【研究成果】Preconception maternal gut dysbiosis affects enteric nervous system development and disease susceptibility in offspring via the GPR41–GDNF/RET/SOX10 signaling pathway

Maternal preconception antibiotic exposure leads to ENS dysplasia in juvenile offspring

To explore how the preconception of maternal gut microbiota dysbiosis might affect ENS development in offspring, we administered antibiotics (ABX) to female mice. The mice were subjected to ABX treatment for 1 week, and the mice were mated with male mice in a 1:1 ratio during the subsequent week. Then, they were access to food and water ad libitum until the birth of their pups (Figure 1A). Compared with the control (CON) group, the ABX group exhibited no significant differences in pregnancy rate per cage, number of pups per dam, maternal weight during pregnancy, and weight of offspring, suggesting that the ABX regimen did not compromise dam fertility and nutritional status of offspring (Figure 1B, Figure S2A).

Lactation is pivotal for mammalian growth and development; therefore, we focused our analysis on the colonic MP of 3-week-old pups, including both sexes. Our observations revealed a noticeable difference in the proximal colonic MP of the ABX group, which appeared to be more dispersed with HuC/D+ staining in juvenile offspring (Figure 1C). We proceeded to count the HuC/D+ neurons, S100β+ EGCs, and two key neuronal subpopulations (neuronal nitric oxide synthase (nNOS)+ and choline acetyltransferase (ChAT)+) within each ganglion. Remarkably, we noted a reduction in HuC/D+ neurons and S100β+ EGCs in the colonic MP ganglia of juvenile mice from the ABX group (Figure 1C), along with a decrease in the proportion of ChAT+ neurons. However, the proportion of nNOS+ neurons was unaffected (Figure 1D). Notably, the colonic MP of ABX pups showed a marked increase in nestin+ cells, suggesting a heightened response to ENS damage (Figure 1D). It did not reveal any significant impact of sex on the results (Figure S2B).

Considering the developmental disturbances observed in colonic enteric neurons and their subtypes, we employed an ex vivo colonic muscle strip perfusion system to assess motor function impairments in ABX pups. Our findings showed that juvenile mice in the ABX group exhibited impaired cholinergic neurotransmission indicated by attenuated colonic muscle strip contractions in response to carbachol (p = 0.07) and diminished muscle strip relaxation upon the administration of atropine. l-NAME and l-arginine were used to assess nitric acid neurotransmission. The data demonstrated no notable differences in either l-NAME-induced contractions or l-arginine-induced relaxation between juvenile mice in the CON and ABX groups (Figure 1E). These results suggest a more pronounced impairment of the cholinergic pathway responsible for colonic contractions, potentially causing weakened colonic contractions and slowed transit. Likewise, sex was found to have no significant impact on the outcomes (Figure S2C).

Furthermore, we examined the effects of preconception ABX treatment on intestinal mucosal development in the 3-week-old offspring. Histological staining of the ileum and colon tissues showed that ABX pups had reduced villus length in the ileum, decreased crypt depth in the colon, and fewer goblet cells in the colonic crypt (Figure 1F). No significant disparities were noted among offspring of different sexes (Figure S2D). These findings indicate that preconception maternal ABX treatment not only led to aberrant ENS development and motor function in juvenile offspring but also impacted intestinal mucosal development.

Offspring of dams treated with antibiotics exhibit persistent ENS dysplasia and increased disease susceptibility

To explore whether the developmental abnormalities of the ENS observed in juvenile offspring persisted into adulthood, we examined the colonic MP of adult offspring of dams treated with antibiotics during the preconception period. Given the known association between ENS abnormalities and gastrointestinal function disorders, we employed WAS to evaluate the susceptibility of these offspring to gut–brain interaction disorders (Figure 2A). To eliminate potential biases introduced by female sex hormones in the stress models, we limited the WAS study to male mice. We found that the number of HuC/D+ neurons and S100β+ EGCs remained reduced in the adult offspring of the ABX group, while the proportion of ChAT+ and nNOS+ neurons no longer showed significant differences. Exposure to WAS did not alter the counts of HuC/D+ neurons and S100β+ EGCs; however, it did induce an increase in the proportion of ChAT+ neurons in both the CON and ABX groups (Figure 2B). Furthermore, we found a significant increase in the mRNA expression of Tgf-β2 in the colon of adult offspring in the ABX–WAS group compared to the CON–WAS group (Figure 2C), suggesting a possible mechanism for the more severe ENS alterations observed. Ex vivo colon motor experiments revealed no significant differences in the cholinergic neural pathway; however, the NOergic neural pathway was impaired in the ABX offspring. Following WAS, the ABX offspring exhibited a more pronounced cholinergic response and a weaker NOergic response than the CON group (Figure 2D). These findings suggest that preconception ABX-induced ENS dysplasia in the offspring may underlie the disrupted colon motility observed in response to WAS. Furthermore, colonic transit was slower in ABX offspring than in CON offspring, whereas the small intestine transit and fecal water content showed no significant differences (Figure 2E,F). Following WAS, both adult offspring displayed a reduction in colonic transit, with a greater effect observed in the ABX offspring (Figure 2E,F). These results confirm the presence of persistent colonic ENS abnormalities in the offspring of ABX dams, which contribute to weakened colonic contractions and a significant acceleration in response to stress.

