Optimized chemogenetic ablation and regeneration of enteric nervous system neurons in zebrafish

Summary

The enteric nervous system (ENS) is the intrinsic nervous system of the gut and regulates essential functions, including motility. ENS loss is associated with gastrointestinal disorders such as Hirschsprung disease. Restoring missing neurons by stimulating neuronal regeneration presents a promising therapeutic approach for ENS disorders. To study ENS regeneration, we developed a targeted chemogenetic neuron ablation model in zebrafish using the Gal4/UAS-NTR 2.0 system. Spatially and temporally controlled neuronal death was confirmed by morphological changes, neuron loss quantification, and TUNEL assay. We observed an acute immune response that normalizes at 1 day post-treatment (dpt). We found complete restoration of ENS neuron numbers to control levels by 9 dpt, with recovery of gut motility. We identify neuronal subtype-specific differences in regenerative capacity and/or timing of cell replacement. Our study establishes a robust platform for dissecting the cellular-molecular mechanisms of ENS regeneration to develop potential treatment approaches for ENS-related disorders.

Highlights

• •Spatiotemporal controlled robust enteric neuron ablation assay
• •Neuron ablation triggers cell death and elicits an immune response
• •Neurons regenerate to control levels by 9 dpt
• •Regeneration of neurons restores gut motility defects

Ganz and colleagues established a targeted chemogenetic enteric nervous system (ENS) neuron ablation model in zebrafish. Neuronal loss is accompanied by an acute immune response. They observe robust restoration of neurons—with subtype-specific differences in regenerative capacity and/or timing—and recovery of gut motility parameters. Their study establishes a platform for dissecting the cellular-molecular mechanisms of ENS regeneration.

Introduction

The enteric nervous system (ENS), the largest subdivision of the peripheral nervous system, consists of an intricate network of neurons and glial cells derived from neural crest progenitors that migrate into the developing gastrointestinal (GI) tract (Ganz, 2018; Hao and Young, 2009; Taylor et al., 2016). This complex network plays a pivotal role in regulating essential GI functions such as gut motility, nutrient absorption, and epithelial secretion (Kuil et al., 2021; Kunze and Furness, 1999; Niesler et al., 2021). Dysfunction or loss of the ENS has been linked to various GI disorders such as Hirschsprung disease (HSCR), achalasia, and autism spectrum disorder (Niesler et al., 2021). HSCR is characterized by variable degrees of ENS neuron loss in the distal gut. HSCR is one of the most common congenital anomalies of the ENS, leading to severe intestinal obstruction in neonates, with an estimated incidence of 1 in 5,000 live births (Butler Tjaden and Trainor, 2013). To date, ENS disorders can only be treated symptomatically or by surgical removal of the area with ENS neuron deficits. The current standard treatment for HSCR is the surgical resection of the affected gut segment. However, patient outcomes vary widely, and most continue to experience lifelong GI defects, highlighting the need for alternative treatments. One of the therapeutic approaches for HSCR is the stimulation of resident stem cells to regenerate missing ENS cells, with promising results but varying efficacy depending on the age and the specific gut region (Burns et al., 2016; Joseph et al., 2011; Soret et al., 2020). The successful translation of these approaches requires a comprehensive understanding of the cellular and molecular mechanisms that underlie effective ENS regeneration. Elucidating these mechanisms necessitates the development of a reliable animal model that exhibits robust and reproducible regeneration of ENS neurons following injury.

So far, only a few studies - mostly in mammalian models - have investigated ENS regeneration using chemical, mechanical, or chemical-genetic (chemogenetic) injury. Although these approaches have provided valuable insights, mammalian models show limited and incomplete neuronal regeneration in the ENS (Rueckert and Ganz, 2022; Stavely et al., 2024). In contrast, zebrafish (Danio rerio) is a robust model for studying organ regeneration, including various parts of the nervous system, such as the brain (Kizil et al., 2012), retina (Hanovice et al., 2019), ENS (El-Nachef and Bronner, 2020; Ohno et al., 2021), but also other organs, such as the kidney (Diep et al., 2011), heart (Cao and Poss, 2018), and fin (Romero et al., 2018). Zebrafish also offer numerous experimental advantages, including rapid development, large numbers of offspring, optical transparency that enables live imaging, and a high degree of genetic homology with humans (Howe et al., 2013). Despite these experimental advantages (Kuil et al., 2021; Ganz, 2018), so far, only two studies have studied ENS regeneration in zebrafish using a laser ablation approach (El-Nachef and Bronner, 2020; Ohno et al., 2021). Those studies provided the first evidence that the ENS can regenerate in zebrafish following focal neuronal loss. In particular, Ohno et al. showed that neural crest-derived cells migrate into the ablated region, proliferate, and give rise to new neurons, partially restoring ENS neuron numbers. However, several important issues remain to demonstrate that the ENS fully regenerates in zebrafish: First, the number of regenerated neurons did not reach control levels in the studied time frame. Second, it remains to be shown whether different ENS neuronal subtypes are able to regenerate. Third, it remains unclear whether the regenerated neurons functionally integrated into existing neuronal circuits to re-establish normal gut function. In addition, focal laser ablation has several technical limitations, including collateral damage to surrounding tissues, labor-intensive protocols, and restricted ablation of just a few cells within a small gut region, hindering downstream analysis for functional recovery assessment (Ohno et al., 2021).

To overcome these challenges, we leveraged an established approach for spatiotemporally controlled cell ablation: the chemogenetic bacterial nitroreductase (NTR) system. This method utilizes the NTR enzyme, which converts prodrugs, e.g., metronidazole (MTZ), into cytotoxic agents, thereby selectively killing cells that express NTR (Bridgewater et al., 1997; Medico et al., 2001). In zebrafish, this chemogenetic method has been successfully applied in regenerative studies by ablating specific cell populations in heart, pancreas, liver, retina, and other organs (Ariga et al., 2010; Curado et al., 2007; Sharrock et al., 2022; White et al., 2017). However, this method has not yet been tested for the ablation of ENS neurons.

In this study, we took advantage of the precise spatial-temporal control of the Gal4/UAS system (Asakawa et al., 2008; Zhang et al., 2019) for the chemogenetic ablation of ENS neurons in postembryonic zebrafish larvae (Figure 1A). We additionally employed the most advanced iteration of the NTR technology, NTR 2.0, which provides effective ablation of target cells at ∼10- to 100-fold lower concentrations of MTZ than the original NTR system (Sharrock et al., 2022). Using this approach, we establish a robust experimental paradigm of ENS neuron regeneration by ablating only NTR-expressing neurons through MTZ treatment. Morphological and TUNEL analysis confirmed that the ablated cells underwent cell death with a concomitant acute immune response that normalizes within 1 day post-treatment (dpt). By 9 dpt, neuronal recovery was complete, with full recovery of most neuronal subtypes, including nitrergic, cholinergic, and VIPergic neurons. Our results demonstrate, for the first time, functional recovery of ENS neuron numbers, including the restoration of gut motility. These findings pave the way for understanding the cellular and molecular mechanisms of ENS regeneration, a first step for developing new therapeutic approaches for GI disorders associated with ENS dysfunction in the future.

Figure 1.

