Escherichia coli Nissle Improves Short-Chain Fatty Acid Absorption and Barrier Function in a Mouse Model for Chronic Inflammatory Diarrhea

Abstract

Background

Defects in SLC26A3, the major colonic Cl⁻/HCO₃⁻ exchanger, result in chloride-rich diarrhea, a reduction in short-chain fatty acid (SCFA)-producing bacteria, and a high incidence of inflammatory bowel disease in humans and in mice. Slc26a3^−/− mice are, therefore, an interesting animal model for spontaneous but mild colonic inflammation and for testing strategies to reverse or prevent the inflammation. This study investigates the effect of Escherichia coli Nissle (EcN) application on the microbiome, SCFA production, barrier integrity, and mucosal inflammation in slc26a3^−/− mice.

Methods

In vivo fluid absorption and bicarbonate secretion were assessed in the gut of slc26a3^+/+ and slc26a3^−/− mice before and during luminal perfusion with 100 mM sodium acetate. Age-matched slc26a3^+/+ and slc26a3^−/− mice were intragastrically gavaged twice daily with 2 × 10⁸ CFU/100 µL of EcN for 21 days. Body weight and stool water content were assessed daily, and stool and tissues were collected for further analysis.

Results

Addition of sodium acetate to the lumen of the proximal colon significantly increased fluid absorption and luminal alkalinization in the slc26a3^−/− mice. Gavage with EcN resulted in a significant increase in SCFA levels and the expression of SCFA transporters in the slc26a3^−/− cecum, the predominant habitat of EcN in mice. This was accompanied by an increase in mucus-producing goblet cells and a decrease in the expression of inflammatory markers as well as host defense antimicrobial peptides. EcN did not improve the overall diversity of the luminal microbiome but resulted in a significant increase in SCFA producers Lachnospiraceae and Ruminococcaceae in the slc26a3^−/− feces.

Conclusions

These findings suggest that EcN is able to proliferate in the inflamed cecum, resulting in increased microbial SCFA production, decreased inflammation, and improved gut barrier properties. In sufficient dosage, probiotics may thus be an effective anti-inflammatory strategy in the diseased gut.

Graphical Abstract

Open in new tabDownload slide

Introduction

Patients with inflammatory bowel disease (IBD) suffer from diarrhea and microbial dysbiosis with a reduction in the bacterial families that produce short-chain fatty acids (SCFAs) by the fermentation of dietary fibers, and with alterations in SCFA content in the gut.^1-3 Because SCFAs play a crucial role in helping the host conserve fluid and electrolytes and help in adjusting the gut luminal pH,⁴ disrupting the production of SCFAs in the colon can lead to diarrhea,⁵ while increasing SCFA production by adding starch to oral rehydration solutions may enhance colonic fluid absorption and mitigate the dehydration associated with acute diarrhea.⁶ Moreover, SCFAs regulate gut barrier function by promoting the expression of genes related to the tight junction proteins, boosting mucus production via epigenetic regulation of MUC2 expression in goblet cells, and exerting a variety of anti-inflammatory and homeostatic effects on the colonic mucosa.^7-10 Therefore, strategies to increase SCFA production in the inflamed gut are discussed as a potential new treatment option for IBD.¹¹

While the importance of SCFAs for colonic fluid absorption and a healthy gut barrier is well known, the problem of how to apply SCFAs to the colonic epithelium is not solved. SCFA salts are absorbed in the stomach and small intestine and, therefore, are not likely to reach the colonic lumen in high enough concentrations for a fluid stimulatory and anti-inflammatory effect. Encapsulation for colonic release is inefficient; SCFAs have a bad odor and a short shelf life, and oral application is accompanied with high rates of adverse events and incompliance.¹² The high concentrations of SCFAs that effectively stimulate colonic fluid absorption are produced by the conversion of starch by the healthy gut microbiome. Our study therefore asked the following questions: (1) Is acetate, the main SCFA produced by Escherichia coli Nissle (EcN), able to effectively restore the fluid absorptive (FA) capacity in the colon of slc26a3^−/− mice? (2) Does an oral delivery of EcN result in live and metabolically active EcN in the gut of slc26a3^−/− mice? (3) What are the overall effects of this probiotic, which is known for its protective function on the gut mucosa, and for stimulating SCFA production of bacterial communities,^13,14 on the microbiome of the slc26a3^−/− colon?