The ENS is closely associated with gastrointestinal disorders. Based on our observations of ENS dysplasia and abnormal intestinal mucosa in the juvenile offspring of ABX dams, we further investigated visceral sensitivity, colonic barrier function, and colonic immunity in adult offspring. The application of WAS significantly increased colorectal distension–electromyography (CRD–EMG) activity in adult offspring, with a more pronounced response observed in the ABX group (Figures S3A,B, S4A,B). Offspring from the ABX group had shorter microvilli, wider tight junctions (TJs) between colonic epithelial cells, shorter TJ lengths, and a lower density of electron-dense material (Figure S3C, Figure S4C,D). WAS resulted in the shortening of the microvilli and the widening of TJs in colonic epithelial cells. Notably, the ABX–WAS offspring displayed even shorter TJ lengths and decreased areas of electron-dense material, along with indications of bridge granule fragmentation and microvillus deterioration (Figure S3C, Figure S4C,D). Further analysis revealed a significant downregulation of occludin mRNA expression in the ABX–sham group compared to the CON–sham group, confirming the presence of abnormal colonic barrier function in the adult offspring of ABX dams (Figure S3D). An impaired gut barrier serves as a crucial structural basis for inflammation. The ABX–WAS offspring exhibited significantly greater changes in mRNA expression of inflammatory markers, including Tnf-α and Csf-1, compared to the CON–WAS offspring (Figure S3E). Mast cells contribute to visceral hypersensitivity. Although the number of colonic mucosal mast cells was not significantly different between the CON and ABX offspring, the ABX offspring displayed a more pronounced response to WAS (Figure S3F). Furthermore, we analyzed the expression of inflammatory factors in the colons of the adult offspring; TNF-α and IL-17A presented significant differences between the ABX and CON groups (Figure S3G). These results are consistent with those of previous studies on irritable bowel syndrome (IBS) and its associated mechanisms [20, 21]. Based on these findings, we propose that the adult offspring of ABX dams exemplify the underdeveloped physiological foundations of the colonic ENS and colonic mucosal barrier. When exposed to WAS, they exhibit a striking ENS response, barrier disruption, and inflammation within the colon, ultimately leading to susceptibility and more severe abnormalities in intestinal motor, sensory, and barrier functions.

ENS dysplasia emerges during embryonic development in the offspring of dams treated with antibiotics via the RET/GDNF/SOX10 signaling pathway

Given the profound and enduring developmental abnormalities observed in the ENS and intestinal mucosa of the offspring of ABX dams, we postulated that these aberrations might have originated during the embryonic stage, which is a crucial period for growth and development. Therefore, we procured intestinal tissues from embryonic day (E) 13.5 and E18.5 from the CON and ABX groups to assess the migration and proliferation of the ENS (Figure 3A). E13.5 signifies the time point at which the ENS migrates from the foregut to the anus, while E18.5 marks the final day of embryonic development for the ENS. As expected, the E13.5 embryos from the ABX group exhibited a notably shorter intestine length from the pylorus to the anus, although the colon width remained unaffected (Figure 3B). Additionally, the migration distance percentage of Tuj1+ nerve fibers in the embryos of the E13.5 embryos from the ABX group was significantly lower. When the intestine was further segmented into 10 equal parts, a discernible decrease in the density of Tuj1+ staining was observed in the ABX group, which was particularly significant in the colon (Figure 3B). Subsequently, Tuj1+ and SOX10+ staining was performed on the colon tissue at E18.5, revealing a substantial reduction in the Tuj1+ staining area and SOX10+ cell count in the ABX group (Figure 3C), indicating the compromised proliferation efficiency of the ENS.

To delve deeper into the potential abnormalities in gene expression and the underlying mechanisms associated with colonic ENS development in embryos of the ABX group, we conducted a transcriptome analysis of colon tissues derived from E18.5. Our findings revealed 567 differentially expressed genes in the ABX group, of which 154 were upregulated, and 413 were downregulated (Figure 3D, Figure S5A). Several genes known for their close association with ENS development, such as Ntrk2, Gfra3, and Dcc, were significantly downregulated in the ABX group (Figure 3D). The Gene Ontology (GO) enrichment analysis revealed that the upregulated genes were linked to developmental processes in the ABX group (Figure S5B). Interestingly, both “Digestive system development” and “Nervous system development” were among the enriched pathways (Figure 3E). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis ed the significant enrichment of the “Wnt signaling pathway” (Figure 3F), a crucial signaling pathway for hindgut development, corroborating the observed colon developmental abnormalities in the ABX offspring. Further gene set enrichment analysis (GSEA) demonstrated significant downregulation of multiple pathways enriched in both the GO and KEGG databases, encompassing “Nervous system development,” “Muscle tissue development,” and the “Wnt signaling pathway” in the ABX embryos (Figure 3G, Figure S5C–F).

The proliferation, migration, and differentiation of the ENS predominantly involve the RET/GDNF and EDN3/EDNRB signaling pathways, with SOX10 playing a pivotal role as a transcription factor in both. Analysis of the mRNA expression of these key molecules in the E18.5 embryonic colon revealed the downregulation of Ret, Gdnf, and Sox10 in the ABX group (Figure 3H). Interestingly, the mRNA expression of Ednrb was significantly upregulated in the colons of ABX fetuses (Figure 3H), suggesting a compensatory adaptation to counteract the downregulation of Ret/Gdnf signaling. These observations suggest that preconception ABX treatment can disrupt normal embryonic colon ENS development, primarily through signaling pathways involving RET/GDNF/SOX10.