Identification of a Gal4 driver line for efficient ENS neuron ablation using a chemogenetic system

(A) Pictorial representation of chemogenetic ablation of ENS neurons. A Gal4 driver line carrier, Tg(Gal4), which drives expression in ENS neurons, was crossed with a responder/effector line carrier Tg(UAS:YFP-NTR 2.0). The Gal4 protein binds to the upstream activating sequence (UAS) element, initiating expression of both yellow fluorescence protein (YFP, reporter protein) and nitroreductase (NTR 2.0). Left side: Schematic representation of untreated control intestine, where ENS neurons co-expressing the Gal4 and UAS transgenes also express both YFP and NTR 2.0 (green). Right side: Schematic representation of intestine following treatment with 400 μM metronidazole (MTZ). The prodrug MTZ is enzymatically converted by NTR 2.0 into a cytotoxic compound, leading to the selective ablation of YFP-NTR 2.0-expressing ENS neurons.

(B and C) The bar graphs show percent colocalization of the pan-neuronal marker HuC/D with Reporter (Rep) expression in the gut for each Gal4 driver. (B) For the 234A:Gal4 line, most Rep+ cells are HuC/D+ at 5 dpf (81.3% HuC/D+/Rep+). The remaining cells are 12.7% HuC/D+/Rep- and 6.0% HuC/D-/Rep+ (n = 12 larvae per group, 2 exp). For the 495A:Gal4 line, only 55% cells are HuC/D+/Rep+ at 5 dpf, the remaining cells are 23.5% HuC/D+/Rep- and 21.2% HuC/D-/Rep+ (n = 12 larvae per group, 2 exp). (C) For the 217B:Gal4 line, there is a high degree of overlap between Rep and HuC/D expression [(85.3% (5 dpf), 85.7% (6 dpf), and 90.2% (7 dpf)]. The remaining cells consist of HuC/D+/Rep- [14% (5 dpf) 12.3% (6 dpf), and 9.0% (7 dpf)] and HuC/D-/Rep+ [0.7% (5 dpf), 2.0% (6 dpf), and 0.8% (7 dpf)], n = 13 (5 dpf), 15 (6 dpf), and 14 (7 dpf) larvae per group, 2 exp per time point. dpf = days post fertilization; exp = experiments.

Results

Identification of the Gal4 driver line for specific ablation of ENS neurons

To identify a Gal4 driver line for maximizing the breadth and selectivity of ENS neuron ablation, we obtained the three Gal4 gene-trap driver lines (495A:Gal4; 234A:Gal4; 217B:Gal4) based on their reported expression patterns in the gut and surrounding tissues from the National Institute of Genetics zebrafish Gal4 collection (https://nig-zebra.nbrp.jp; Asakawa et al., 2008). We crossed these lines with a UAS reporter line to analyze the extent and specificity of reporter gene expression in the ENS and surrounding tissues. To determine if reporter gene expression within the ENS is primarily restricted to neurons, thereby minimizing off-target ablation, we investigated reporter gene co-expression with the pan-neuronal marker HuC/D in the ENS at 5 dpf (495A:Gal4xUAS:GFP, 234A:Gal4xUAS:NTR:dsRed) and at 5, 6, and 7 dpf (217B:Gal4xUAS:GFP) (Figure 1).

Our analysis revealed notable differences in the co-expression profiles among the driver lines. The 495A:Gal4 line has previously been reported to show largely gut-restricted expression, but which gut cells express the reporter had not been determined (Asakawa et al., 2008). We found that the intestines of the 495A:Gal4 line displayed only about half HuC/D+/Reporter+ cells, with many reporter+ cells being HuC/D- and HuC/D+/Reporter-cells (Figure 1B). The expression of 234A:Gal4 has been reported to be expressed in the ENS, but also broadly outside the gut (Kawakami et al., 2010; Heanue et al., 2016). In contrast, the 217B:Gal4 has been reported to have reporter expression in the gut with minimal expression outside the gut compared to the 234A:Gal4 line. Consistent with these reports, we found that both 234A:Gal4 and 217B:Gal4 driver lines exhibited high co-expression of HuC/D+ and Reporter with only a small population of HuC/D-/Reporter+ and HuC/D+/Reporter-cells (Figures 1B and 1C). Both the 234A:Gal4 and 217B:Gal4 lines had sufficient Reporter+ cells in the gut. However, 234A:Gal4 drives extensive reporter gene expression outside the ENS and in ENS progenitor cells (Heanue et al., 2016), which would lead to the ablation of additional, non-targeted cells. Therefore, we chose the 217B:Gal4 line to maximize the extent and specificity of ENS neuron ablation (Figure 1C).

Establishing a robust chemogenetic ablation assay for ENS neurons

To establish a robust ENS neuron ablation assay, we used offspring from a cross between the 217B:Gal4 and UAS:YFP-NTR 2.0 transgenic lines that co-express YFP and NTR 2.0 in cells in the gut (YFP+). To ablate YFP-NTR 2.0 expressing cells, we treated YFP+ larvae with 400 μM MTZ for 24 h (Sharrock et al., 2022) starting at 5 or 12 dpf. To assess the extent of ENS cell loss, we quantified YFP+ cells at 6 dpf and 13 dpf (1 dpt for MTZ-treated larvae). In untreated controls, no significant difference in YFP+ cells was observed between 5 and 6 dpf, or 12 and 13 dpf. Conversely, MTZ-treated groups showed a near-complete loss of YFP+ cell numbers at 1 dpt compared to 0 dpt (before treatment) and untreated control (Figures 2A–2C).

Figure 2.

Establishment of a robust ENS neuron ablation assay

(A) The timeline of dpf and dpt illustrates the experimental procedure for MTZ treatments initiated at either 5 or 12 dpf.

(B) Left: YFP+ neurons (white dots) are present before treatment (0 dpt) at 5 or 12 dpf. Right: after 24 h treatment with 400 μM metronidazole (MTZ), YFP+ cells are completely absent in MTZ-treated fish (YFP+ MTZ) but present in untreated controls (YFP+ control).

(C) Quantification of YFP+ neurons in control and MTZ-treated groups. Data are shown as mean ± SD (n = 10 larvae per group, 2 exp); statistical significance ^∗∗∗∗p < 0.0001; ns = not significant.

(D) At 12 dpf, YFP+ cells [(anti-GFP immunostaining (top, green)] and the pan-neuronal marker HuC/D+ cells (middle, magenta) highly colocalize (bottom, overlay).

(E) Quantification of GFP+/HuC/D+, GFP+/HuC/D-, and GFP-/HuC/D+ cells. Data are shown as mean (n = 10 larvae per group, 2 exp).

(F) Percentage of YFP+ cells that are HuC/D+ and HuC/D-cells in E. (B) Whole-mount side views of live larval guts at the stage indicated. (D) Maximum projection confocal images of dissected intestine at 12 dpf; dpf = days post fertilization; dpt = days post treatment; exp = experiment; SD = standard deviation; scale bars, 100 μm.

As 24 h of MTZ treatment showed robust YFP+ cell ablation at both developmental stages, we decided to move forward with MTZ treatment at 12 dpf for further analyses. Embryonic neurogenesis is still very active at 5 dpf (Olden et al., 2008; Roy-Carson et al., 2017). Therefore, ablating neurons at 5 dpf might include processes of developmental neurogenesis rather than of neuronal regeneration, which we aimed to avoid.

To confirm that YFP+ cells are neurons at 12 dpf, we analyzed co-expression of YFP and HuC/D in 217B:Gal4; UAS:YFP-NTR 2.0 fish. Quantification revealed that, on average, 44.4 YFP+ neurons were observed among a total of 204.6 HuC/D+ neurons. We found that 96.3% of YFP+ cells were HuC/D+, indicating that the vast majority of YFP+ cells are ENS neurons, while only 3.7% YFP+ cells are HuC/D-. Therefore, we will hereafter refer to YFP+ cells in these lines as YFP+ neurons (Figures 2D–2F).