We chose the slc26a3^−/− mouse model of chronic but mild colonic inflammation for this study for several reasons:

1. This mouse model recapitulates many features/symptoms observed in patients with the rare genetic disorder chloride-losing diarrhea (CCD or CLD).¹⁵ CLD patients have a very high risk of developing IBD,^16,17 and the slc26a3^−/− mice develop spontaneous mild colonic inflammation in adulthood.¹⁸

2. Both CLD patients and slc26a3^−/− mice have gut microbial dysbiosis with a depletion of SCFA producers similar to that described for patients with ulcerative colitis.^1,12

A comprehensive investigation of the impact of EcN on gut function was performed, including its effects on the cecal and colonic microbiome, stool water content, and gastrointestinal (GI) transit time, and its ability to stimulate SCFA production by the gut microbiota and attenuate the observed pathological features in the gut of slc26a3^−/− mice, such as the reduced Muc2 production by goblet cells, increased proinflammatory cytokines, and altered host defense mechanisms.

Methods

Animals

Slc26a3^+/+ and Slc26a3^−/− mice¹⁵ were bred and maintained at Hannover Medical School under standard temperature and light conditions. To prolong survival and prevent dehydration, the mice were fed a specialized diet (Altromin, Lage, Germany, Cat. no. C0197) along with the normal chow, and a half-maximal Pedialyte drinking solution containing glucose (69.38 mM), sodium chloride (17.45 mM), sodium citrate (1.46 mM), and potassium citrate (3.14 mM). All mice used in this study were age-matched and used between 19 and 26 weeks old. Animal studies are reported in accordance with ARRIVE guidelines 2.0.¹⁹ All animal experiments were approved by the Hannover Medical School Committee on investigations involving animals and an independent committee established by the local authorities (authorization number: 33.12-42502-04-19/3197 for breeding and 33.19-42502-04-20/3561 for experiment). Supplementary Table 1 presents the sex match in the experiments. Because they also needed to be age-matched and cohoused littermates, the sex match was good but not 100% identical in each experimental group.

In Vivo Fluid Transport Measurements by Single-Pass Perfusion of Proximal Colon

Mice were anesthetized with isoflurane (Forene, Abbott, Wiesbaden, Germany) via tracheal intubation connecting to a mechanical ventilator. The surgical procedure was performed as previously described²⁰ with minor modifications. Briefly, 2 incisions were made at the proximal and distal ends of the proximal colon. The proximal colon had an intact blood supply and was gently flushed before perfusion. Thirty minutes were allowed for the stabilization of cardiovascular, respiratory, and intestinal functions before the experiments were commenced. The cecum was perfused with an unbuffered solution, pH titrated to 7.4, 37 °C, consisting of NaCl 145.5 mM, KCl 4.0 mM, CaCl₂ 1.2 mM, at a rate of 30 mL/h. For selected experiments, a 5% CO₂/95% O₂-gassed, HCO₃-buffered perfusate (NaCl 121 mM; NaHCO₃ 24 mM; KCl 4.0 mM; CaCl₂ 1.2 mM; C₂H₃NaO₂ 100 mM), or an O₂-gassed, HEPES/Tris-buffered perfusate (17 mM HEPES/3 mM Tris, 125 mM NaCl, KCl 4.0 mM; CaCl₂ 1.2mM; C₂H₃NaO₂ 100 mM) was utilized. All effluents were visually free of blood throughout all experiments. The rate of fluid absorption was calculated according to the weight of the influx and the outflow (effluent). The rate of luminal alkalinization was determined by the back titration of the effluents to pH 5.0 with 5 mM HCl.

EcN Colonization

EcN was cultured overnight in Luria Broth media, and then harvested by centrifugation (5 minutes at 8000 g). The resulting pellet was resuspended in sterile phosphate-buffered saline (PBS)/glycerol (80/20%), adjusted to a concentration of 2 × 10⁸ colony-forming unit (CFU)/100 µL, and stored at −80 °C until further use. To assess E. coli or EcN loads under baseline conditions and after EcN gavage, fecal and cecal content was weighed, homogenized in 500 µL 1× PBS, and serial dilutions were plated on Enterobacteriaceae-selective MacConkey agar (Sifin, Berlin, Germany). EcN with a known concentration was also plated on MacConkey agar as a positive control under the same conditions. Single colonies on the MacConkey agar were picked, and PCR was performed to confirm the presence of EcN in the colonies, as previously described.²¹ PCR with E. coli was also performed as a negative control, and the primers for EcN and total bacteria are listed in Supplementary Table 2. The number of colonies in the concentration-known positive EcN control was counted, and EcN CFUs/mg of fecal or cecal content were estimated for each sample after colony enumeration. The total bacterial quantification in the fecal and cecal content was determined by quantitative flow cytometry (MACSQuant Analyzer, Miltenyi Biotec, Germany) using fluorescence staining of microbial cells with SYBR Green I (Thermo Fisher Scientific, Waltham, MA, USA).²² The ratio of EcN to total bacteria was then calculated.

