Fecal microbiota transplantation enhances cell therapy in a rat model of hypoganglionosis by SCFA‐induced MEK1/2 signaling pathway

Abstract

Hirschsprung disease (HSCR), one of several neurocristopathies in children, is characterized by nerve loss in the large intestine and is mainly treated by surgery, which causes severe complications. Enteric neural crest‐derived cell (ENCC) transplantation is a potential therapeutic strategy; however, so far with poor efficacy. Here, we assessed whether and how fecal microbiota transplantation (FMT) could improve ENCC transplantation in a rat model of hypoganglionosis; a condition similar to HSCR, with less intestinal innervation. We found that the hypoganglionosis intestinal microenvironment negatively influenced the ENCC functional phenotype in vitro and in vivo. Combining 16S rDNA sequencing and targeted mass spectrometry revealed microbial dysbiosis and reduced short‐chain fatty acid (SCFA) production in the hypoganglionic gut. FMT increased the abundance of Bacteroides and Clostridium, SCFA production, and improved outcomes following ENCC transplantation. SCFAs alone stimulated ENCC proliferation, migration, and supported ENCC transplantation. Transcriptome‐wide mRNA sequencing identified MAPK signaling as the top differentially regulated pathway in response to SCFA exposure, and inhibition of MEK1/2 signaling abrogated the SCFA‐mediated effects on ENCC. This study demonstrates that FMT improves cell therapy for hypoganglionosis via short‐chain fatty acid metabolism‐induced MEK1/2 signaling.

Fecal microbiota transplantation enhances enteric neural crest‐derived cell transplantation in a rat model of hypoganglionosis supporting a non‐surgical treatment option of neurocristopathies including Hirschsprung disease.

Introduction

The gastrointestinal tract relies on the enteric nervous system (ENS) for its physiological function and to maintain homeostasis (Nagy & Goldstein, 2017). Interruptions in ENS development during embryogenesis can cause neurocristopathies, such as Hirschsprung disease (HSCR). Surgical excision is currently being used as the primary therapy for HSCR, which alleviates abdominal distension and constipation but can cause medical complications (Heuckeroth, 2018; Pan et al, 2021). Therefore, effective alternatives should be developed.

Enteric neural crest‐derived cells (ENCCs) can self‐renew and differentiate into neurons and glia to rebuild the ENS. Emerging studies have reported the effects of ENCC transplantation on the HSCR model (Workman et al, 2017; Yu et al, 2017). However, ENCC transplantation has some bottlenecks, including insufficient proliferation and migration pre‐transplantation, large‐scale apoptosis during short‐term transplantation, and ineffective reconstruction of the neural circuit, which may be due to the incompatibility between the ENCCs and intestinal niche. ENCCs treated with cytokines, drugs, and signaling pathways regulators to ameliorate these side effects have also failed to repair the ENS completely (Fattahi et al, 2016; Zhang et al, 2017).

The physiological functioning of the gastrointestinal tract relies on ENS integrity, immunity, and microbiota harmony (Hao et al, 2016). The immune system and microbiota have been proposed to impact the ENCC phenotype and development, and the organization and maturation of the ENS (Yoo & Mazmanian, 2017; Yarandi et al, 2020). Thus, the abnormalities present in HSCR, including intestinal inflammation, microbiota disorders, and metabolite imbalance, may create a hostile environment that leads to incompatibility between ENCCs and the intestinal microenvironment (Neuvonen et al, 2018; Yu et al, 2018). Although intestinal inflammation has been shown to be selectively cytotoxic to neurons, its effects on ENCCs are still unclear (Venkataramana et al, 2015).

Gut microbiota composition is critical during the immature postnatal phase of the ENS because microbiota and their metabolites, such as short‐chain fatty acids (SCFA) and serotonin (5‐hydroxytryptamine, 5‐HT), participate in neurogenesis and neuroprotection (Kabouridis & Pachnis, 2015; Spohn et al, 2016; De Vadder et al, 2018; Foong et al, 2020). For example, indigenous spore‐forming bacteria from mouse and human gut microbiomes promote 5‐HT biosynthesis, which indicates that certain bacteria produce important modulators for gut homeostasis and can exert neuroactive effects on the body. SCFAs are hypothesized to play essential roles in ENS development by regulating GPR43/41 and promoting the secretion of 5‐HT from enterochromaffin cells (Yano et al, 2015; Spohn et al, 2016; Liu et al, 2021).

Emerging research have suggested dysbiosis and inflamed intestines in HSCR, which may interfere with ENCCs (Pierre et al, 2014; Venkataramana et al, 2015; Li et al, 2016; Neuvonen et al, 2018; Yu et al, 2018). Fecal microbiota transplantation (FMT) seems an effective therapeutic option for inflamed bowel diseases as it can reshape a recipient's gut microbiome (El‐Salhy et al, 2020; Quraishi et al, 2020). Whether FMT can relieve inflammation and promote ENCC transplantation in treating HSCR remain unknown. In this study, we performed 16S rDNA sequencing and targeted mass spectrometry to develop a detailed profile of the gut microbiome and its metabolites to determine their roles in ENCC phenotype and transplantation. We found that SCFAs, especially from Bacteroides and Clostridium, regulated the MAPK‐signaling pathway to enhance the therapeutic effect of ENCC transplantation, providing a proof of concept for their use as an intervention for diseases that affect ENS development.

Results

Hypoganglionosis negatively influences the phenotype of ENCCs

ENCC transplantation into the hypoganglionic segment of the host gut demonstrated potential yet limited capacity to cure HSCR, which suggests that the intestinal microenvironment in hypoganglionosis may negatively affect ENCCs. Therefore, to address the influence of the microenvironment of the hypoganglionic intestine on ENCCs, we established an in vitro system in which ENCCs were cultured in the presence of extracts derived from the hypoganglionic gut.

We first isolated ENCCs and generated ENCC‐derived neurospheres. ENCCs cultured in vitro could proliferate and form characteristic free‐floating neurospheres (Appendix Fig S1A) with a compact appearance and positive neuronal stem cell markers (Appendix Fig S1B) and yielded the derivatives positive for glial and neuronal markers (Appendix Fig S1C).

Then, a rat model of hypoganglionosis was generated using a serosal application of 0.1%–30‐min benzalkonium chloride (BAC) (Appendix Fig S2A). Neuronal markers, Tuj1 and PGP9.5, and the number of ganglia decreased significantly after BAC treatment (Appendix Fig S2B–D).

Next, we derived soluble extracts from BAC‐ and sham‐treated rats' colons (referred to as “intestinal extracts”). These sterile‐filtered intestinal extracts were added at the concentration of 10%, as defined by the dose–response curve (Appendix Fig S3), to the medium for ENCC culture to simulate how the hypoganglionosis intestinal microenvironment influences ENCC phenotype, including the generation and proliferation of neurospheres, as well as the viability, proliferation, migration, and apoptosis of ENCCs constituting the neurospheres. The ENCC‐derived neurospheres cultured with intestinal extracts from BAC‐treated colons were characterized by smaller sizes (Fig 1A and B) and changes in diameters (Fig 1C) than those cultured with sham extracts, suggesting significantly lower proliferative capacity. Moreover, the intestinal extracts from the hypoganglionic gut significantly decreased ENCC viability (Fig 1D), proliferation (Fig 1E), migration (Fig 1F), and increased apoptosis (Fig 1G). Overall, these data indicate that the intestinal microenvironment in hypoganglionosis might exert a potentially negative effect on the ENCC phenotype.

Figure 1. The intestinal microenvironment in hypoganglionosis negatively influences the phenotype of enteric neural crest‐derived cells.

Comparison of the phenotype of ENCCs treated with the intestinal extract from hypoganglionosis rats at weeks 2, 4, and 6 after BAC treatment. The intestinal extracts from the sham‐treated guts were used as controls.

• A
  Morphology of the ENCC‐derived neurospheres on days 2 and 4. Scale bars represent 200 μm.
• B, C
  The proliferation of ENCC‐derived neurospheres is represented by neurosphere diameter (B) and change in diameter on day 4 compared with day 2 (C).
• D
  ENCC viability assessed by CCK‐8 assay.
• E
  ENCCs number analyzed by nuclear staining with DAPI.
• F
  Migration of ENCCs evaluated by transwell plate assay and stain with crystal violet, scale bar, 200 μm.
• G
  ENCC apoptosis determined by Annexin V‐FITC/PI stain and flow cytometry.

