The Role of Morphine-Induced Impairment of Intestinal Epithelial Antibacterial Activity in Dysbiosis and its Impact on the Microbiota-Gut-Brain Axis

Recent evidence suggests that chronic exposure to opioid analgesics such as morphine disrupt the intestinal epithelial layer and cause intestinal dysbiosis. Inhibiting opioid-induced dysbiosis can preclude the development of tolerance to opioid-induced antinociception, suggesting an important role of the gut-brain axis in mediating opioid effects. However, the mechanism underlying opioid-induced dysbiosis remains unclear. Host-produced antimicrobial peptides (AMPs) are critical for the integrity of the intestinal epithelial barrier as they prevent the pathogenesis of the enteric microbiota. Here, we report that chronic morphine exposure reduces expression of the antimicrobial peptide, Regenerating islet-derived 3 gamma (Reg3γ), in the ileum resulting in reduced intestinal antimicrobial activity against Gram-positive bacteria, L. reuteri. Fecal samples from morphine-treated mice had reduced levels of the phylum, Firmicutes, concomitant with reduced levels of short-chain fatty acid, butyrate. Fecal microbial transplant (FMT) from morphine-naïve mice restored the antimicrobial activity, the expression of Reg3γ, and prevented the increase in intestinal permeability and the development of antinociceptive tolerance in morphine-dependent mice. Similarly, oral gavage with sodium butyrate dose-dependently reduced the development of antinociceptive tolerance, and prevented the downregulation of Reg3γ and the reduction in antimicrobial activity. The alpha diversity of the microbiome was also restored by oral butyrate in morphine-dependent mice. These data implicate impairment of the antimicrobial activity of the intestinal epithelium as a mechanism by which morphine disrupts the microbiota-gut-brain axis.


INTRODUCTION
It is well established that the chronic use of µ-opioid analgesics such as morphine results in the development of tolerance to the antinociceptive effects. Thus, a larger dose may be required for adequate pain control. Dose escalation can increase the propensity for unwanted effects such as addiction, constipation, and death due to overdose in severe cases since tolerance to opioid-induced respiratory depression does not develop at the same rate as tolerance to antinociception 1 . Despite the known risks, opioids remain the gold standard for managing pain in the clinic. It is, therefore, important to 1) understand the mechanisms that form the basis for the development of tolerance and 2) apply this information to formulate novel strategies to confront the ongoing opioid crisis.
Traditionally, nausea, vomiting, and constipation are reported as common gastrointestinal effects of µopioid analgesics 2 . Recent clinical evidence indicates that opioid use is associated with intestinal dysbiosis [3][4][5][6][7] . While nutritional de ciencies, poor hygiene, and comorbidities can contribute to an altered microbiota in opioid users [8][9][10] , the absence of a microbial shift in opioid users treated with opioid receptor antagonists, naloxone or naltrexone 4 , suggests that the dysbiosis may be an opioid receptordependent effect.
Several studies have demonstrated that the disruptive effects of opioids on the intestinal microbiota and the intestinal epithelium contribute to the development of tolerance to opioid-induced antinociception in vivo 11,12,16 and in primary afferent dorsal root ganglia neurons 11,12 as noted by the reversal of in vivo and cellular tolerance to morphine after antibiotic treatment. However, there is a signi cant gap in our current understanding of how opioids induce dysbiosis and its impact on antinociceptive tolerance.
The host and autochthonous microbiota maintain a commensal relationship through several mechanisms, including the secretion of antimicrobial proteins and peptides (AMPs) and IgA 19,20 . AMPs are released primarily by specialized epithelial cells called Paneth cells that reside within small intestinal crypts 20 . Enterocytes in the villi also secrete AMPs 21  Interestingly, fecal samples of opioid users were de cient in the short-chain fatty acid (SCFA), butyrate, and butyrate-producing bacteria 4,7 . Anerobic bacteria in the intestines produce butyrate by fermenting dietary ber and resistant starches 22 . Butyrate is utilized as a signi cant energy source by colonocytes; it improves epithelial barrier function by inducing the expression of tight junction proteins, increasing AMP production, and inhibiting pro-in ammatory cytokines 22 . We previously reported that exacerbating intestinal in ammation in an experimental model of colitis increased the rate at which tolerance developed to morphine's antinociceptive effects 18 . Since butyrate is critical for maintaining the physiology of the intestinal mucosa and inhibits intestinal in ammation, in the present study, we tested the effect of butyrate on epithelial barrier function, and on the development of tolerance to morphineinduced antinociception.