Preconception antibiotic exposure profoundly influences the maternal gut microbiome and metabolome

To delineate the mechanisms leading to abnormal ENS development in fetal mice in the ABX group, we conducted a comprehensive assessment of the maternal gut microbiota and its metabolites. Fecal samples were collected from dams at E18.5 for metagenomic analysis. Our findings revealed a notable decrease in alpha diversity in the microbiota of preconception ABX dams, as indicated by the Shannon and Simpson indices (ABX–DAM), compared to that in the control group (CON–DAM) (Figure 4A, Figure S6A). Principal coordinate analysis (PCoA) based on the Bray–Curtis distance further confirmed the distinct clustering between the ABX–DAM and CON–DAM groups (Figure 4B, Figure S6B). Specifically, at the phylum level, significant differences were observed in the abundances of Firmicutes and Bacteroidetes between the two groups, resulting in a significantly decreased F/B ratio in the ABX–DAM group (Figure S6C,D). Linear discriminant analysis effect size (LEfSe) analysis identified Lactobacillus and Paramuribaculum as key genera distinguishing the CON–DAM group, whereas Bacteroides and Parabacteroides had the most significant impact on the ABX–DAM group (Figure S6E–G). The ABX–DAM group harbored a greater number of unique species, including potentially harmful pathogens, such as Escherichia coli and Klebsiella oxytoca (Figure 4C,D). Among the common species, many beneficial bacteria were depleted in the ABX–DAM group, including Lactobacillus intestinalis, Limosilactobacillus reuteri (L. reuteri), and Muribaculum gordoncarteri (M. gordoncarteri) (Figure 4E,F).

To gain deeper insight into the interrelationships within the maternal gut microbiota, we employed a co-occurrence network analysis. Despite the decreased alpha diversity in the fecal samples from the ABX–DAM group, their internal microbial interactions were more complex. Within the predominant species, interactions in the ABX–DAM group remained intricate and were dominated by negative correlations (Figure 4G). This underscores the competitive relationships and nuanced regulatory mechanisms involved in microbial reconstitution in the ABX–DAM group. Moreover, our analysis revealed the depletion of functional pathways involved in amino acid and nucleotide synthesis in the ABX–DAM group, whereas those pentose phosphate pathways were enriched (Figure S7A–D). These findings suggest that preconception ABX treatment significantly alters the maternal gut microbiome and metabolome, potentially contributing to the observed abnormalities in ENS development in the offspring.

Due to variations in the abundance of bacteria capable of producing short-chain fatty acids (SCFAs) between the CON–DAM and ABX–DAM groups, we evaluated SCFA levels in the cecal contents and serum of maternal mice at E18.5. In the ABX–DAM group, the concentrations of acetate, propionate, butyrate, and valerate in the cecal contents were notably lower than those in the CON–DAM group, with propionate and valerate showing particularly large fold changes. However, no appreciable differences were observed in the serum levels of these SCFAs (Figure 4H,I; Figure S8A,B). To investigate whether there were metabolite differences beyond SCFAs, we employed a targeted metabolome approach (Figure S8C, Table S3) to further analyze the cecal contents and serum of maternal mice at E18.5, which revealed a distinct separation between the groups based on orthogonal partial least squares discriminant analysis (OPLS-DA) (Figure S8D). Using volcano and bar plots, we identified key metabolites in the cecal contents, including indoleacetic acid and N6-methyladenosine, both of which were consistently downregulated in the ABX–DAM group and strongly correlated with each other (Figure 4J,K, Figure S8E). Additionally, KEGG pathway enrichment analysis demonstrated distinct metabolic capabilities between the two groups (Figure S8F). Analysis of the serum metabolome also showed a clear separation between the groups, with tighter intragroup clustering compared to that of the cecal contents (Figure S8G). Volcano and bar plots ed corticosterone and cortexolone as the two compounds with the largest downregulated fold-changes in the ABX–DAM group, exhibiting a robust correlation (Figure 4L,M, Figure S8H). Phenol and caffeine concentrations were markedly upregulated in the ABX–DAM group. KEGG pathway enrichment analysis indicated several significantly downregulated metabolic pathways in the ABX–DAM group (Figure S8I). Although there was minimal overlap in the differential metabolites between the cecal contents and serum, the trends in concentration changes between the two were consistent (Figure 4L). Collectively, our findings the substantial modifications in the gut microbiota and its metabolites in the ABX group, which may contribute to the emergence of ENS abnormalities.

Multi-omics analysis revealed that L. reuteri, M. gordoncarteri, and propionate significantly influence ENS development

Using a multi-omics approach, we investigated the key microbial species and metabolites that are associated with the development of the ENS. Initially, we employed Spearman's rank correlation test to identify components with high connectivity and top rankings within the interaction network between species and metabolites, considering these to be of significant importance due to their central position and close interactions. We designate these as core differential species and metabolites with tightly associated relationships. Specifically, M. gordoncarteri, L. reuteri, Neglectibacter sp. X4, and Anaerotruncus sp. 1XD42-93 were identified as core species (Figure 5A,B). Additionally, hyocholic acid, uridine, N6-methyladenosine, 1-methyladenosine, propionate, and valerate in the cecal contents, as well as corticosterone and cortexolone in the serum, were identified as core metabolites (Figure 5C).

