A Rapid F0 CRISPR Screen in Zebrafish to Identify Regulator Genes of Neuronal Development in the Enteric Nervous System

ABSTRACT

Background

The neural crest‐derived enteric nervous system (ENS) provides the intrinsic innervation of the gut with diverse neuronal subtypes and glial cells. The ENS regulates all essential gut functions, such as motility, nutrient uptake, immune response, and microbiota colonization. Deficits in ENS neuron numbers and composition cause debilitating gut dysfunction. Yet, few studies have identified genes that control neuronal differentiation and the generation of the diverse neuronal subtypes in the ENS.

Methods

Utilizing existing CRISPR/Cas9 genome editing technology in zebrafish, we have developed a rapid and scalable screening approach for identifying genes that regulate ENS neurogenesis.

Key Results

As a proof‐of‐concept, F0 guide RNA‐injected larvae (F0 crispants) targeting the known ENS regulator genes sox10, ret, or phox2bb phenocopied known ENS phenotypes with high efficiency. We evaluated 10 transcription factor candidate genes as regulators of ENS neurogenesis and function. F0 crispants for five of the tested genes have fewer ENS neurons. Secondary assays in F0 crispants for a subset of the genes that had fewer neurons reveal no effect on enteric progenitor cell migration but differential changes in gut motility.

Conclusions

Our multistep, yet straightforward CRISPR screening approach in zebrafish tests the genetic basis of ENS developmental and disease gene functions that will facilitate the high‐throughput evaluation of candidate genes from transcriptomic, genome‐wide association, or other ENS‐omics studies. Such in vivo ENS F0 crispant screens will contribute to a better understanding of ENS neuronal development regulation in vertebrates and what goes awry in ENS disorders.

Our rapid and cost‐effective F0 CRISPR/Cas9‐based screening approach provides a comprehensive framework for the functional evaluation of candidate genes for their role in enteric nervous system (ENS) development and function. This screening approach identifies new potential regulator genes of ENS neurogenesis and provides important first insights into their underlying biology.

Summary.

• Established F0 CRISPR screening approach in zebrafish to test the role of developmental and disease gene functions in ENS neuronal development, enteric progenitor cell migration, and intestinal transit.

• As a proof‐of‐concept, targeting the known ENS regulator genes sox10, ret, or phox2bb phenocopied known ENS phenotypes with high efficiency using F0 CRISPR approaches.

• We identified five potential new regulator genes of ENS neurogenesis, some of which affected intestinal transit.

1. Introduction

The enteric nervous system (ENS) is the largest and most complex division of the peripheral nervous system [1]. Developmentally derived from the neural crest cell population, the ENS forms a vast network of diverse neuron types and glial cells that innervate the different gut regions [1]. ENS development can be subdivided into two main stages: during early ENS development in zebrafish, enteric progenitor cells (EPCs) enter the developing gut at 32 h post fertilization (hpf), migrate to and along the developing gut, and reach the end of the gut at about 68 hpf [2]. EPC migration is linked to EPC proliferation during migration, as altered EPC proliferation impairs migration and vice versa [3, 4, 5, 6]. Once EPCs are in place, they differentiate into neurons and glial cells from anterior to posterior, starting at 54 hpf [2, 7]. At 5 days post fertilization (dpf), a functional neuronal network has formed, and zebrafish larvae start to feed externally [7]. The later stage of ENS development includes neuronal and glial EPC proliferation, differentiation, and specification [7]. Changes in neuron numbers can arise from effects on either stage of ENS development. If EPC migration is impaired, EPCs are not present in sufficient numbers for normal neuronal differentiation and specification [3, 4, 5, 6]. Alternatively, EPC migration may be normal, but later steps of ENS development may be affected, resulting in fewer ENS neurons.