MATERIALS AND METHODS
Animals: Male Swiss Webster mice (Harlan Sprague Dawley, Inc. Fredrick, MD, USA) six-eight weeks old, weighing 25-30 g, were housed ve to a cage with ad libitum access to food and water in animal care quarters maintained under a 12-hour light/dark cycle (lights on from 7 am to 7 pm). Animals were randomly assigned to control and treatment groups. All animal procedures were in accordance with the protocols reviewed and approved by the Institutional Animal Care and Use Committee at Virginia Commonwealth University (VCU IACUC). Results of the animal experiments were reported in accordance with the recommendations of the ARRIVE 2.0 guidelines.

Test Drugs:
Morphine: 1. 75-mg morphine or placebo pellets, obtained from the National Institute on Drug Abuse (NIDA, Bethesda MD), were aseptically implanted in the subcutaneous cavity on the dorsum under iso urane (2.5%) anesthesia as described previously 11 . Mice were allowed to recover in their home cages. On test day 7, the mice were subjected to antinociceptive response experiments (Figures 1 and 3B); used for fecal microbiome analysis (Figures 1, S1, and S2); for harvesting blood, stool, and colon tissue for evaluating butyrate concentration ( Figure 2); for measuring intestinal epithelial permeability ( Figure 4A); in antibacterial activity experiments ( Figures 5 and 7); or for evaluating Regenerating islet-derived 3 gamma (Reg3γ) gene expression in the ileum ( Figure 6A) as described in the subsequent methods sections. Each animal was used only once.
2. Morphine sulfate (National Institute on Drug Abuse Drug Supply Program, Bethesda, MD) was diluted in saline to 1, 2, 4, and 8 mg/mL. Mice were injected intraperitoneally twice daily with saline or increasing doses of morphine as follows: Day 1-20 mg/kg morphine, Day 2-40 mg/kg morphine, Day 3-40 mg/kg morphine, and Day 4-80 mg/kg morphine. Mice were used on test day 5 in the warm-water tailwithdrawal experiment in Figure 3A, for evaluating intestinal epithelial permeability in Figure 4B, for measuring the expression of Reg3γ in Figure 6B, and in antibacterial activity experiments in Figure 8 as described in the following methods. Each animal was used only once.
Sodium butyrate: Sodium butyrate (ThermoFisher Scienti c, Waltham, MA) was prepared in saline at concentrations of 0.125, 0.250, 0.500, and 1.000 M and administered twice daily by oral gavage. For the dose-response experiment in Figure 3A, mice injected with ramping doses of morphine (as described above) were orally administered saline or different concentrations of sodium butyrate (0.125, 0.250, 0.500, or 1.000 M) for four days. Antinociception was measured on day 5 using the warm-water tailwithdrawal test described below. In all subsequent experiments, 0.250 M sodium butyrate or its vehicle, saline, was administered twice daily through oral gavage. The number of treatments with sodium butyrate was contingent on the duration of exposure to morphine, such that mice subcutaneously implanted with pellets were administered 0.250 M sodium butyrate for six days ( Figures 3B, 5, 6B, 7, and S2), and mice injected with ramping doses of morphine received 0.250 M sodium butyrate for four days ( Figures 3A, 4B, and 6B). Naloxone HCl: Naloxone HCl (Sigma-Aldrich, St. Louis, MO) was prepared in saline at concentrations of 0.2, 0.4, and 0.8 mg/mL and injected intraperitoneally twice daily at escalating doses 10 minutes before the administration of morphine sulfate in the following manner: Day 1-2 mg/kg naloxone, Day 2-4 mg/kg naloxone, Day 3-4 mg/kg naloxone, and Day 4-8 mg/kg naloxone. The doses of naloxone were 1/10 th of the doses of morphine sulfate. Ileum tissue was collected on day 5 for use in the antibacterial activity assay in Figure 8.