YFP+ neurons undergo DNA damage-induced cell death after MTZ treatment that elicits acute immune cell activation

To determine the timeline of neuronal ablation, we quantified the number of YFP+ neurons at 6, 12, and 24 hpt. We find that YFP+ neurons are essentially absent at 24 hpt, indicating that the MTZ-induced ablation of YFP+ neurons is completed within 24 h post-treatment (Figures S1A and S1B). To test if YFP+ neurons undergo cell death after MTZ treatment, we evaluated changes in neuronal morphology and performed TUNEL assays to detect cells undergoing cell death. YFP+ neurons in untreated control larvae displayed normal morphology with no visible signs of morphological changes. In contrast, YFP+ neurons in MTZ-treated samples exhibited significant morphological changes, including cell body fragmentation consistent with neurons undergoing cell death [Figure S1C (Cusack et al., 2013; Ham et al., 2010)]. Additionally, we conducted TUNEL assays combined with immunostaining for YFP to assess DNA damage in control and MTZ-treated fish at 6 hpt. As expected, we found GFP+/TUNEL+ cells in both treated and control groups since apoptotic cell death is an ongoing cellular process at 12 dpf (Lee et al., 2023). However, significantly higher numbers of GFP+/TUNEL+ cells were observed in the MTZ-treated group compared to the YFP+ untreated control (Figures 3A and 3B). In addition, GFP+/TUNEL+ neurons exhibited clear morphological changes, indicating that YFP+ neurons were undergoing cell death (Figure 3C). These results indicate that MTZ treatment induces DNA damage, which leads to cell death with pronounced structural alterations attending this process, consistent with prior studies using this system for selective cell death (Kulkarni et al., 2018; White and Mumm, 2013).

Figure 3.

MTZ treatment induces DNA-damage cell death in YFP+ neurons, and L-plastin+ immune cells increase rapidly after YFP+ cell death in the gut

(A) At 6 hpt, there is an increase in TUNEL staining (magenta) in YFP+ neurons (anti-GFP immunostaining) in MTZ-treated larvae compared to untreated controls. In the YFP+ control group, GFP+ neurons have normal morphology. In contrast, GFP+ neurons in the MTZ-treated group show fragmented morphology indicative of cell death.

(B) Quantification of GFP+/TUNEL+ cells shows a significant increase in YFP+ MTZ-treated larvae compared to YFP+ untreated controls. Data are shown as mean ± SD (n = 9 larvae per group, 2 exp); statistical significance ^∗∗∗p < 0.001.

(C) Magnified images of GFP+/TUNEL+ cells (white arrowhead) in the YFP+ MTZ-treated group.

(D) L-plastin+ immune cells (magenta) in YFP- MTZ (non-ablated) and YFP+ MTZ (ablated) larvae at 6 hpt, 12 hpt, and 1 dpt.

(E) Quantification of L-plastin+ cells at the indicated time points. At 6 hpt, YFP+ MTZ groups show a significant increase in L-plastin+ cells in the gut compared with YFP- MTZ. No significant difference (ns) was detected at 12 hpt and 1 dpt. Data are shown as mean ± SD (n = 10 larvae per group, 2 exp); Statistical significance: ^∗∗p < 0.01; ns = not significant.

(F) Fold-change analysis of L-plastin+ cell numbers in YFP+ MTZ larvae relative to YFP- MTZ larvae across time points shows a peak at 6 hpt followed by a decline to baseline levels by 1 dpt. (A, C, D) Maximum projection confocal images of dissected intestines at the time point indicated. dpf = days post fertilization; exp = experiment; hpt = hours post treatment; scale bars, 20 μm in A; 100 μm in B; 50 μm in D and E.

Immune cell activation plays a crucial role in neuronal regeneration (Kyritsis et al., 2012; White et al., 2017). To determine whether MTZ-induced cell ablation stimulates an immune response, we analyzed immune cell dynamics using the pan-leukocyte marker L-plastin (Hiroto Shinomiya, 2012). In YFP- MTZ-treated (non-ablated) and YFP+ MTZ-treated (ablated) larvae, we quantified L-plastin+ cell numbers in the gut across the time points of neuronal cell death (6 hpt, 12 hpt, and 1 dpt; see Figure S1). At 6 hpt, YFP+ MTZ-treated larvae exhibited a significant increase in L-plastin+ cells compared with YFP- MTZ-treated larvae. By 12 hpt and 1 dpt, no significant differences were detected (Figures 3D and 3E). Consistently, fold-change analysis revealed a peak in L-plastin+ cell numbers at 6 hpt, which steadily declined back to normal levels by 1 dpt (Figure 3F). These findings indicate that the MTZ-induced ablation of YFP+ cells triggers a transient immune response in the gut, characterized by an acute accumulation of L-plastin+ cells that resolves within 24 h.

YFP+ neurons regenerate following targeted ablation

Next, we investigated whether YFP+ neuron numbers were restored to control levels, which would indicate that they regenerate after MTZ treatment. We quantified YFP+ neurons along the whole gut at different time points before and after treatment in YFP+ MTZ and YFP+ control larvae. YFP+ control larvae exhibited an obvious neuronal network from 0 to 15 dpt (Figures 4A and 4B). In contrast, YFP+ neuronal numbers were reduced in the YFP+ MTZ larvae at 1 dpt, confirming successful ablation of YFP+ neurons (Figures 4A and 4B). By 2 dpt, we find a small number of YFP+ neurons along with developing neurites that bridge across the gut (Figure S2). YFP+ neuron numbers continue to increase, but through 6 dpt, there was a significant difference in YFP+ numbers compared to the control group. By 9 dpt, regenerated YFP+ neuron numbers reached levels comparable to the controls and remained at control levels through 12 and 15 dpt (Figures 4A and 4B). We noticed that approximately half of the YFP+ MTZ larvae exhibited abnormal body posture after 9 dpt. This phenotype might be caused by off-target ablation of reporter-positive cells in the spinal cord area (Figure S3). To investigate whether this affected YFP+ neuron regeneration, we compared regenerated YFP+ neuron counts in larvae with normal posture and those with defective posture at 12 dpt but found no significant difference (Figure S4). This indicates that postural defects did not impact neuron regeneration. To determine if there were differences in the timeline of neuron regeneration within discrete gut regions compared to the whole gut, we quantified YFP+ neurons in three main gut regions - anterior, midgut, and posterior (Jianlong et al., 2020; Wallace et al., 2005). We found that the timeline of YFP+ neuron regeneration was the same in the three gut regions as in the whole gut (Figures 4B–4F).

Figure 4.

Timeline of YFP+ neuron regeneration in the whole intestine and defined gut regions

(A) YFP+ neurons (anti-GFP immunostaining) and the neuronal network in YFP+ control and YFP+ MTZ zebrafish larvae at 0 (before treatment) and 1, 3, 6, 9, 12, and 15 dpt.

(B) YFP+ control larvae exhibited consistent GFP+ neuron numbers till 6 dpt with a slight decrease by 9 through 15 dpt. In contrast, the YFP+ MTZ group showed a dramatic decrease in YFP+ neurons, confirming complete ablation of YFP+ neurons (p < 0.001) at 1 dpt. A small number of YFP+ neurons appeared by 3 and 6 dpt, but numbers were still significantly decreased compared to the control. By 9 dpt, neuronal number fully recovered with no significant difference between treated and controls at 9, 12, and 15 dpt.