Experimental Protocol

slc26a3^+/+ and slc26a3^−/− mice were intragastrically gavaged twice daily (with a gap of 10-12 hours) with EcN (2 × 10⁸ CFU/100 µL) for 21 days (Figure 2A). The mice were monitored, and the body weight was assessed daily. Following 21 days, the mice were euthanized, and samples were collected for further analysis. Ungavaged slc26a3^+/+ and slc26a3^−/− mice served as baseline controls.

Stool Water Content and Total GI Transit Time

Stool water content was measured daily prior to gavage as previously described.^18,23 Total GI transit time was measured once per week using Carmine red-stained food pellets as previously described.²³

Stool Chloride Assay

Fecal samples were collected and dried at 37 °C for 3 days. The dried samples were weighed and then resuspended in ultrapure water. The suspension was incubated at 60 °C for 40 minutes, followed by centrifugation at 21 000 g for 10 minutes. The chloride concentration in the resulting supernatant was determined using a chloride assay kit (MAK023, Sigma-Aldrich) according to the manufacturer’s instructions. The chloride concentrations in the fecal samples were then back-calculated.

SCFA Measurement

Concentrations of SCFAs, including but not limited to, acetate, butyrate, and propionate, were quantified in cecal and fecal content. This was done at the RCU Metabolomics of Hannover Medical School on the basis of a gas chromatography-mass spectrometry method including a derivatization step and addition of a labeled standard as previously described.²²

Histology

Cecal and colonic tissues from each group were collected and fixed using 4% paraformaldehyde. Paraffin-embedded sections (3 μm) were stained with hematoxylin and eosin to determine the crypt depth and muscle layer thickness as previously described.²³ Sections were also stained with Alcian blue/periodic acid-Schiff (AB/PAS) to detect the different mucins.¹⁸ The total number of AB/PAS⁺ cell thecae was counted in at least 5 crypts and averaged for each section. All the images were acquired under the Leica DFC295 light microscope.

Quantitative PCR

mRNA expression of a variety of genes was assessed in different intestinal segments (cecum and distal colon) by quantitative PCR using ribosomal protein S9 (Rps9) as a reference gene. RNA extraction, cDNA transcription, and quantitative PCR were performed as per manufacturer’s instructions and as previously described.²⁰ Briefly, total RNA was extracted using RNeasy Mini Kit (Qiagen, Hilden, Germany), and 1 µg of RNA was reverse transcribed with the QuantiTect Reverse Transcription Kit (Qiagen GmbH). The resulting cDNA was diluted 1:20 with DNase-free water, and 4 μL of the dilution was used as a template for PCR. Each reaction also contained 5 μL of 2× qPCRBIO SyGreen Mix Lo-ROX (PCR Biosystems) and an appropriate amount of primers (Supplementary Table 3).

Microbial 16S rRNA Gene Analyses

DNA in cecal and fecal content was extracted using ZymoBIOMICS Miniprep Kit (ZYMO, Irvine, CA, USA). Libraries for 16S rRNA gene sequencing were prepared as described previously.²² Amplification of the V3V4 region using primers 341F/805R was performed as described previously.²² The obtained amplicons were sequenced using Illumina MiSeq (2 × 300 bp). Sequences were processed via the DADA2 pipeline (v1.20) and annotated based on RDP’s taxonomy.²⁴ Chimeras were eliminated, and the sequences with counts >5 that were annotated at the phylum level were included. Subsequent analyses, α-diversities and nonmetric multidimensional scaling based on relative abundance data, were conducted using the phyloseq package.²⁵

Statistics

The data analysis was conducted using GraphPad Prism version 8.0.2 (GraphPad Software Inc., San Diego, CA). The results are presented as means ± SEM. Unpaired Student’s t-tests (for parametric data with normal distribution) or nonparametric Mann-Whitney U-tests were used for the comparisons within genotypes and/or between control and EcN-treated groups Daily body weight and stool water content were analyzed using the area under the curve. Significant differences were indicated by the symbols # and * for P <.05. The symbol # represents comparisons between the 2 different genotypes, while the symbol * represents comparisons between baseline and after EcN gavage within the same genotype.