The neuronal network of the ENS controls all essential gut functions, such as motility, nutrient uptake, immune response, and microbiota colonization [1, 8, 9, 10, 11]. Reduction in neuron numbers results in altered gut function; for example, impaired intestinal transit of food moving through the gut [11, 12, 13]. Key regulator genes promoting ENS neuronal development include components of the Ret signaling pathway and Phox2b [7, 14, 15]. Yet, despite the critical role of the ENS for gut development and function, few studies have identified genes that guide the generation of diverse neuronal types in the ENS; for example, the transcription factor genes Sox6, Ascl1, and Pbx3 [16, 17, 18, 19, 20]. The recent surge in transcriptomic data has generated a wealth of novel but functionally untested candidate genetic regulators of neuronal development in the ENS [16, 18, 21, 22, 23, 24]. However, the classical approaches for generating F2 mutant screens to functionally test the role of multiple candidate genes in neuronal development are time‐ and resource‐intensive [25]. In addition, there is no screening approach to test if genes regulate early or late stages of ENS neuronal development. CRISPR/Cas9 technology has facilitated rapid gene‐specific editing in zebrafish to generate F0 and germline mutants for studying development and disease [26, 27, 28, 29, 30, 31]. F0 CRISPR screens allow for prioritizing genes before generating stable mutant lines and subsequently analyzing phenotypes in‐depth. They have recently become an effective approach to prioritizing gene candidates with a focus on various phenotypes, including synapse formation, retinogenesis, spinal cord regeneration, macrophage development, and behavior through adulthood [32, 33, 34, 35, 36]. Additionally, the highly efficient CRISPR/Cas9 has driven multiplex F0 screens, inducing genome modifications by targeting multiple genes at once [26, 28, 30]. F0 analysis in zebrafish targeting Hirschsprung Disease (HSCR) candidate genes has successfully demonstrated changes in ENS neuron numbers along the gut in F0 crispants and thereby validated the role of HSCR candidate genes in ENS development [37, 38, 39]. Also, targeting single gene candidates regulating ENS development showed effects on neuron numbers in zebrafish [40]. Building on this, our F0 CRISPR screen described here uses a multiplex screening approach in a combination with analysis of ENS neuron numbers, EPC migration, and intestinal transit in F0 crispants to test genes for their role in neuronal ENS development and function.

Here, we present a rapid and cost‐effective CRISPR/Cas9‐based screening pipeline in zebrafish to functionally test candidate genes of ENS neurogenesis. The small, externally developing zebrafish larvae exhibit a fully functional and transparent gut by 5 days, allowing for the rapid evaluation of ENS phenotypes and GI function in live animals [7, 41]. Our screening approach provides a comprehensive framework for the functional evaluation of candidate genes for their role in EPC migration, neuronal differentiation, and ENS function. Our proof‐of‐principle experiments targeting the critical ENS developmental regulator genes ret, sox10, and phox2bb showed that we can phenocopy known ENS phenotypes in F0 guide‐RNA injected zebrafish larvae (F0 crispants). Then, we tested if functional loss of 10 candidate transcription factor genes leads to changes in ENS neuron numbers. We find that F0 crispants for five genes have fewer ENS neurons, identifying them as potential new regulators of ENS neurogenesis. We performed secondary assays that evaluate EPC migration and intestinal transit for three genes with the strongest reduction in ENS neuron numbers. F0 crispants showed no difference in EPC migration compared to controls, suggesting that early stages of ENS development were not affected. Further, evaluating ENS function in this gene subset showed decreased intestinal transit in F0 crispants, indicating effects on essential ENS‐regulated gut functions. Our F0 CRISPR screening approach thus identifies new potential regulators of ENS neurogenesis and provides important first insights into the underlying biology of these transcription factor genes.

2. Material and Methods

2.1. Zebrafish Husbandry

All experiments were carried out in accordance with animal welfare laws, guidelines, and policies and were approved by the Michigan State University Institutional Animal Care and Use Committee. The Tg(phox2bb:EGFP) ^w37 line (referred to as phox2bb:GFP from here forward) was maintained in a laboratory breeding colony according to established protocols [42]. Adult zebrafish were bred naturally in system water, and fertilized eggs were transferred to 100 mm Petri dishes containing ~25 mL of embryo medium. Embryos were allowed to develop at 28.5°C and staged by hours post fertilization according to morphological criteria [43].

2.2. gRNA Design and Synthesis

Gene‐specific gRNA sequences and corresponding genotyping primer sets were designed using Chop Chop (http://chopchop.cbu.uib.no/). The gRNA oligo was annealed to the scaffold gRNA primer containing a T7 RNA polymerase binding site (Table S1). The resulting dsDNA template was amplified by Phusion High‐Fidelity DNA polymerase (Thermo Fisher Scientific, F530S). The resulting dsDNA template was purified with a DNA clean and concentrate kit (Zymo Research, D4013), in vitro transcribed with the MEGAscript T7 kit (Invitrogen, AM1334), and purified with an RNA clean and concentrate kit (Zymo Research, R1013).