Data information: All data are represented as means ± SEM (n = 6 biological replicates per group). *P < 0.05, **P < 0.01, and ***P < 0.001, defined by one‐way ANOVA or Kruskal–Wallis test (> 2 groups) followed by multiple comparisons with Bonferroni post hoc test. BAC, benzalkonium chloride; ENCCs, enteric neural crest‐derived cells.

Source data are available online for this figure.

Microbiota dysbiosis and inflammation in hypoganglionosis

The negative effect of the hypoganglionosis intestinal microenvironment on ENCCs may be due to various factors, including microbiota, metabolites, inflammation, toxins or remnants from epithelial cells (Pierre et al, 2014; Venkataramana et al, 2015; Li et al, 2016; Neuvonen et al, 2018; Yu et al, 2018).

To investigate the alterations in intestinal microbiota in hypoganglionosis, we performed 16S rDNA sequencing. We compared the fecal microbial communities of the hypoganglionic colon (referred to as “expanded colon segment,” ES), segment proximal to the hypoganglionic colon (referred to as “narrow colon segment,” NS), and sham‐treated segment (referred to as “sham”; Fig 2A and C). Firmicutes, Bacteroidetes, and Proteobacteria were the top three dominant populations. Compared with the sham group, Firmicutes phylum and Lactobacillus genus were significantly enriched, while Proteobacteria phylum and Stenotrophomonas, Delftia, Bacteroides, and Romboutsia genus were decreased in the hypoganglionosis group. There were no significant changes at the phylum and genus levels between NS and ES groups (Fig 2B and D).

Figure 2. Microbial dysbiosis and inflammation in hypoganglionosis.

Comparison of fecal microbial species and inflammation in the hypoganglionosis and sham rats. For hypoganglionosis, fecal samples from the narrow segments (NS, BAC‐treated, hypoganglionic) and expanded segments (ES, proximal‐to‐narrow segments, with abnormal ganglion) colon are analyzed, respectively.

• A–D
  Cluster analysis at the (A) phylum and (C) genus levels. A column accumulation diagram of the top 10 most abundant (B) phyla and (D) genera.
• E, F
  Alpha‐diversity indices demonstrate species abundance (E: OTUs, ACE, and Chao1) and diversity (F: Shannon and Simpson).
• G
  The functional distribution of intestinal microbiota using principal component analysis.
• H, I
  Inter‐group diversity analysis and metabolic mass spectrometry to compare the differences in (H) intestinal microbiota and (I) SCFAs metabolites between the hypoganglionosis and sham rats. The species indicated in red are linked to SCFAs production.
• J
  Representative HE staining and quantification to assess the intestinal mucosal injury. Scale bar, 100 μm.
• K
  Representative IHC stain and quantification of CRP and TNF‐α levels. Scale bar, 100 μm.
• L
  Representative immunofluorescence stain and quantification of β‐catenin levels. Nuclei were stained blue with DAPI. Scale bar, 200 μm.

Data information: Data (in A–I) are represented as means ± SEM (n = 3 biological replicates per group). Data (in J–L) are represented as means ± SEM (n = 6 rats per group) and presented relative to values from sham rats. *P < 0.05, **P < 0.01, and ***P < 0.001 defined by two‐tailed Student's t‐test (2 groups) or one‐way ANOVA and Kruskal–Wallis test (> 2 groups) followed by multiple comparisons with Bonferroni post hoc test. CRP, C‐reactive protein; ES, expanded segments; NS, narrow segments; PC, principal component; SCFAs, short‐chain fatty acids; TNF‐α, tumor necrosis factor α.

Source data are available online for this figure.

In addition, alpha‐diversity indices including OTUs, ACE, Chao1, and Shannon significantly decreased in hypoganglionosis rats (Fig 2E and F), suggesting that hypoganglionosis might contribute to a significant reduction in the intestinal microbiota diversity and abundance. Principal component analysis (PCA) demonstrated that the functional distribution of intestinal microbiota overlapped between the narrow and expanded segments in the hypoganglionosis group but not with the sham group (Fig 2G). These results suggest a unique composition of the intestinal microbiome in hypoganglionosis.

Next, inter‐group diversity analysis was performed to explore the differences in bacterial abundance. Lactobacillus was enriched, while Stenotrophomonas, Prevotellaceae, Delftia, Romboutsia, Bacteroides, Clostridium, Paraprevotella, Faecalibacterium, Alistipes, Negativibacillus, and Sutterella were reduced in the hypoganglionosis group when compared to the sham group. As most of the genera whose abundance varied in hypoganglionosis were reported to be the producers of SCFAs (marked in red in Fig 2H; Morrison & Preston, 2016; Zhao et al, 2018), we performed metabolic targeted mass spectrometry of SCFAs in fecal samples. Lower levels of acetic, propionic, and butyric acids and higher levels of hexanoic acid were found in the hypoganglionosis group compared with the sham group (Fig 2I). These results indicate that changes in SCFA levels coincide with microbiota dysbiosis in hypoganglionosis, suggesting a possible role of SCFA metabolism by gut microbes in ENCC transplantation.

Then, we evaluated inflammation and mucosal damage in hypoganglionosis. H&E staining showed mucosal damage in the hypoganglionic segments, accompanied by local hyperplasia, crypt damage, and muscle layer thickening (Fig 2J). Moreover, higher scores of intestinal mucosal injury were found in hypoganglionosis rats than in sham rats (Fig 2J). In addition, hypoganglionosis samples had significantly higher TNF‐α and lower β‐catenin expression than sham samples (Fig 2K and L). These results indicated inflammation in hypoganglionosis.

Altogether, these data suggest altered microbiota and inflammation in hypoganglionosis.

FMT enhances the ENCC phenotype and promotes cell therapy for hypoganglionosis

Since the hypoganglionosis intestinal microenvironment negatively affected ENCCs (Fig 1) and hypoganglionosis showed altered microbiota and inflammation (Fig 2), we hypothesized that imbalance in gut flora and inflammation might limit the success of ENCC transplantation, which could be improved by FMT.

To test whether FMT could improve ENCC transplantation, hypoganglionosis rats were treated with broad‐spectrum antibiotics followed by transplantation of fecal microbiota from pooled feces of healthy littermates. Two weeks after FMT, intestinal extracts were collected, filtered, and added to the medium for ENCC culture. ENCCs cultured with intestinal extracts from FMT‐treated hypoganglionosis exhibited significantly better neurosphere formation (Fig 3A) and viability (Fig 3B), better proliferation (Fig 3B), higher migration (Fig 3C and D), and lower apoptosis (Fig 3E and F) than those cultured with intestinal extracts from untreated hypoganglionosis. These results suggest that FMT could ameliorate the negative effect of hypoganglionosis intestinal microenvironment on the ENCC phenotype.

Figure 3. Fecal microbiota transplantation enhances the ENCC phenotype and promotes cell therapy for hypoganglionosis.

Comparison of ENCC‐derived neurospheres treated with the intestinal extract from hypoganglionosis rats, hypoganglionosis treated with fecal microbiota transplantation (FMT), and sham rat.

• A, B
  The diameter of ENCC‐derived neurospheres on day 2 (A). ENCC viability assessed by CCK‐8 assay (B).
• C, D
  Migration of ENCCs using transwell plate assay and stain with crystal violet (C), scale bar, 200 μm. (D) Quantification of (C).
• E, F
  ENCC apoptosis determined by Annexin V‐FITC/PI stain and flow cytometry (E). (F) Quantification of (E).
• G
  Graphical protocol and weight change in the hypoganglionosis: untreated and treated with ENCCs only, or ENCCs+FMT.
• H
  Laparoscopic analysis indicating that the same segments were treated with BAC in all the groups.
• I
  Delay in distal colonic transit time.
• J, K
  Comparison of neuronal markers. (J) Immunofluorescence stain and number of PGP9.5⁺ cells. Nuclei were stained blue with DAPI; triangles indicate the PGP9.5⁺ cell. Scale bars represent 200 μm. (K) Representative immunofluorescence images of Tuj1 and GFP localization at the injection and adjacent sides. Asterisks and triangles indicate the Tuj1⁺/GFP⁺ cells and Tuj1⁺/GFP⁻ cells, respectively. Nuclei were stained blue with DAPI. Scale bar, 200 μm.
• L, M
  Number of Tuj1⁺, Tuj1⁺/GFP⁺ cells (L), and Tuj1⁺/GFP⁻ cells (M).