Results

Sodium Acetate Increases Fluid Absorption and HCO₃⁻ Output Rates in slc26a3^−/− Gut In Vivo

The basal FA and HCO₃⁻ output rates (J_HCO3⁻) were significantly lower in the slc26a3^−/− proximal colon in comparison to the slc26a3^+/+ mice (Figure 1A and B). To determine if a SCFA⁻/HCO₃⁻ transporter can replace the absent SLC26A3, 100 mM sodium acetate (SA) (pH 7.4) was luminally applied to the proximal colon. This elicited a robust response, with the FA and J_HCO3⁻ rates significantly increasing in the proximal colon of both WT and slc26a3^−/− mice (Figure 1A and B). Therefore, the supplementation of SCFA directly to the colonic mucosa is able to rescue fluid absorption and luminal alkalinization in the gut of the slc26a3^−/− mice.

Effective EcN Colonization in slc26a3^−/− and slc26a3^+/+ Cecum

Slc26a3^+/+ and slc26a3^−/− mice were intragastrically gavaged with 2 × 10⁸ CFU/100 µL EcN, a SCFA producer with predominant acetate production, twice a day for 21 days (Figure 2A). This dose was chosen because it had strongly reduced the destructive effects of concomitant DSS application on the colonic barrier.¹⁴ We collected the cecal and fecal content from non-gavaged mice and EcN-gavaged mice after 21 days of gavage with EcN, homogenized the samples, and serially diluted them before plating the different dilution steps on MacConkey agar. This permitted the determination of viable E. coli in the fecal sample. We then picked individual colonies from the samples without gavage, and after EcN gavage, and performed PCR with EcN-specific primers to confirm the presence of EcN in the cecum and colon (Supplementary Figure 1A and B). Before EcN gavage, the presence of E. coli in the cecum and colon of slc26a3^+/+ and slc26a3^−/− mice was nonexistent to extremely low, confirming results from our previous microbiome studies in this strain.^18,26 All the picked colonies after gavage consisted of EcN. To assess whether EcN actually proliferated in the gut, we assessed the total bacteria count, as well as the percentage of CFU of EcN in the cecal and fecal content. The total bacteria count was slightly but significantly lower in the slc26a3^−/− cecal content. A higher CFU load of EcN in gut lumen than in the daily gavaged EcN content was observed. The EcN load was higher in slc26a3^−/− mice in comparison to the slc26a3^+/+ mice, especially in the cecal content (Figure 2B and C). The percentage of EcN of the total bacterial load was approx. 0.4% in slc26a3^+/+ and 2% in slc26a3^−/− luminal content.

Improved Cecal SCFA Concentration Upon Treatment With EcN in slc26a3^−/− Mice

EcN is known to generate predominantly acetate and relatively minor amounts of propionate and butyrate.¹³ We therefore next assessed whether the levels of SCFA were affected after gavage with EcN. In non-gavaged mice, the total SCFA concentration, as well as that for acetate, propionate, and butyrate individually, were all significantly lower in the slc26a3^−/− cecal and fecal content compared to wild-type (WT) mice (Figure 3A and B). After the 3-week period with twice-daily EcN application, the concentration of total SCFAs and the 3 dominant SCFA species were significantly increased both in the slc26a3^+/+ and in the slc26a3^−/− cecal content (Figure 3A). In the feces, both total SCFAs and propionate increased after EcN gavage in the slc26a3^+/+ mice, while only propionate showed a significant increase in the feces of slc26a3^−/− mice (Figure 3B). Interestingly, the low SCFA concentrations in the slc26a3^−/− cecum were accompanied by much lower mRNA expression levels of the SCFA transporters proton-coupled monocarboxylate transporter 1 (Mct1) and sodium-coupled monocarboxylate transporter 1 (Smct1) in the slc26a3^−/− compared to slc26a3^+/+ cecum. After EcN treatment, the mucosal mRNA expression levels of Mct1 had increased to almost the levels of the WT (Figure 3C). In the distal colon, the expression of Mct1 was 4-fold lower than that in the cecal mucosa, did not differ between slc26a3^−/− and WT mucosa (Figure 3D), and did not increase after EcN gavage. The Smct1 expression levels were significantly lower than Mct1 expression levels in the cecum, but similar to Mct1 in the distal colon, with no difference between the groups (Figure 3D) and no increase after EcN gavage. This suggests that EcN application is able to increase the SCFA production in the cecum (its physiological habitat), and that this is accompanied by an increase in the expression of the SCFA transporters in the epithelium.

