A Novel Probiotic Combination Ameliorates Crohn’s Disease–Like Ileitis by Increasing Short-Chain Fatty Acid Production and Modulating Essential Adaptive Immune Pathways

Abstract

BACKGROUND

Crohn’s disease (CD) represents a significant public health challenge. We identified a combination of beneficial probiotic strains (Saccharomyces boulardii, Lactobacillus rhamnosus, Lactobacillus acidophilus, and Bifidobacterium breve) and amylase that may antagonize elevated bacterial pathogens in the inflamed gut. Our aim was to characterize the effect(s) of this novel probiotic supplement in SAMP1/YitFc (SAMP) mice with CD-like ileitis.

METHODS

Three groups of 7-week-old SAMP mice were used in this study. The first experimental group was administered 1 dose of the probiotic supplement (probiotic strains + amylase) diluted in sterile phosphate-buffered saline (PBS) (0.25 mg in 100 µL of PBS) every day for 56 days through the gavage technique, the second group had a probiotic supplement (probiotic strains without amylase), and the third group was a control group in which animals were administered sterile PBS. At the end of the treatment, mice were sacrificed and ilea were collected for histological scoring of ileitis and NanoString analysis. Stool samples were evaluated by 16S ribosomal RNA and gas chromatography–mass spectrometry analyses.

RESULTS

Histology scores showed that mice treated with probiotics + amylase had a significant decrease of ileitis severity compared with the other 2 groups. 16S ribosomal RNA and gas chromatography–mass spectrometry analysis showed that abundance of species belonging to genus Lachnoclostridium and Mucispirillum schaedleri were significantly increased compared with the other 2 groups, and this increase was associated with augmented production of short-chain fatty acids. NanoString data showed that 21 genes involved in B memory cell development and T cell infiltration were significantly upregulated in probiotic-treated mice and that 3 genes were significantly downregulated.

CONCLUSIONS

Our data provide experimental proof for a beneficial effect of the designed probiotic formulation on the severity of CD-like ileitis in the SAMP mouse model, involving both alteration of intestinal genetic pathways and microbial rearrangements. Thus, we propose that this novel probiotic mixture should be further tested as an adjuvant therapy in the treatment of biofilm-associated disorders such as CD, in which it has been proven that polymicrobial imbalance plays a critical role in dysbiosis and gut inflammation.

KEY MESSAGES.

What is already known?

Dysbiosis of the gut microbiome has been implicated in inflammatory bowel diseases (IBDs). One potential strategy to prevent and treat IBD is gut microbiota modulation by probiotics.

What is new here?

We identified beneficial probiotic strains combined with the hydrolytic enzyme amylase that antagonize the proinflammatory activity of elevated bacterial pathogens. Probiotic-treated mice developed a less severe ileitis compared with phosphate-buffered saline–treated and probiotic without amylase–treated groups, demonstrating structural microbiome changes as well as increased short-chain fatty acid production.

How can this study help patient care?

These findings provide the rationale for clinical testing of this novel probiotic combination in patients with IBD.

Introduction

Inflammatory bowel diseases (IBDs) are gastrointestinal disorders that can cause chronic inflammation in any part of the gastrointestinal tract.¹ IBDs encompass 2 main forms: ulcerative colitis and Crohn’s disease (CD). While symptoms may differ from patient to patient, the most common symptoms caused by CD are abdominal cramps and pain, persistent diarrhea, constipation leading to bowel obstruction, and rectal bleeding.² Because a complete cure for CD is not available, the main therapies are directed at maintaining CD patients in a stage of remission by administration of steroids and biological therapies.³ The major problem associated with steroids and other prescribed immunosuppressants is the occurrence of multiple side effects including opportunistic infections.⁴ Hence, novel therapeutic approaches are needed to manage CD.

Several studies demonstrated that the microbiome (bacterial community also referred to as the bacteriome) and the mycobiome (fungal community) play a pivotal role in human health and disease.^5-7 Previous research in diabetic nonhealing foot ulcers demonstrated that bacteria belonging to the order Alcaligenaceae and the fungus Candida albicans formed polymicrobial biofilms (PMBs) in the wounds of affected patients.⁸ Moreover, PMBs involving bacterial phyla Fusobacteria and Bacteroidetes and fungi belonging to the phylum Zygomycota were also detected in patients affected by oral cancer.⁹ These associations provide evidence that the mycobiome is an important player in the pathogenesis of several diseases, and that these 2 polymicrobial communities (ie, bacterial and fungal) interact together and negatively impact the host, as these microorganisms produce extracellular enzymes that cause tissue damage, resulting in an increase in proinflammatory cytokines, culminating in oxidative damage and apoptotic cell death.¹⁰

Recently, Ghannoum et al¹⁰ demonstrated that bacterial (Escherichia coli, Serratia marcescens) and fungal (C. tropicalis) abundance of these pathogens was significantly elevated in CD patients, and enabled thicker PMBs. Polymicrobial culture of these species formed more robust biofilms compared with single-species biofilms,¹¹ and the combination biofilms is capable of inducing damage to the intestinal epithelial cells provoking an inflammatory response.¹²