Data information: Data (in A–M) are represented as means ± SEM (n = 6 biological replicates per group) and presented relative to values of the hypoganglionosis group. *P < 0.05, **P < 0.01, and ***P < 0.001, defined by two‐tailed Student's t‐test (2 groups) or one‐way ANOVA and Kruskal–Wallis test (>2 groups) followed by multiple comparisons with Bonferroni post hoc test. BAC, benzalkonium chloride; ENCCs, enteric neural crest‐derived cells; FMT, fecal microbiota transplantation; GFP, green fluorescent protein; PGP9.5, protein gene product; Tuj1, neuron‐specific class III beta‐tubulin.

Source data are available online for this figure.

We next asked whether FMT could increase the success of ENCC transplantation. ENCCs were labeled with an enhanced green fluorescent protein reporter (eGFP) using lentiviral transduction (Appendix Fig S4) to follow transplanted cells. Hypoganglionosis rats were treated with broad‐spectrum antibiotics followed by FMT and ENCC transplantation (Fig 3G, upper part). The treatment did not affect the weight change (Fig 3G, lower part). Exploratory laparotomy showed the most severe intestinal dilation in hypoganglionosis, while reduced intestinal dilation in hypoganglionosis treated with ENCCs and ENCCs+FMT (Fig 3H). In addition, ENCCs+FMT treatment significantly shortened distal colonic transit time in the hypoganglionosis rats, compared with no treatment or ENCC transplantation (Fig 3I). Together, these results suggest that FMT enhances the effect of ENCC transplantation in restoring intestinal peristalsis in hypoganglionosis.

To understand the basis of the observed improvement, we next assessed how the FMT affected ENCC engraftment in the hypoganglionic gut. ENCCs+FMT treatment significantly increased Tuj1 expression in the hypoganglionic gut compared with ENCC transplantation and no treatment groups (Fig 3K). Precisely, we observed more Tuj1⁺/GFP⁺ and Tuj1⁺/GFP⁻ cells in ENCCs+FMT and ENCC transplantation groups (Fig 3K–M), indicating enhanced neurogenesis from both transplanted (Tuj1⁺/GFP⁺ cells) and endogenous (increased Tuj1⁺/GFP⁻ cells) ENCCs. Moreover, increased Tuj1⁺/GFP⁺ cells were seen at both the injection and adjacent sides (Fig 3K and L), demonstrating that transplanted ENCCs migrated and colonized the hypoganglionic gut. Higher Tuj1⁺/GFP⁻ cells were only observed at the injection sites but not the adjacent sites, indicating that FMT partially affected endogenous neurogenesis after ENCC transplantation (Fig 3L and M). Additionally, the expression of PGP9.5 further supported the role of FMT in ENCC transplantation (Fig 3J).

In summary, these data demonstrate that FMT could increase the success of ENCC engraftment, colonization, and functional integration in hypoganglionosis.

FMT improves microbial dysbiosis, which coincides with increased short‐chain fatty acid production and reduced IL‐17 levels

To determine the potential factors involved in the positive effect of FMT on ENCC transplantation, we conducted a systematic analysis of microbiota, SCFAs, and inflammatory factors in the serum, intestinal extract, and feces from the hypoganglionosis: untreated (Hypoganglionosis) or treated with broad‐range antibiotics (Antibiotic), followed by FMT (FMT) and sham rats, by 16S rDNA sequencing, metabolomics, and ELISA.

The composition of the intestinal bacteria at the phylum and genus levels significantly varied in the groups (Fig 4A and C). The top most abundant phyla in all groups were Firmicutes, Bacteroidetes, and Proteobacteria. In hypoganglionosis, the abundance of Firmicutes and Bacteroidetes was significantly reduced, while Verrucomicrobiota increased. After FMT, bacterial abundance was similar to that in the sham group. Specifically, the abundance of Firmicutes and Bacteroidetes significantly increased, and Verrucomicrobiota decreased (Fig 4B). The top most abundant genera were Escherichia‐Shigella, Akkermansia, and Lactobacillus. Escherichia‐Shigella was the most abundant in the Antibiotic group. The abundance of Lactobacillus increased, and that of Akkermansia decreased after FMT (Fig 4D). Moreover, the sparse curve and PCA results showed a partial overlap between the FMT and sham groups (Fig 4E and F), indicating that FMT treatment restored gut microbiota in hypoganglionosis. Next, we explored the differences in genera and species of gut microflora in response to FMT. The abundance of Akkermansia, related to intestinal epithelial barrier function, was significantly decreased at both genus and species levels after FMT (Appendix Fig S5). Conversely, the relative abundances of Lactobacillus, Bacteroides, Clostridium, Bifidobacterium, Lachnospiraceae, and Romboutsia, which have been reported as the producers of SCFAs, were significantly increased in hypoganglionosis after FMT to the similar or higher levels as seen in sham rats (Appendix Fig S5).

Figure 4. FMT improves microbial dysbiosis, which coincides with increased short‐chain fatty acid production and reduced IL‐17 levels.

Comparison of fecal microbial species, short‐chain fatty acids, and inflammation between hypoganglionosis: untreated (Hypoganglionosis) or treated with broad‐range antibiotics (Antibiotic), broad‐range antibiotics followed by FMT (FMT), and sham groups.

• A–D
  Cluster analysis at the (A) phylum and (C) genus. Column accumulation diagrams of the top abundant microbiota at the (B) phylum and (D) genus levels.
• E, F
  Sparse curve (E) shows the relative abundance, and (F) principal component analysis (PCA) demonstrates the functional distribution of microbiota.
• G
  SCFAs subsets in the feces and intestinal extracts, analyzed by targeted mass spectrometry.
• H
  IL‐10, IL‐17, and TNF‐α levels in serum and intestinal extracts, assessed by ELISA.
• I
  Correlation analysis of acetate, propionate, or butyrate with IL‐17 in the intestinal extracts.
• J, K
  Environmental association analysis shows Bacteroides and Clostridium (K: circled in red) correlated significantly with SCFAs at the species level.

Data information: Data are represented as means ± SEM (n = 6 rats/samples per group). *P < 0.05, **P < 0.01, and ***P < 0.001, “#” indicates significant difference compared with the sham group, defined by one‐way ANOVA or Kruskal–Wallis test (>2 groups) followed by multiple comparisons with Bonferroni post hoc test. FMT, fecal microbiota transplantation; IL, interleukin; PCA, principal component analysis; RDA, redundancy analysis; SCFAs, short‐chain fatty acids.

Source data are available online for this figure.

The increased abundance of bacterial species, known as SCFA producers, motivated us to assess SCFA levels in feces and intestinal extracts by targeted mass spectrometry, which showed that FMT significantly promoted butyrate production (Fig 4G). Likewise, when addressing intestinal and systemic inflammation, FMT treatment significantly reduced the levels of IL‐17, a key pro‐inflammatory cytokine, both in the intestinal extract and serum of hypoganglionosis rats (Fig 4H). However, no significant change was observed in the levels of TNF‐α or the anti‐inflammatory cytokine IL‐10 (Fig 4H).

Subsequently, we studied the relationship between SCFAs and IL‐17, as well as SCFA levels and gut microbiota composition. Butyrate negatively correlated with IL‐17 levels in the gastrointestinal tract (Fig 4I). Correlation analysis using SCFAs as environmental factors showed that acetate and butyrate were the most closely correlated with intestinal microbiota composition at genus and species levels (Fig 4J), while Bacteroides and Clostridium were the most closely correlated with SCFAs before and after FMT (Fig 4K).

In summary, these data highlight that FMT might promote cytoprotection and support ENCC transplantation by ameliorating microbial dysbiosis, which increases SCFAs and suppresses IL‐17 production.

SCFAs and IL‐17 differentially affect the ENCC phenotype

Since the positive effect of FMT on microbial dysbiosis in hypoganglionosis was related to SCFAs and IL‐17, we next assessed the direct effect of these factors on ENCC phenotype.