Scatter plots illustrate the relationship between propionate, M. gordoncarteri and L. reuteri (Figure 5D,E). Furthermore, our analysis uncovered PWY-6527, PWY6317, 1CMET2-PWY, and PWY7977 as the core metabolic pathways (Figure S9A–E). Notably, L. reuteri and other Lactobacillus species significantly contributed to the depleted pathways in the antibiotic-treated dams, whereas opportunistic pathogens, such as Escherichia coli, contributed more to the enriched pathways, ing the distinct microbial profiles between the two groups (Figure S9A–E).

To identify the crucial metabolites and core bacterial species influencing embryonic ENS development, we correlated the differentially abundant maternal bacteria, cecal metabolites, and mRNA expression levels of five crucial ENS developmental genes in the embryonic colon (previously examined in Figure 3H). The RET/GDNF and SOX10 pathways exhibited stronger correlations than the EDN3/EDNRB pathway (Figure 5F). Significant positive correlations were observed between M. gordoncarteri, L. reuteri, propionate, valerate, with Ret, Gdnf, Sox10, and Edn3 expression (Figure 5F). Given the lower concentration of valerate than propionate, we identified propionate in the cecal content as a key metabolite. Additionally, serum corticosterone and cortexolone levels were positively correlated with the mRNA expression of Ret and Gdnf in the embryonic colon (Figure S10A). Analysis of metabolite receptors in the fetal colon revealed significant downregulation or decreasing trends in the expression of Tgr5, Tlr2, Nr3c1, and Nr3c2 in the ABX group, with the exception of Tlr4 (Figure 5I, Figure S10B). Exploring the mechanisms by which maternal metabolites exert their effects, we found a consistent positive correlation between metabolite levels and the mRNA expression of corresponding receptors in the embryonic colon. Propionate and valerate showed the strongest correlation with Gpr41 (Figure 5J and Figure S10C). The relationship between Gpr41 expression and Ret, Gdnf, and Sox10 expression was particularly pronounced (Figure 5K, Figure S10D), suggesting the critical role of GPR41 in embryonic ENS development. Based on these findings, we propose that the core bacterial species M. gordoncarteri and L. reuteri influence propionate production and facilitate ENS development via GPR41, which is impeded by ABX treatment.

L. reuteri plays a crucial role in ENS development via the production of propionate

Considering that M. gordoncarteri is exclusively observed in rodents and may exhibit pathogenicity, we focused on L. reuteri because of its beneficial metabolites, such as SCFAs and tryptophan derivatives, and its profound effect on microbial community structure. To identify probiotics that are beneficial for ENS development and have potential clinical applications, we chose L. reuteri as the intervention throughout gestation. This was performed to mitigate any adverse effects of preconception ABX treatment on embryonic ENS development as early as possible (Figure 6A). Preconception ABX treatment influenced colon length, and the Tuj1+ fiber migration percentage and density at E13.5. However, gestational supplementation with L. reuteri had a remarkable ameliorative effect on ENS migration and density (Figure 6B). Subsequent staining of E18.5-mouse embryo colons for Tuj1+ fibers and Sox10+ cells revealed that L. reuteri effectively rescued the Tuj1+ staining area affected by preconception ABX treatment, although its impact on Sox10+ cell counts was less pronounced (Figure 6C). mRNA expression analysis of embryonic colons demonstrated that L. reuteri supplementation restored the downregulation of the RET/GDNF pathway induced by ABX preconception treatment (Figure 6D).

We analyzed SCFA levels in the maternal cecal contents and serum during the late gestational period. Significantly higher levels of acetate and propionate were observed in the ABXreuteri group, whereas butyrate and valerate levels exhibited less significant changes (Figure 6F, Figure S11A). Serum levels of SCFAs showed an upward trend (Figure S11B). In vitro cultivation experiments have also demonstrated that L. reuteri is capable of producing acetate and propionate (Figure S11C). Correlation analysis of the mRNA expression of Ret, Gdnf, and Sox10 revealed significant correlations with propionate and valerate levels, surpassing those with acetate and butyrate levels (Figure 6G, Figure S11D). Notably, embryonic colon Gpr41 mRNA expression showed a strong correlation with maternal cecum propionate and demonstrated a significant recovery in the ABXreuteri group (Figure 6H, Figure S11E). Additionally, treatment with L. reuteri rescued the expression of Tgr5 (Figure S11F). Furthermore, embryonic colon Gpr41 exhibited significant positive correlations with Ret, Gdnf, and Sox10 mRNA expression, emphasizing the vital role of propionate in embryonic ENS development (Figure S11G).