Synergistic interaction of bacterial and fungal microorganisms in the dysbiotic state often leads to biofilm formation in the intestine (referred to as digestive plaque), which has been shown to produce detrimental effects on the host. Therefore, it is logical to suggest that inhibition of digestive biofilms formed by microbial pathogens may have a beneficial effect on the host. One approach to inhibiting these biofilms and the consequent dysbiosis could be through the use of probiotics (defined by the World Health Organization as microorganisms capable of improving the health of the host when administered in suitable amounts).¹³ Multiple studies demonstrated that probiotics use can stimulate the host immune system and consequently ameliorate gastrointestinal symptoms. For instance, bacterial species belonging to the genus Bifidobacterium have been proven to increase the production of immunoglobulin A by mesenteric lymph nodes,¹⁴ which is involved in immune exclusion of extraneous antigens by inhibiting binding to the epithelial cells and consequent penetration of pathogenic bacteria and fungi.¹⁵ Furthermore, Galdeano et al¹⁶ demonstrated that Lactobacillus casei stimulated gut-associated immune cells causing increase in T and immunoglobulin A⁺ B lymphocytes, mucus-producing goblet cells, and macrophages. To date, very few studies have investigated the effect of probiotics on IBDs.¹⁷ In order to identify effective probiotic strains that are capable of reducing intestinal dysbiosis caused by E. coli, S. marcescens, and C. tropicalis found elevated in CD patients, Hager et al¹⁸ conducted a series of bacterial-bacterial, fungal-fungal, and bacterial-fungal correlation analyses using microbiome data of individuals with CD and their healthy relatives and developed a novel probiotic blend composed of bacteria and yeast strains. Specifically, the novel probiotic mix consists of Saccharomyces boulardii, Lactobacillus rhamnosus, L. acidophilus, and B. breve. To ensure that this nutritional supplement is effective against PMBs formed by these pathogens, amylase, an hydrolytic enzyme that possesses anti-biofilm activity,¹⁹ was incorporated as an ingredient in this formulation. Next, they evaluated the effect of this formulation using an in vitro biofilm assay and showed that the probiotic combination prevented PMBs formed by C. tropicalis, S. marcescens, and E. coli.¹⁸ Although the probiotic blend has been shown to be effective against PMBs in vitro, its in vivo effect in an animal experimental IBD model has not been performed.

Based on its in vitro effect, we hypothesized that administration of this novel probiotic combination may alter the intestinal microbiome of the SAMP1/YitFc (SAMP) mice and that the resulting microbial modification may benefit the host in terms of reduced intestinal inflammation. Hence, in this study we investigated how the probiotic administration affects the severity of ileitis in SAMP mice. The SAMP mouse model is a well-established model of CD-like ileitis that spontaneously develops inflammation of the small intestine after 10 weeks of age that significantly resembles human disease pathology.²⁰ Thus, SAMP mice represent a valuable tool for developing and evaluating novel IBD therapies. Additionally, to determine whether consumption of the probiotic formulation could lead to significant microbial alterations in mice, we characterized the microbiome of mice before and after probiotic administration using 16S ribosomal RNA (rRNA) analysis.

We report here that SAMP mice treated with the probiotic mix had significant attenuation of ileal disease severity compared with phosphate-buffered saline (PBS)–treated control animals. Moreover, 16S rRNA analysis indicated that probiotic consumption led to increased α- and β-diversity of the gut microbiome compared with PBS-treated mice, and that bacteria belonging to the genus Lachnoclostridium and Mucispirillum schaedleri were significantly more abundant in the intestine of treated mice compared with the control animals. Furthermore, gas chromatography–mass spectrometry (GC/MS) analysis demonstrated that several short-chain fatty acids (SCFAs) were significantly increased after probiotic administration, including butyric acid, valeric acid, hexanoic acid, and heptanoic acid. Additionally, we show that probiotic administration affects the expression of several genes involved in multiple cellular mechanisms, such as B cell development, lymphocytic infiltration, and immune response to bacterial infection.

Finally, in order to provide a mechanistic explanation of the reduced level of ileitis following probiotic consumption, we show that SAMP mice treated with the probiotic combination in the absence of the hydrolytic enzyme amylase do not exhibit reduced intestinal inflammation compared with the control animals. This observation suggests that combining amylase with the probiotic strains is critical for the amelioration of digestive inflammatory symptoms.

In summary, this study highlights the beneficial effect of the designed probiotic formulation on the severity of CD-like ileitis in the SAMP mouse model, involving both changes related to intestinal genetic pathways and microbial alterations. These findings provide the rationale for clinical testing of this novel probiotic combination as an adjuvant therapy in the treatment of IBD patients.

Methods

Experimental animals

SAMP mice (species: Mus musculus) were propagated in the Animal Resource Center at Case Western Reserve University. All mice used in this study were 7 weeks of age and were age and sex matched among experimental groups.

Experimental mice were housed in ventilated micro-isolator cages (Allentown Inc) with cotton nestlets for environmental enrichment (Envigo) and 1/8-inch corn bedding, and kept on 12-hour light/dark cycles.

All mice had ad libitum access to water and standard laboratory rodent diet P3000 (Harlan Teklad) throughout the experiments. Harem breeding was set up by co-housing one 8-week-old male with two 8-week-old females. All procedures were approved by the Case Western Reserve University Institutional Animal Care and Use Committee and were in accordance with the Association for Assessment and Accreditation of Laboratory Animal Care guidelines. All experiments were conducted in a blinded manner, without prior knowledge of treatments and mouse groups by the experimenter. Mice were randomized to different interventions using a progressive numerical number. The code for each mouse was known only to the animal caretaker and was revealed at the end of the study.

Test materials

The test materials including the probiotic with and without amylase were provided by BIOHM Health, LLC.

Homogenization of the fecal pellets

Because recent studies have called attention to a concerning variability related to microbiome results between different laboratories,²¹ we followed a very conscientious approach to ensure the accuracy of our microbiome data and to ensure homogeneous microbial composition at baseline: in order to control for cage-to-cage variability that could be attributable to gut microbiome changes, corn bedding from each cage was collected and mixed for 3 minutes, and then an equal amount of pellet was redistributed in each experimental cage a week before the probiotics administration. At the same time, 20 mg of fecal pellets were collected from each mouse in sterile PBS. All the pellets were homogenized in PBS using glass beads of 0.5-mm diameter (catalog number: 11079105; Biospec Products) and of 1.0-mm diameter (catalog number: 11079110; Biospec Products). The fecal homogenate was then gavaged (200 µL/ mouse for 3 times in 3 consecutive days [once per day]).