Increase in Goblet Cell Granules Upon Treatment With EcN

We have previously shown that the absence of Slc26a3 resulted in a loss of goblet cell theca and downregulation of Muc2 expression in the distal colon.¹⁸ The present study extends this observation to the slc26a3^−/− cecum, as evidenced by the AB-PAS staining, counting of the total AB/PAS⁺ goblet cell theca, and Muc2 mRNA expression. Posttreatment with EcN, this loss of goblet cell theca was reversed, and Muc2 mRNA expression increased significantly (Figure 4A and B). On the other hand, while there was a trend towards an improvement both in the count and Muc2 expression, it was not significant in the slc26a3^−/− distal colon (Figure 4A and C). The histological assessment revealed no significant difference in cecum crypt depth between slc26a3^+/+ and slc26a3^−/− mice. However, the slc26a3^−/− mice had a longer crypt in the distal colon, and the EcN treatment failed to reverse this hyperplasia (Supplementary Figure 2).

Reduced Inflammatory State of the Cecal Mucosa After EcN Treatment

The absence of Slc26a3 expression resulted in spontaneous mild inflammation in the distal colon at a relatively late stage in their lifespan.¹⁸ In the present study, a similar increase in the mRNA expression of proinflammatory markers Ly6g, Lcn2, and Il13 was observed in the cecum of slc26a3^−/− mice, which was significantly lower after EcN treatment (Figure 5A). Il22 mRNA expression, a cytokine with both pro- and anti-inflammatory actions (because it upregulates antimicrobial peptides), was also strongly upregulated in slc26a3^−/− colonic mucosa, and normalized in the cecal mucosa after EcN gavage (Figure 6A). However, these changes were observed only in the cecum after EcN treatment, while the proinflammatory markers remained high in the slc26a3^−/− distal colon (Figure 5B). In contrast, TNFα and Il1β mRNA expression levels were not different between slc26a3^+/+ and slc26a3^−/− cecal mucosa, while it was significantly increased in the slc26a3^−/− distal colonic mucosa compared to slc26a3^+/+ (Figure 5A and B).

Normalization of Antimicrobial Peptide Expression After EcN Treatment

Another important previous observation has been the increase in host defense strategies in the form of increased expression of a panel of antimicrobial peptides in the slc26a3^−/− distal colon.²⁶ This is recapitulated in the slc26a3^−/− cecum as well (Figure 6A). Interestingly, after EcN treatment, this increased host defense responses, seen by the shifts in the mRNA expression of the antimicrobial Reg3 lectins Reg3b and Reg3g, and phospholipase A2 (Pla2g2a), had normalized, but only in the slc26a3^−/− cecum and not in the distal colon (Figure 6A and B).

Effect of EcN Treatment on the Cecal and Fecal Microbiome in slc26a3^−/− Mice and slc26a3^+/+ Littermates

We and others have shown that the absence of Slc26a3 resulted in a significantly dysbiotic cecal and fecal microbiome with shifts in multiple taxonomic families, most importantly a decrease in SCFA-producing bacteria.^18,26,27 After 3 weeks of oral EcN application, the microbiome remained distinct from, and of lower diversity in the slc26a3^−/− compared to slc26a3^+/+ in the cecal luminal content and in the feces (Figure 7A–D). However, slc26a3^−/− mice showed a reduction in the mucolytic Deferribacteraceae and an increase in Prevotellaceae after EcN gavage (Figure 7E) in the cecal contents. In addition, a significant increase in the Lachnospiraceae and Ruminococcaceae was seen in the fecal content after EcN gavage (Figure 7F). This suggests that despite the fact that the EcN bacteria in the total microbiome were only 1%-2%, it did affect the microbiome composition. It is therefore likely that the change in SCFA concentration in the distal gut lumen is a combination of the SCFA production by EcN as well as of the other SCFA producers with a significant increase after gavage. The latter may be the key event in increasing overall SCFA production. Several classic SCFA-producing taxa in the human microbiome, such as Roseburia, Faecalibacterium, Coprococcus, Ruminococcus, and Oscillospira, were detected at very low levels or not at all, and their contribution to the observed changes in SCFA concentration in our mouse cohort is therefore unlikely.