2.3. Protein Loading

A 2X Cas9 protein buffer solution was made containing 3200 pg/nL Alt‐R S.p. Cas9 Nuclease V3, 500 μg (IDT, 1081059), 600 mM KCl, and 8 mM HEPES pH 7.5. Protein loading was achieved by mixing up to 2 μL gRNA (200 ng/ul) with the 2X Cas9 buffer to achieve a final injection solution of 100 ng/μl (25 ng of each individual gRNA and 1600 pg/nL Cas9). The injection needle was calibrated to deliver ~1 nL of the injection solution into the yolk/cell interface of one‐cell stage zebrafish embryos.

2.4. Immunostaining

GFP immunostaining of embryos for GFP (1:1000, Thermo Fisher Scientific, Eugene, OR, catalog number A‐11120) was performed as previously described [44] with minor modifications. Briefly, embryos were fixed in 4% paraformaldehyde, 0.15 mM CaCl₂, and 4% sucrose in 0.1 M PO₄ buffer for 2 h at room temperature. After removing the fix, embryos were washed 3 × 10 min in 0.5% PBSTx. The pigmentation was removed by incubating for approximately 15 min in bleaching solution (3% H₂O₂/ 0.5% KOH). Subsequently, embryos were washed 3 × 10 min in 0.5% PBSTx. Embryos were incubated in a blocking solution for an hour at room temperature. The blocking solution was prepared with 0.5% PBS‐Triton X‐100, 1% DMSO, 2% bovine serum albumin (BSA, Sigma‐Aldrich, catalog number A3059), and 5% normal growth serum (NGS, ThermoFisher Scientific, catalog number PCN5000). Embryos were then incubated for 16–18 h in primary antibody in blocking solution. After removing the primary antibody, embryos were washed three times in 0.5% PBS Triton X‐100 for a minimum of 2 h and incubated for 16–18 h in secondary antibody in blocking solution. Finally, embryos were washed 3 × 10 min each and stored in the refrigerator in 0.3% PBS Triton X‐100.

2.5. Analysis of ENS Neuron Numbers (Including Genotyping)

2.5.1. Qualitative Assessment of ENS Reduction

On a Zeiss Axio Zoom v16 stereo microscope, 6 dpf F0 crispants previously identified with a golden phenotype were screened for ENS reduction. Anesthetized larvae were positioned laterally, and focusing throughout all focal planes, crispants with reduced ENS were identified as having large gaps in continuous GFP expression.

2.5.2. Quantification of Neuron Number in Anterior Part of the Gut in Ret Crispants

At 6 dpf, images of ret and slc24a5 crispants were acquired as described below, and quantitative analysis was performed by counting the number of GFP⁺ cells in the whole gut region spanning 200 μM in length. To determine the 200 μM segment, a vertical line was drawn at the point where the posterior swim bladder meets the gut tube, serving as the midpoint (Figure 1B). Using the established midpoint, a 200 μM scale bar (Zen software) was drawn, and GFP⁺ cells within this region were counted.

FIGURE 1.

F0 crispants targeting ret, sox10, or phox2bb recapitulate known ENS phenotypes. (A–E) F0 crispants targeting ret phenocopy the lack of phox2bb:GFP ⁺ ENS cells (green) in stable ret mutant larvae (C) in comparison to wildtype‐level ENS cells in control larvae (A). Compare ENS phenotypes between (C) and (D), arrows point to remaining ENS neurons in the anterior gut. (E) A small portion of ret F0 crispants show ENS cell reduction (B) Schematic of zebrafish larvae, the gray boxed area indicates the area shown in A‐M. Line indicates length used for quantification of anterior neuron numbers in ret crispants in (F). (F) Significantly fewer GFP⁺ cells are present within 200 um of anterior gut in ret F0 crispants compared to control injected larvae (p < 0.0001, 2 experiments, % shown as mean ± SEM, each data point equals one larva). (G) % of ENS phenotypes in ret F0 crispants (2 experiments, % shown as mean ± SEM). (H–J) F0 crispants targeting sox10 (I) phenocopy the lack of GFP⁺ ENS cells in stable sox10 mutant larvae (H). A small portion of F0 crispants targeting sox10 show ENS cell reduction (J). (K) % of ENS phenotypes in control slc24a5, sox10, and phox2bb F0 crispants (≥ 2 experiments, % shown as mean ± SEM). (L, M) The majority of F0 crispants lack GFP⁺ ENS cells (L). A small portion have reduction in ENS cells (M). (A, C–E, H–J, L–M) Whole‐mount side views at 6 days post fertilization (dpf). Asterisks: autofluorescence in gut epithelium; dashed line: Gut outline. Scale bar = 100 μm.