Histology

Ileums from probiotic-treated and PBS-treated mice were removed, flushed of fecal contents, opened longitudinally, and placed in Bouin’s fixative. Tissues were embedded in paraffin and stained with hematoxylin and eosin. Inflammation was evaluated by a trained pathologist in a blinded fashion using a previously described scoring system.²² Briefly, scores ranging from 0 (normal histology) to 3 (maximum severity of histologic changes) were used to evaluate 4 individual histologic subindices for (1) active inflammation (infiltration with neutrophils), (2) chronic inflammation (lymphocytes, plasma cells, and macrophages in the mucosa and submucosa), (3) villus distortion, and (4) percent ulceration.

Stereomicroscopy

Stereomicroscopic analysis and pattern profiling serve to quantify the intestinal health in experimental mouse models and to determine the extent of mucosal inflamed areas that occurs with chronic inflammation. To analyze the patterns of stereomicroscopic abnormalities, micrographs obtained with a JEOL microscope were analyzed using ImageJ public domain software (Version 1.53t; National Institutes of Health). We measured single abnormal mucosal lesions and plotted the areas for visualization and circular dimension analysis.

Microbiome analyses

After 3 days of fecal homogenization, 1 fecal pellet from each mouse was collected and stored at -80 °C. A week after the fecal homogenization, 1 group of SAMP mice was administered with 1 dose of probiotic diluted in sterile PBS (0.25 mg in 100 µL of PBS) every day for 60 days through gavage technique. The control group received sterile PBS every day for 60 days through gavage technique. A second sample of fecal pellets from each mouse was then collected at the end of the treatment. Stool samples were analyzed by 16S rRNA gene amplicon sequencing investigating the gut microbiome. Metagenomic DNA was extracted from fecal aliquots thawed on ice and resuspended in 600 mL DNA stabilization buffer (Stratec Biomedical) and 400 mL phenol/chloroform/isoamyl alcohol (25:24:1, by volume; Sigma-Aldrich). Cells were mechanically lysed (3 × 6.5 m/s for 40 seconds) with 500 mg 0.1-mm glass beads (Roth) using a bead-beater (MP Biomedicals) fitted with a cooling adapter. After heat treatment (95 °C, 8 minutes) and centrifugation (16 000 at 3 g; 5 minutes; 4°C), 150 mL of supernatant was incubated with 15 mL of ribonuclease (0.1 mg/mL; Amresco) at 37°C and centrifuged (550 at 3 g; 30 minutes). DNA was purified with the NucleoSpin gDNA Clean-up Kit (Macherey-Nagel), following the manufacturer’s instructions. Concentrations and purity were determined with the NanoDrop and Qubit system (Thermo Fisher Scientific) and stored at -20 °C. Preparation of amplicon libraries (V3–V4 region) and sequencing was performed as described in detail previously.²³ After purification with the AMPure XP system, sequencing was carried out in paired-end mode (275 bp) with pooled samples containing 25% (vol:vol) PhiX standard library in a MiSeq system (Illumina Inc) prepared according to the manufacturer’s instructions. Raw reads were processed with the Integrated Microbial Next Generation Sequencing pipeline, based on the UPARSE approach.²⁴ In brief, sequences were demultiplexed, trimmed to the first base with a quality score < 3, and then paired. Sequences with <350 and >500 nucleotides and paired reads with an expected error >3 were excluded from the analysis. Remaining reads were trimmed by 10 nucleotides on each end to prevent analysis of the regions with distorted base composition observed at the start of sequences. The presence of chimeras was tested with UCHIME. Operational taxonomic units were clustered at 97% sequence similarity, and only those with a relative abundance >0.5% in ≥1 sample were kept. Taxonomies were assigned at 80% confidence level using the RDP classifier, the SILVA database applying SINA, and EzBioCloud. Bray-Curtis dissimilarity between microbial communities were visualized using the CosmosID-HUB application. A principal coordinate analysis (PCoA) plot was used for visualization of the data in the distance matrix in a 2-dimensional plot. β-diversity indices were obtained through permutational multivariate analysis of variance test, which is based on the prior calculation of the distance between the cohorts included to the PCoA of β-diversity. The F test was used to compare within-cohort to between-cohort variance. A P value ≤.05 means that the spread of the samples is different between the cohorts.

GC/MS analysis

SCFAs were extracted from mouse feces and analyzed through the GC/MS technique as previously described.²⁵ Briefly, 50 mg of mouse feces were collected in a 1.5-mL microfuge tube containing 300 μL of water. Two 3.2-mm stainless beads were then added in each tube. Feces were then homogenized with a bullet blender using MP homogenizer (MP Biomedicals). Samples were then centrifuged at 14 000 g for 10 minutes. The supernatant was then transferred to a new 1.5-mL microfuge tube. For calibration samples, 250 μL of acetone, 10 μL and 1 μg/mL of IS solution, and 10 μL of the calibration series SCFA standards were added. A total of 100 μL of 172 mM PFBBr in acetone was added to each sample. Samples were then heated at 60 °C for 30 minutes in laboratory stove. A total of 500 μL of n-hexane and 250 μL of water were then added to each sample. Next, 250 μL of the n-hexane (upper layer) was transferred into a new empty glass autosampler vial with a glass insert for every sample. A total of 1 μL of each sample was injected in the GC/MS (5977B GC/MSD with a VF-5 ms column [25 m, 0.25 mm, 0.25 µm; Agilent; Cat#: CP8941]), spitless at 280 °C. Helium was used as carrier gas at a constant flow rate of 1.20 mL/min. The following temperature gradient was used: 1 minute at 40 °C, linear increase at 40 °C per minute to 60 °C, held for 30 minutes at 60 °C, linear increase at 25 °C per minute to 210 °C, linear increase at 40 °C per minute to 315 °C, and held for 3 minutes at 315 °C. The transfer line temperature was set at 280 °C. Methane was used as chemical ionization gas at approximately 15 psi. Ions obtained were detected in the negative mode using SIM, the obtained signals were then integrated, and the relative retention time and area ratios were calculated using the respective IS. Next, the slope and the LLOQ for every SCFA were determined by performing linear regression. Finally, the SCFA concentrations were calculated by using the area ratios obtained from the biological samples and the slopes obtained from the analysis of the calibration series samples.