(C) Schematic representation of the zebrafish gut regions shows the three analyzed gut regions (magenta = anterior, green = midgut, blue = posterior) at 0–9 dpt and 12–18 dpt based on Jianlong et al. (2020) and Wallace et al. (2005). Anterior (D), mid (E), and posterior (F) gut regions show the same timeline of YFP+ neuron regeneration as the whole gut (B). (A) Maximum projection confocal images of dissected intestines at the time point indicated; Data are presented as mean ± SD (10 larvae per group, 2 experiments); Statistical significance: ^∗∗∗p < 0.001, ^∗∗p < 0.01, and ^∗p < 0.05; dpt = days post treatment; scale bars, 100 μm.

As shown above, YFP+ neurons colocalized with a small subset of HuC/D+ neurons, indicating that only a fraction of ENS neurons were ablated (Figures 2D–2F). To assess if the ablation of YFP+ neurons has any effects on the YFP- ENS neuronal population, we quantified HuC/D+ neurons in YFP+ MTZ and YFP+ control groups at 0 dpt (before treatment) and 1 and 3 dpt. We found no significant difference in HuC/D+ neuron numbers between treated and control groups after the ablation of YFP+ neurons (Figures S5A–S5D). This lack of significance may be due to the relatively small proportion of YFP+ neurons within the total HuC/D+ population—on average, 44.4 out of 204.6 neurons (Figure 2E). Additionally, MTZ treatment alone in YFP- MTZ-treated larvae did not significantly decrease overall HuC/D numbers (Figure S5E).

Specific co-expressed neuronal subtypes regenerate following ablation

The ENS is composed of various neuron subtypes that control the different functions of the gut, including gut motility (Holmberg et al., 2006; Roach et al., 2013; Uyttebroek et al., 2010). The subtypes can be grouped into three main classes, which include nitrergic/VIPergic inhibitory motor neurons [neuronal nitric oxide synthase+ (nNOS+); vasoactive intestinal peptide+ (vip+)], cholinergic excitatory motor neurons [solute carrier family 18 member 3a+ slc18a3a+)], and serotonergic neurons [5-hydroxytryptamine+, 5-HT+)] (Kuil et al., 2023; Li et al., 2025; Uyttebroek et al., 2010). To determine the neuronal subtype composition within the YFP+ neuronal population, we analyzed co-expression of these subtype markers with YFP at 12 dpf, prior to MTZ treatment. At 0 dpt, we found that 57.54%, 12.26%, 13.33%, and 6.58% of YFP+ neurons co-expressed nNOS, 5-HT, vip, and slc18a3a, respectively (Figure S6), indicating the presence of distinct YFP+ neuronal subtypes prior to ablation, which fall into the three major classes of ENS neurons. This distribution indicates that the YFP+ population is enriched for nitrergic inhibitory neurons relative to other ENS subtypes (Kuil et al., 2023; Li et al., 2025; Uyttebroek et al., 2010).

To examine if all neuronal subtypes within the YFP+ neuron population regenerated, we analyzed the percentage of YFP+ and subtype+ co-expressing cells within the YFP+ neuron population (% [YFP+/subtype+]/YFP+ neurons) in the MTZ-treated group at 9 dpt and compared them to the untreated control (Figure 5). Notably, YFP+/nNOS+ neurons showed no significant difference between the MTZ-treated and control group, suggesting robust stability and regenerative capacity for YFP+/nNOS+ neurons (Figures 5A and 5B). Analysis of YFP+/5-HT+ neurons showed a significant difference between the MTZ-treated and control group, indicating a limited regenerative capacity or delayed regenerative kinetics for this neuronal subtype (Figures 5C and 5D). Using HCR to detect vip and slc18a3a expression, we found that YFP+/vip+ and YFP+/slc18a3a+ neurons showed no significant difference between the MTZ-treated and control group, suggesting that these neuronal subtypes also had successfully restored following regeneration (Figures 5E–5H). Overall, these findings suggest that the majority of YFP+ cells are nitrergic neurons before ablation and after regeneration (Figures 5A, 5B, and S6). After regeneration, the percentage of YFP+/nNOS+, YFP+/vip+, and YFP+/slc18a3a+ neurons is successfully restored. In contrast, YFP+/5-HT+ neurons exhibited lower and/or slower regenerative capacity compared to the other subtypes.

Figure 5.

YFP+/nNOS+, YFP+/vip+ and YFP+/slc18a3a+ but not YFP+/5-HT+ neurons regenerate following ablation

(A) Example of nNOS+ (magenta) and YFP+ (anti-GFP immunostaining, green) neurons that co-localize (overlay) at 9 dpt in YFP+ control and YFP+ MTZ groups (white arrowheads).

(B) Quantification of percent [GFP+/nNOS+]/GFP+ neurons at 9 dpt in YFP+ control and YFP+ MTZ shows no significant difference.

(C) Example of 5-HT (magenta) and YFP+ (anti-GFP immunostaining, green) neurons that co-localize (overlay) at 9 dpt in YFP+ control and YFP+ MTZ groups (white arrowheads).

(D) Quantification of ([GFP+/5-HT+]/GFP+) neurons at 9 dpt in YFP+ control and YFP+ MTZ. The % [GFP+/nNOS+]/GFP+ neurons are significantly reduced in the MTZ-treated group compared to the control group at 9 dpt.

(E) Example of vip (magenta) and YFP+ (anti-GFP immunostaining, green) neurons that co-localize (overlay) at 9 dpt in YFP+ control and YFP+ MTZ (white arrowheads).

(F) Quantification of [GFP+/vip+]/GFP+ neurons at 9 dpt in both YFP+ control and YFP+ MTZ groups shows no significant difference.

(G) Example of slc18a3a (magenta) and YFP+ (anti-GFP immunostaining, green) neurons that co-localize (overlay) at 0 dpt control and 9 dpt of YFP+ control and YFP+ MTZ (white arrowheads).

(H) Quantification of [GFP+/slc18a3a+]/GFP+ neurons shows no significant difference between the groups. (A, C, E, G) Maximum projection confocal images of dissected intestines at the indicated stage. Data are presented as mean ± SD [n = 9 larvae (YFP+/nNOS+ and YFP+/5-HT+) and n = 10 larvae (YFP+/vip+ and YFP+/slc18a3a+) per group, 2 experiments), each dot represents one fish]; Statistical significance = ^∗p < 0.05; ns = not significant; dpt = days post treatment; scale bars, 10 μm.

Decreased gut wave movements recover to control levels after neuronal regeneration

ENS neurons play a critical role in regulating gut motility. Excitatory motor neurons mediate muscle contractions and inhibitory motor neurons induce muscle relaxation (Furness, 2012; Kunze and Furness, 1999). We find that the major portion of YFP+ neurons co-express inhibitory and excitatory motor neuron markers (Figure S6). Thus, we aimed to determine 1) whether the loss of YFP+ neurons affects gut motility and 2) whether the regenerated neurons restore gut function after ablation.