To further validate the crucial role of propionate in ENS development, we conducted gestational intervention with propionate in maternal mice (Figure 7A). After the administration of sodium propionate, a significant increase in propionate levels was observed in the cecum of the dams, which was 16.45 times higher than in dams supplemented with L. reuteri (Figure S12A). We observed that supplementation with propionate restored colon shortening as well as the migration and proliferation of the ENS, which were affected by preconception antibiotic exposure at E13.5 (Figure 7B). Subsequent staining of Tuj1+ fibers and Sox10+ cells in the colon at E18.5 revealed that propionate significantly rescued the Tuj1+ staining area and that Sox10+ cell counts were affected by preconception antibiotic exposure (Figure 7C). The mRNA expression analysis of embryonic colons showed that propionate supplementation restored the ABX-induced downregulation of Ret, Gdnf, and Sox10 (Figure 7D). Preconception antibiotic exposure led to the downregulation of the mRNA expression levels of SCFA receptors in maternal mice, and propionate intervention restored the expression of Gpr41 and Gpr43 (Figure 7E). Furthermore, the mRNA levels of Gpr41 exhibited a significant positive correlation with the mRNA expression of Ret, Gdnf, and Sox10, suggesting that propionate regulates the expression of ENS development-related genes through the GPR41 pathway (Figure 7F). In contrast, valerate had a relatively weak effect on ENS development (Figure S12B–E).

Collectively, our findings suggest that supplementation with L. reuteri after ABX treatment during pregnancy can increase cecal propionate levels. This modulated the expression of Ret, Gdnf, and Sox10 through the GPR41 signaling pathway, promoting the amelioration of ENS dysplasia induced by preconception ABX treatment.

Animals

Female (7 weeks old) and male (9 weeks old) C57BL/6J mice were used from the Department of Laboratory Animal Science, Peking University Health Science Center, Beijing, China, and reared as specific pathogen-free (SPF) mice. The animals were housed in SPF facilities with a 12-h light–dark cycle (lights on from 6:00 AM to 6:00 PM). Upon entering the SPF facility, all animals underwent a 1-week acclimation period before proceeding to the next stage of the experiment. Unless otherwise specified, all the mice had ad libitum access to sterilized food (Xietong Shengwu) and water.

For the experimental investigation of the impact of gut microbiota dysbiosis on the development of ENS in juvenile and adult offspring, the ABX group of female mice received a daily oral gavage of a mixture containing neomycin (Innochem) (100 mg/kg), metronidazole (Aladdin) (100 mg/kg), and vancomycin (Innochem) (50 mg/kg) in the morning and evening, along with ad libitum access to ampicillin (Aladdin) (1 mg/mL) in the drinking water, to simulate the preconception gut microbiota dysbiosis state. The CON group received an oral gavage of the vehicle (sterile water). Oral gavage was continued for 1 week, and the mice were mated with male mice in a 1:1 ratio during the subsequent week. The appearance of the vaginal plug was designated as E0.5. After the vaginal plug was observed, the pregnant mice were transferred to new SPF cages and provided ad libitum access to food and water. Body weight was measured weekly. After parturition, the offspring were raised on SPF bedding, and their body weights were recorded weekly with ad libitum access to food and water. A subset of the offspring (unspecified sex) was killed at 3 weeks of age for tissue collection. Considering the influence of hormonal fluctuations on visceral sensitivity, subsequent WAS modeling, testing, and tissue collection were performed only on non-euthanized male mice. Each offspring was considered as one sample size.

To experimentally investigate the impact of gut microbiota dysbiosis on embryonic ENS development, female mice in the ABX and CON groups underwent ABX or vehicle treatment and were mated accordingly. Fetal mice were collected at E13.5 and E18.5. Additionally, serum, cecal content, and fecal samples were obtained from the dams. Data from all fetuses (unspecified for sex) derived from a single dam were averaged to represent one sample size.

For experimental intervention with L. reuteri in the context of gut microbiota dysbiosis, treatment procedures for the ABX and CON groups remained unchanged. In the ABXreuteri group, female mice underwent the aforementioned ABX treatment, and starting from E0.5, they received daily oral gavage of 200 μL of L. reuteri (BNCC) at a concentration of 1 × 109 CFU/mL until the completion of tissue collection. Sterile water was administered to the CON and ABX dams by gavage. Housing conditions, time of tissue collection, and sample size calculations were performed as described previously.

For experimental intervention with propionate in the context of gut microbiota dysbiosis, treatment procedures for the ABX and CON groups remained unchanged. In the ABXPro group, female mice underwent the aforementioned ABX treatment; starting from E0.5, they were provided with ad libitum access to a solution of sodium propionate (Sigma-Aldrich) at a concentration of 200 mmol/L until E13.5 or E18.5, at which point the dams were killed and tissue collection was carried out. CON and ABX dams were provided with sterile water at libitum. Housing conditions, time of tissue collection, and sample size calculations were performed as described previously.

Water avoidance stress (WAS)

The WAS paradigm was conducted by placing the offspring from the WAS group on a platform (3 × 6 cm) elevated 10 cm above the ground and fixed in the center of a box (60 × 50 cm) filled with room-temperature water, with the water level reaching 1 cm below the top of the platform. The sham group was not subjected to water filling, whereas all other conditions remained the same. Each mouse in the WAS group remained on the platform for 1 h at a fixed time in the afternoon for 10 consecutive days. During the modeling period, no interventions were made to the behavior of the mice, and measures were taken to minimize noise, intense light, and other stimuli. At the end of each WAS or sham session, all fecal pellets in the collection container were counted to measure colonic motility [47].

Fecal water content

The experiment was conducted on the day of the final WAS session. The mice were individually placed in dry bedding-free cages for 1 h. Fecal pellets were collected every 15 min and immediately covered with centrifuge tube caps. After 1 h, the collected fecal pellets were weighed. Subsequently, the fecal pellets were dried in an oven at 60°C for 24 h and weighed again to determine the dry weight. Fecal water content was calculated using wet and dry weights to determine the proportion of water in the fecal pellets.