NanoString gene expression analysis

Total RNA was extracted from ileums of probiotic-treated mice at the end of the treatment. Extracted RNA was incubated with a custom panel of 795 bar-coded probes (NanoString Technologies) specific for genes associated with immune function and inflammation (NanoString Technologies). Genes were selected from KEGG pathways for chemokine signaling (04062), B cell receptor signaling (04662), T helper 1 and T helper 2 cell differentiation (04658), IBD (05321), systemic lupus erythematosus (05322), and other pathways. NanoString technology is characterized by high reproducibility and sensitivity. As criteria of differential expression, we used P ≤ .05. Six biological replicates per condition were evaluated. Standard NanoString protocols were followed. Reporter probes, hybridization solution, sample, and capture probes were mixed together and hybridized overnight at 65 °C. After hybridization, samples were transferred and processed in the NanoString nCounter Prep Station. The Prep Station washed away excess probes and purified the target/probe complexes using magnetic beads. Briefly, the hybridization solution containing the target/probe complexes is mixed with magnetic beads complementary to sequences on the capture probe. This process is followed by a washing step to remove the excess reporter probes. The target/probe complexes then were hybridized to magnetic beads complementary to sequences on the reporter probe. A final washing step was performed to remove excess capture probes. The purified target/probe complexes were deposited in a cartridge, laid flat, and immobilized for data collection. Data were analyzed by ROSALIND (https://rosalind.onramp.bio/), with a HyperScale architecture developed by OnRamp BioInformatics, Inc. Read Distribution percentages, violin plots, identity heatmaps, and sample Multidimensional scaling plots were generated as part of the quality control step. Normalization was done by dividing counts within a lane by the geometric mean of the normalizer probes from the same lane. The NormqPCR R library²⁶ was used to select normalizer probes using the geNorm algorithm. Fold changes and P values were calculated using criteria provided by NanoString. Clustering of genes for the final heatmap of differentially expressed genes was done using the PAM (Partitioning Around Medoids) method using the fpc R library² that takes into consideration the direction and type of all signals on a pathway, the position, role and type of every gene, etc. Hypergeometric distribution was used to analyze the enrichment of pathways, gene ontology, domain structure, and other ontologies. The topGO R library was used to determine local similarities and dependencies between Gene Ontology terms in order to perform Elim pruning correction. Several database sources were referenced for enrichment analysis, including Interpro, National Center for Biotechnology Information, MSigDB, REACTOME, and WikiPathways. Enrichment was calculated relative to a set of background genes relevant for the experiment.

Microbial function prediction through 16S rRNA gene sequences

Predictive microbial functional profile was generated using the PICRUSt software, which infers the pathway content of the microbiome by assigning bacterial functional genes for the operational taxonomic units, normalized for 16S rRNA gene copy number, using the MetaCyc database.²⁷ Functional enrichment analysis and visualization plots were generated using the CosmosID-HUB application. Differences in the relative abundance of microbial features were determined by linear discriminant analysis (LDA) effect size using the CosmosID-HUB Web application (https://app.cosmosid.com). Specifically, the nonparametric factorial Kruskal-Wallis sum-rank test was used to detect features with significant differential abundance with respect to the class of interest. As a final step, LDA was used to estimate the effect size of each differentially abundant feature and rank the feature accordingly. A logarithmic LDA score ≥2.0 was used as a threshold for nonparametric factorial Kruskal-Wallis test, which was used to judge whether a test statistic was statistically significant when comparing the 2 cohorts.

Statistical analysis

Experiments were conducted at a minimum in duplicate. Univariate and multivariate analyses were conducted using the collective data from replicated experiments. When the data fulfilled the assumptions for parametric statistics, comparisons of continuous data across experimental groups were conducted using Student’s t tests. Alternative nonparametric tests were used for data with unfulfilled distribution assumptions regarding the normality of the data. Data were expressed as SEM, and 95% confidence intervals were reported when appropriate. An alpha level of 0.05 was considered significant. All analyses were conducted with GraphPad software Version 9.3.1 (350).

All authors had access to the study data and had reviewed and approved the final manuscript.

Study approval

All experiments were approved by the Institutional Animal Care and Use Committee at Case Western Reserve University and conducted following the Association for Assessment and Accreditation of Laboratory Animal Care guidelines.

Results

Oral administration of the probiotic blend decreased spontaneous ileitis in SAMP mice

We hypothesized that oral administration of the aforementioned probiotic mix would alter susceptibility to spontaneous ileitis in SAMP mice. We tested this hypothesis by analyzing the severity of ileitis in 7-week-old SAMP mice that were treated with either oral gavage of the probiotic mix (experimental group) or PBS (control group) for 56 days. Histological analysis of mice ilea demonstrated that probiotic-treated mice had significant attenuation of ileal disease severity compared with control animals (P < .001). In this regard, decreased active inflammation, villous distortion, mononuclear inflammation, and transmural inflammation were observed (Figure 1A and 1B).

Figure 1.

Oral administration of the probiotic blend decreased spontaneous ileitis in SAMP1/YitFc mice A, Histological analysis shows decreased ileal inflammation in probiotic-treated mice compared with the control group (unpaired t test, 8.83 ± 0.48 vs 13.51 ± 0.65; P < .0001; n ≥ 40 per group). B, Representative histopathological sections of hematoxylin and eosin–stained ilea show improved villous architecture in probiotic-treated mice (left) compared with the control group (right) and increased immune cell infiltrates (highlighted in the images obtained at 40× magnification) into mucosal layers of control animals compared with the probiotic-treated mice. C, Affected areas expressed as a percentage of the whole ileum show that probiotic-treated mice develop significantly less cobblestone ulcers in comparison with control mice (unpaired t test, 24.14 ± 1.82 vs 30.57 ± 1.77; P < .05; n = 7 per group). D, Images from stereomicroscopic analysis of the distal part of ileum show a decreased number of affected areas (indicated by yellow arrows) in comparison with unaffected areas (indicated by black arrows) in the probiotic-treated mice compared with the control group. Imaging equipment used: Zeiss Axiophot/Axiocam; JEOL microscope. Data are shown as mean ± SEM and are representative of 3 independent experiments; *P < .05, ***P < .001.