Fecal Microbiota Transplant (FMT): Fresh fecal pellets (100 mg) from placebo or morphine-pelleted mice were collected on day 7 and suspended in 1.2 mL of cold (4°C) phosphate-buffered saline (PBS) containing 10% glycerol. The suspension was homogenized and then centrifuged at 800xg for three minutes. The supernatant was transferred to a separate tube and stored at -80°C. The concentration of total bacteria was determined by measuring optical density (OD), such that OD = 0.5 represented 1x10 8 cells. 100 µL of the fecal supernatant (1 x 10 9 cells/dose) was then administered twice daily for six days to recipient mice groups via oral gavage according to the following scheme: 1. Placebo-pelleted mice that did not receive fecal microbiota transplants (PP-Sham), 2. Morphine-pelleted mice that did not receive fecal microbiota transplants (MP-Sham), 3. Placebo-pelleted mice that received fecal microbiota from placebo-pelleted donor mice (PP + PP-FMT), 4. Morphine-pelleted mice that received fecal microbiota from placebo-pelleted donor mice (MP + PP-FMT), 5. Placebo-pelleted mice that received fecal microbiota transplants from morphine-pelleted donor mice (PP + MP-FMT), and 6. Morphine-pelleted mice that received fecal microbiota transplants from morphine-pelleted donor mice (MP + MP-FMT) ( Figure 1C).
Antinociceptive response tests: The warm-water tail-withdrawal and hot-plate assays were used in the present study. In the warm-water tail-withdrawal test, the distal 1/3 tail was immersed in a water bath at 56°C. The latency to withdraw the tail from the warm water was recorded. A maximum cutoff of 10 seconds was set to prevent tissue damage. On test day 7 in pelleted mice ( Figures 1D and 3B) and test day 5 in injected mice ( Figure 3A), baseline responses were recorded, following which the mice received acute morphine (10 mg/kg s.c.). 25 minutes later, tail-ick latencies were recorded to test for the development of tolerance to the 10 mg/kg morphine challenge. Antinociception induced by 10 mg/kg morphine was quanti ed as the percentage of maximum possible effect (%MPE), such that: %MPE = [(challenge latency−baseline latency) / (Maximum cutoff−baseline latency)] x 100.
In the hot-plate assay, individual mice were placed on a Syscom Model 35D hot-plate set at 56°C, and the latency to lick their hind paw or jump was recorded. A maximum cutoff of 30 seconds was set to prevent tissue damage. Baseline responses were recorded on test day 7 in pelleted mice ( Figure 1E), following which the mice were injected with acute morphine (10 mg/kg s.c.). 25 minutes later, hot-plate responses were measured again to test for the development of tolerance to the acute morphine challenge.
Antinociception induced by 10 mg/kg morphine was quanti ed as %MPE as described above.
Detection of butyrate in blood, colon, and stool: Blood, colon tissue samples, and fecal material were collected from morphine or placebo-pelleted mice on day 7. Samples were immediately homogenized 1:4 with deionized water and stored at -30°C. Seven-point calibration curves of 10-1000 µg/g butyrate (Sigma-Aldrich, St. Louis, MO), a butyrate-free control, and a negative control free of butyrate and the internal standard (ISTD) were prepared. Butyrate was extracted and analyzed using a modi ed previously published method 23 . In brief, 100 μL of methanol containing 20 µg butyrate-1,2-13 C 2 (Sigma-Aldrich, St. Louis, MO), the ISTDs, was added to 0.20 g aliquots of each calibrator, control, or specimen except the negative control. Samples were mixed for ve minutes, centrifuged for 30 minutes, and then left for 30 minutes at 4°C. 100 μL of the clear supernatant was transferred into a new tube and washed with 100 μL propyl formate. Samples were mixed for ve minutes and centrifuged for 30 minutes before transferring 50 µL of the organic layer to GC vials for analysis. Gas chromatography-mass spectrometer (GC-MS) analysis was performed on a Shimadzu GC/MS-QP2020 NX Single Quadrupole GC-MS (Shimazu, Kyoto, Japan) controlled by GCMS solution software (Shimadzu, Kyoto, Japan). Chromatographic separation was performed using a ZB-FFAP column, 30 m x 32mm, 0.25 µm (Phenomenex, Torrance, CA). A sample volume of 2 µL was injected in splitless mode with an injector temperature of 200°C. The carrier gas was Helium with a 2 mL/minute ow rate. The initial oven temperature of 55°C was held for four minutes, then ramped to 130°C at 50°C/minute and held for 3.7 minutes. Finally, the temperature was raised to 250°C at 30°C/minute and held for two minutes. Linear regression of the peak area of ratios of the quanti cation ion for butyrate (72 m/z) and the ISTD quanti cation ion (75 m/z) was used to construct the calibration curves. For each analytical run, the coe cient of determination (r 2 ) was higher than 0.996.
The concentrations of each calibrator were determined to be within ± 20% of their expected concentration.