Colonic transit time

The colonic transit speed in mice was assessed using the latent bead test [48]. The experiment was conducted on the second day after the WAS. Briefly, mice were fasted (with access to water) for 12 h. They were then lightly anesthetized using isoflurane (RWD), and a 2.5 mm spherical plastic bead coated with glycerol was gently inserted into the distal colon using a glass rod positioned 2 cm from the anus. Subsequently, the mice were individually placed in bedding-free cages, and the time from bead insertion to expulsion was recorded. The experiment was repeated twice with a 6-h interval. The average of the two experiments was calculated as the bead expulsion latency.

Small intestinal transit time

An experiment was conducted using a bead expulsion test. Each mouse was administered 200 μL of carmine red solution (6% carmine (Macklin) dissolved in a 0.5% carboxymethylcellulose (Bide) water solution via oral gavage. After 45 min, the mice were euthanized by cervical dislocation. The abdominal cavity was opened and the small intestine was carefully removed from the pylorus to the ileocecal junction, minimizing any stretching of the intestinal tract. The small intestine was laid out in a straight line, and the proportion of the distance traveled by the carmine red solution to the total length of the small intestine was measured to assess intestinal transit function.

Visceral sensitivity test

CRD–EMG was performed according to established methods with some modifications [48]. A latex balloon (1 cm wide, 1.5 cm long) attached to a thin catheter (1 mm inner diameter, 1.5 mm outer diameter) at the head was prepared as a pressure-sensitive balloon. The mice were fasted (with access to water) for 24 h before visceral sensitivity testing. After light isoflurane anesthesia, a glycerol-coated balloon was inserted into the rectum of the mouse, positioning its distal end approximately 1 cm from the anus. The catheter connecting the balloon was secured to the base of the mouse tail using adhesive tape. The distal end of the balloon catheter was connected to a pressure gauge (World Precision Instruments) and a syringe via a three-way stop coat. After a 20-min adaptation period, the balloon was rapidly inflated at pressures of 20, 40, and 60 mmHg in a random order, with each inflation maintained for 20 s and a 4-min interval between inflations. Each pressure measurement was repeated thrice. Signals were collected and analyzed using a Biological Signal Acquisition System software (Taimeng). Changes in visceral sensation were represented by subtracting the area under the curve (AUC) for 20 s, corresponding to the pre-distension baseline, from the AUC for 20 s of distension.

Histology, electron microscopy, and immunofluorescence staining

Hematoxylin and eosin (H&E) staining was used to assess intestinal mucosal development in mouse pups. The ileum and colon were fixed in 10% formaldehyde–phosphate-buffered saline (PBS) solution for 24 h, dehydrated, embedded in paraffin, and sectioned into 5 μm-thick slices. After deparaffinization and hydration, sections were stained with H&E (Sigma-Aldrich), mounted with neutral gum (ZSGB–BIO), and promptly imaged.

Periodic acid-Schiff (PAS) staining was performed to evaluate the development of goblet cells in the colons of the offspring. Paraffin-embedded colon sections were deparaffinized and incubated in 1% periodic acid solution (Sigma-Aldrich) for 10 min. Subsequently, the sections were stained with Schiff's reagent (Sigma-Aldrich) for 40 min and counterstained with Harris hematoxylin solution (Sigma-Aldrich) for 25 min. After each step, the stained sections were rinsed with PBS. Finally, sections were mounted with neutral gum and immediately imaged.

For transmission electron microscopy (TEM), colon slices measuring 3 × 3 mm were fixed in 4% paraformaldehyde for 24 h, followed by fixation with 1% osmium tetroxide. The tissues were embedded in neutral resin and sectioned into thin slices. The sections were stained with uranyl acetate and lead citrate, and then stored under dry conditions for imaging purposes, which were promptly conducted.

For whole-mount staining of the longitudinal muscle-containing myenteric plexus (LMMP), colon tissue was opened along the mesentery and fixed with a 4% paraformaldehyde solution in a culture dish containing silica gel, with the serosal side facing down. Fixation was performed at 4°C for no longer than 6 h. Then, a 20% sucrose solution (Aladdin) was added for dehydration, and the samples were stored at 4°C until further processing. Subsequently, under a dissecting microscope, the samples were dissected by gently scraping off the mucosa and separating the submucosal and circular muscle layers to obtain the LMMP. The samples were then incubated at room temperature for 2 h in a blocking solution containing 0.3% Triton X-100 (Aladdin, Shanghai, China) and 10% donkey serum (Solarbio) in PBS. Following blocking, the primary antibody was added and incubated at 4°C for 48 h. After removing the primary antibody solution, the samples were washed thrice with PBS (three times, 15 min each) and incubated with the secondary antibody at room temperature for 4 h. Subsequently, the samples were co-incubated with Hochest33342 (Solarbio) for 5 min to label the cell nuclei, followed by PBS washes (five times, 10 min each). Whole mounts were placed on adhesive-coated glass slides, covered with a fluorescent mounting medium, and stored in the dark at –20°C for further analysis. For whole fetal mouse intestinal nerve mounts, no microdissection steps were required. The primary antibody was incubated at 4°C for 24 h, and the remaining steps were the same as for the myenteric plexus whole mounts. All the antibodies were diluted in PBS containing 0.3% Triton X-100 and 10% donkey serum.