These histological results were corroborated by stereomicroscopy analysis, which was employed to evaluate the phenotype of the intestinal surface.²⁸ This analysis revealed a potent reduction of ileal disease severity in probiotic-treated mice compared with control animals. Probiotic administration led to a healthier 3-dimensional ileal structure and a reduced distribution of abnormal mucosa evaluated by cobblestone ulcers (P < .05) (Figure 1C and 1D). Overall, these data indicate that the probiotic administration has an anti-inflammatory role during chronic ileitis in the SAMP mouse model.

Consumption of the probiotic and amylase blend leads to discrete alterations of the mouse microbiome

The microbiome structure of the 4 cohorts was analyzed by PCoA (Figure 2A). Various degrees of overlap were observed for the cohort clusters. In addition, we calculated the Bray-Curtis dissimilarities from different cohorts. Pairwise permutational multivariate analysis of variance analysis of β-diversity indices showed that the microbiome of the probiotic-treated mice post–probiotic administration was significantly different compared with the microbiome of the same mice before the treatment (F statistic: 2.954; P = .005), with the microbiome of the control group before the PBS administration (F statistic: 3.876; P = .020), and with the microbiome of the control group at the end of the PBS administration (F statistic: 2.635; P = .041). Interestingly, no significant difference was observed between the microbiome of the 2 groups at day 1 of the experiment, or between the microbiomes of the control group before and after the PBS administration (Table 1). Furthermore, the richness of the bacteriome in fecal samples of probiotic-treated mice was significantly higher compared with the richness in the control group posttreatment (P < .02) (Figure 2B). These data demonstrate that probiotic consumption increased the α-diversity in the gut of treated mice. Therefore, comparison of the gut microbiome within the probiotic-treated and PBS-treated groups may provide insights on organisms and dysbiosis linked to IBDs. As expected, no differences in α-diversity were seen between the 2 groups at baseline (ie, before gavage), indicating that the observed diversity by the end of the probiotic treatment is a specific effect of the probiotic administration. Thus, in subsequent studies, we performed 16S rRNA analysis to depict the microbiome differences between the 2 groups of mice. Our data show that probiotic administration was associated with specific alterations of bacterial abundance at the genus and species levels. Specifically, the abundance of species belonging to the genus Lachnoclostridium was significantly increased after probiotic administration compared with the control group (P < .02) (Figure 2C). Moreover, abundance of M. schaedleri was also significantly higher after probiotic administration compared with the control group (P ≤ .05) (Figure 2D). In contrast, abundance of bacterial species belonging to the family Lactobacillaceae were significantly decreased following probiotic administration compared with the control group (P < .05) (Figure 2E). Finally, abundance of Selenomonas lacticifex was significantly decreased after probiotic administration compared with the control group (P ≤ .05) (Figure 2F). Taken together, these data indicate that probiotic blend administration leads to increased α-diversity and β-diversity, significantly affecting the abundance of specific bacterial populations.

Figure 2.

Probiotic administration induces discrete alteration of the gut microbiome. A, Principal coordinate analysis (PCoA) based on the structure of the microbiome in the 4 cohorts was calculated using Bray-Curtis distances with a multivariate t distribution. Each data point represents a single mouse. Ellipses represent the 95% confidence level. Color is indicative of cohort. B, The α-diversity was significantly higher in the probiotic-administered group at the end of the treatment compared with the same group before probiotic administration (Simpson’s Diversity Index: 0.83 ± 0.04 vs 0.89 ± 0.03; P < .02; n = 6 per group). On the contrary, there was no difference in the α-diversity in the phosphate-buffered saline–treated control group (Simpson’s Diversity Index: 0.84 ± 0.04 vs 0.86 ± 0.03; P = ns). C, 16S ribosomal RNA analysis of the percentage of bacterial species in the feces of the probiotic-treated vs phosphate-buffered saline–treated mice showed that abundance of species belonging to the genus Lachnoclostridium was significantly increased after probiotic administration compared with the control group (unpaired t test, percentage gut microbiome: 0.08 ± 0.02 vs 0.30 + 0.10; P < .02). D, Abundance of Mucispirillum schaedleri was significantly higher after probiotic administration compared with the control group (unpaired t test, percentage gut microbiome: 0.13 ± 0.07 vs 0.48 + 0.18; P ≤ .05). E, Abundance of bacterial species belonging to the family Lactobacillaceae was significantly decreased after probiotic administration compared with the control group (unpaired t test, percentage gut microbiome: 0.35 ± 0.05 vs 0.16 + 0.05; P < .05). F, Abundance of Selenomonas lacticifex was significantly decreased after probiotic administration compared with the control group (unpaired t test, percentage gut microbiome: 0.50 ± 0.08 vs 0.24 + 0.08; P ≤ .05). Data are shown as mean ± SEM and are representative of 3 independent experiments; *P ≤ .05, **P < .02. OTU, operational taxonomic units; PC, principal component; 2D, 2-dimensional.

Table 1.

β-diversity indices between cohorts.

Cohort	Permutations	Statistic	P Value
Probiotic day 1–control day 1	999	1.282	.237
Probiotic day 1–probiotic day 56	999	3.876	.02
Probiotic day 1–control day 56	999	2.227	.07
Control day 1–probiotic day 56	999	4.945	.011
Control day 1–control day 56	999	2.888	.062
Probiotic day 56–control day 56	999	2.635	.041

A statistical difference was present only between the probiotic-treated cohort at the end of the treatment compared with the other 3 cohorts examined.