2.5.3. Genotyping F0 Crispants

Genomic DNA was isolated from individual 6–7 dpf crispants with reduced ENS to analyze insertions and deletions. Larvae were placed in lysis buffer containing 1X Thermo pol reaction buffer (New England Biolabs, B9004S) and 1 μg/μl Proteinase K (Invitrogen, 25,530,049) in nuclease‐free water, incubating for 50 min at 55°C and then 98°C for 10 min.

2.6. Analysis of EPC Migration

GFP immunostaining was performed on F0 crispants. The posterior end of the gut that includes the EPC migration front was imaged using a Zeiss Axio Zoom V16 stereo microscope. Using the ZEN software, the distance between the end of the gut and the EPC migration front (“migration progress”) was measured (Figure 3A).

FIGURE 3.

Secondary assays analyzing enteric progenitor cell migration and intestinal transit in F0 crispants. (A, top) slc24a5 F0 crispant control and embryo schematic (A, bottom) show the bilateral streams of GFP⁺ enteric progenitor cells (EPCs, green) migrating posteriorly to populate the gut. The gray box outlines the region of the embryo (A, top) and scale bar denotes the distance measured between the EPC migration front and the end of the gut (migration progress). (B) Quantification of migration progress in F0 crispants for jarid2a, dlx1a, and mycn compared to F0 slc24a5 crispant controls showed no significant (ns) difference (two experiments, each data point equals one embryo). (C) Experimental set‐up for the intestinal transit assay. Phenotypically sorted larvae are fed fluorescently labeled food at 6 and 7dpf. On day 7, 4 h post‐feeding, larvae that have eaten are selected (“feeders”) and placed in a new dish without food. Sixteen hours later, clearance of labeled food from the gut is determined. In sox10 mutants (E) and jarid2a F0 crispants (F) fluorescently labeled food (green) is still visible in the guts compared to the cleared guts of controls (D). (G) Quantification of percent larvae with cleared guts (≥ 2 experiments, % shown as mean ± SEM, each data point equals one experiment, each experiment ≥ 6 larvae). sox10 mutants and jarid2a F0 crispants have significantly lower percentages of larvae with cleared guts compared to controls. Asterisks show a significant difference to controls using an unpaired t‐test (p < 0.05) or no significant (ns) difference. dlx1a and mycn F0 crispants trend towards fewer larvae with cleared guts. Whole‐mount side views of 66 h post fertilization (hpf) embryo (A) and 8 days post fertilization (dpf) larvae (D–F) in the area indicated in the schematics. Scale bar = 100 μm.

2.7. Preparation of Fluorescent Tracer

Fluorescent green tracer was prepared as previously described [12] by mixing 100 mg of powdered larval feed (Larval AP100 100 μm, Zeigler Bros), 150 μL of yellow‐green fluorescent 2.0 μM polystyrene microspheres (FluoSpheres carboxylate modified microspheres 2% solid solution, Invitrogen) and 50 μL of deionized water. The ingredients were mixed on a watch glass to form a paste and then spread into a thin layer. The mixture was allowed to dry at room temperature in the dark. Once dry, the mixture was scraped off and crushed into a fine powder between two sheets of weighing paper.

2.8. Intestinal Transit Assay

At 6 dpf, embryos were sorted for reduced or normal ENS by fluorescence and transferred to a petri dish with embryo media in groups of 25 embryos. The intestinal transit assay was adapted from that previously described [12]. Starting on 6 dpf, embryos were fed 2 mg of the fluorescent powder per dish. At 7 dpf, embryos were transferred to fresh embryo media and fed 2 mg of fluorescent tracer. In the afternoon of 7 dpf, embryos were screened for feeders by the presence of fluorescent food in the gut. The feeders were then transferred to a fresh petri dish with embryo media. On the morning of 8 dpf, embryos were transferred to a fresh dish, anesthetized, and examined for clearance of fluorescent food from the gut using the Zeiss Axio Zoom V16 stereo microscope.

2.9. Image Acquisition

Images were acquired using a Zeiss Axio Zoom V16 stereo microscope and ZEN software. Images were processed and analyzed using Photoshop 2023 (Version 24.6.0, Adobe Systems Inc., San Jose, CA, USA) and figures were assembled using Adobe Illustrator 2023 (Version 27.5, Adobe Systems Inc., San Jose, CA, USA).