We found that ENCC viability changed in a dose‐dependent manner (Fig 5A), suggesting a specific range of SCFA concentrations beneficial to the ENCCs. Then, we selected SCFA concentrations that yielded maximal ENCC viability, 2.0 mM for acetate, 0.125 mM for propionate, and 0.065 mM for butyrate (Fig 5A, denoted as m4), and these that caused the lowest ENCC viability, 30.0 mM for acetate, 2.0 mM for propionate, and 1.0 mM for butyrate (Fig 5A, denoted as m8). When addressing their impact on ENCC migration and apoptosis, higher migration for low (m4) concentrations and lower migration for high (m8) concentrations of SCFAs were observed (Fig 5B and C). In addition, low concentrations of SCFAs did not affect apoptosis, while high concentrations significantly induced ENCC apoptosis (Fig 5G and H).

Figure 5. SCFAs and IL‐17 differentially affect the ENCC phenotype.

Phenotype analysis of ENCCs treated with SCFAs.

• A
  ENCC viability assessed by CCK‐8 assay. Data are normalized to the NC group.
• B, C
  Migration of ENCCs evaluated using transwell plate assay and stain with crystal violet (B); Scale bars represent 200 μm. (C) Quantitative analysis of (B).
• D–F
  Analysis of different IL‐17 concentrations on ENCC viability (D), and 50 ng/ml IL‐17 on ENCC migration (E, F); scale bars represent 200 μm.
• G, H
  ENCC apoptosis determined by Annexin V‐FITC/PI stain and flow cytometry (G). (H). Quantification of (G).

Data information: Data are represented as means ± SEM (n = 6 biological replicates per group). *P < 0.05, **P < 0.01, and ***P < 0.001, “#” indicates significant difference compared with NC group, defined by two‐tailed Student's t‐test (2 groups) or one‐way ANOVA and Kruskal–Wallis test (>2 groups) followed by multiple comparisons with Bonferroni post hoc test. CCK‐8, cell counting kit‐8; ENCCs, enteric neural crest‐derived cells; IL, interleukin; NC, normal control; OD, optical density; PI, propidium iodide; SCFAs, short‐chain fatty acids.

Source data are available online for this figure.

Last, we focused on whether SCFAs could affect the ENCC phenotype by altering IL‐17 expression. IL‐17 significantly reduced ENCC proliferation and migration compared with the control group (Fig 5D and F). However, IL‐17 did not affect ENCC apoptosis (Fig 5G and H).

Altogether, these findings indicate that SCFAs may enhance and IL‐17 suppresses the neurogenic potential of ENCCs.

SCFAs synergize with ENCCs in inducing neurogenesis in the hypoganglionic gut

To further verify the positive effect of SCFAs on the neurogenic potential of ENCCs in vivo, GFP‐labeled ENCCs were transplanted into hypoganglionosis rats, followed by an enema with SCFAs (Fig 6A). We found that the ENCCs+SCFAs treatment resulted in a significantly shorter colonic transit time than only ENCC transplantation, SCFA enema, or no treatment (Fig 6B). Besides, more PGP9.5‐positive cells were observed in response to ENCCs+SCFAs treatment than ENCCs or SCFAs alone (Fig 6C and D). These results indicate the synergistic effect of the SCFAs and ENCC transplantation on neurogenesis in the hypoganglionic gut. In addition, there were more Tuj1⁺ and Tuj1⁺/GFP⁺ cells in the injection and adjacent sites and more Tuj1⁺/GFP⁻ cells in the injection site after ENCCs+SCFAs treatment compared with ENCCs or SCFAs alone, indicating that SCFA‐induced neurogenesis mainly relied on local ENCC transplantation (Fig 6E–G). These results suggest that future studies should focus on activating a larger range of residual neurogenesis.

Figure 6. SCFAs synergize with ENCCs in inducing neurogenesis in hypoganglionic gut.

• A
  Graphical protocol of ENCC transplantation and SCFA enema in the hypoganglionosis rat model.
• B
  Delay in distal colonic transit time.
• C, D
  Representative immunofluorescent images (C) and (D) number of PGP9.5⁺ cells. Nuclei were stained blue with DAPI; Triangles indicate the PGP9.5⁺ cell. Scale bars represent 200 μm.
• E
  Representative immunofluorescent images of Tuj1 and GFP localization at the injection and adjacent sides. Asterisks and triangles indicate the Tuj1⁺/GFP⁺ cells and Tuj1⁺/GFP⁻ cells, respectively. Scale bar, 200 μm.
• F, G
  Number of Tuj1⁺ and Tuj1⁺/GFP⁺ cells (F) and Tuj1⁺/GFP⁻ cells(G).

Data information: Data are represented as means ± SEM (n = 6 biological replicates per group) presented relative to values of the sham group. *P < 0.05, **P < 0.01, and ***P < 0.001. “#” Indicates significant difference compared with the sham group, defined by two‐tailed Student's t‐test (2 groups) or one‐way ANOVA and Kruskal–Wallis test (>2 groups) followed by multiple comparisons with Bonferroni post hoc test. FMT, fecal microbiota transplantation; ENCCs, enteric neural crest‐derived cells; GFP, green fluorescent protein; PGP9.5, protein gene product; SCFAs, Short‐chain fatty acids; Tuj1, neuron‐specific class III beta‐tubulin.

Source data are available online for this figure.

SCFAs modify neurogenesis via the MEK1/2 signaling pathway

To decipher the potential mechanism of SCFA‐dependent neurogenesis, we treated ENCCs with optimized concentrations of acetate (2.0 mM), propionate (0.125 mM), and butyrate (0.065 mM) (as determined in Fig 5) and examined gene expression using RNA‐seq. We identified 94 (35 up‐regulated and 59 down‐regulated), 140 (56 up‐regulated and 84 down‐regulated), and 684 (301 up‐regulated and 383 down‐regulated) differentially expressed genes (DEGs) in ENCCs treated with acetate, propionate, and butyrate, respectively (Fig 7A). Similarly, among 24 DEGs identified in ENCCs after IL‐17A intervention, 14 were up‐regulated, and 10 were down‐regulated (Fig 7A). ClusterProfiler packages were used for gene ontology/Kyoto Encyclopedia of Genes and Genomes enrichment analysis, which showed that these genes were significantly enriched in the MAPK and PI3K/AKT signaling pathways (Fig 7B). Moreover, some cytokine receptor interaction and signal transduction pathways were significantly enriched, suggesting they may be related to the anti‐inflammatory and neuroprotective functions of SCFAs (Fig 7B). Furthermore, Gene Set Enrichment Analysis, the interacting networks analysis of SCFAs target‐binding genes from Comparative Toxicogenomics Database and BindingDB Databases, also suggested that SCFAs modulate the ENCC phenotype through signal transduction, including interleukin, interferon, the complement, and MAPK signaling pathways (Fig 7C and D). However, these DEGs in ENCCs after IL‐17A intervention were not significantly enriched.

Figure 7. SCFAs modify neurogenesis via the MEK1/2 signaling pathway.

• A
  Volcano diagram shows differentially expressed genes (DEGs) in ENCCs after SCFAs and IL‐17A intervention, with up‐regulated and down‐regulated genes shown in yellow and blue, respectively (¦log fold change¦ ≥ 0.5 and adjusted P < 0.05).
• B
  Venn diagram shows 24 potential downstream pathways of SCFAs in ENCCs shared among acetate, propionate, and butyrate (left). Gene ontology/Kyoto Encyclopedia of Genes and Genomes enrichment analyses are detailed in the histogram, MAPK, PI3K‐Akt signaling pathways (indicated in red), as well as the cytokine signaling pathways (indicated in blue) are significantly enriched.
• C
  Venn diagram shows 19 shared downstream DEGs of acetate, propionate, and butyrate by Gene Set Enrichment Analysis, details in the histogram suggest that SCFAs are associated with cytokine signal transduction (indicated in blue).
• D
  Interaction networks of the potential target genes of acetate, propionate, and butyrate from the Comparative Toxicogenomic Database, and BindingDB databases present MAPK (circled in purple) and cytokine (circled in red) signaling pathways.
• E
  ENCC viability assessed by CCK‐8 assay after intervention with MEK1/2 inhibitor (U0126, 50 μM), p38 inhibitor (SB203580, 1 μM), JNK inhibitor (SP600125, 50 μM), and MER5 inhibitor (BIX 02189, 10 μM).
• F–H
  Diameter analyses of the ENCC‐derived neurospheres (F), migration of ENCCs evaluated using transwell plate assay, and stain with crystal violet on day 2 (G, H) show that MEK1/2 inhibitor U0126 suppressed SCFA‐induced ENCC proliferation and migration. Scale bars, 200 μm.