Probiotic administration increases production of SCFAs by gut microbiome

Next, to gain insight into the mechanisms by which the probiotic blend positively affected treated animals, we examined the metabolic basis for the described anti-inflammatory effect of the probiotics during chronic ileitis. We measured SCFAs in the feces and in the cecum content collected from posttreatment probiotic- and PBS-treated mice. Our data showed a significant increase of butyric acid (P < .02) (Figure 3A), valeric acid (P < .02) (Figure 3B), and hexanoic acid (P < .05) (Figure 3C) in the fecal samples of the probiotic-treated mice compared with the PBS-treated control animals. Also, analysis of the cecum content revealed an increase of valeric acid (P < .05) (Figure 3D) and heptanoic acid (P < .05) (Figure 3E) in probiotic-treated mice compared with the control group. Overall, these results indicate that the microbiome shift caused by the probiotic administration led to significant alterations related to SCFAs production.

Figure 3.

Probiotic administration leads to increased production of short-chain fatty acids by the gut microbiome. Gas chromatography–mass spectrometry analysis shows increased levels of (A) butyric acid (unpaired t test: 2581 ± 1959 vs 708 ± 455; P < .02), (B) valeric acid (unpaired t test: 79.00 ± 55.40 vs 34.83 ± 21.30; P < .02), and (C) hexanoic acid (unpaired t test: 15.38 ± 2.50 vs 12.85 ± 3.05; P < .05) in fecal samples of probiotic-treated mice compared with the control group. Levels of (D) valeric acid (unpaired t test: 259.00 ± 96.37 vs 155.50 ± 80.74; P < .05) and (E) heptanoic acid (unpaired t test: 56.00 ± 63.48 vs 30.62 ± 43.52; P < .05) were also increased in cecum samples of probiotic-treated mice compared with the control group. Gas chromatography–mass spectrometry equipment used was 5977B GC/MSD with a VF-5 ms column (25 m, 0.25 mm, 0.25 µm; Agilent; Cat#: CP8941). Data are shown as mean ± SEM and are representative of 2 independent experiments; *P < .05, **P < .02.

Probiotic administration increased the expression of genes involved in immune functions

In order to address mechanisms associated with decreased level of ileitis and potential immunologic alterations in probiotic-treated mice, we used 2 Mouse NanoString Panels focused on 795 genes affecting inflammatory and immunological pathways (NanoString Technologies). Figure 4A provides a visualization of the captured data in a heatmap chart. This plot shows the clustering of genes significantly decreased or increased in probiotic-treated mice compared with the PBS-treated control animals. We used PBS-treated mice to define baseline gene expression and observed significant differences compared with probiotic-treated mice. Specifically, 21 genes were significantly upregulated in probiotic-treated mice and 3 were downregulated (fold change >±2; P < .05).

Figure 4.

Probiotic administration is associated with immune system dysfunction. A, Heatmap expressing NanoString data for 795 genes was obtained with ROSALIND software and showing that the ileums of probiotic-treated mice had 21 genes significantly upregulated and 3 genes significantly downregulated compared with control mice (n = 6 per group). B, Metabolic pathways enrichment analysis based on MetaCyc database between the probiotic-treated group and control group at the end of the treatment (linear discriminant analysis threshold >2; P < .05). Increased expression of (C) Cr2 (unpaired t test: 6.05-fold vs 3.61; P < .05; n = 6 per group), (D) Aicda (unpaired t test: 5.72-fold vs 3.53; P < .05), (E) Cd19 (unpaired t test: 7.72-fold vs 5.42-fold; P < .05), (F) Cxcl13 (unpaired t test: 8.72-fold vs 6.07-fold; P < .05), (G) Ccr7 (unpaired t test: 5.93-fold vs 4.97-fold; P ≤ .05), and (H) Ccl19 (unpaired t test: 6.05-fold vs 3.62-fold; P < .05) in probiotic-treated mice compared with control animals. Decreased expression of (I) Cxcr4 (unpaired t test: 7.76-fold vs 6.83-fold; P ≤ .02) and (L) Tnfrsf13c (unpaired t test: 5.35-fold vs 3.97; P < .05) in probiotic-treated mice compared with control animals. Data are shown as mean ± SEM and are representative of 3 independent experiments; *P ≤ .05, **P ≤ .02. C, control; P, probiotic.

Furthermore, to study the possible role of transcriptionally active microorganisms in the gut, we used the computational approach PICRUSt to predict the functional composition of the gut microbiome, and subsequently we derived MetaCyc pathways differently expressed between the probiotic-treated group and the control group (Figure 4B). Functional sorting of the 24 genes that were different between probiotic-treated and PBS-treated mice showed that the majority belonged to families of genes involved in B cell development, increased susceptibility to bacterial infection, and lymphocytic infiltration, including Cr2, Aicda, Cd19, Cxcl13, Ccr7, Ccl19, Cxcr4, and Tnfrsf13c (Figure 4C-4L). In summary, these data indicate that the probiotic administration led to significantly altered expression of genes involved in multiple immune functions.

Removing amylase from the probiotic combination negates its ameliorating effect on ileitis

In order to confirm our hypothesis that the amylase enzyme present in the probiotic combination is necessary to ameliorate gastrointestinal manifestations, we dosed SAMP mice with the probiotic combination without amylase and compared them with control animals (PBS treated). Histological analysis showed that there was no significant difference in ileitis between mice dosed with probiotics without amylase and the control group (P = ns) (Figure 5A and 5B). Additionally, PCoA showed no difference in terms of α-diversity in fecal samples collected from mice treated with probiotic without amylase compared with PBS-treated control animals (P = ns) (Figure 5C). Moreover, microbiome profiling showed no differences in the abundance of species belonging to the genus Lachnoclostridium and the species M. schaedleri (P = ns) in mice treated with probiotics without amylase compared with the PBS-treated group (Figure 5D and 5E). Next, we measured SCFAs in the feces collected from mice dosed with probiotics without amylase vs PBS-treated mice. Our data showed no significant difference related to butyric acid, valeric acid, hexanoic acid, and heptanoic acid (P = ns) between the 2 groups (Figure 5F-5I). Finally, NanoString analysis on messenger RNA extracted from the ileum of the 2 experimental groups (probiotic without amylase and PBS) showed a dramatic reduction in the number of genes altered in animals treated with the probiotic blend without amylase, compared with the number of genes altered after administration of the probiotic combination that contained amylase (previously shown in Figure 4A). Specifically, only 1 gene was significantly upregulated (Cd36), and 4 genes were downregulated compared with the control group (P < .05) (Figure 5L). Interestingly, sorting of the 5 genes that were different between mice treated with probiotics without amylase and PBS-treated mice showed no functional correlation with families of genes involved in B cell development, increased susceptibility to bacterial infection, and lymphocytic infiltration (as was previously shown after complete probiotic administration).

