor or activator, to affect gene expression [70]. Loss of chd7 via morpholino-knockdown in zebrafish results in widespread phenotypes, mimicking the human congenital condition known as CHARGE syndrome [71], which affects many tissue and organ systems, including the eye, craniofacial tissues [72]. Zebrafish chd7 loss of function embryos suffer from altered NCC specification, exhibiting decreased expression of NCC specifiers foxd3 and sox10 within cranial and vagal domains, as well as reduced crestin⁺ ENCC en route to and along the foregut [71]. Not surprisingly, chd7 loss of function embryos display complete ENS loss by the larval stage [71,73], and disrupted gut motility [74].

Recently one factor associated with chromatin modification, Ring finger protein 2 (Rnf2), has been discovered to impact zebrafish ENS formation [75]. Modifying histones to control chromatin state is another prominent way regulators exert epigenetic control over gene expression. The mammalian Polycomb repressive complex 1 contains the ubiquitin E3 ligase RING1, as either RING1A or RING1B, catalyzing the ubiquitination of Lys119 of histone H2A, largely a repressive mark [76]. Zebrafish mutants for rnf2, with predicted homology to Ring1b, present colonic aganglionosis, intestinal hypoganglionosis, and a reduction in gut tube ENCCs [75]. Examination during early NCC specification stages in rnf2 mutants revealed that the zebrafish NCC specifiers foxd3 and sox9b were differentially reduced in expression amongst cranial, vagal, and trunk NCC domains, where sox9b exhibited a strong loss of expression in cranial NCC, while foxd3 was globally restricted. During gut pre-entry stages, the ENCC expression of phox2bb, ret, and hand2 was reduced [75], suggesting that Rnf2 regulates these developmental ENS genes.

Within the past several years, zebrafish have been leveraged to determine if candidate HSCR genes identified from large-scale studies are required for ENS development. Specifically, several large studies analyzed HSCR-patient sequencing datasets and prioritized candidates for knockdown using the zebrafish model. These recent studies have identified many conserved, novel genes not known to have previous roles in ENS development. The identified genes include; dennd3, ncln, nup98, and tbata [77]; ufd1l, tbx2b, slc18a1, and mapk8 [60]; and acss2, sh3pxd2aa, eno3, and ubr4 [59] (Table 1). With this exciting field of zebrafish ENS research still relatively new, it will be interesting to see how the identified candidate genes functionally affect specific aspects of enteric development.

Over the past 20 years, there have been an increasing number of studies that utilize zebrafish to delineate the genetic networks, signaling landscapes, and epigenetic influences that underpin early ENS formation. To date, these studies have illuminated that at least 65 genes are essential for zebrafish ENS development (Table 1). As such, it is important to provide a timely overview of the GRN underlying zebrafish enteric neuron development, which I have synthesized from numerous studies and depicted using the GRN Biotapestry model [78] (Figure 2). Additionally, many recent transcriptomic studies have expanded our knowledge of the genes expressed within developing zebrafish ENS cells [36,40–42,79,80]. Looking to the future, it will be essential to validate connections between gene nodes within the GRN and incorporate how other newly discovered factors affect ENS ontogenesis.

• Zebrafish have been used as a robust and relevant model to understand vertebrate ENS development for over 20 years. To date, the field has discovered 65 genes necessary for proper ENS development.

• Genetic, epigenetic, and signaling inputs direct the early development of enteric neurons from vagal and ENCCs during ENS development. Validation of GRN links will be critical to resolve the zebrafish enteric GRN.

• Many novel candidate genes have been identified from HSCR-patient data sets as important for ENS development. Moving forward, it will be important to decipher how the newly identified genes regulate specific aspects of enteric development, how they interact with enteric GRN modules, and for purposes of furthering HSCR research, it will be important to validate findings in zebrafish with mammalian and human models of enteric development.

The author declares that there are no competing interests associated with this manuscript.

Funding is provided by the National Science Foundation CAREER Award 1942019 and the National Institutes of Health DK124804 awarded to R.A.U.

I would like to thank members of the Uribe lab and neural crest community for many fruitful discussions over the years.

dpf

days post fertilization

ENCC

enteric neural crest cell

ENS

enteric nervous system

GDNF

glial-derived neurotrophic factor

GRN

gene regulatory network

hpf

hours post fertilization

NCC

neural crest cell

PNS

peripheral nervous system

RA

retinoic acid