2.10. Sequence Analysis

Sequence analysis was performed on larvae after phenotypic analysis. PCR products containing the target site were amplified with gene‐specific primers (Table S1). The percentage of insertions and/or Deletions around the target site was determined using the free bioinformatics analysis tool (ICE CRISPR analysis) from Synthego, using the standard settings (https://www.synthego.com/).

2.11. Statistical Analysis

To determine significant differences, we performed an unpaired t‐test using GraphPad Prism 9.1.2.

3. Results

3.1. F0 Guide RNA‐Injected Larvae Fully Recapitulate Known ENS Phenotypes

First, we established that we can phenocopy known ENS phenotypes in F0 crispant zebrafish. As proof‐of‐principle, we co‐injected Cas9 protein and a guide RNA (gRNA) targeting the known ENS developmental regulator genes ret, sox10, or phox2bb in phox2bb:GFP transgenic zebrafish embryos. At larval stages, GFP⁺ cells are mostly ENS neurons [45]. F0 crispants were examined for qualitative reduction of GFP⁺ cells as a direct readout in live larvae. Stable ret mutants have no or only a few ENS neurons in the anterior part of the gut; stable sox10 mutants completely lack enteric neurons, and larvae injected with morpholino targeting phox2bb (phox2bb morphants) result in a significant reduction in ENS neurons [14, 15, 46, 47]. Likewise, ret F0 crispants have mostly no GFP⁺ cells in the ENS or only a few GFP⁺ cells remaining in the anterior gut, phenocopying stable ret mutant phenotypes (Figure 1A–G). We find a significant reduction of GFP⁺ cells in the anterior gut in ret F0 crispants (Figure 1G). In sox10 F0 crispants, we find fewer GFP⁺ cells in the ENS, with a phenotypic range from complete lack of GFP⁺ cells to ENS reduction, with most larvae showing strong to complete loss of GFP⁺ ENS cells (Figure 1H–K). This indicates that sox10 crispants phenocopy stable sox10 mutant phenotypes. We see a correlation between the ENS and pigment phenotype in sox10 stable mutants. In F0 crispants with complete pigmentation loss, we find a complete absence of GFP⁺ ENS cells (Figure S1A). F0 crispants with less reduction in pigment cells also have more variable ENS phenotypes (Figure S1A). F0 crispants targeting phox2bb show significant reduction of GFP⁺ ENS cells, phenocopying ENS defects in phox2bb morphants (Figure 1K–M). For sox10, ret, and phox2bb F0 crispants, we find robust efficiency of insertions and/or deletions (indels) at the target site (Figure S1B–E).

3.2. Reverse Genetic Screen Identifies New Regulators of ENS Neurogenesis

To identify new regulators of ENS neurogenesis, we performed a CRISPR F0 screen targeting 10 candidate genes (Table S1) selected from previous transcriptome analyses in the ENS [16, 18]. These genes are predicted to regulate different aspects of ENS neurogenesis, for example, stem cell maintenance, neuronal specification, and differentiation [16, 18], so our candidate genes may play a role in different steps of neuronal development. In addition, all genes are expressed in the zebrafish ENS [23, 24, 48]. For higher screening efficiency, we grouped the genes in four pools targeting two to three genes per pool (Table S1). Each pool consisted of genes with similar proposed roles in ENS neurogenesis. For pool injections, individual embryos were injected with a pool of gRNAs that consisted of a single gRNA targeting each gene in the pool with two to three candidate genes per pool. Screening gene pools allows the elimination of several candidate genes at once if the pool does not show changes in neuron numbers. A proportion of zebrafish genes have duplicate copies due to a whole‐genome duplication in the teleost fish lineage [49]. In such cases, both duplicates were combined in one pool to account for potential functional redundancy, for example, jarid2a and jarid2b. A gRNA targeting the slc24a5 gene was co‐injected in the phox2bb:GFP ⁺ one‐cell stage embryos to provide positive visual control of injection efficiency. At 2 days, slc24a5 F0 crispants have a highly consistent reduction of pigmentation (golden phenotype) that allows for easy visual screening [26]. As outlined (Figure 2, timeline), we screened each pool for larvae with complete or partial golden phenotypes, then screened these larvae for qualitative changes in GFP⁺ cells at 6 days when ENS neurons are well‐established. A proportion of F0 crispants have fewer GFP⁺ ENS cells along the gut in all pools except pool 2 (Figure 2A–H and Figure S2). For all genes, we found a robust indel efficiency between 64% and 91% in F0 crispant genomic DNA and good survival after injection (Figure 2J; Figures S1F and S3). To test which gene(s) are responsible for the ENS phenotypes, each gRNA was then injected individually, except for pool 2, which had no larvae with phenotypes. We found that mutations in jarid2a, mycn, foxj3, dlx1a, and phox2a genes resulted in a consistent and significantly increased percentage of larvae with fewer GFP⁺ cells compared to control larvae injected with slc24a5 gRNA. In contrast, targeting jarid2b, foxj2, and foxn3 did not consistently reduce GFP⁺ cell numbers (Figure 2D′D″,F′–F‴,G′–G‴,H). To test that ENS phenotypes are not due to off‐target effects of gRNAs, we designed two additional gRNAs resulting in deletions for each of the 5 genes that showed ENS phenotypes in F0 crispants. For each gene, we find a consistently significantly increased percentage of larvae with fewer GFP⁺ cells compared to control larvae injected with slc24a5 gRNA, suggesting that the ENS phenotypes are not due to off‐target effects (Figure 2I). Also, for our second set of gRNAs, we found a robust indel efficiency between 94% and 100% in F0 crispant genomic DNA (Figure 2K and Figure S4).