To assess this, we examined one gut motility measure, the time of the peristaltic wave interval (PWI) of anterograde movements (Holmberg et al., 2006). We measured the PWI at 1, 3, and 9 dpt in YFP+ and YFP- MTZ-treated fish, and YFP+ untreated controls. We included the YFP-MTZ-treated group to determine if MTZ treatment alone impacted PWI. At 1 dpt, we found no significant difference in the PWI time between control and treated groups (Figure 6A, Videos S1, S2, and S3). However, at 3 dpt, the PWI time in the YFP+ MTZ-treated group increased significantly compared to both control groups, suggesting slower peristaltic movement in the YFP+ MTZ group (Figure 6B, Videos S4, S5, and S6). At 9 dpt, the PWI time in the YFP+ MTZ-treated group was back to control levels (Figure 6C, Videos S7, S8, and S9). Overall, our results suggest that gut motility function is affected after YFP+ neuronal loss and then restored after regeneration, indicating functional recovery. Although YFP+ neurons represent only a subset of the HuC/D + ENS population (Figures 2D–2F), they are enriched for nNOS+ inhibitory motor neurons (Figure S6). Because nitrergic neurons regulate gut relaxation and peristaltic timing, their loss may explain the increased PWI despite targeting only a small fraction of ENS neurons (Furness, 2012; Kunze and Furness, 1999).

Figure 6.

Peristaltic interval wave time recovers at 9 dpt after YFP+ neuron ablation

(A) At 1dpt, there is no significant difference in gut PWI time across all experimental groups: YFP+ MTZ treated, YFP+ untreated control, YFP- MTZ treated.

(B) At 3dpt, YFP+MTZ-treated larvae showed significantly slower gut PWI time compared to both control groups.

(C) At 9dpt, there was no significant difference in PWI across all experimental groups. Data are presented as mean ± SD (n = 6 larvae per group, 2 experiments); statistical significance: ^∗∗p < 0.01; ns = not significant; dpt = days post treatment.

Video S1. YFP- MTZ

Representative video of YFP- metronidazole (MTZ)-treated zebrafish at 1 day post treatment (1 dpt), showing anterograde peristaltic waves of the gut.

Video S2. YFP+ control

Representative video of YFP+ control zebrafish at 1 dpt, showing anterograde peristaltic waves of the gut.

Video S3. YFP+ MTZ

Representative video of YFP+ MTZ-treated zebrafish at 1 dpt, showing anterograde peristaltic waves of the gut.

Video S4. YFP- MTZ

Representative video of YFP- MTZ-treated zebrafish at 3 dpt, showing anterograde peristaltic waves of the gut.

Video S5. YFP+ control

Representative video of YFP+ control zebrafish at 3 dpt, showing anterograde peristaltic waves of the gut.

Video S6. YFP+ MTZ

Representative video of YFP+ MTZ-treated zebrafish at 3 dpt, showing anterograde peristaltic waves of the gut.

Video S7. YFP- MTZ

Representative video of YFP- MTZ-treated zebrafish at 9 dpt, showing anterograde peristaltic waves of the gut.

Video S8. YFP+ control

Representative video of YFP+ control zebrafish at 9 dpt, showing anterograde peristaltic waves of the gut.

Video S9. YFP+ MTZ

Representative video of YFP+ MTZ-treated zebrafish at 9 dpt, showing anterograde peristaltic waves of the gut.

Discussion

The topic of reparative regeneration has gained significant attention over the past decade due to its potential therapeutic implications, for example, to treat ENS diseases with loss of ENS neurons. Yet, to date, no study has demonstrated functional ENS neuron regeneration either in mammalian models or in zebrafish. In the present study, we show targeted chemogenetic ablation and functional regeneration of ENS neurons for the first time in zebrafish (Figure 7). Our results determined the spatial and temporal cellular dynamics of targeted neuron ablation as demonstrated by increased TUNEL labeling at 6hpt. Additionally, we observed an acute L-plastin+ immune cell response at 6 hpt that resolves within 1 dpt (Figure 7). We find full recovery of ablated neuron numbers by 9 dpt for the majority of ENS neuron subtypes investigated. Specifically, we found full recovery of nitrergic, cholinergic, and VIPergic subtypes, but only partial or delayed regeneration of serotonergic neurons. Importantly, we also showed the impacts of neuronal ablation on a gut motility parameter and its recovery, suggesting functional regeneration in our system (Figure 7).

Figure 7.

Timeline of events after targeted ablation and regeneration of ENS neurons in zebrafish

Timeline from 0 dpt/12 dpf to 9 dpt/21 dpf depicts the different events: (i) neuronal subtype composition of YFP+ neuron population at 0 dpt; (ii) chemogenetic ablation approach activates cell death of YFP+ neurons to completely ablate YFP+ neurons; (iii) neuronal regeneration at 3 and 6 dpt, which affects gut peristalsis at 3dpt; and (iv) complete regeneration of YFP+ cells and functional recovery at 9 dpt. Panel 1 (0dpt/12dpf): YFP+ cells co-localize with the pan-neuronal marker HuC/D and various neuronal subtypes (nNOS, 5-HT, vip, and slc18a31a). Panel 2 (1dpt/13dpf): Set-up of Gal4-UAS:NTR 2.0-based ablation system. MTZ treatment induces DNA-damage cell death in YFP+ neurons, as confirmed by TUNEL assay at 6hpt. Additionally, there is an acute accumulation of L-plastin+ immune cells at 6hpt that resolves within 24 h. Panel 3 (3dpt/15 dpf): Neuronal ablation resulted in increased peristaltic wave interval time, reflecting reduced gut motility. Panel 4 (9dpt/21dpf): Complete recovery of YFP+ neurons and most neuronal subtypes (nNOS+, vip+, slc18a31+), except 5-HT+ neurons, and full functional recovery of gut motility. dpf = days post fertilization, dpt = days post treatment.

Extensive ENS neuron ablation and their complete regeneration in zebrafish

In zebrafish, the only two studies to study ENS neuron regeneration used focal laser ablation to target a few ENS neurons in a small gut region. Both studies reported the generation of new neurons following ablation (El-Nachef and Bronner, 2020; Ohno et al., 2021). However, quantification of regenerated neurons showed that they did not return to control levels in the analyzed time frame, nor were impacts on gut function investigated (Ohno et al., 2021). In our study, using chemogenetic ablation of ∼44.4 ENS neurons, we find robust regeneration of ablated cells to control levels by 9 dpt, which stay at control levels through 15 dpt, the last time point that we analyzed. We find that different neuronal subtypes within the three main ENS neuron classes are able to regenerate, but identify interesting subtype-specific differences in regenerative capacity or timing (Figure 7). Our work also shows, for the first time, recovery of a parameter of gut function - gut wave propagation - indicative of successful functional regeneration (Figure 7). In addition, our approach has several experimental advantages compared to laser ablation. First, laser ablation is less efficient, as it can only ablate a limited number of neurons within specific gut regions (El-Nachef and Bronner, 2020; Ohno et al., 2021). For instance, one study only ablated ∼16 neurons within one-somite-width of the posterior gut at 10–15 dpf (Ohno et al., 2021). In contrast, our chemogenetic ablation system successfully ablated on average 44 neurons throughout the whole gut, significantly improving efficiency. Second, laser ablation is time-consuming as it requires individual larval preparation and precise manual targeting (Ohno et al., 2021). In contrast, our methods enable the simultaneous ablation of ENS neurons in several hundred larvae. Because neuronal ablation can be induced simultaneously in large numbers of larvae, this chemogenetic system could also be adapted for chemical or genetic screening approaches to identify modulators of ENS regeneration for future studies. Third, laser irradiation can cause unintended damage to non-neuronal, surrounding tissues, including smooth muscle and epithelial cells, which may impact regeneration and complicate downstream analyses such as gut motility. In contrast, our approach is a spatiotemporally controlled ablation that precisely targets only specific ENS neurons and has no deleterious effects on the overall neuronal population in the gut (Figure 7).