Intestinal permeability assay: On the test day, i.e., day 7 in pelleted mice and day 5 in intermittently injected mice, animals were orally gavaged with FITC-conjugated dextran (100 mg/ml in PBS, Sigma-Aldrich, St. Louis, MO) at a dose of 44 mg/100 g body weight of FITC-labeled dextran. After four hours, mice were anesthetized with iso urane, and 300-500 µl of blood was collected by cardiac puncture. Serum collected from blood samples by centrifugation for 15 minutes at 1500xg and 4°C was diluted with an equal volume of PBS. 100 µl of diluted serum was transferred to a 96-well plate, and FITC concentration was uorometrically quanti ed by emission spectrometry (Promega, Madison, WI) at 528 nm using an excitation wavelength of 485 nm. Serum from mice not administered FITC-dextran was used to determine background. All concentrations were measured against a standard curve of serially diluted FITC-dextran.
Bactericidal activity assay: The bactericidal activity assay was performed based on the procedure described by Udden et al. 24 .
Preparation of conditioned media from ileum tissue samples: 4-5 cm of the distal ileum was resected and immediately ushed with sterile-ltered ice-cold PBS to remove digesta. Ileum tissue samples were cut longitudinally, rinsed in sterile-ltered ice-cold PBS, and weighed. Tissue samples were disinfected in 5 ml DMEM/F12 media supplemented with 5% FBS and 1x antibiotics (penicillin, streptomycin, and vancomycin) for two hours at 37°C in a 95%O 2 /5%CO 2 incubator. After disinfection, residual antibiotics were washed off by rinsing the samples three times with 5 ml antibiotic-free DMEM/F12 media supplemented with 5% FBS. Rinsed ileum tissue samples were then cut into 1 cm pieces using sterile scissors and transferred to 12-well cell culture plates containing fresh antibiotic-free DMEM/F12 media supplemented with 5% FBS. 1 ml of DMEM/F12 supplemented with 5% FBS was used per 100 mg of tissue. Samples were incubated at 37°C in an incubator with 5% CO 2 and 95% O 2 for 12 hours. Tissue supernatants (or conditioned media) were subsequently transferred into sterile 1.5 ml centrifuge tubes. Tissue debris was sedimented by centrifugation at 12,000 x g at 4°C for ve minutes, and the conditioned media was used for the antibacterial activity assay.
Antibacterial activity assay: The prototypical Gram-negative bacteria, Escherichia coli (E. coli strain HB101), was inoculated in 5 ml of Luria-Bertani (LB) broth and incubated overnight at 37°C with constant shaking at 250 rpm. Cultured bacteria were collected by centrifugation at 1,200 x g for 10 minutes at 4°C and then resuspended in fresh LB broth at a nal concentration of 1x10 5 cells/mL. 20 µL of the diluted bacteria were added to 500 µL of the ileum-derived conditioned media and incubated for one hour in an incubator maintained at 37°C and 95% O 2 /5% CO 2 . An additional 20 µL of the diluted bacteria were incubated with 500 µL of DMEM/F12 media supplemented with 5% FBS and 1 x antibiotics (penicillin, streptomycin, and vancomycin; positive internal control) and with 500 µL antibiotic-free DMEM/F12 media containing 5% FBS (negative internal control). 500 µL of the ileum-derived conditioned media supplemented with 20 µL of fresh LB broth was also incubated along with the other samples to check for the presence of contamination (sham control). Thus, an experiment with each ileum-derived conditioned media constituted four groups: A) ileum supernatant + E. coli, B) ileum supernatant + LB broth, C) negative control, and D) positive control. A BHI agar plate was divided into four quadrants, and 50 µL per group was evenly applied to each quadrant. The agar plate was incubated at 37°C, and bacterial colonies were counted 15-18 hours later. The experiment was repeated for ileum supernatants prepared from different mice.
The prototypical Gram-positive bacteria, Lactobacillus reuteri (L.reuteri strain ATCC 53608), was cultured in de Man, Rogosa, and Sharpe (MRS) broth supplemented with 0.001% Tween 80 in a 95% O 2 /5% CO 2 incubator maintained at 37°C. The cultured bacteria were collected by centrifugation, resuspended in fresh MRS broth, and serially diluted to a nal concentration of 1 x 10 5 cells/mL. The activity of the ileum tissue supernatants against L. reuteri was tested using the methodology described above for E. coli. Brie y, each ileum tissue supernatant experiment consisted of four groups: A) ileum supernatant + L.reuteri, B) ileum supernatant + MRS broth, C) DMEM/F12 media + 5% FBS + L.reuteri (negative control), and D) DMEM/F12 media + 5% FBS + 1x antibiotics (penicillin, streptomycin, and vancomycin) + L.reuteri (positive control). 50 µL per group was uniformly smeared onto MRS agar plates divided into four quadrants, and the total number of bacterial colonies formed was determined after incubation for 15-18 hours.