FIGURE 2.

F0 CRISPR screen identifies new regulators of ENS neurogenesis. The screening pipeline is designed to be performed within 10 days. (A) Schematic of zebrafish larvae, the boxed area indicates the area shown in B–G‴. (B) Uninjected control and (C) slc24a5 F0 crispants show wildtype levels of GFP⁺ ENS neurons (green) at 6 dpf. (D, F, G) F0 crispants at 6 dpf with fewer GFP⁺ cells targeting jarid2a + jarid2b (pool 1), mycn + foxj2 + foxj3 (pool3), dlx1a + foxn3 + phox2a (pool 4), but not (E) homeza + homezb (pool 2). F0 crispants, which target each gene individually from the three pools with ENS phenotypes, show fewer GFP⁺ cells for jarid2a (D′), mycn (F′), foxj3 (F‴), dlx1a (G′) and phox2a (G‴), but not jarid2b (D″), foxj2 (F″), and foxn3 (G″). (H) Average percentage of F0 crispants with and without reduced ENS phenotypes for each pool and each individual gene (≥ 2 experiments, % shown as mean ± SEM). The dashed line shows the percent phenotypes for the slc24a5 control F0 crispants. Asterisks show a consistent and significant difference using an unpaired t‐test (p < 0.05) to slc24a5 F0 crispant controls. (I) Average percentage of F0 crispants with and without reduced ENS phenotypes for second screen using two additional gRNAs for the individual genes with a phenotype (≥ 2 experiments, % shown as mean ± SEM). The dashed line shows the percent phenotypes for the slc24a5 control F0 crispants. Asterisks show a consistent and significant difference to slc24a5 F0 crispant controls using an unpaired t‐test (p < 0.05). (J) F0 crispants for each gene show a high percentage of indels in PCR amplicons of the target area (≥ 2 experiments, % shown as mean ± SEM, each data point equals one larva). (K) F0 crispants for each gene show a high percentage of indels in PCR amplicons of the target area (2 experiments, % shown as mean ± SEM, each data point equals one larva). Asterisks: Autofluorescence in gut epithelium; dashed line: Gut outline. Scale bar = 100 μm.

3.3. F0 Crispants Show no Change in Enteric Progenitor Cell Migration

Reduction in ENS neuron numbers at larval stages can result from impaired EPC migration during early steps of ENS development. To test if EPC migration was affected, we chose one gene from each pool with the highest levels of ENS phenotypes for further analysis. We quantified the migration progress of EPCs towards the end of EPC migration in jarid2a, mycn, and dlx1a F0 crispants (Figure 3A,B). We found no significant difference in EPC migration progress compared to controls injected with slc24a5 gRNA (Figure 3A,B), suggesting that EPCs migrate normally in all tested F0 crispants. Thus, altered EPC migration to and along the gut is likely not the cause of ENS neuron reduction. This result suggests that jarid2a, mycn, and dlx1a regulate later steps of ENS neuronal development.