Zebrafish are a powerful model system to study ENS regeneration

So far, ENS regeneration research has mostly focused on mammalian model systems with only limited work in non-mammalian models such as zebrafish (Corvetti et al., 2001; El-Nachef and Bronner, 2020; Gabella and Trigg, 1984; Galligan et al., 1989; Hanani et al., 2003; Jew et al., 1989; Joseph et al., 2011; Karaosmanoǧlu et al., 1996; Laranjeira et al., 2011; Luck et al., 1993; Ohno et al., 2021; Stavely et al., 2024; Tokui et al., 1994). Previous studies in mammalian models involving surgical (Galligan et al., 1989; Karaosmanoǧlu et al., 1996; Tokui et al., 1994), stenosis (Corvetti et al., 2001; Gabella and Trigg, 1984; Jew et al., 1989; Joseph et al., 2011), and chemical (Hanani et al., 2003; Laranjeira et al., 2011; Luck et al., 1993) ablation methods consistently report incomplete neuronal regeneration, and gut region- or species-specific differences in regenerative outcomes of ENS neurons (Rueckert and Ganz, 2022). Notably, all previous approaches also reported structural abnormalities at the injury site by collateral damage of surrounding tissues. To address this, recent works in adult mice reported focal ablation specifically of ENS neurons using a diphtheria toxin-based chemogenetic ablation approach (Bhave et al., 2019; Stavely et al., 2024). Even though this study reported that mature, post-mitotic ENS neurons in adult mice can regenerate axons and reinnervate the gut after injury, there was still limited regeneration of neurons even three months after ablation (Stavely et al., 2024). This indicates that mammals have generally limited ENS regenerative abilities. In contrast, the present study demonstrates that zebrafish are a powerful model to study the signals and cellular-molecular processes underlying functional ENS regeneration. Our spatio-temporally controlled approach ablates approximately one-fifth of all ENS neurons by inducing cell death with subsequent full recovery after only 9 days. Ablation along the entire gut will enable the analysis of cellular responses and gene regulatory programs active during ENS regeneration, as whole guts can be analyzed. We find an acute immune response during the time of cell death, which normalizes at 1 dpt. Studies on adult zebrafish brain traumatic injury and selective rod photoreceptor ablation in the retina have demonstrated that immune responses are crucial components of regenerative processes (Kyritsis et al., 2012; White et al., 2017). In our study, selective ablation of ENS neurons (Figure 7) led to a marked accumulation of L-plastin+ cells as early as 6 h after MTZ treatment. Given that L-plastin is strongly expressed in leukocytes, including macrophages, monocytes, and neutrophils (Hiroto Shinomiya, 2012), this finding suggests that selective neuronal death in the ENS triggers an acute immune cell recruitment, similar to what has been described in models of traumatic brain injury and retinal regeneration (Kyritsis et al., 2012; White et al., 2017). Future studies will be needed to provide mechanistic insights into how immune-neural interactions (Graves et al., 2021) regulate ENS regeneration. Our approach thus provides a rapid experimental setup to identify the signals and cellular responses important for neuronal regeneration in the ENS in the future (Figure 7).

Regeneration of neurons in the zebrafish CNS vs. ENS

The zebrafish central nervous system has been extensively studied with regard to its regenerative capacity (Kaslin et al., 2009; 2013; März et al., 2010) while little is known about ENS regeneration (Ohno et al., 2021). In the central nervous system, brain regions show different capacities to regenerate neuronal subtypes. In the zebrafish telencephalon, all ablated neuron types regenerate (Barbosa et al., 2015; Kroehne et al., 2011; Reimer et al., 2008), whereas the cerebellum is capable of regenerating many, but not all, neuron types (Kaslin et al., 2017). This variable regenerative capacity of the cerebellum has been attributed to the varying ability of cerebellar stem and progenitor cells to generate neurons and glia throughout the lifespan (Kaslin et al., 2009; 2013). We also find differences in the regeneration of neuron types in the ENS, where some subtypes (nitrergic, cholinergic, and VIPergic) regenerate fully, whereas serotonergic neurons do not reach control levels (Figure 7). There could be three possibilities for this difference in regeneration: (1) serotonergic neurons take longer to regenerate, (2) the progenitor cells or cellular mechanisms to regenerate serotonergic neurons are not activated at this stage of post-embryonic development, and (3) the low level of serotonergic cell loss incurred is insufficient to trigger serotonergic cell replacement. Further work will determine which cellular responses are triggered after cell ablation and the progenitor population(s) that generate new ENS neurons and neuronal subtypes in zebrafish.

Resource availability

Lead contact

Requests for further information and resources should be directed to the lead contact, Julia Ganz (ganz@msu.edu).

Materials availability

This study did not generate unique reagents.

Data and code availability

All data reported in this paper will be shared by the lead contact upon request. This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This work was supported by NSF CAREER #2143267 to J.G., NIH R21NS123629 to J.G., a Research Grant from the American Neurogastroenterology and Motility Society to J.G., A.V.S. was supported by a Royal Society of New Zealand Marsden grant awarded to A.V.S, D.F.A. and J.S.M. (contract 24-VUW-106), and NIH R01OD020376 to J.S.M. and D.F.A., supported development of the improved NTR 2.0 enzyme. We want to thank Rachel Alcorn, Taylor Lawrence, Giovanna Haddock, Kristen Lounsbury, Anthony Turner, Samantha Lampman, and Sharon An for excellent fish care.

Author contributions

M.A.S, K.M., H.R., and J.G. designed the study; M.A.S, K.M., and H.R. performed the experiments, data curation, validation, analysis, investigation, and visualization; M.A.S. wrote the manuscript; M.A.S, K.M., H.R., A.V.S., D.F.A., J.S.M., and J.G. revised and edited the manuscript; D.F.A., J.S.M., and J.G. supervised the project; All authors have read and agreed to the published version of the manuscript.