Data information: Data are represented as means ± SEM (n = 6 biological replicates per group). *P < 0.05, **P < 0.01 and ***P < 0.001, “#” indicates significant difference compared with NC group, defined by one‐way ANOVA or Kruskal–Wallis test (>2 groups) followed by multiple comparisons with Bonferroni post hoc test. AceNC, control to acetate; AceP, treated with acetate; ButNC, control to butyrate; ButP, treated with butyrate; CCK‐8, cell counting kit‐8; DEGs, differentially expressed genes; ENCCs, enteric neural crest‐derived cells; IL, interleukin; NC, normal control; NES, normal enrichment score; ProNC, control to propionate; ProP, treated with propionate; SCFAs, short‐chain fatty acids.

Source data are available online for this figure.

As MAPK signaling pathway was the most affected after treatments with SCFAs, we examined the influence of four MAPK downstream signaling molecules (MEK1/2, p38, JNK, and MEK5) on ENCC proliferation and migration. The optimized concentration of MEK1/2, p38, JNK, and MEK5 inhibitors (U0126, SB203580, SP600125, and BIX02198, respectively) were determined to be 50 μM (U0126), 1 μM (SB203580), 50 μM (SP600125), and 10 μM (BIX02198) (Appendix Fig S6). MEK1/2 inhibitor U0126 suppressed ENCC proliferation (Fig 7E and F) and migration (Fig 7G and H) induced by acetate, propionate, or butyrate treatment, which indicated that SCFAs might modulate the neurogenic potential of ENCCs via the MEK1/2 signaling pathway.

Discussion

HSCR is a neurocristopathy in children which requires surgical treatment while generating recurrent or life‐threatening complications (Heuckeroth, 2018; Pan et al, 2021). ENCCs can maintain their stemness to reestablish ENS and support bowel function, suggesting that ENCC transplantation might be a promising therapeutic strategy for HSCR patients (Fattahi et al, 2016; Workman et al, 2017; Yu et al, 2017). However, ENCC engraftment in hypoganglionic gut encounters interference from shifted microbial populations and intestinal inflammation. These obstacles may generate adverse effects on neurons, so they require further characterization before targeting in the clinical application (Venkataramana et al, 2015; Neuvonen et al, 2018; Yu et al, 2018). This study revealed that FMT reduced the harmful effects of hypoganglionic intestine on ENCC transplantation and improved intestinal function in a hypoganglionosis rat model. The potential mechanism of therapeutic function of FMT included the production of Bacteroides‐ and Clostridium‐derived SCFAs, which enhanced neurogenesis in vitro and in vivo. We hypothesized that FMT could improve cell replacement therapy for hypoganglionosis by promoting the production of SCFAs. Our results showed that the neurogenesis induction by SCFAs involved the MEK1/2 signaling pathway.

The bacterial species and metabolites in the hypoganglionic gut identified in this study were consistent with previous reports (Pierre et al, 2014; Li et al, 2016; Neuvonen et al, 2018). Abnormalities in the gut microbiome coincided with the decreased SCFAs and intestinal inflammation, which negatively influenced the success of ENCC transplantation. Accumulating evidence highlights the functional role of gut microbiota and its metabolites in ENS development, organization, and maturation, especially during the postnatal synchronized developmental window when initial colonization of gut microbiota occurs and ENCCs are developing. During this window, any perturbation, including intestinal inflammation and SCFA deficiency, could cause long‐lasting effects on the ENCCs (Hao et al, 2016; Yoo & Mazmanian, 2017; Yarandi et al, 2020). Gut microbiota is also important for physiological processes in adult animals (De Vadder et al, 2018; Yarandi et al, 2020). These results suggest that gut microbiota affects the adaptation of the ENS to its surrounding environment (Foong et al, 2020). Conversely, a functional ENS can modulate and maintain the balance and dynamics of pro‐inflammatory and anti‐inflammatory microbial communities to promote human health (Rolig et al, 2017). Based on these results and given the proximity of the intestinal microbiota to the ENS, these mutually beneficial interactions could influence ENCC transplantation methods and be considered when optimizing the cell therapy for hypoganglionosis. Moreover, interactions between intestinal microbiota and the ENS before, during, and after ENCC transplantation could create a better microenvironment for repairing ENS injuries (Ji et al, 2021). Indeed, the SCFA‐producing Bacteroides and Clostridium were missing in the hypoganglionosis model and were reestablished after FMT, which was consistent with improved ENCC transplantation, further indicating a positive role of FMT in ENCC transplantation in HSCR therapies (Morrison & Preston, 2016; Zhao et al, 2018).

Interestingly, our findings demonstrated that ENCC transplantation shortened colonic transit time by around 64%, while combined treatment with ENCC and FMT by around 77%. The effect is modest but significant. It may be because we used fecal microbiota that contains various microorganisms that produce various metabolites, and the concentration of the factors improving ENCC was limiting for the more robust effect. When using purified SFCAs to support ENCC transplantation, the difference between ENCCs and ENCCs+SCFAs was more pronounced. SCFAs enhanced the ENCC phenotype and improved ENCC transplantation success, indicating SCFA metabolism as a crucial element of the FMT effect on ENCC transplantation. SCFAs could be potential functional mediators in the intestinal, nervous, endocrine, and blood systems (Dalile et al, 2019; Mirzaei et al, 2021). A large body of evidence has shown that SCFAs are essential for maintaining intestinal homeostasis, including energy supply and electrolyte balance, anti‐inflammatory and antibacterial activity, epithelial barrier maintenance, and hormone signaling, which are all consistent with the neurogenic and neuroprotective properties of SCFAs (Dalile et al, 2019; Mirzaei et al, 2021). Our study has demonstrated that SCFAs promote neurogenic and neuroprotective effects via activation of the MEK1/2 signaling pathway. This work provides evidence that intestinal bacteria, their metabolites, and the downstream pathways have promising novel therapeutic potential for patients with HSCR (Cheng et al, 2021). Moreover, SCFAs were proposed to be involved in ENCC transplantation. The SCFA‐derivative 5‐hydroxytryptamine (5‐HT) and its receptor 5‐HT₄R are linked to neurogenesis and neuroprotection (Liu et al, 2009). SCFAs can also stimulate IL‐10 immunosuppression and maintain intestinal homeostasis (Sun et al, 2018). Multiple interactions among SCFAs, 5‐HT, IL‐10/IL‐17, and ENCCs, as well as the micro‐homeostasis of intestinal bacteria, metabolites, the host immune system, and ENS, require further investigation (Chen et al, 2019; Ahrends et al, 2021; Vicentini et al, 2021). Other beneficial strains with therapeutic effects and their metabolites from the microbiome should also be studied.

Using FMT to rebuild healthy microbiomes after dysbiosis has been a significant endeavor in treating infection, metabolic and immune dysfunction, cancer, and other diseases (El‐Salhy et al, 2020; Quraishi et al, 2020). In addition to reconstructing the microbiome and intestinal immune balance, FMT can regulate intestinal metabolite production, and promote neuronal and mucosal 5‐HT biosynthesis (Yano et al, 2015). We found that FMT could promote the microbial colonization of germ‐free mice and produce a normal microbiome, which modified the neuroanatomy of the impaired ENS in SCFAs‐dependent manner and increased intestinal motility. This study shows that FMT can enhance the HSCR intestinal microenvironment by increasing the abundance and diversity of microbiota, as well as their associated SCFAs. In addition, FMT also decreased inflammation in the gut. The role of inflammation in ENS regeneration is uncertain. The inflammation could negatively impact the ENCC phenotype (Sommer et al, 2018), had a positive effect (von Boyen et al, 2006), or exerted dual effects on ENCCs (Rühl et al, 2001). Previous publications show that inflammation impacted ENCCs in a highly complex way, with different inflammatory factors or cells producing diverse effects. Our study found that the pro‐inflammatory factor IL‐17 significantly inhibited ENCC proliferation and migration. Moreover, higher IL‐17 levels in hypoganglionosis rats were reduced after FMT. In summary, our study provides a new potential strategy for improving ENCC transplantation as a therapy for neurocristopathies, including HSCR, achalasia, slow‐transit constipation, and others, which proved to be related to ENS disorder (Niesler et al, 2021).