Figure 5.

Removing amylase from the probiotic negates the ameliorating effect on ileitis. A, Histologic analysis shows no significant differences between mice treated with the probiotic minus amylase mix and the control group (unpaired t test, 13.30 ± 1.29 vs 14.54 ± 1.16; P = ns; n ≥ 10 per group). B, Representative histopathological sections of hematoxylin and eosin–stained ilea show no statistical difference in terms of villous architecture, number of immune cell infiltrates, and thickness of the muscularis layer. C, No differences related to the α-diversity were found in the amylase negative probiotic-treated group before and after the probiotic administration compared with the control group (Simpson’s Diversity Index: 0.86 ± 0.03 vs 0.84 ± 0.04; P = ns). D, 16S ribosomal RNA analysis of the percentage of bacterial species in the feces of the amylase negative probiotic-treated vs phosphate-buffered saline–treated mice showed no difference in the abundance of species belonging to the genus Lachnoclostridium compared with the control group (unpaired t test, percentage gut microbiome: 0.18 ± 0.03 vs 0.21 + 0.11; P = ns). E, Abundance of Mucispirillum schaedleri was non significantly different after amylase negative probiotic administration compared with the control group (unpaired t test, percentage gut microbiome: 0.19 ± 0.10 vs 0.14 + 0.13; P = ns). Gas chromatography–mass spectrometry analysis shows no significant differences in levels of (F) butyric acid (unpaired t test: 157.7 ± 44.3 vs 140.8 ± 58.4; P = ns; n ≥ 5 per group), (G) valeric acid (unpaired t test: 124.2 ± 6.3 vs 105.8 ± 17.49; P = ns), (H) hexanoic acid (unpaired t test: 6.6 ± 0.4 vs 5.4 ± 0.9; P = ns), and (I) heptanoic acid (unpaired t test: 35.8 ± 0.2 vs 27.8 ± 4.8; P = ns) in fecal samples of mice treated with probiotic blend without amylase compared with the control group. L, Heatmap expressing NanoString data for 795 genes was obtained with ROSALIND software and showed that ileums of mice dosed with probiotics without amylase have 1 gene upregulated and 4 genes downregulated compared with control mice (n ≥ 6 per group). Data are shown as mean ± SEM and are representative of 3 independent experiments; P = ns. C, control; P, probiotic; PwA, probiotic without amylase.

Taken together, these data indicate that the critical mechanism for the anti-inflammatory effect of the probiotic combination during chronic ileitis is based on having a product that contains probiotic strains combined with amylase. This combination is likely to possess anti-biofilm activity that protects the gut lining from the detrimental effects caused by bacterial and fungal pathogens such as leaky gut.

Discussion

In this study, we report that consumption of a novel probiotic blend reduced intestinal inflammation in vivo in SAMP mice. Specifically, when dosed with the probiotic mix, SAMP mice exhibited significantly less villous distortion, less transmural inflammation, and a reduced number of mucosal surfaces affected by cobblestones compared with the control group. Furthermore, we show that an optimal outcome was achieved via the probiotics (with amylase) due to its antibiofilm activity. Several studies focused on the interactions of microbiome communities have shown that microbial members interact and develop interactive cooperative strategies that often lead to biofilm formation.²⁹ During dysbiosis, biofilms create a favorable environment for fungal and bacterial communities, as both increase their tolerance to antimicrobial treatments as a result of residing in a biofilm environment, in which ability to avoid immune cells and attenuated penetration of antimicrobials have been observed, compared with a planktonic setting.³⁰ Moreover, biofilms negatively affect the host, as bacteria and fungi increase production of enzymes that can cause epithelial tissue damage, which as a consequence leads apoptosis and oxidative damage. Moreover, Candida changes its morphology from yeast to hyphal form (a known virulence factor), allowing it to cause damage to the gut epithelial lining. Because of the aforementioned tendency of intestinal microbes to form biofilms, we tested whether addition of amylase, an enzyme previously reported to disrupt biofilms, would enhance the efficacy of the probiotic mixture.

In this study, we report that disruption of the biofilm matrix provides access to the probiotic strains that subsequently inhibit the growth of pathogenic fungi and bacteria and allow interactions between gut microbiome and mycobiome. Disrupting these interactions results in distinct alterations of the fecal microbial population and protects the integrity of the intestinal barrier, which renders SAMP mice more resistant to colonization by pathogenic microorganisms. This result was dependent on the presence of both the probiotic strains as well as the antibiofilm amylase.

Large-scale sequencing analysis was used to observe the gene differences between the intestinal microbiome of mice administered probiotics compared with PBS-treated mice. Heterogeneity analysis of the fecal microbiome revealed a diverse microbial composition between probiotic-treated vs PBS-treated control mice. Specifically, there was no difference in the α- and β-diversity between the 2 groups prior to probiotic administration (ie, at baseline). In contrast, following probiotic administration, a significant shift of the microbial population in probiotic-treated mice was observed.

Multiple probiotic studies have demonstrated a positive correlation between increased richness of the gut microbiome (α-diversity) and amelioration of symptoms in several disease states.³¹ These observations convincingly support the concept that the intestinal microbial alteration is possible through probiotic administration and suggest that the main driving force for the decreased level of ileitis in probiotic-treated mice is through a probiotic effect.

