The early life microbiota protects neonatal mice from pathological small intestinal epithelial cell shedding

2.1. Animals

All mouse experiments were performed under the UK Regulation of Animals (Scientific Procedures) Act of 1986. The project licences PDADA1B0C and PPL80/2545 under which these studies were carried out was approved by the UK Home Office and the UEA Animal Welfare Ethical Review Body. Time‐mated female C57BL/6 Jax mice and female adult C57BL/6 Jax mice (aged 6‐10 weeks) were obtained from Charles River UK. All animals were maintained in the UEA Disease Modeling Unit under a standard 12 hours light/dark cycle and fed standard rodent chow and water ad libitum under specific pathogen‐free (SPF) conditions.

2.2. Induction of cell shedding and tissue collections

At 14, 21, and 29 days postpartum, pups (n = 6 per group) or adult control mice (n = 6) received an intraperitoneal (IP) injection of 1.25mg/kg liposaccharide (LPS) from Escherichia coli 0111:B4 (Sigma), 0.5mg/kg murine Tumor Necrosis Factor Alpha (TNF‐ɑ: 315‐01A, Peprotech) or sterile saline (control). Mice were sacrificed 90 minutes post‐challenge with LPS/ TNF‐ɑ. At autopsy, proximal small intestine was collected in 10% neutral buffered formalin (Sigma) and fixed for 24 hours. Samples of proximal small intestine and spleen were also collected into RNA later for transcriptome analysis or frozen on dry ice for subsequent analysis (Figure 1A).

2.3. Fecal microbiota transplant

To determine the effects of the microbiota in the protection against LPS shedding we altered the gut microbiota with a combination of antibiotics and FMT from adult mice as per Table 1 and Figure 5A. Ciprofloxacin and ampicillin provide a broad range spectrum combination that targets Gram‐positive and Gram‐negative bacteria, thereby effectively knocking down the wider murine resident microbiota. Dosing was determined based on previous published studies, and because both can also transfer to the infant (ie, pup) via breast milk (Drugs and Lactation database). ²⁰ , ²¹ At 10 days postpartum (D10) group 2 and 3 had 100 µg/mL of ciprofloxacin and 500 µg/mL of ampicillin added to the cage drinking water for 5 days to deplete the microbiota. After 5 days, both groups received normal drinking water for the rest of the study. For the FMT, two to three fecal pellets were collected from the control group Dam. Contents were gently homogenized using sterile glass beads and sterile PBS and passed through a 0.7 µm filter, with 50 µL of solution given to each pup by oral gavage. FMT’s were prepared daily, and pups were orally gavaged once a day from D16 to D20.

2.4. Histology

Formalin fixed tissue was dehydrated through a graded ethanol and xylene series and embedded in paraffin wax. About 5 μm sections were used for immunohistochemistry. Following deparaffinization and rehydration, tissue sections were treated with 1% hydrogen peroxide in methanol to block endogenous peroxidases. Subsequently, slides were treated using heat‐induced antigen retrieval in 0.01M citrate acid buffer (pH 6) followed by antibody staining. CC3 staining was performed using rabbit polyclonal anti‐active Caspase‐3 antibody (AF835, R&D systems) and peroxidase‐labeled anti‐rabbit EnVision secondary antibody (Dako) or DAB or mouse polyclonal Caspase‐3 antibody (NB100‐56708H Novus Biologicals LTD) and visualized with DAB peroxidase (HRP) substrate kit (SK‐4100, Vector Laboratories LTD). Rabbit polyclonal anti‐active Caspase‐3 antibody BrdU staining was performed using Sheep anti‐BrdU‐biotin (ab2284, Abcam) and neutravidin‐HRP (31030, Pierce). TNF‐R1 staining was performed using rabbit Anti‐TNF‐R1 antibody (sc‐7895, Santa Cruz) and peroxidase‐labeled anti‐rabbit EnVision secondary antibody (Dako). Following secondary antibody binding, signal was detected using 3,3′‐diaminobenzidine (Vector) followed by counterstaining with hematoxylin.