Our study has certain limitations. First, the hypoganglionosis intestinal microenvironment is a “black box” with a broad range of metabolites, toxins, or remnants from epithelial cells. Previous studies have shown the limited capacity of only ENCC transplantation to cure HSCR, which implies undefined negative factors (Pan et al, 2021). Although we did not address all the possible harmful elements, our work indicates that microbial dysbiosis and inflammation may be two factors negatively affecting the success of ENCC transplantation. Moreover, we propose further research on the key individual strains, specific metabolites, and the underlying molecular mechanisms. Second, the BAC‐induced model may have certain limitations compared to the genetic models of HSCR. However, it has been applied as one of the classical animal models and a transplantation platform for almost a decade worldwide (Burns et al, 2016). This is because it harbors similar symptoms and pathological features of hypoganglionosis as HSCR and can be established by a relatively simple process that limits the harm done to the animals. Besides, the fecal microbiota transplantation, ENCC transplantation, and short‐chain fatty acid treatment experiments (Figs 3G and 6A) lasted for at least 7 weeks, so the gene‐knockout models might be inapplicable. Furthermore, the gene‐knockout models, next to inducing abnormal ENS, often show dysfunction of the immune, urinary, reproductive, and digestive systems and have too short survival time (e.g., Hol^Tg/Tg mice die before postnatal day 28, and around 85% of Ednrb^s‐l/s‐l mice die before postnatal day 40; Soret et al, 2020) to study the long‐term effects of HSCR treatment. Thus, we used the BAC‐induced model in this study. In addition, the permeability or activity range of BAC is limited to seromuscular layers, we assumed that the “hypoganglionic intestinal extract” and microbiota disorder could directly result from the lower number of enteric neurons/ganglia but not BAC treatment, which needs further study (Fujiwara et al, 2019; Soret et al, 2020).

In conclusion, our study provides a new horizon that FMT promotes cell therapy for hypoganglionosis, which involves gut microbiota metabolite short‐chain fatty acid‐induced MEK1/2 signaling pathway.

Materials and Methods

Animal experiments

Experimental animals

Sprague–Dawley (SD) rats were obtained from the Experimental Animals Center of Xi'an Jiaotong University. Enteric neural crest‐derived cells (ENCCs) were harvested from pregnant SD rats on embryonic day 14.5. SD rats weighted approximately 250 g and had ad libitum access to food and water. Rats were fasted for 12 h before abdominal surgery to prepare the hypoganglionosis model, which is approximate to Hirschsprung disease (HSCR). After the operation, all rats were housed in polypropylene boxes under a 12‐h light/dark cycle. All the animal procedures were handled in accordance with the guidelines outlined by the Animal Care and Ethics Committee of Xi'an Jiaotong University (No. 2018‐2148).

Hypoganglionosis model and collection of an intestinal extract

BAC‐induced model of hypoganglionosis has been used, as it is one of the classical animal models, harbors similar symptoms and pathological features of hypoganglionosis as HSCR, and can be established by a relatively simple process that limits the harm done to the animals (Sato et al, 1978; Burns et al, 2016). Twenty‐four rats were randomly divided into sham and BAC‐treated (2, 4, and 6 weeks) groups, with six rats per group.

All procedures used to establish the model have been described previously (Sato et al, 1978; Yu et al, 2017). Briefly, after full anesthesia and a laparotomy, a 1 cm segment of the descending rectovesical colon, which was also defined as 1 cm at the distal colon behind fallopian tubes and about 5 cm from the anal orifice, was exteriorized and wrapped tightly with filter paper soaked thoroughly with 0.1% benzalkonium chloride (BAC) for 30 min; PBS was used for sham group. All rats were sacrificed by cervical dislocation at 2, 4, and 6 weeks after BAC treatment; then, a laparotomy was employed to exteriorize the BAC‐treated colons. Next, the colons were used to obtain the intestinal extracts. First, cleaning of the intestinal fecal material was performed for each colon. Next, samples were vortexed in 2 ml medium in an ice bath for 5 min, followed by centrifugation at 150 g for 10 min to collect the intestinal extract. Finally, all extracts were passed through a 0.22 μm sterilizing grade filter and stored at −80°C for later use in the experiments, where the effect of intestinal extracts on ENCC culture was studied to evaluate the influence of hypoganglionosis intestinal microenvironment on developmental characteristics of ENCCs. Meanwhile, the intestinal tissues were subjected to fixation in 4% formaldehyde and paraffin embedding for immunofluorescence staining. Moreover, fresh fecal pellets were collected directly from the anal orifices, immediately snap‐frozen in liquid nitrogen and stored at −80°C for later analyses. Last, the blood was collected in coagulation‐promoting tubes (#JN‐RYL‐042, JUNNUO, CHN) from the heart, and the serum was obtained after centrifugation at 1,200 g for 12 min and stored at −80°C.

ENCC culture and phenotype analysis

ENCC isolation and culture

ENCC isolation was performed under aseptic conditions as described in our previous study (Yu et al, 2017). Briefly, pregnant SD rats (embryonic day 14.5) were sacrificed by cervical dislocation. Then, a cesarean section was conducted to extract the entire fetal digestive tract under an anatomical microscope. After dissociation by trituration and trypsinization by incubation in 0.25% trypsin/EDTA at 37°C for 12 min, cells were centrifuged at 100 g for 5 min at room temperature. Next, cells were resuspended in sterile D‐PBS and passed through a 40 μm filter mesh to produce a single‐cell suspension of ENCCs.

The ENCCs were then cultured in Dulbecco's modified Eagle's medium supplemented with 0.5% N2, 1% B27, 10 ng/ml recombinant basic fibroblast growth factor, 10 ng/ml recombinant epidermal growth factor, and 100 U/ml each penicillin and streptomycin. The cultured ENCCs were seeded at 5 × 10⁵ cells/ml and maintained at 37°C in a humidified atmosphere with 5% CO₂. The ENCCs were purified by subculture several times and cultured for 5 days to form neurospheres. The ENCC‐derived neurospheres were harvested by centrifuging at 80 g for 2 min, then stained with mouse monoclonal antibody against nestin and rabbit monoclonal antibody against neurotrophin receptor P75 (p75^NTR) overnight at 4°C, followed by goat anti‐mouse IgG H&L conjugated with Cy3 and goat anti‐rabbit IgG H&L conjugated with Alexa Fluor 488 (the antibody information is provided in Appendix Table S1), subsequently identified by immunofluorescence microscopy (Appendix Fig S1B). The culture medium was changed every other day.

Proliferation analysis by measuring the diameter of neurospheres, cell counting, and CCK‐8 assay

To analyze cell proliferation, the ENCC single‐cell suspensions were reseeded at the cell density of 5 × 10⁵ cells/ml. Secondary neurospheres derived from the reseeded ENCCs were collected on day 5 and transferred to a 24‐well plate in 200 μl of medium with or without intervention (10% intestinal extract, SCFAs, IL‐17, or inhibitors of MAPK). After 2 days, images of neurospheres were taken in 10 random fields using an inverted microscope equipped with a DP70 digital camera and Manager software. The diameter of a neurosphere was measured as an average of its horizontal and vertical diameters. Each field was analyzed three times.

To assess ENCC viability, 90 μl of ENCC single‐cell suspensions dissociated from secondary neurospheres was seeded at the cell density of 1 × 10⁵ cells/ml in a 96‐well plate, with or without intervention (10% intestinal extract, SCFAs, IL‐17, or inhibitors of MAPK), at 37°C in a humidified atmosphere with 5% CO₂ for 2 days. Then, the cells were treated with 10 μl CCK‐8 solution for 3 h of culture, and the optical density (OD) was recorded at 450 nm using a microplate reader. The equivalent assessment of culture medium was used as the negative control. Cell counting was routinely performed by nuclear staining with 4,6‐diamidino‐2‐phenylindole (DAPI) and counting under a fluorescent microscope. All assays were performed in triplicate.