Declaration of interests

JSM has been awarded patents for the creation (US patent #7,514,595) and use of zebrafish expressing nitroreductase enzymes for gene (US patent #8,071,838) and drug discovery (US patent #8,431,768) applications.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-5-HT (Serotonin) antibody	IMMUNOSTAR	Cat# 20080; RRID: AB_572263
Mouse anti-HuC/HuD monoclonal antibody (16A11)	Thermo Fisher	Cat# A-21271; RRID: AB_221448
Mouse anti-GFP monoclonal antibody (3E6)	Thermo Fisher	Cat# A11120; RRID: AB_221568
Rabbit anti-GFP polyclonal antibody	Thermo Fisher	Cat# A11122; RRID: AB_221569
Rabbit anti-nNOS (neuronal nitric oxide synthase) polyclonal antibody	GeneTex	Cat# GTX133407; RRID: AB_2886970
Rabbit anti-Lcp1 (L-plastin) polyclonal antibody	GeneTex	Cat# GTX124420; RRID: AB_11167454
Chemicals, peptides, and recombinant proteins
Bovine Serum Albumin	Sigma	Cat# A-9647
DNase I	Thermo Fisher	Cat # 18068-015
Formalin solution, neutral buffered, 10%	VWR	MSPP-66312-0
Fluoromount-G® mounting media	Southern Biotech	Cat # 0100-01
GOTAQ(R) GREEN MASTER MIX	M7122	Cat #M7122
Hybridization chain reaction (HCR) buffers Probe Hybridization buffer Probe Wash buffer Amplifier buffer	Molecular Instruments	N/A
HCR Hairpins – 488 HCR Hairpins – 546	Molecular Instruments	N/A
Metronidazole BioXtra	Sigma	M1547-25G
Paraformaldehyde 8% Aqueous Solution, EM Grade	Electron Microscopy Sciences	https://www.emsdiasum.com/8-formaldehyde-aqueous-solution-3?srsltid=AfmBOooMCqUIuBX3UJGSrDn8Sz6y9lDrmJhi6pf5lViF2L4nXYtwB1Mj
Proteinase K Solution	Thermo Fisher	Cat# AM2548
vip HCR probe GenBank: NM_001114553.3	Molecular Instruments	N/A
slc18a3a HCR probe GenBank: NM_001077550.1	Molecular Instruments	N/A
Critical commercial assays
Click-iT™ Plus TUNEL Assay Kit	Thermo Fisher	Cat# C10618
Experimental models: Organisms/strains
Zebrafish: Tg(SAGFF(LF)217B)	Asakawa et al. (2008), https://nig-zebra.nbrp.jp	N/A
Zebrafish: Tg(SAGFF(LF)234A)	Asakawa et al. (2008), https://nig-zebra.nbrp.jp	N/A
Zebrafish: Tg(gSAlzGFFM495A)	Asakawa et al. (2008), https://nig-zebra.nbrp.jp	N/A
Zebrafish: Tg(5xUAS:EGFP)nkuasgfp	Kawakami et al. (2010)	N/A
Zebrafish: Tg(UAS-E1b:NfsB-mCherry)c264	Davison et al. (2007)	N/A
Zebrafish: Tg(5xUAS:GAP-TagYFP-2A-NTR2.0)jh513Tg; Et(2xNRSE-cfos KALT4)gmc617; Tg(5xUAS:GAP-ECFP, he1.1:GAP-ECFP)gmc1913	Sharrock et al. (2022); Xie et al., 2012	N/A
Oligonucleotides
GAL4 Forward primer 5′-CGCTACTCTCCCAAAACCAAAAGG-3′	IDT	N/A
GAL4 Reverse primer 5′- TCTCTTCCGATGATGATGTCGCAC -3′	IDT	N/A
Software and algorithms
GraphPad	GraphPad Software, San Diego, California, USA, https://www.graphpad.com	N/A
Slidebook6 software Version 6.0.13	3i′s SlideBook microscopy, https://www.intelligent-imaging.com/slidebook	N/A
ZEN Imaging Software, ZEN 2.3 blue edition	ZEISS, https://www.zeiss.com/microscopy/en/products/software/zeiss-zen.html#LanguageSwitchOverlayCloseButton	N/A
Biorender	https://www.biorender.com/	N/A
Other
Water, Molecular Biology Grade	Fisher BioReagents	CAS 7732-18-5
Invitrogen™ PBS - Phosphate-Buffered Saline (10X) pH 7.4, RNase-free	Fisher BioReagents	CAT# AM9624

Experimental model and study participant details

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. Zebrafish lines were maintained in a laboratory breeding colony according to established protocols (Westerfield, 2000). Adult zebrafish were bred naturally in system water, and fertilized eggs were collected and transferred to 100 mm petri dishes containing embryo medium. Embryos were allowed to develop at 28.5°C and staged by days post fertilization (dpf) as described (Kimmel et al., 1995; Webb et al., 2009). The zebrafish lines utilized in this study are listed in the key resources table. Double transgenic larvae were established to co-express YFP and NTR 2.0 in the ENS neurons. For this, the Tg(5xUAS:GAP-TagYFP-2A-NTR2.0)jh513Tg reporter/effector line was maintained together with the Et(2xNRSE-cfos KALT4)gmc617; Tg(5xUAS:GAP-ECFP,he1.1:GAP-ECFP)gmc1913 transgenic fish line. The Et(2xNRSE:cfos:KalTA4)gmc617 line expresses KalTA4, a zebrafish-optimized Gal4/VP16 fusion protein with neuronal specificity (Xie et al., 2012; Sharrock et al., 2022). We confirmed that the Tg(SAGFF(LF)217B) drives reporter expression in ENS neurons (described in the results section), as an incross of Tg(5xUAS:GAP-TagYFP-2A-NTR2.0); Et(2xNRSE-cfos KALT4)gmc617; Tg(5xUAS:GAP-ECFP,he1.1:GAP-ECFP)gmc1913 did not show YFP expression in ENS neurons. The Tg(5xUAS:GAP-ECFP,he1.1:GAP-ECFP)gmc1913 line expresses cyan fluorescent protein (CFP), our analysis was focused on YFP expression driven by the Gal4/UAS system. Genotyping of Gal4 driver lines was performed with the following primers for Gal4 (GAL4F: 5′-CGCTACTCTCCCAAAACCAAAAGG-3’; GAL4R: 5′- TCTCTTCCGATGATGATGTCGCAC -3′ using the PCR protocol: 95 ° × 3 min - [95 ° × 30 s - 60 ° × 30 s - 72 ° × 30 s] x 35°–72° x 5′ 12° ∞.

Method details

Chemogenetic ablation using metronidazole

The 217B:Gal4 driver line was crossed with the UAS:YFP-NTR 2.0 responder line. Offspring from this cross were sorted based on gut YFP+ cells into YFP+ and YFP− groups at 5 dpf using a Zeiss Axio ZoomV16 fluorescent stereo microscope. The sorted larvae were then treated at 5 dpf or for the later time points transferred to the nursery until the day of the assay. Prior to MTZ treatment, larvae - except for the 5 dpf groups - were starved for 24 h to clear the gut of autofluorescent matter. Approximately 50 larvae were transferred to a Petri dish (100 mm × 15 mm) containing 40 mL of 400 μM MTZ (M1547 Sigma Aldrich) (Sharrock et al., 2022) in zebrafish system water (diluted from a 5 mM stock solution). The following experimental groups were used: 1) YFP+ MTZ (YFP+ larvae treated with MTZ; 2) YFP+ control (YFP+ larvae, not treated with MTZ); and 3) YFP- MTZ (YFP- larvae, treated with MTZ). Larvae were incubated in the dark for 24 h to ensure maximal drug exposure as MTZ is light sensitive. After 24 h, the effects of MTZ treatment on YFP+ cell ablation were assessed using a Zeiss Axio Zoom.V16 fluorescent stereo microscope. After confirmation of cell loss, the MTZ solution was discarded and larvae were returned to the nursery, placed in 2.8-L tanks (∼40 larvae per tank), and raised until the 21 dpt/33 dpf.

ENS neuron regeneration analysis

The regeneration of MTZ-treated YFP+ cells in the gut was assessed at 12 dpf (0 dpt, before ablation) and subsequently at three-day intervals from 13 dpf (1 dpt) to 27 dpf (15 dpt) and compared to YFP+ untreated controls. At each regeneration time point, both control and treated larvae were checked briefly using a Zeiss Axio Zoom.V16 fluorescent stereo microscope, then fixed and immunoassayed (as described below) by confocal microscopy imaging for data acquisition.