2.5. Caspase‐3 quantification

IECs were counted on a cell positional basis from villus tip (Cell position (CP) 1) down toward the crypts under 400× magnification. 20 well‐orientated hemi‐villi were counted per mouse and analyzed using the Score, WinCrypts ²² and PRISM (version 8) analysis software. IECs were defined as “normal” in cases where staining for active Caspase‐3 was absent. Immuno‐abeled cells with either unaltered or shedding morphology were treated as Caspase‐3 positive. IEC's stained with mouse polyclonal anti‐active Caspase‐3 antibody (NB100‐56708H) were visualized with DAB and staining intensity was measured using FIJI (ImageJ 1.52h, National Institutes of Health, USA). Imaging was with an Olympus BX60 microscope and C10plus digital camera.

2.6. ELISAs

Murine TNF‐ɑ or Murine IL‐10 Ready set go enzyme‐linked immunosorbent assay (ELISA) kits were used (eBioscience). Protocol was performed as per manufacturer's instructions. In brief, Costar 9018 ELISA plates were coated with 100 µL/well of capture antibody diluted in Coating Buffer overnight at 4°C followed by blocking with ELISASPOT diluent. Small intestinal samples were homogenized in Cell Lytic MT buffer (Sigma) using Lysing matrix E beads (MP Biomedicals) and homogenates cleared by centrifugation at 10 000 rpm for 15 minutes. After quantification using BCA kit (Pierce), equivalent amounts of test samples or standard material were applied to wells in duplicate and incubated overnight at 4°C. Detection Antibody, Streptavidin‐HRP and substrate were then applied, and plates were read at 450 nm with 570 nm correction.

2.7. Western blots

Spleens were homogenized in Cell Lytic buffer using Lysing matrix E beads (MP Biomedicals) and homogenates cleared by centrifugation at 10 000 rpm for 15 minutes. Homogenates were quantified by BCA analysis and 10 mg of total protein separated by SDS‐PAGE with 10% acrylamide resolving gel. Proteins were then transferred to Hybond‐P PVDF membrane (GE Healthcare). Immunostaining was performed with 1/100 anti‐MYD88 antibody (ab2064, Abcam) and 1/2000 Goat anti‐Rabbit IgG HRP conjugate (Millipore) or 1/5000 Rabbit Anti‐Beta‐Actin antibody (ab8227, Abcam) and 1/5000 Goat anti‐rabbit HRP. Signal was detected using Immobilon^TM Western chemiluminescent HRP substrate (Millipore) and the FluorChem E imaging system (Protein Simple).

2.8. Cytokine antibody array

Small intestinal samples were homogenized in Cell Lytic MT buffer (Sigma) using Lysing Matrix E beads (MP Biomedicals) at a concentration of 50mg/ml. Homogenates were then cleared by centrifugation at 10 000 rpm for 15 minutes before quantification of protein using BCA kit (Pierce). About 250 mg of total protein was then used to probe Murine antibody inflammation arrays (Abcam; ab133999) as per manufacturer's instructions. In brief, membranes were for 30 minutes before application of homogenized whole small intestinal lysates in blocking buffer overnight at 4°C. Following washing, biotin conjugated anti cytokine antibodies were applied for 2 hours before further washing and application of HRP conjugated streptavidin. Chemiluminescent detection was then performed using supplied detection reagents and visualization performed using FluorChem E imaging system (Protein Simple). See Table S1 for plate layout.

2.9. Isolation of IECs

IECs were isolated using a modification of the Weisser method. ²³ Briefly, whole small intestine was opened longitudinally, chopped into c.a. 5 mm² pieces and washed in 0.154 M sodium chloride (NaCl) containing 1 mM Dithiothreitol (DTT). Intestine was then incubated in a salt solution containing 1.5 mM Potassium Chloride (KCl), 96 mM NaCl, 27 mM Tri‐sodium citrate (Na₃C₆H₅O₇), 8mM sodium di‐hydrogen phosphate (NaH₂PO₄), and 5.6mM di‐sodium hydrogen phosphate (Na₂HPO₄), pH 7.3 for 15 minutes at 37°C with gentle shaking (140 rpm). Epithelial cells were then stripped from mucosa by further incubation in PBS containing 1.5mM Ethylene diamine tetraacetic acid (EDTA) and 0.5 mM DTT for 15 minutes at 37°C, 140 rpm. Epithelial cells were then recovered and stabilized by pelleting at 1000 rpm for 5 minutes followed by lysis in buffer RLT RNA lysis solution (Qiagen).