Transwell cell migration analysis

Cell migration was analyzed using Millicell Hanging Cell Culture Inserts in a 24‐well plate as described in our previous study (Yu et al, 2017). Neurospheres derived from ENCCs were collected and dissociated into single‐cell suspensions. Briefly, 1 × 10⁵ cells were suspended in 500 μl medium with or without intervention (10% intestinal extract, SCFAs, IL‐17, or inhibitors of MAPK) and added to the top chamber of the Millicell Hanging Cell Culture Insert, whereas the bottom chamber was filled with 750 μl medium containing 1% fetal calf serum as a chemo‐attractant. After 72 h of incubation, cells that had migrated through the membrane were fixed with methanol, stained with crystal violet, and counted under a microscope. Assays were performed in triplicate.

Apoptosis analysis by flow cytometry

ENCCs were collected as neurospheres and dissociated into single‐cell suspensions, as described above. ENCCs were seeded at 1 × 10⁵ cells/well in a 24‐well plate in 200 μl medium with or without intervention (10% intestinal extract, SCFAs, or IL‐17) for 5 days. The single ENCC suspensions were collected and passed through the cell strainer. Apoptosis was assessed by flow cytometry after staining with an Annexin V‐FITC/PI Apoptosis Detection Kit as described previously (Yu et al, 2017). Briefly, 1 × 10⁵ cells were suspended in 500 μl binding buffer and mixed with 5 μl Annexin‐FITC and 5 μl propidium iodide, followed by incubation for 10 min at room temperature. Apoptosis was measured within 1 h after staining with a NovoCyte^® flow cytometer and NovoExpress software (v1.4.1), according to the manufacturer's instructions.

Differentiation analysis by immunofluorescence staining

ENCC‐derived neurospheres were harvested and dissociated into single‐cell suspensions. The cells were replated at the cell density of 5 × 10⁵ cells/ml in a polylysine‐coated 24‐well plate in 200 μl 1% FBS‐containing medium with or without intervention (10% intestinal extract, SCFAs, IL‐17, or inhibitors of MAPK) and cultured at 37°C in a humidified atmosphere with 5% CO₂. After 3‐day culture, cells were fixed with 1% paraformaldehyde followed by permeabilization with 0.1% TritonX‐100 for 10 min and blocking with 10% fetal bovine serum (FBS) for 1 h at room temperature. Next, cells were stained with primary antibodies, mouse monoclonal antibody against beta‐tubulin III (Tuj 1), and rabbit monoclonal antibody to the glial fibrillary acidic protein (GFAP), followed by goat anti‐mouse IgG H&L conjugated with Cy3 and goat anti‐rabbit IgG H&L conjugated with Alexa Fluor 488 (the antibodies information is provided in Appendix Table S1). Nuclei were counterstained with DAPI. After acquiring images using a fluorescence microscope equipped with a DP70 digital camera, Tuj1‐, GFAP‐, and DAPI‐positive cells were counted in five random fields at 20× magnification. ENCC differentiation was scored by the ratio of Tuj1‐, GFAP‐, and DAPI‐positive cells.

ENCC transplantation procedures and related treatments

Transduction of ENCCs with LV‐eGFP and transplantation

The single‐cell suspensions of ENCCs, prepared as described above, were transduced with lentivirus‐expressing green fluorescent protein (GFP) at 100, 50, 25, and 12.5 times multiplicity of infection (MOI) (Yu et al, 2017). After 24 h, all medium was replaced with fresh medium, and the cells were seeded at 5.0 × 10⁵ cells/well in 24‐well plates at 37°C in a humidified atmosphere with 5% CO₂. After culturing for 7 days, transduced ENCCs were observed under a fluorescence microscope and analyzed with a NovoCyte^® flow cytometer and NovoExpress software version 1.4.1 to determine the infection efficiency. Next, the ENCCs transduced with the selected MOI of lentivirus were transplanted into hypoganglionosis rats with or without intervention (FMT and SCFAs). Transplantation was performed by subserosal injecting into the hypoganglionic intestine, defined as the BAC‐treated intestinal segment, at the position of 0, 3, 6, and 9 o'clock. On the 2^nd week after ENCC transplantation, all rats were sacrificed by cervical dislocation and the intestinal tissue, feces, and serum were harvested. All rats were randomly divided into subgroups, with six rats per group.

Broad‐spectrum antibiotic treatment

Antibiotic treatment was administered according to previously published protocols (Muller et al, 2020; Vicentini et al, 2021). Broad‐spectrum antibiotic mix, which contained 1 g/l ampicillin, 1 g/l neomycin, 0.5 g/l vancomycin, and 1 g/l metronidazole, was diluted in sterile water. Rats were given free access to the antibiotic solution for 14 days. The sterile water was used as the control. The solution was pH controlled (7.4–7.6) and replaced every 2 days. The bottles of antibiotics were swapped for regular autoclaved water on day 14.

Fecal microbiota transplantation

Fecal microbiota transplantation (FMT) was performed according to the published method (Josefsdottir et al, 2017; Lleal et al, 2019). Briefly, six pairs of littermate female rats that weighted approximately 250 g were selected as fecal donors, which stool contents were harvested and pooled, diluted 1:5 (W/V) in sterile PBS, then filtered by three‐layer gauze, and centrifuged at 100 g for 5 min, thereupon the supernatant was centrifuged at 6,500 g for 5 min. The sediment was retained and resuspended in sterile PBS with 20% glycerol at 0.5 g/ml. The fecal supernatants were stored at −80°C for further application. In the 2^nd week of modeling, the recipient hypoganglionosis rats were administered broad‐spectrum antibiotic treatment for 2 consecutive weeks, followed by an enema with donor fecal supernatants (1 ml) every 2 days for another 2 weeks. The sterile PBS was used as the control. Finally, the intestinal tissue, intestinal extracts, feces, and serum of all rats were harvested. All rats were randomly divided into subgroups, with six rats per group.

Short‐chain fatty acid mix supplementation

A mixture of SCFAs was administered according to the previous report (Liu et al, 2021). SCFA mixture, composed of acetate, propionate, and butyrate at a ratio of 3:1:1, was administrated by enema once a day at 500 mg/kg body weight for 2 weeks before ENCC transplantation and 1 week after that. The sterile PBS was used as the control. The solution was pH controlled (7.4–7.6) and changed weekly. All rats were randomly divided into subgroups, with six rats per group.

Analysis of gut microbiota

16S rDNA sequencing

To characterize the bacterial community taxonomically, we performed 16S rDNA amplicon sequencing. First, genomic DNA was extracted from feces using QIAamp Fast DNA Stool Mini Kit. Illumina compatible primers (338F, 5’‐ACTCCTACGGGAGGCAGCAG‐3′; 806R, 5’‐GGACTACHVGGGTWTCTAAT‐3′) were then used to amplify the 16S ribosomal DNA V3‐V4 regions by conventional PCR. The amplification products were subjected to gel purification prior to quantification by QuantiFluor‐ST. After reverse transcription, the cDNAs library was processed by fragmentation, end repair, and A‐tailing using the TruSeq^® DNA PCR‐free sample preparation kit. Then, the pooled library was run and sequenced on the NovaSeq600.

According to QIIM, the effective Tags were obtained after chimera picking and quality filtering. Clustering by Uparse v7.0.1001 was performed to generate a list of open reference operational taxonomic units (OTUs, with identity 97%). The taxonomic assignment was subsequently achieved with the SSUrRNA database. Additional alpha‐diversity and beta‐diversity analyses were conducted with QIIME. Finally, we performed a correlation analysis of the gut flora and host environmental factors (SCFAs) using the R (version 4.1.0) with psych and vegan packages. The function of microbiota genes was predicted by Tax4Fun.