Sample fixation

To ensure optimal imaging conditions, larvae were fasted for 24 h prior to sampling at the required stages to allow for gut clearance of autofluorescent matter. Then, larvae were euthanized with an overdose of Tricaine Methanesulfonate (MS-222). For immunolabeling, euthanized samples were fixed in 4% paraformaldehyde (PFA) prepared in 1x sweet buffer (8% sucrose, 0.2 M CaCl₂, 0.2 M PO₄ buffer, pH 7.3) for 2 h at room temperature. Then, they were washed subsequently 2 times in 0.5% Phosphate buffer saline-Triton X-100 (PBSTX) for 15 min at room temperature, and stored in PBSTX at 4°C. For TUNEL assay, euthanized samples were fixed overnight in 10% Natural buffer saline (NBF) at 4°C, then washed three times in PBS for 5 min at room temperature and stored in PBS at 4°C. For Hybridization Chain Reaction (HCR), euthanized larvae were rinsed thoroughly in RNAase free water, and fixed in RNAase free 4% PFA in PBS at 4°C for ∼16 h, washed three times in PBS for 5 min at room temperature, dehydrated in a graded methanol series (50, 70, 90, 95 and 100% methanol in nuclease free water), and stored in 100% methanol at −20°C for a minimum of 16 h or up to six months (Ibarra-García-Padilla et al., 2021).

TUNEL assays were performed using the Click-iT Plus TUNEL Assay Kits (C10618, Thermo Fisher Scientific) for In Situ Apoptosis Detection. For this, PBS was removed from fixed samples and permeabilization was carried out using Proteinase K (20 μg/mL) for 50 min at room temperature, followed by three PBS washes and a deionized water rinse. Then samples were re-fixed in 10% NBF in PBS for 5 min at 37°C, followed by three additional PBS washes and another deionized water rinse. To detect DNA strand breaks (i.e., for the positive control), the larvae were incubated in 1 unit of DNase I (Cat. No. 18068-015) diluted into 1× DNase I Reaction Buffer for 30 min at room temperature (Figure S7). After that, larvae were preincubated in 150 μL of Terminal deoxynucleotidyl transferase (TdT) buffer for 10 min at 37°C, incubated in the 100 μL TdT reaction mixture (94 μL TdT Reaction Buffer, 2 μL 5-ethynyl-2′-deoxyuridine-5′-triphosphate (EdUTP), and 4 μL TdT enzyme) for 30 min at 37°C, rinsed in deionized water, washed twice in 3% bovine serum albumin (BSA) in 0.1% PBS-Triton X-100 for 5 min each, and one time in PBS. Then the larvae were incubated with the reaction cocktail (90 μL TUNEL Supermix and 10 μL 10× Click-iT Plus TUNEL Reaction Buffer Additive) for 30 min at 37°C in the dark, followed by two 3% BSA in PBS washes and two final PBS washes. TUNEL staining was assessed using a Zeiss Axio Zoom.V16 fluorescence stereomicroscope. Immunostaining was subsequently performed under dark conditions, as described below, and larvae were stored in PBS until imaging.

Immunostaining

Whole-mount immunostaining was performed as previously described (Bandla et al., 2022; Uyttebroek et al., 2010). The following primary antibodies: 5-HT (1:5,000), HuC/D (1:1,00), mouse GFP (1:300), rabbit GFP (1:1,000), nNOS (1:1,000) and L-plastin, Lcp1 (1:100) were utilized to perform immunostaining (key resources table). The primary antibodies were detected using Alexa Fluor 488 and/or 568 secondary antibodies, followed by washes with 0.3% PBSTX. The staining was assessed using Zeiss Axio Zoom.V16 fluorescent stereo microscope and samples were stored in 0.3% PBSTX at 4°C until imaging.

Hybridization chain reaction

The (HCR) kit, probes, vip (Accession number: NM_001114553.3) and slc18a3a (Accession number: NM_001077550.1), and hairpins were from Molecular Instruments, Inc. (https://www.molecularinstruments.com) and the HCR protocol was performed as previously described (Ibarra-García-Padilla et al., 2021) with minor modifications. The HCR protocol was followed by immunostaining. The fixed samples stored in 100% methanol were rehydrated in graded methanol series, permeabilized with pre-chilled acetone at −20°C for 25 min, treated with 20 μg of Proteinase K for 1 h at room temperature, refixed in 4% PFA in PBS for 20 min, and washed with PBS-Tween before hybridization. For HCR hybridization, samples were pre-hybridized at 37°C, incubated with a probe mixture (4 μL of each 1 μM probe set stock) for 12–16 h, washed to remove excess probes, and amplified for 12–16 h, followed by sequential washes (Ibarra-García-Padilla et al., 2021). For antibody staining, samples were blocked in 5% goat serum, incubated with primary anti-mouse GFP (1:300) antibodies overnight at 4°C, washed extensively, and incubated with secondary antibodies overnight (Ibarra-García-Padilla et al., 2021). Throughout the protocol, samples were protected from light to prevent photobleaching, ensuring optimal visualization of hybridization and staining results. Afterward, stained samples were stored in PBST at 4°C until imaging.

Imaging, data acquisition, and quantification

For analysis of colocalization and quantification, intestines were dissected with forceps and mounted on glass slides in 0.3% PBSTX or Fluoromount-G mounting media (Southern Biotech, catalog number 0100-01) and covered with a coverslip. For the analysis of Gal4-based reporter co-expression with the pan-neuronal marker HuC/D, mounted intestines were imaged on a Zeiss Axio Imager microscope operated through the ZEN software to determine colocalization. Confocal z-stacks images were captured using the 3i spinning disk confocal microscope with a 10, 20, or 60× objectives, operated through Slidebook6 software (Version 6.0.13). For cell quantification, z-stacks were generated from confocal stacks in Slidebook6, which were used for manual cell counts and colocalization analysis. For analysis within gut regions, the images of the intestine were divided into three sections - anterior, middle, and posterior - as previously described (Jianlong et al., 2020; Wallace et al., 2005). Cell counts were normalized to 200 μm to account for varying lengths of each of gut regions between larvae and to 1000 μm for the whole gut.

Analysis of neuron morphology

To analyze morphological changes in neurons after MTZ exposure, control and MTZ-treated larvae were euthanized at the timepoint indicated, their intestines were dissected, mounted on glass slide in PBS and imaged on the 3i spinning disk confocal microscope with a 60× objective, operated through Slidebook6 software (Version 6.0.13).

Gut anterograde movements analysis

Larvae from control and treated groups were fasted 24 h prior to analysis, anesthetized with MS-222, and embedded in 0.8% low-melting-point agarose. Videos of the gut region from midgut to posterior were recorded on the Zeiss Axio Zoom.V16 fluorescent stereo microscope with ZEN Imaging Software (ZEN 2.3 blue edition). All time-lapse videos were recorded using the same conditions, including exposure time (206s), frame rate (100 frames/206 s), and pixels (2752 x 2208). The total number of anterograde, peristaltic wave movements was analyzed from the recorded videos. The Peristaltic Wave Interval (PWI) time was calculated as: PWI = T/N where PWI represents the average time interval (in seconds) between two consecutive anterograde peristaltic movements. The total duration of each video (T, in seconds) and the total number of peristaltic movements observed (N) were recorded.

Statistical analysis

For fractional analysis of neuronal colocalization patterns, descriptive statistics were applied. For multiple comparisons, one-way and two-way ANOVA were applied to normally distributed data, while the non-parametric Kruskal–Wallis test was used for non-normally distributed data. For pairwise comparisons, unpaired t-tests were used for normally distributed data, and the Mann–Whitney test was used for non-normally distributed data. To evaluate temporal changes across multiple time points in regeneration studies, multiple unpaired t-tests were performed between MTZ-treated and control larvae. Fold-change analysis was calculated based on the average number of cells in each group. All statistical analyses were performed using GraphPad Prism 9.1 (GraphPad Software, San Diego, CA, USA). A significance threshold of p < 0.05 was used to determine statistically significant differences.

Published: July 14, 2026