2.10. Isolation of DNA and 16S rRNA sequencing and analysis

Mouse fecal samples were collected into Lysis Matrix D tubes from the same mice at Day 14, 21, and 29. Adult fecal samples were collected from both parents at the same time as Day 14 samples. DNA extraction was carried out using a FastDNA Spin Kit for Soil (SKU 116560200‐CF, MP Bioscience) following the manufacturers protocol with the exception of extending the bead‐beating step to 3 minutes. DNA was quantified using a Qubit® 2.0 fluorometer (Invitrogen). 16S rRNA region (V1‐V2) primers were used for library construction (Table S2), with the following PCR conditions; cycle of 94°C 3 minutes and 25 cycles of 94°C for 45 seconds, 55°C for 15 seconds, and 72°C for 30 seconds. Sequencing of the 16S rRNA gene libraries was performed using Illumina MiSeq platform with 300 bp paired‐end reads. Raw reads were filtered through quality control using trim galore (version 0.4.3), minimum quality threshold of phred 33, and minimum read length of 60 bp. Reads that passed threshold were aligned against SILVA database (version: SILVA_132_SSURef_tax_silva) using BLASTN (ncbi‐blast‐2.2.25+; Max e‐value 10e−3) separately for both pairs. After performing BLASTN alignment, all output files were imported and annotated using the paired‐end protocol of MEGAN on default Lowest Common Ancestor (LCA) parameters. ²⁴ For analysis of microbiota composition, the samples were normalized by subsampling each sample to 3224 reads (the number of reads of the lowest sample) to equalize library sizes using the Phyloseq package version 1.24.2 in RStudio version 1.1.463.

2.11. Isolation of RNA, cDNA synthesis, and real‐time PCR

RNeasy plus mini‐spin columns were used to isolate total RNA (Qiagen). In brief, samples were homogenized using a rotor stator hand‐held homogenizer in buffer RLT (tissue samples) or reconstituted in buffer RLT by pipetting (isolated IECs). Lysates were processed through QIAshreddor columns (Qiagen) and subsequently RNeasy mini‐spin columns. Purified RNA was eluted into RNAase free water.

For TNF‐R1 real‐time PCR analysis of isolated IECs, cDNA was synthesized using Quantitect reverse transcription kit (Qiagen) and real‐time PCR performed using Quantitect SYBR green mastermix (Qiagen) and Quantitect murine TNF‐R1 primer assay (Qiagen) or hypoxanthine–guanine phosphoribosyltransferase (HPRT) 5′‐GACCAGTCAACAGGGGACAT‐3′ (sense) and 5′‐AGGTTTCTACCAGTTCCAGC‐3′ (antisense) primers. Cycling was performed on a Roche LightCycler 480 using the following conditions: 95°C, 5 minutes then 40 cycles of 95°C, 10 seconds; 60°C, 35 seconds. Relative quantification of levels of transcript expression was calculated using the Pfaffl method by comparing cycle threshold (CT) value of each target gene to the CT value of housekeeper. ²⁵ Data are presented as a “fold change” in expression (normalized against control untreated mice per cells).

For PCR array analysis, reverse transcription was performed using the Transcriptor first strand cDNA synthesis kit (Roche) and cDNA analyzed on custom real‐time ready 384 well custom PCR arrays (Roche). Arrays contained the following murine targets: IL‐1B, Ml2, Arrdc2, Bid, Card6, Casp4, Cycs, Ikbe, Rel, Tlr2, Traf5, Tnfrs8, Tnfrsf1b, Tnfrs19, Bd2l1, Rn18s, Il6, Alpi, Pdzd7, Casp12, Card11, Casp7, Fadd, Nkb1, Rela, Traf6, Traf2, Tnfrs9, Tnfrs11b, Tnfrs15, Bd2l10, Rpl13a, Il10, Pdx1, Apaf1, Card10, Casp3, Casp9, Fas, Nkb2, Relb, Traf3, Traf7, Tnfsf11, Tnfsf25, Tnfsf21, Bfar, Ptgds, Ifng, Ceacam10, Bad, Casp1, Caps6, Cd40lg, Fasl, Nkbia, Serpinb2, Tnf, Tnfsf18, Tnfsf8, Tnfsf4, Tnfsf14, Ephb3, Ptgs2, Col18a1, Pzp, Bax, Card9, Casp2, Cd40, Faslk, Nkbib, Serpinb9, Traf1, Tnfsf10, Tnfrsf10b, Tnfrsf1a, Tnfsf18, Hprt1, Ddk1, Gasdma, Fgl1, Bd2, Casp14, Casp8, Cflar, Ikbkb, Nod2, Actb, B2m, Gapdh.