Discussion

The present study elucidates the effect of an oral application of the probiotic EcN for 3 weeks on the low SCFA production and transporter expression, the mucus barrier dysregulation, the inflammatory phenotype, the strong upregulation of antimicrobial defense mechanisms, and the fluid and chloride absorption defect in the distal gut of slc26a3^−/− mice.

Patients with defective SLC26A3 develop CLD, which results in systemic alkalosis, acidic diarrhea, and a proinflammatory phenotype in the gut. These features are also seen in slc26a3^−/− mice.^18,28 In a previous publication, we had observed that the microbiome of the slc26a3^−/− mice had extremely low numbers of the major SCFA-producing families, both in the cecum and in the colon.²⁶ A similar decrease in SCFA producers was also seen in the feces of CLD patients¹² and is known to occur in patients with IBD.²⁹ Since SCFA anions reach high concentrations in the cecum and proximal colon and are major metabolites to drive fluid absorption in these gut segments,⁶ we assumed that the diarrheal phenotype of slc26a3^−/− mice, as well as of CLD and IBD patients, may be aggravated by a low SCFA concentration in the lumen of the colon. We therefore assessed the effect of acetate (as well as other SCFA entities, not shown here) on the FA capacity of the slc26a3^−/− and slc26a3^+/+ proximal-mid colonic mucosa in vivo and found that switching a luminal perfusion with an electrolyte solution containing NaCl to a solution containing 100 mM acetate in exchange for 100 mM Cl⁻ dramatically increased the FA rate in the slc26a3^−/− mice to almost the same levels as in the slc26a3^+/+ mice. HCO₃⁻ output rates were also stimulated, but there was still a significant difference between slc26a3^−/− and slc26a3^+/+ HCO₃⁻ output rates.

Thus, the challenge was to achieve the generation of acetate in the distal gut of the slc26a3^−/− mice by the application of an acetate-producing probiotic. We chose EcN as the bacterium for gavage for several reasons: Firstly, it is not known if probiotics can colonize the slc26a3^−/− intestine, and EcN offers the possibility to exactly determine the colonization of this bacterium in the gut lumen.²¹ Secondly, EcN is an avid producer of acetate,³⁰ and acetate effectively increased fluid absorption in WT mice with luminal concentrations as low as 10 mM (Figure 1, and Tan Qinghai, unpublished observations). Thirdly, there is a large body of literature related to the anti-inflammatory effects of EcN in the gut. EcN has been extensively studied for its biosafety aspects,³¹ given its notable absence of known pathogenic properties associated with other E. coli strains.^21,32 Instead, it exhibits direct antagonistic effects through colicins and microcins against other Enterobacteria.^33,34 While EcN has demonstrated its ability to produce SCFA in vitro,³¹ there is a lack of reports on its in vivo capabilities. Since we have found that SA can directly induce fluid absorption and HCO₃⁻ secretion in the colon, we hypothesized that the application of an acetate producer EcN could alleviate diarrhea in slc26a3^−/− mice but may also strengthen the weakened gut barrier and exert an anti-inflammatory effect in the distal gut of the slc26a3^−/− mice either by itself, or by a beneficial action on the gut microbiome and therefore by other SCFA producers.

Our primary objective was to establish a gavage strategy that would ensure the stable colonization of EcN within the gut, enabling consistent SCFA production. Quantification results from our study indicate the viability of EcN within the gut, with a higher concentration observed in the cecum. Due to the colonization resistance, most probiotics are excreted out of the colon with the stool after oral administration, thereby being undetectable.³⁵ And given the short transit time in the gut,²⁶ we were aware of potential difficulties for EcN to establish colonization. The cecum is the likely reservoir for live EcN and emerges as a logical site for prolonged EcN presence. A study has shown that EcN can concentrate in the cecum of chickens, protecting it from the challenge of Campylobacter jejuni, while the EcN treatment did not impact the evenness and richness of the cecal microbiota.³⁶