RNA isolation and qRT-PCR: Total RNA was extracted from the ileum of placebo or morphine-pelleted mice orally administered with FMT on day 7 and from the ileum of mice injected repeatedly with saline or morphine and orally administered with 0.250 M sodium butyrate or its vehicle, saline, on day 5 using TRIzol reagent (ThermoFisher Scienti c, Waltham, MA). RNA samples were treated with DNase 1 (RNasefree, ThermoFisher Scienti c, Waltham, MA) to remove DNA contamination. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed on a Mini-Opticon real-time PCR system (Bio-Rad, Hercules, CA) by using the iTaq Universal SYBR Green One-Step kit (Bio-Rad, Hercules, CA) as described previously 25 . Gapdh was used as the internal control. Primers used in this study were: murine Reg3γ forward, 5′-CGTGCCTATGGCTCCTATTGCT-3′; murine Reg3γ reverse, 5′-TTCAGCGCCACTGAGCACAGAC-3′; murine Gapdh forward 5′-CCATGGAGAAGGCTGGGG-3′; and murine Gapdh reverse 5′-CAAAGTTGTCATGGATGACC-3′ (Integrated DNA Technologies, Inc., Skokie, Illinois).
Microbiome pro ling: Fecal pellets were collected from saline-treated placebo-pelleted, butyrate-treated placebo-pelleted, saline-treated morphine-pelleted, and butyrate-treated morphine-pelleted mice. DNA was extracted using the QIAamp Fast DNA Stool Mini Kit according to the manufacturer's protocols, and DNA was sent to CosmosID (Cosmosid Inc, Rockville, MD) and subjected to whole shotgun sequencing using the Illumina platform. An average of 5.921M reads per sample was achieved with a minimum of 4.514M reads and a maximum of 8.816M reads. CosmosID's k-mer based approach was used for taxonomic identi cation by comparing sequences to an in-house database. Pro les were analyzed using the ltered species-level data containing 297 bacterial species. Counts were renormalized to the mean number of reads with a pseudo count added to each bacterial species count and the counts were log10 transformed.
Blinding: Experimenters were not blinded while performing experiments. However, separate investigators conducted the experiments to ensure reliability of results.

Data analysis:
The threshold for statistical signi cance was P < 0.05. Post hoc analysis of the ANOVA was performed only for signi cant main effects or signi cant interactions. GraphPad Prism (version 9.4.1) was used for data analysis. Data are presented as mean ± SEM.
Warm-water tail-withdrawal test and hot-plate test: Data represented as %MPE in Figure 1 were evaluated by two-way ANOVA with FMT treatment and morphine treatment as the two independent variables. The Tukey's multiple comparisons test was used for post hoc analysis. Data represented as %MPE in Figure  3A were evaluated by two-way ANOVA with butyrate dose and morphine treatment as the two independent variables, and in Figure 3B, by two-way ANOVA with butyrate treatment and morphine treatment as the two independent variables. Tukey's post hoc test was used for multiple comparisons between groups.
Butyrate levels in blood, colon, and stool: Data in Figure 2A were analyzed by two-way ANOVA with butyrate treatment and morphine treatment as the independent variables. Tukey's multiple comparisons test was used for post hoc analysis. Data in Figures 2B and 2C were analyzed by unpaired two-tailed Student's t-test.
Intestinal permeability assay: Serum concentrations of FITC-dextran were evaluated in Figure 4A by twoway ANOVA with FMT treatment and morphine treatment as the independent variables and in Figure 4B by two-way ANOVA with butyrate treatment and morphine treatment as the independent variables. Tukey's post hoc test was used for multiple comparisons between groups.