Cycling was performed using Lightcycler 480 probe master mix and a Roche LightCycler 480 using the following conditions: 95°C, 10 minutes then 40 cycles of 95°C, 10 seconds; 60°C, 30 seconds, 72°C, 1 seconds. Final analysis of CT values was using the DDCT method as described above.

2.12. ¹H NMR metabolomics

Stomach contents were collected from mice at autopsy and immediately frozen on dry ice before transfer to −80°C prior to analysis. Small intestinal contents were flushed from the duodenum using sterile water and frozen similarly prior to analysis. ¹H nuclear magnetic resonance (NMR) was used to identify the presence and concentration of several metabolites. Supernatant samples were thawed at room temperature and prepared for ¹H NMR spectroscopy by mixing ~50 mg (FW) of stomach content samples with 600 µL of NMR buffer (0.26 g NaH₂PO₄ and 1.41 g K₂HPO₄) made up in 100% D₂O (100 mL), containing 0.05% NaN₃ (50 mg), and 1 mM sodium 3‐(Trimethylsilyl)‐propionate‐d4, (TSP) (17 mg) as a chemical shift reference. The sample was mixed, and 500 µL was transferred into a 5‐mm NMR tube for spectral acquisition. The ¹H NMR spectra were recorded at 600 MHz on aBruker Avance spectrometer (Bruker BioSpin GmbH, Rheinstetten, Germany) running Topspin 2.0 software and fitted with a 5 mm TCI cryoprobe. Sample temperature was controlled at 300 K. Each spectrum consisted of 512 scans of 65,536 complex data points with a spectral width of 12.3 ppm (acquisition time 2.66 seconds). The noesypr1d pre‐saturation sequence was used to suppress the residual water signal with low‐power selective irradiation at the water frequency during the recycle delay (D1 = 3 s) and mixing time (D8 = 0.01 s). A 90° pulse length of 12.9 μs was set for all samples. The “noesypr1d” pre‐saturation sequence was used to suppress the residual water signal with a low‐power selective irradiation at the water frequency during the recycle delay. Spectra were transformed with a 0.3‐Hz line broadening, manually phased, baseline corrected, and referenced by setting the TSP methyl signal to 0 ppm. The metabolites were quantified using the software Chenomx NMR suite 7.6. Metabolites were identified using information found in the literature (references) or on the web (Human Metabolome Database, http://www.hmdb.ca/) and by the use of the 2D‐NMR methods, COSY, HSQC, and HMBC.

2.13. Statistics

Results were plotted and analyzed using PRISM version 8 software (GraphPad). Data (mean + SD) were analyzed using ordinary one‐way ANOVA with Tukeys posttest with a 95% confidence interval. Cleaved caspase‐3 (CC3) was measured using FIJI (Fiji Contributions) and plotted as mean fluorescence intensity (MFI) from three sections per mouse (n = 7). Microbiota and metabolome data analysis was carried out in R version 3.5.0 using RStudio (version 1.1.463). The “vegdist” function in the “vegan” (version 2.5.4) R package was used to calculate Bray‐Curtis dissimilarity matrices for the microbiota abundance and metabolomics data. NMDS plots of Bray‐Curtis matrices were generated for microbiota data using the vegan package. Permutational multivariate analysis of variance (PERMANOVA) was performed on Bray‐Curtis matrices using the “pairwiseAdonis” package. Diversity analysis of microbial abundance data was performed using the “vegan” package. Statistical analysis of number of genera and Shannon diversity between groups was performed using Kruskal‐Wallis one‐way analysis of variance with Pairwise Wilcoxon Rank Sum Tests using Benjamini‐Hochberg stepwise adjustment.

3.1. Neonatal mice are refractory to LPS‐induced pathological cell shedding

We previously demonstrated that IP administration of LPS‐induced potent cell shedding in adult mice with peak response 90 minutes post injection as determined by CC3 staining. ¹⁷ , ²⁶ To assess cell shedding response of neonatal mice, 14‐ and 21‐day‐old mice were administered with either low (1.25 mg kg‐1) or high (10 mg kg‐1) dose LPS (or PBS vehicle control) and sacrificed 90 minutes post‐administration (Figure 1A). Surprisingly, IECs in 14‐ and 21‐day‐old mice were CC3 negative following low‐and high‐dose LPS and villi exhibited normal histology at the macroscopic level (Figure 1B,C). We next treated 29‐day‐old neonatal mice with IP LPS and observed the re‐emergence of cell shedding response (Figure 1B,D). As expected, adult mice (8‐10 weeks) showed enhanced cell shedding responses as determined by CC3 positive cells and associated pathology within the small intestine (Figure 1B,E).