After we had established a stable colonization of the distal gut by EcN, the SCFA levels in the cecal content and in the feces of the slc26a3^−/− and slc26a3^+/+ mice were measured in non-gavaged mice (baseline) and after 3 weeks of EcN gavage. SCFA levels were indeed much higher in the cecal content and in the feces of the slc26a3^+/+ than in the slc26a3^−/− mice, showing that the previously observed reduced amount of bacterial species that were known to produce SCFA²⁶ did indeed impact the SCFA production in the cecum and feces of the slc26a3^−/− mice. In addition, the SCFA content was higher in the cecal content than in the feces, suggesting significant absorption as the luminal content passes through the distal intestine. EcN gavage had resulted in a significant increase in fecal content of total SCFAs and of acetate in pilot experiments with WT mice, as described in the literature, which we had interpreted as a direct effect of the EcN metabolism. However, significantly increased concentrations for acetate, propionate, and butyrate were observed in both the slc26a3^+/+ as well as slc26a3^−/− cecal content (Figure 3), suggesting that the increased SCFA production may be due to other alterations in the bacterial composition and not just by EcN alone.

Another interesting observation was the strongly reduced expression levels of the SCFA transporters MCT1, which is a proton-coupled organic anion transporter,³⁷ and SMCT1, a Na⁺-coupled organic anion transporter in the cecal mucosa of the mice. Interestingly, the cecal expression levels for both transporters increased after EcN gavage. In the distal colon, Mct1 expression levels were approx. 4-fold lower than that in the cecum, not different between slc26a3^−/− and slc26a3^+/+ mucosa. This suggests that either the luminal SCFA concentrations themselves or consequences of the EcN colonization upregulate the SCFA transporter expression. We may thus underestimate the actual SCFA production after EcN gavage.

The cecum of the slc26a3^−/− mice showed an increase in goblet cell mucus thecae count and MUC2 expression after EcN treatment, to almost the levels of the WT cecum (Figure 4). SCFA has been shown to upregulate Muc2 gene expression through enhanced acetylation of histone H3 and H4, along with increased methylation of H3 at the MUC2 promoter.⁷ The reported effects of EcN on mucin production are variable: In HT29 intestinal cells, EcN increased the mRNA and protein levels of multiple MUC proteins including MUC2. This effect was more distinct in the polarized cells in comparison to the nonpolarized cells.³⁸ It has also been speculated that EcN could influence the intestinal epithelial cell differentiation towards the goblet cell lineage in vitro and in vivo.³⁹ Another study investigated the influence of EcN on mucin gene expression in human intestinal cells and found that EcN did not significantly alter mucin gene expression in these cells.⁴⁰ We previously observed a reduced prominence and count of the goblet cell thecae in the slc26a3^−/− mid-distal colon, as well as altered Muc2 gene expression.¹⁸ Similarly, reduced goblet cell granules and Muc2 mRNA expression were found in the cecum of the slc26a3^−/− mice, and interestingly, significant increases in both parameters were seen after EcN gavage in the slc26a3^−/− but not the WT cecum.

As previously reported for the mid-distal colon,^18,26 the expression levels for a panel of proinflammatory cytokines were increased in the slc26a3^−/− cecum. After EcN gavage, the expression values were completely normalized in the cecal, but not in the distal colonic mucosa (Figure 5). Interestingly, the increased expression of Il-22, antimicrobial peptides Reg3b/c, and phospholipase 2g2a (Pla2g2a) was completely normalized in the cecum, while it remained high in the distal colon (Figure 6). This is an important finding, since a recent finding has demonstrated that inflammation self-perpetuates by causing aberrant antimicrobial activity that disrupts symbiotic relationships with gut microbes.⁴¹

The results show a significant anti-inflammatory effect of EcN in the part of the gut in which they preferably reside. We assume that the short GI transit time of the mice (Supplementary Figure 1D) does not permit a sufficient contact time between the EcN and the distal colonic mucosa for a local anti-inflammatory effect. It is surprising that EcN gavage did inhibit inflammation induced by DNBS in mice, which causes a distal colitis.⁴²