3.4. F0 Crispants Show Changes in Intestinal Transit

Lack of ENS neurons as observed in sox10 mutants affects gut motility and intestinal transit of food through the gut [11, 13, 50], but whether a milder reduction in ENS neuron numbers as seen in our F0 crispants impairs gut motility remains to be tested. We analyzed if the reduction in ENS neurons is indicative of impacts on gut motility in F0 crispants targeting jarid2a, mycn, and dlx1a. For this, we used the intestinal transit assay that measures if food travels through the gut and is an efficient measure of impacts on gut motility (Figure 3C [11, 12]). In sox10 mutant larvae and jarid2a F0 crispants, the percentage of larvae with cleared guts was significantly lower compared to controls, suggesting that gut motility is impaired (Figure 3D–G). In dlx1a and mycn F0 crispants, we saw a trend of reduced average clearance rate compared to controls (Figure 3G).

4. Discussion

This study presents an F0 CRISPR/Cas9 genome editing screen that is a rapid, scalable, and inexpensive approach to functionally test candidate genes for a role in ENS neuronal development and function. Using our F0 CRISPR screening approach, we have three main findings: (1) our proof‐of‐principle experiment confirmed that ENS neuronal phenotypes are detected in F0 crispants and phenocopy known ENS phenotypes connected to ret, sox10, or phox2bb; (2) we have identified five potential new regulators of ENS neurogenesis; (3) F0 jarid2a, mycn, and dlx1a crispants show altered intestinal transit.

4.1. Identification of New Transcription Factor Genes Regulating ENS Neurogenesis

In our F0 CRISPR screen, we have identified five transcription factor genes (jarid2a, mycn, foxj3, dlx1a, and phox2a) for which F0 crispants show fewer ENS neurons along the gut. The secondary assays testing changes in EPC migration and intestinal transit aim to determine possible functions of the genes with ENS phenotypes during ENS development. Taking the results from our F0 screen together with the known roles of the five genes in neuronal development, we propose a working model of how these five genes might regulate the different stages of ENS neuronal development and function (Figure 4). In the central nervous system (CNS), Mycn is expressed in neuronal progenitor cells and controls neuron numbers by promoting progenitor cell proliferation and inhibiting neuronal differentiation [51, 52, 53, 54, 55, 56]. We propose that mycn promotes neuronal EPC proliferation and/or inhibits the differentiation of neuronal EPCs. In cell culture, Foxj3 inhibits neural differentiation of embryonic stem cell populations [57]. This suggests that foxj3 may inhibit neuronal differentiation in the ENS. In the CNS, Jarid2 has been suggested to maintain the postmitotic state of differentiated neurons and inhibit re‐entry into the cell cycle [58]. Similarly, jarid2a may keep ENS neurons in a differentiated state. Phox2a promotes neuronal differentiation in the CNS [59, 60, 61]. In mouse Phox2a mutants, early ENS development was not affected and later stages were not analyzed [59]. Our F0 screen suggests that phox2a promotes later stages of ENS neuronal development, consistent with the expression of phox2a in differentiating ENS neurons in zebrafish [62]. Dlx1 is expressed in CNS progenitor cells and has been suggested to promote neuronal differentiation [63, 64]. In contrast to our findings, adult mouse Dlx1 mutants have normal ENS neuronal density but significantly slower intestinal motility [65]. Further analysis of dlx1a stable mutants in zebrafish will be important to determine if the roles of Dlx1 during ENS development diverge between mice and zebrafish. As a next step, in‐depth analyses in stable mutants will determine which steps of ENS development are regulated by mycn, foxj3, jarid2a, phox2a, and dlx1a.

FIGURE 4.

Working model for functions of candidate genes with reduced ENS neurons. The working model combines our F0 screen results with known roles in neuronal development. Based on this, we propose which steps and stages of ENS neuronal development (migration or differentiation) and ENS function the different genes regulate.