Immunofluorescence staining of cryosections was performed to assess the number of mast cells in the adult mouse offspring. Intestinal tissue was fixed in 4% paraformaldehyde for 6 h at 4°C. After washing with PBS, the tissue was immersed in a 20% sucrose solution in PBS at 4°C for 24 h, followed by transfer to a 30% sucrose solution at 4°C for another 24 h. Subsequently, the tissue was embedded in optimal cutting temperature compound (SAKURA) and stored at –80°C. Cryosections of 8 μm thickness were cut and mounted on adhesive-coated glass slides, equilibrated to room temperature, and incubated at room temperature for 1 h for blocking. Then, the primary antibody was added to the blocking solution and incubated overnight at 4°C. The next day, the sections were washed three times with PBS (5 min each) and incubated with secondary antibody at room temperature for 1 h. The sections were then co-incubated with Hoechst33342 for 5 min to label the cell nuclei, washed with PBS (three times, 5 min each), and finally covered with an anti-fade mounting medium for fluorescence and stored at –20°C in the dark for further analysis.

Antibodies and their working concentrations are shown in Table S1.

Image acquisition and analysis

Whole-slide scanning of H&E and PAS-stained sections was performed using a NanoZoomer (Hamamatsu). Image acquisition of the E18.5 fetus and offspring ENS staining was performed using a confocal microscope (Zeiss). Fluorescence imaging of mast cells from the E13.5 fetus and adult offspring ENS was performed using an inverted fluorescence microscope (Zeiss). The TEM images of ultrathin sections of adult offspring were acquired using TEM (JEOL) at 120 kV. All samples were exposed to the same equipment for the same duration. To assess postnatal intestinal development, the length and cell count of 20 randomly selected villi or crypts in each animal section were measured. To count postnatal neurons and glial cells, 10 randomly selected ganglia were counted on each slide. To calculate the positive area and positive cell count, 10 images were selected at 20× objective magnification for each sample, and ImageJ software (RRID:SCR_002285) was used to analyze and measure the pixel number or cell count of positive immunolabeled regions of interest. For electron microscopy, at least five fields of view were studied for each sample at a magnification of 5000×. The experimental results were evaluated by an observer who was blinded to the experimental conditions.

In vitro intestinal motility assessment

The in vitro assessment of intestinal motility was performed using a validated method with some modifications [49]. Juvenile or adult mouse colonic segments (approximately 1 cm) were obtained by cervical dislocation after euthanasia. The segments were mounted in an organ bath filled with Krebs buffer solution and continuously introduced with 95% O2 and 5% CO2 at 37°C for 20 min or until a stable baseline was reached, followed by the application of 0.3 g tension. Muscle contraction and relaxation levels were measured before and after drug treatment using a tension sensor system (Taimeng). Between each drug treatment, the colonic segments were washed thrice with Krebs solution. Each colonic segment was subjected to two tests. During this process, the Krebs solution was changed every 15 min. The following drug concentrations were used: 10 μM carbachol (Aladdin), 50 μM atropine (Aladdin), 2 mM l-NAME (Macklin), and 15 mM l-arginine (Macklin). Results were calculated as g tension/g dry tissue weight.

Quantitative real-time polymerase chain reaction (qRT-PCR)

qRT-PCR was performed to analyze the total RNA extracted from embryonic or offspring colons using a Tiangen RNA Extraction Kit (TIANGEN). Following the determination of the RNA concentration, cDNA amplification was performed using a cDNA synthesis kit (Absin). The cDNA, qPCR mix (Toyobo), and the corresponding primers were mixed and subjected to a reaction in QuantStudio 5 (Thermo Fisher Scientific). mRNA expression was normalized using Gapdh as the reference gene, and the mRNA levels were calculated using the 2−ΔΔCt method. The primer sequences are listed in Table S2.

Transcriptome sequencing

The dams in the CON and ABX groups were euthanized at E18.5 via cervical dislocation, and the colons were dissected from their embryos. Total RNA was extracted using the Tiangen RNA Extraction Kit. Subsequently, cDNA amplification was performed using a cDNA synthesis kit, followed by the construction of an mRNA library. After initial library construction, the concentration of the library was detected using a fluorescence quantifier (Qubit 4.0), and the fragment size was assessed using a Qsep400 high-throughput biofragment analyzer. Finally, the effective concentration of the library was quantified using qPCR. Sequencing was performed by Metware Biotechnology Co., Ltd. using a NovaSeq. 6000 platform (Illumina) to generate 150-bp paired-end reads. Fastp v0.23.2 and HiSAT2 v2.2.1 were used for quality filtering and sequence alignment, respectively. Differential expression analysis was performed using DESeq. 2 v1.38.3. The differentially expressed genes were subjected to GO enrichment analysis using DAVID v2023q4. KEGG pathway enrichment analysis and GSEA for KEGG and GO analyses were performed using clusterProfiler v4.6.0.

Luminex liquid suspension array chip

A MILLIPLEX Mouse High-Sensitivity T-Cell Magnetic Bead Panel kit (Millipore) was used according to the manufacturer's instructions. Briefly, the protein supernatant extracted from the colonic tissues of adult offspring was incubated overnight in a 96-well plate embedded with microbeads. Subsequently, the plate was incubated with detection antibodies for 1 h. Streptavidin Phycoerythrin was added to each well for 30 min, and the values were read using a calibrated Luminex 200 system (Luminex Corporation).