We also performed 16S sequencing of the total microbiome in order to find out if the presence of EcN changes the microbial composition. Recent reports described a large impact of EcN outer membrane vesicles, or intact EcN on the microbial diversity and on specific families in mice,^43,44 while such changes were not observed in chicken.⁴⁰ We did not observe a significant increase in the strongly reduced microbial diversity of the slc26a3^−/− cecal content and feces after EcN gavage, nor was the diversity significantly altered in the slc26a3^+/+ mice after EcN gavage. However, there were some potential meaningful changes: Firstly, the mucolytic Deferribacteraceae, which were significantly increased in the slc26a3^−/− cecum, decreased after EcN gavage. This may in part explain the increased cecal mucus content. Secondly, the abundance of the SCFA producers Lachnospiraceae and Ruminococcaceae was significantly higher in the feces after EcN gavage. This may explain why not just acetate, but also propionate and butyrate had increased in the cecal content and feces of the slc26a3^−/− mice, and to some extent in the WT mice.

We did not observe a decrease in stool water content in the slc26a3^−/− feces despite the increased SCFA levels (Supplementary Figure 3). The most likely explanation is that the EcN colonized predominantly in the cecum. While this is likely a major site of murine colonic SCFA absorption due to the high MCT1 expression levels, it is possible that the lack of Cl⁻ absorption via Slc26a3 is decisive for the water content in the feces, and that the SCFA concentration in the colonic content remains too low to be able to combat the lack of Slc26a3-mediated Cl⁻ absorption.

Lastly, do these findings help in paving the way to develop probiotic treatment for human colonic disease, given the fact that the EcN effects were seen only in the cecum of the slc26a3^−/− mice? The murine distal intestine differs functionally and morphologically from that of humans in several aspects: Firstly, mice have a very short GI transit time compared to humans. Secondly, they have a prominent cecum, with the highest rate of microbial SCFA production taking place in the cecum, while SCFA and fluid absorption is strongest in the proximal colon, possibly because of the high NHE3 expression, not found in the murine cecum. The firm pellets are formed during the short passage through the proximal colon and rapidly pass through the colon. In contrast, the human cecum is short, while the rectum has a large reservoir function. SCFA production and absorption likely occur in most of the human colon as long as the stool is soft. Therefore, a longer contact time with probiotic bacteria may occur in a higher percentage of the human compared to murine colonic mucosa. However, it is likely that different probiotics than EcN,⁴⁵ higher doses than currently suggested for EcN, genetically altered or optimally encapsulated probiota,^46,47 or a mixture of different strains⁴⁸ may be required for the prevention or treatment of colonic inflammation in humans.

Conclusions

A 3-week oral administration of EcN was able to induce stable colonization with live EcN in the inflamed slc26a3^−/− mice and resulted in an increase in SCFA-producing microbial species, an increase in SCFA levels in the slc26a3^−/− gut, a decrease in the inflammatory signature in the cecal mucosa, and an increase in the mucus content. While this was a proof of principle study, it gives rise to the hope that well-controlled probiotic treatment may decrease the high incidence of IBD in CLD patients, and may also be of benefit to patients with IBD, in particular for preventing new flares in the setting of residual inflammation.

Supplementary Data

Supplementary data are available at Inflammatory Bowel Diseases online.

Acknowledgments

We gratefully thank and acknowledge the help of all members of the Institute for Animal Research of the MHH for their help with the animal breeding and maintenance. Thanks to Katrin Künnemann for the preparation of the bacterial cultures. We also would like to thank Dr. Heike Bähre and Frank-Mathias Gutzki from Research Core Unit (RCU) Metabolomics, MHH, for their help with measuring the short-chain fatty acids in our samples.

Author Contributions

U.S. conceived and provided financial support for the study. A.K. and U.S. supervised the study. Z.Y., A.K., and Q.T. conducted the experiments, collected and analyzed data. S.W., X.T., D.R., G.A.G., and M.V. provided technical support and critical advice. Z.Y., A.K., and U.S. wrote the original manuscript. Z. Y. and U.S. reviewed and edited the article. All authors read and approved the final manuscript.

Funding

This work was funded by Deutsche Forschungsgemeinschaft (DFG) SE260/19-1, 21-1, and 22-1 and China Scholarship Council (CSC) stipend #202006160035 (to Z.Y.).

Conflicts of Interest

None declared.

Data Availability

The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary Materials. Raw data for the 16S rRNA sequencing were generated at the Institute for Medical Microbiology and Hospital Epidemiology, MHH, Hannover Germany. Sequence data are available at the European Nucleotide Archive (PRJEB71393). Derived data supporting the findings of this study are available from the corresponding author on request.

References