4.2. Advantages of F0 CRISPR Screens to Prioritize Gene Functions in ENS Development

Determining the most promising targets among a large list of candidate genes is important since the establishment of homozygous mutant phenotypes requires a considerable amount of resources and time (approximately 9 months). Notably, F0 screens in zebrafish for ENS phenotypes are advantageous over other animal model systems due to the accessibility for manipulation and visualization at early developmental stages. Analyzing ENS phenotypes and effects on ENS function in live animals allows for faster screening, completing neuronal phenotyping and secondary assays within 10 days. Our setup thus enables the functional evaluation of 15 gRNA pools (three genes per pool) in about 4 weeks. The ability to target multiple genes in a single F0 crispant eliminates the time and resources required to test multiple separate genes. It is important to consider possible interactions between gene candidates within one pool, as in rare cases, the functional loss of one gene candidate may rescue the effect of the functional loss of a second candidate gene within one pool. If the expected positivity rate within the candidate genes is relatively low, targeting several genes in one pool can efficiently eliminate several genes with one experiment and thus reduce time. Our screening method is scalable by multiplexing larger pools of candidate genes with up to eight target genes per injection depending on the expected positivity rate within the candidate gene list [28]. In addition, the CRISPR/Cas9 system is rapidly evolving with emerging advances that create new screening opportunities including F0 adult screens in zebrafish [25, 32, 34, 66, 67, 68]. Another consideration is the presence of F0 crispants with and without phenotypes. Phenotypic variability might be due to mosaicism of mutations in the targeted gene within the tissue of interest within each F0 crispant. Alternatively, mutations in the targeted gene might result in a range of phenotypes. Analysis of stable mutant lines for each gene will help determine the extent of phenotypes connected to the functional loss of each gene. Our F0 CRISPR screening approach is also amenable to incorporating neuron number quantification, additional aspects of ENS development, ENS function, gut development, or additional time points including juveniles or adults, providing exciting opportunities for advancing ENS development and function.

4.3. Evaluation of Human Gut Disease Candidate Genes by F0 CRISPR Screens

F0 zebrafish crispants targeting Hirschsprung Disease candidate genes successfully identified several candidate genes with a reduction in ENS neurons similar to the Hirschsprung Disease phenotype [37]. A subset of the genes identified in our screen is also candidate genes for human disorders with gastrointestinal symptoms. Jarid2 and Dlx1 are candidate genes for autism spectrum disorder [69, 70]—a neurodevelopmental disorder with a high prevalence of gastrointestinal symptoms [71, 72, 73, 74]. Mycn is a candidate gene for Feingold Syndrome that includes as a symptom congenital intestinal obstruction [75]. It has been suggested that ENS neuron density is altered in cases of intestinal obstruction [76]. Our results suggest that jarid2a, mycn, and dlx1a affect ENS neuronal development and might consequently contribute to gastrointestinal dysfunction. This illustrates that our F0 screening approach can be used to efficiently screen for ENS phenotypes of human gut disease candidate genes.

4.4. Limitations to Modeling ENS Disorders in Zebrafish

Even though many gene functions in ENS development are conserved between mammalian species and zebrafish, there are some key anatomical and developmental differences that need to be taken into consideration when using zebrafish as an animal model for ENS disorders. In zebrafish, enteric progenitor cells migrate in two bilateral streams, whereas in mammals, enteric progenitor cells migrate in chains streams along the developing gut [2, 77, 78]. In mice, ENS neurons are arranged into two layers, forming (1) the myenteric plexus and (2) the submucosal plexus [1]. Each layer consists of ENS neurons clustering into ganglia. In zebrafish, ENS neurons are sandwiched between the circular and longitudinal smooth muscle layers in an equivalent position to the mammalian myenteric plexus but lack a second layer of ENS neurons equivalent to the submucosal plexus in mammals. Further, ENS neurons do not form ganglia, but form a neuronal network that consists of single cells or pairs of cells [41, 79].

5. Conclusion

Our multistep, yet straightforward CRISPR screening approach in zebrafish tests the genetic basis of ENS developmental and disease gene functions that will facilitate the high‐throughput evaluation of the manifold candidate genes emerging from transcriptomic, genome‐wide association, or other ENS‐omics studies. Such in vivo ENS F0 crispant screens will contribute to a better understanding of ENS neuronal development regulation in vertebrates and what goes awry in ENS disorders.

Author Contributions

Conceptualization: A.E.D., J.G. Formal analysis: A.E.D., J.G. Funding acquisition: J.G. Investigation: A.E.D., N.R.W.S., S.A.C., C.G.L., C.N., M.C.S., J.G. Methodology: A.E.D., J.G. Project administration: A.E.D., J.G. Supervision: A.E.D., J.G. Validation: A.E.D., N.R.W.S., J.G. Visualization: A.E.D., J.G. Writing – original draft: A.E.D., J.G. Writing – review and editing: A.E.D., N.R.W.S., S.A.C., C.G.L., C.N., M.C.S., J.G.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1.

Table S1.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.