Metagenomic sequencing

Total microbial DNA from preconception dams was extracted using the QIAamp PowerFecal Pro DNA Kit (Qiagen), and the DNA concentration was measured. The NEBNext Ultra DNA Library Preparation Kit (NEB) was used to generate sequencing libraries. DNA samples were sonicated to 350 bp, followed by end polishing, A-tailing, and ligation. Subsequently, sequencing was performed using Novogene (Novogene) on the NovaSeq. 6000 platform.

The quality of the filtered read pairs was assessed using FastQC (V.0.11.8). Bowtie2 (V.2.5.1) [50] was used to remove the host DNA. The remaining high-quality reads were used for further analyses. All samples were subjected to species-level analysis using MetaPhlAn4 (V.4.0.6) [51] and mpa (V.Oct22. CHOCOPhlAnSGB.202212) marker gene databases. Functional analysis was performed using HUMAnN3 (V.3.7) [51] in UniRef90 mode. The detected genes were further grouped into gene families and pathways and then sum-normalized within HUMAnN3.

Alpha diversity was estimated using Shannon and Simpson indices. Beta diversity was analyzed using the Bray–Curtis distance algorithm and visualized by PCoA. These analyses were conducted using the corresponding computations and plotting scripts within the Parallel-Meta Suite [52]. LEfSe was applied to the relative abundance of species and pathways to identify group-associated biomarkers [53]. Features with a linear discriminant analysis score >2.0 and p-value <0.05 were considered statistically significant. Spearman's rank correlation test was used to calculate the correlation between the species.

M650 targeted metabolome analysis

M650 targeted metabolome analysis was conducted following the standardized workflow of APExBIO (APExBIO). In brief, serum samples (50 μL) or cecal contents (50 mg) were mixed with a pre-cooled solution of methanol/acetonitrile/water (2:2:1, v/v), vortexed, subjected to 30 min of low-temperature ultrasonication, and then incubated at –20°C for 10 min. After centrifugation at 14,000 g and 4°C for 15 min, the supernatant was collected for further analysis. Following the preparation of external and internal standards for 623 metabolites (Table S3), separation was performed using an Agilent 1290 Infinity LC Ultrahigh-Performance Liquid Chromatography system (UHPLC) (Agilent) with HILIC (2.1 mm × 100 mm × 1.7 μm) and C18 (2.1 mm × 100 mm × 1.7 μm) columns. Mass spectrometry was performed using an AB 6500 QTRAP mass spectrometer (AB SCIEX). The raw MRM data were processed using MultiQuant software (AB SCIEX) for peak extraction, obtaining peak areas, and determining the ratio of the internal standard peak areas. Quantification was performed based on the standard curves. OPLS-DA and KEGG pathway enrichment analyses were conducted using MetaboAnalyst v6.0.

Measurement of SCFAs

Four SCFAs (acetate, propionate, butyrate, and valerate) were prepared as standard solutions in water, 15% phosphoric acid, l-ethyl ether, and internal standards to create ten concentration gradients. Serum samples (50 μL) or cecal contents (50 mg) were mixed with 15% phosphoric acid and a combination of internal standards and ethyl ether. The mixture was centrifuged at 12,000 rpm and 4°C for 10 min to obtain the supernatant. Separation was performed using a Thermo Trace 1300 gas chromatography system (Thermo Fisher Scientific) equipped with an Agilent HP-INNOWAX capillary column (30 mm × 0.25 mm × 0.25 μm). Mass spectrometry analysis was conducted using a Thermo ISQ 7000 mass spectrometer (Thermo Fisher Scientific). Proteo Wizard v.3.0.8789 software was used to calculate the SCFA content in the samples.

In vitro cultivation of L. reuteri

The concentration of L. reuteri was adjusted to 1 × 109 CFU/mL. Subsequently, 1 mL of this L. reuteri solution was inoculated into 10 mL of a liquid culture medium containing peptone, lipids, glucose, propanediol, glycerol, and the necessary electrolytes for L. reuteri growth (n = 3). The cultivation was maintained at 37°C under anaerobic conditions. Supernatants were collected at 0, 3, 6, and 24 h, centrifuged at 12,000 rpm for 5 min, and then the levels of SCFAs were measured.

Statistical analysis

The data are presented as the mean ± standard deviation (SD). The Shapiro–Wilk and Kolmogorov–Smirnov tests were used to assess the normal distribution of the data. Unpaired two-tailed Student's t-test or non-parametric tests (Mann–Whitney U test) were used for comparisons between the two groups. For data involving more than two groups, a one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test or non-parametric tests (Kruskal–Wallis test followed by Dunn's multiple comparison test) was performed. A two-way ANOVA was used for analyses involving two variables, followed by Tukey's or Sidak's multiple comparison test. Correlation analysis was conducted using Spearman's rank correlation test. The specific statistical tests employed for each panel are described in the figure legends. The value of n was varied in the experiments and is specified in each figure legend. The bar graphs represent individual biological replicates as single points. Statistical significance was set at p < 0.05. Statistical analyses were performed using GraphPad Prism 10 software (GraphPad Software; RRID:SCR_002798).