We reasoned that absence of CC3 staining in neonates may be the result of delayed IEC apoptosis. To test this hypothesis, 14‐day‐old neonatal mice were administered low‐or high‐dose LPS and tissue collected up to 180 minutes post‐LPS administration. In contrast to adult mice which show extensive CC3 staining at 90 minutes coupled with associated villus atrophy at later time points, neonatal mice showed no evidence of CC3 staining up to 180 minutes post‐LPS, nor evidence of pathological changes that normally accompany the cell shedding response (data not shown). Taken together, mice are largely refractory to LPS‐induced cell shedding during the neonatal period, with intermediate responses observed in Day 29 old mice, with adult mice very susceptible to pathological cell shedding. These distinctions cannot be explained by altered kinetics of cell death responses during the early life developmental period.

3.2. Cell signalling pathways that drive pathological cell shedding are expressed in neonatal mice

Previous work suggests that systemically delivered LPS signals via mononuclear cells which express TLR4. This signaling cascade is mediated by the TLR‐adaptor protein MYD88, which drives production of TNF‐ɑ, which in turn binds TNFR1 expressed on the surface of IECs and triggers apoptosis and cell shedding. ²⁶ We reasoned that this signaling cascade may be diminished in neonatal mice, thus preventing initiation of the cell shedding response. IECs were isolated from the mucosa of neonatal (Day 14), and adult mice and expression of TNF‐R1 determined by real‐time PCR. Levels of TNFR1 transcript was significantly increased in control neonatal mice compared to adult mice, but neonates dosed with LPS had significantly reduced TNFR1 (Figure 2A). We next reasoned that production of TNF‐ɑ in neonatal mice may be impaired. However, ELISA analysis of whole small intestinal homogenates from Day 14 neonatal mice exposed to low‐and high‐dose LPS confirmed production of TNF‐ɑ at levels comparable to adult mice (Figure 2B).

Previous studies have shown reduced expression of MYD88 protein in human neonatal peripheral blood leukocytes, and we hypothesized that altered MYD88 expression in neonatal mice may drive reduced cell shedding. ²⁷ However, there were no significant differences in the levels of MYD88 protein in homogenized spleen samples taken from 14‐day‐old and adult mice (Figure 2C). To confirm our findings indicating there are no neonatal defects in cell shedding signaling pathways we also challenged mice IP with TNF‐ɑ. Whilst adult mice showed the expected potent cell shedding response, neonatal mice were largely refractory to cell shedding, with only isolated cell shedding events observed (Figure 2D). Taken together, this suggests that core components of the cell shedding signaling pathways are intact in neonatal mice and that blunting of the cell shedding response cannot be explained by altered expression profiles of these components at the transcriptional and translational level.

3.3. Inflammatory responses are altered in neonatal versus adult mice during pathological cell shedding

PCR and antibody arrays were used to explore if differing cell shedding outcomes in neonatal mice may be driven by alterations in immunologic/inflammatory responses and/or in expression of bacterial recognition markers. We observed reduced expression of pro‐inflammatory markers such as IFN‐γ but increased anti‐inflammatory cytokines such as IL‐10 in neonatal mice versus adults following LPS stimulation (Figure 3A). We also observed altered expression of transcripts associated with bacterial recognition including nucleotide‐binding oligomerization domain‐like receptors (NODs) and members of TNF receptor‐associated proteins including TRAF1, TNFrsf10, and TNFrsf21, alongside changes to components of the apoptotic cascade including reduced expression of CARD9, 10, 11, and Casp3 (Figure 3B). At a translational level we observed elevation of IL‐10, IL‐7, IL‐12, and TNF‐ɑ in neonatal versus adult mice during pathological cell shedding, with the increased IL‐10 levels in neonatal mice confirmed by ELISA (Figures S1 and S3C). These data indicate that neonatal mice may have a more anti‐inflammatory and anti‐apoptotic milieu when compared to their adult counterparts, which may be in part driving altered cell shedding responses.