Analysis of short‐chain fatty acids

Targeted mass spectrometry

A targeted mass spectrometry assay was developed to detect the SCFA levels in feces and intestinal extracts. Briefly, acetic, propionic, butyric, isobutyric, valeric, isovaleric, and hexanoic acid were purchased from Sigma Aldrich, as standard solutions with 10 concentration gradients (0.02, 0.1, 0.5, 1, 2, 5, 10, 25, 50, and 100 μg/ml). The sample was mixed with 50 μl 15% phosphoric acid, 10 μl 75 μg/ml isohexanoic acid (internal standard), and 140 μl diethyl ether, homogenated for 1 min, then centrifuged at 15,000 g at 4°C for 10 min. Next, the supernatant was collected for the test. Agilent HP‐Innowax column was used for split injection, and the sample volume was 1 μl, the split injection ratio was 10:1, the inlet temperature was 250°C, the ion source temperature was 230°C, the transmission line temperature 250°C, and the quadrupole temperature 150°C. The programmed temperature raised from 90 to 120°C at 10°C/min, next to 150°C at 5°C/min, and finally, to 250°C at 25°C/min for 2 min. Helium carrier gas flow rate was 1.0 ml/min. Mass spectrometry conditions included electron bombardment ionization (EI) source, SIM scanning mode, and electron energy 70 eV. The samples and standard solutions of SCFAs were measured by MS. The area under the curve for standard SCFA solution normalized to the internal standard was used as a single point on that SCFA standard curve. A standard curve for each SCFA was established based on the concentration gradient and served to calculate the concentration of this SCFA in the samples. All the samples were measured repeatedly, at least six times.

Analysis of inflammatory cytokines

Enzyme‐linked immunosorbent assay (ELISA)

The levels of IL‐17, IL‐10, and TNFα in the serum and intestinal extracts were analyzed using ELISA kits [IL‐17 (Rat IL‐17A ELISA Kit, #ab214028, Abcam, USA), IL‐10 (Rat IL‐10 ELISA Kit, #ab100746, Abcam, USA), and TNFα (Rat TNF alpha ELISA Kit, #ab236712, Abcam, USA)] according to the manufacturer's instructions. The results were acquired at 450 nm using a microplate reader.

Analysis of colon tissue

Pathological analysis

Pathological analysis was performed to evaluate the efficacy of ENCCs, FMT, and SCFAs in treating disabled enteric nervous system (ENS) in hypoganglionosis rats. Twenty‐five 5‐μm‐thick transverse sections of intestinal tissues were randomly selected for hematoxylin and eosin staining and immunofluorescence analyses. The sections were blocked with 10% goat serum (#16210064, Thermo Fisher, USA) and incubated with rabbit monoclonal antibody against PGP9.5, mouse monoclonal antibody against Tuj1, and rabbit monoclonal antibody against GFP overnight at 4°C, followed by goat anti‐mouse IgG H&L conjugated with Cy3 and goat anti‐rabbit IgG H&L conjugated with Alexa Fluor 488. Besides, C‐reactive protein antibody and TNFα antibody were used to investigate the inflammation in the hypoganglionic intestine (the antibody information is provided in Appendix Table S1). Nuclei were counterstained with DAPI. PGP9.5‐, Tuj1‐, GFP‐, and DAPI‐positive cells were observed by fluorescence microscopy (Olympus, BX51) under a DP70 digital camera and quantitatively analyzed by ImageJ software. The ganglion was defined as the sites with high expression of Tuj1 and PGP9.5 in the intramuscular or submucosal region (Appendix Fig S2; Kameda, 2005; Yu et al, 2017).

Histological scoring

Histological scoring was performed according to the previous report (Rakoff‐Nahoum et al, 2004). Colon tissue was fixed with 10% neutral formalin, paraffin embedded, sectioned at 3–6 μm, and stained with hematoxylin and eosin. Sections were analyzed in a blinded manner by a trained gastrointestinal pathologist. Inflammatory infiltrate was scored according to the extent and character of infiltrate. The infiltrating leukocyte extent score equals the involved area plus the severity score per each layer of the intestine mucosal, submucosal, and muscularis propria. Epithelial injury score equals % area of section plus mucous membrane ulcer/erosion plus intestinal crypt lesion. Histopathological changes of ulcer/erosion were scored on a scale of 0–4 (where 0 = none; 1 = mild; 2 = moderate; and 3 = severe) for each parameter. Area of intestinal crypt lesion was scored as follows: 0 = no involvement; 1 ≤ 10% of section; 2 ≤ 40%; 3 ≤ 70%; and 4 ≤ 100%. The scores for inflammatory infiltrate and epithelial injury were determined during the histological scoring of colons. A score was determined for each group, with at least 6–8 rats per group.

Analysis of colon function

Distal colonic transit time

We used the bead expulsion test to assess distal colonic propulsion according to the previous report (Kishi et al, 2020). Rats were lightly anesthetized with 10% chloral hydrate. Then, a 4 mm spherical plastic bead, cleaned and disinfected beforehand, coated in paraffin, was gently inserted 6–7 cm into the distal colon using a silicone pusher. Rats were placed in individual, bedding‐free cages, and latency time for bead expulsion was recorded. Each rat was tested at least three times at 2 h intervals, and the mean was considered the time for bead expulsion.

Analysis of genes and pathways

Transcriptome‐wide mRNA sequencing

We performed transcriptome‐wide mRNA sequencing to detect significant differentially expressed genes in ENCCs treated with SCFAs and IL‐17. Total RNA was extracted from ENCCs with TRIzol reagent (Invitrogen Corporation, USA) according to the manufacturer's protocols. RNA concentration was determined spectrophotometrically using NanoPhotometer^® (IMPLEN, USA), and the purity was checked by 260/280 ratio. Then, mRNA was purified by magnetic beads with Oligo (dt) and converted to a cDNA library with the NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (#E7775, New England Biolabs, UK). The cDNA library was determined by Agilent 2,100 (Agilent Technologies, USA) before RNA sequencing was performed on a NextSeq (Illumina, USA).

Subsequently, the Perl script was used to filter raw data, and clean data were aligned to the reference genome GRCh38 from the ENSEMBL database (http://www.ensembl.org/index.html) assembly using HISAT2 v2.1.0 with default parameters. Reads count for each gene was done by HTSeq v0.6.0. TPM (Transcripts Per Kilobase of exon model per Million mapped reads) was calculated, the expression of genes was quantified as log2(TPM + 1), and genes were filtered for log2(TPM + 1) > 2.5. DESeq2 v1.6.3 was designed for differential gene expression analysis, and the P‐value was corrected by the Benjamini–Hochberg procedure. Then, genes with adjusted P‐value ≤0.05 and ¦ log2‐fold change ¦ ≥ 0.5 were identified as differentially expressed genes (DEGs). Finally, several approaches to enrichment, including Gene Ontology (GO, http://geneontology.org/), KEGG (Kyoto Encyclopedia of Genes and Genomes, http://www.kegg.jp/) enrichment, and Gene Set Enrichment Analysis (GSEA), were developed to uncover DEG function annotations.

Target genes of SCFAs

SCFA target‐binding information was collected from the Comparative Toxicogenomics Database (CTD, https://ctdbase.org/; Davis et al, 2021) and the BindingDB databases (https://www.bindingdb.org/bind/info.jsp; Gilson et al, 2016). The target‐binding genes were visualized by Cytoscape 3.8.2 to generate interacting networks.

Statistics

Statistical analysis was performed by Graphpad Prism 8.0 (GraphPad Software, USA). Normally distributed data were presented as means ± standard error of the mean (SEM), and non‐normally distributed data were presented as medians (interquartile range). Depending on the distribution and variances, two‐tailed Student's t‐test (two groups), one‐way ANOVA, and Kruskal–Wallis test (> 2 groups) followed by multiple comparisons with Bonferroni post hoc test were applied to compare differences between groups. Significance was set at P < 0.05. All figure panels represent data independently repeated at least three times, yielding similar results.

Author contributions

Donghao Tian: Data curation; software; methodology; writing—original draft. Wenyao Xu: Data curation; methodology. Weikang Pan: Investigation. Baijun Zheng: Investigation. Weili Yang: Methodology. Wanying Jia: Methodology. Yong Liu: Supervision; visualization. Malgorzata A. Garstka: Formal analysis; writing—original draft; writing—review and editing. Ya Gao: Supervision; funding acquisition; validation; investigation; visualization; project administration. Hui Yu: Conceptualization; resources; data curation; formal analysis; supervision; funding acquisition; validation; investigation; visualization; writing—original draft; project administration; writing—review and editing.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

Source Data for Figure 6

Source Data for Figure 7

The EMBO Journal (2023) 41: e111139

Data availability

All the data presented are available in this study and the 16S rDNA sequencing RawData in Fig 2 were deposited in external repositories, which can be accessed online at https://doi.org/10.6084/m9.figshare.21378693.v1.