In addition to the aforementioned variation in α- and β-diversity, we found significant changes in specific bacterial taxa, such as bacteria belonging to the genus Lachnoclostridium and to the species M. schaedleri and S. lacticifex. Specifically, Lachnoclostridium bacteria and M. schaedleri were consistently increased in stool samples upon completion of the probiotic treatment compared with the PBS-treated control group, and this increase was strongly associated with a steady increase in the production of SCFAs, metabolites produced by certain probiotic bacterial strains, as a product of fermentation of dietary fibers.

The Lachnoclostridium genus includes bacteria such as Clostridium XIVa,³² which is known to constitute a significant part of the human gut microbiome³³; it plays a role in homeostasis and can stimulate anti-inflammatory effects. Furthermore, via its metabolites such as butyric acid, clostridial cluster XIVa maintains gut health by altering the intraepithelial lymphocytes profile in the colon³⁴ and by inducing the differentiation and expansion of regulatory T cells with consequent attenuation of colitis in mice.³⁵ The microbiome of IBD patients has been associated with significantly decreased abundance of Clostridium clusters such as Clostridium XIVa compared with healthy individuals.³⁶ Although the role of bacteria belonging to the genus Lachnoclostridium remains unclear, the aforementioned studies highlight how the organisms are essential for immune homeostasis.

The probiotic combination has been shown to induce a pivotal microbiome shift in the gut. Therefore, we postulate that the decrease in ileitis observed in the SAMP mice may be mediated by secretory products, such as metabolites, produced by the microbial communities as a result of probiotic administration. In this regard, SCFAs may play a critical role in gut homeostasis by acting as ligands for G protein receptors and consequently activating anti-inflammatory pathways and preventing downstream leaky gut syndrome.³⁷ SCFA content in fecal samples has been proposed as a biomarker for several physiological processes and for studying the effect of nutritional interventions.³⁸ Given these proposals, we tested the stool samples of the experimental mice through GC/MS analysis to assess any difference related to SCFA production following probiotic administration. We report that several SCFAs, such as butyric, valeric, hexanoic, and heptanoic acid, were consistently more abundant in the stool samples and cecum content of probiotic-treated mice compared with the control animals. These data provide mechanistic evidence supporting our findings, given that butyric acid in particular gained increasing attention in recent years for supporting digestive health by providing energy to intestinal epithelial cells³⁹ and lowering risk of diseases by its potent anti-inflammatory activity.⁴⁰ The relevance of our results is additionally supported by the observation that bacteria belonging to the genus Lachnoclostridium (found significantly increased in abundance after probiotic treatment) are one of the most prolific sources of butyric acid in the gut.³⁹ Besides Lachnoclostridium, we also reported a positive correlation between butyric acid concentration and abundance of the bacterial species M. schaedleri. Our findings are corroborated by Alrafas et al,⁴¹ who demonstrated a positive correlation between butyric acid concentration, M. schaedleri abundance, and anti-inflammatory effect in a mouse colorectal cancer model. Moreover, our findings in regard to increased level of valeric acid in SAMP mice following probiotic treatment compared with the control group support results reported by Dabek-Drobny et al,⁴² who found higher levels of valeric acid in a healthy cohort control group compared with IBD patients. Valeric acid is a SCFA produced in the intestine that conveys several beneficial properties to the host, such as intestinal epithelial integrity and amelioration of gastrointestinal tract functions.⁴³

Finally, NanoString analysis showed that probiotic administration not only drastically altered the abundance of gut bacteria involved in the homeostasis of the intestinal mucosa, but also critically altered the expression of 24 genes implicated in B cell development, increased susceptibility to bacterial infection, and lymphocytic infiltration. These data are confirmed by numerous studies correlating dysregulation of adaptive immunity⁴⁴ and bacterial infections⁴⁵ with worsening of IBD symptoms. In particular, our findings are in agreement with a previous study by Shi et al⁴⁶ that found a potent association between lower gene expression of Ccr7 in ulcerative colitis and colorectal cancer patients and with the study conducted by Schepp-Berglind et al,⁴⁷ which demonstrated a protective role of Cr2 in a dextran sulfate sodium–colitis mouse model. Moreover, we found an increased expression of the Cxcr4 gene in probiotic-treated mice. CXCR4 is the receptor for the chemokine CXCL12, normally expressed by intestinal epithelial cells in the intestinal mucosa.⁴⁸ It contributes to gut burrier maturation and epithelial cell migration⁴⁹ via cAMP-mediated cellular functions. Conversely, the presence of CXCR4/CXCL12 in the intestine has been also located to resident CXCR4⁺ T cells and to patients affected by IBD. This may be due to differences between model systems, specifically human vs mouse. Differences in human vs mouse chemokines are commonly reported, and might reflect diverse biological functions.

Interestingly, NanoString data also showed how none of the aforementioned genetic pathways were altered when SAMP mice were administered with the incomplete probiotic mix lacking the amylase enzyme. This observation demonstrates the pivotal role of amylase in the antibiofilm activity leading to better penetration of the probiotic strains and consequent production of metabolites and alteration of genes all necessary to maintain the homeostasis of the intestinal mucosa and the overall health of the small intestine.

Conclusions

Taken together, our data provide proof for a beneficial effect of this novel probiotic combination on the severity of CD-like ileitis in the SAMP mouse model. Improvement included both alteration of intestinal genetic pathways and microbial rearrangements. Thus, we propose that this novel probiotic combination should be further studied in clinical trials on IBD patients, in which it has been proven that polymicrobial imbalance plays a critical role in dysbiosis and gut inflammation.⁵⁰

Funding

This work was supported by National Institutes of Health grants DK042191, DK055812, DK091222, and DK097948 to F.C.; DK125526 to L.D.M.; and AI145289 to M.G.; and BIOHM Health, LLC.

Conflict of Interest

M.G. is a co-founder of BIOHM Health LLC.