3.4. Microbiota profiles and microbial‐derived metabolites are significantly altered in neonatal mice when compared to adults

Previous studies have indicated that the neonatal microbiota significantly influences immune and epithelial responses. ²⁸ Moreover, humans and mice show distinct alterations in microbiota composition during the early life developmental window, and our previous work highlighted the importance of (early life) gut microbes in regulating the cell shedding response. Thus, we sought to determine how microbial profiles diversify at key postnatal time points which also correspond to the observed differences in cell shedding responses.

The bacterial composition of fecal samples was determined by 16S rRNA amplicon sequencing and analysis. Genus level clustering of samples using non‐metric multidimensional scaling (NMDS) indicated separation of the microbiota profiles between Day 14 from Day 21, Day 29 and adult mice (PERMANOVA comparisons P = .006), while Day 21 mice showed a more divergent microbiota that was separate from Day 29 mice (PERMANOVA comparison P = .006), which clustered closely with adult (maternal) mouse samples (Figure 4A). The clustering of the microbiota composition in Day 14 mice was driven by the genus Streptococcus while Day 21 mice showed a divergent microbiota composition driven in part by Lactobacillus abundance. Day 29 mice clustered with the adult samples with the composition driven by genera including Roseburia, Bilophila, and Alistipes (Figure 4A).

We observed that number of genera detected, and the Shannon diversity index increased during the neonatal period; with the lowest number of genera in Day 14 mice compared to Day 21, Day 29, and adult mice (Figures S2A,B). We also examined the 30 most abundant genera by relative abundance from Day 14, 21, 29, and adult mice using Bray‐Curtis dissimilarity (Figure 4B), showing that Day 14 mice samples clustered with a distinct microbiota composition comprised of Streptococcus, Staphyloccous, Escherichia, and Enterococcus that were absent from older samples while Lachnoclostridium, Roseburia, Candidatus Arthromitus (or Segmented filamentous bacteria), and Alistipes became dominant by 29 days and in adult mice. This composition was also illustrated with the relative abundance within in mouse sample (Figure 4C). Interestingly, the 21‐day‐old group appeared to have a more varied microbiota profile between animals, possibly linked to the rapidly changing nutritional environment during this period (ie, moving from maternal milk to solid food) (Figure 4C).

We next performed ¹H NMR analysis of duodenal and stomach contents harvested from neonatal and adult mice to profile changes in metabolites (Figures 4D and S2C,D). Day 14 mice duodenum samples clustered separately with a higher proportions of the bile acids cholate and taurocholate (Figures 4D and S2J,K), at which time the presence of SN‐glycerol‐phosphocholine in the stomach contents indicated the milk‐based diet at 14 days (Figure S2E,F). Day 14 mice also showed an absence of leucine and glycine from duodenum contents potentially linked to their milk diet (Figure S2I‐M). The appearance of the microbial metabolites propionate and trimethylamine in stomach contents at Day 21 and 28 indicated the start of coprophagy and the potential exposure and transfer of new microbes between neonatal mice. These changes suggested dramatic shifts in microbiota and metabolic profiles during the neonatal period related to changes in diet, which correlates with distinct responses to LPS‐induced cell shedding.

3.5. Microbiota composition and antibiotic‐induced microbiota disturbances modulate pathological cell shedding responses

Next, we determined if the protective responses observed in neonatal mice were associated with differences in the microbiota and metabolite profiles as seen in Figure 4. We performed an FMT experiment whereby we dosed neonatal mice (Day 16) with adult (dam) fecal samples (8‐10 weeks) for 5 days via oral gavage, with further groups also treated with antibiotics (via drinking water for 5 days) to deplete the resident neonatal microbiota (alone), and in combination with FMT. On D21 all group received LPS by IP injection (Figure 5A). Histopathological analysis of small intestine showed that adult FMT alone was insufficient to significantly disrupt intestinal pathology (Figure 5B). However, disruption of the neonatal microbiota with antibiotics significantly increased damage to the epithelium, which was not reversed by FMT from adult mice (Figure 5B). Neonatal mice treated with antibiotics (Day 10‐14) with and without adult FMT had significantly increased apoptotic cell death (measured by CC3 intensity) (Figure 5B,C). Although we did not observe a significant increase in intestinal pathology/damage scores from FMT alone, we did observe that FMT induced a significant increase in apoptotic cell death (Figure 5D). These data indicate that the neonatal microbiota plays a key role in modulating cell shedding.