Bactericidal activity assay: The total number of bacterial colonies formed on agar plates were converted to colony forming units (CFU)/mL, such that CFU/mL = (number of colonies x dilution factor) / 50 µL. Data in Figures 5A-C and 7A-D were evaluated by repeated-measures one-way ANOVA with the media as the independent variable. Tukey's post hoc test was used for multiple comparisons between groups. CFU/mL data were transformed into percent bactericidal activity in Figures 5D, 6E, and 8, such that % Bactericidal activity = {[(CFU/mL of antibiotic-free DMEM/F12 media + bacteria) -(CFU/mL of Ileum supernatant + bacteria)] / (CFU/mL of antibiotic-free DMEM/F12 media + bacteria)} * 100. Data in Figure  5D was assessed by one-way ANOVA with treatment as the independent variable. The Holm-Sidak post hoc test was used for multiple comparisons between groups. Data in Figure 7E was evaluated by twoway ANOVA with butyrate treatment and morphine treatment as the two independent variables. Multiple comparisons between groups were made using Tukey's post hoc test. Data in Figure 8 was evaluated by one-way ANOVA with treatment as the independent variable. Tukey's multiple comparisons test was used for post hoc analysis.
qRT-PCR: Relative expression of Reg3γ to Gapdh was calculated using the 2 −ΔΔCt method, and values were expressed as fold change. Data were analyzed in Figure 6A by two-way ANOVA with FMT treatment and morphine treatment as the independent variables and in Figure 6B by two-way ANOVA with butyrate treatment and morphine treatment as the independent variables. Multiple comparisons between groups were made using Tukey's post hoc test.
Microbiome analysis: The Permutational multivariate analysis of variance (PERMANOVA) using the Bray-Curtis dissimilarity index was utilized to evaluate the beta diversity of the fecal bacteria between the different groups. Results of the statistical analysis were obtained using the CosmosID Hub. The alpha diversity index, Chao1, for all the groups was determined using the CosmosID Hub. The data were analyzed by unpaired two-tailed Student's t-test or two-way ANOVA, and Tukey's post hoc test was used for pairwise comparisons using GraphPad Prism (version 9.4.1).

RESULTS
Chronic morphine exposure alters the composition of the fecal microbiota. Evaluation of the alpha diversity (within group diversity) of fecal bacterial communities using the Chao1 Index revealed signi cantly increased bacterial abundance in the saline-treated morphine-pelleted (MP) animals compared to the saline-treated placebo-pelleted (PP) animals ( Figure 1A). Principal coordinate analysis of the beta diversity (diversity between groups) of the fecal bacteria measured with the Bray-Curtis Index revealed discrete clustering of the saline-treated MP mice and saline-treated PP mice ( Figure 1B, Table  S1). Further analysis of the bacterial taxa at the phylum level showed contraction of Firmicutes and Actinobacteria, and expansion of Bacteroidetes in saline-treated MP mice ( Figure S1). Altogether, the data indicated that chronic morphine treatment altered the composition of the fecal bacteria.
Fecal microbiota transplants from placebo-pelleted mice inhibited the development of tolerance to morphine-induced antinociception. Opioid-induced dysbiosis has been implicated in the development of tolerance to opioid-induced antnociception 11 . Since chronic morphine exposure induced dysbiosis, we investigated whether replacing the bacteria of morphine-treated mice with that of control mice using fecal microbiota transplants (FMT) altered the development of tolerance to morphine-induced antinociception ( Figure 1C). A two-way ANOVA analysis of the antinociceptive effect produced by 10 mg/kg morphine in the warm-water tail-withdrawal test revealed a signi cant FMT x morphine treatment interaction [F (2, 44) = 19.36, P < 0.001] ( Figure 1D). Chronic morphine-treated mice responded poorly to the 10 mg/kg morphine challenge compared to PP mice (28.5 ± 4.2 %MPE vs. 100 ± 0 %MPE in PP mice; Figure 1D), indicating the development of tolerance. FMT from PP mice (PP-FMT) prevented the development of tolerance to morphine-induced antinociception. 10 mg/kg morphine produced signi cant antinociception in MP mice treated with PP-FMT (88.9 ± 7.0 %MPE; Figure 1D). Interestingly, FMT from MP mice (MP-FMT) did not produce tolerance to morphine-induced antinociception in PP mice.
A two-way ANOVA analysis of the hot-plate data indicated a signi cant interaction of FMT x morphine treatment [F (2, 21) = 6.154, P = 0.008] ( Figure 1E). Chronic morphine-treated mice were tolerant to the 10 mg/kg morphine challenge dose in the hot-plate test (40.1 ± 14.6 %MPE vs. 100 ± 0 %MPE in PP mice; Figure 1E). Consistent with the warm-water tail-withdrawal test results, PP-FMT prevented the development of tolerance to morphine-induced antinociception in MP mice. MP mice treated with PP-FMT exhibited signi cant antinociception to 10 mg/kg morphine (89.2 ± 10.8 %MPE; Figure 1E). Thus, replenishing the fecal microbiota of morphine-treated mice prevented the development of tolerance to morphine-induced antinociception in the warm-water tail-withdrawal and hot-plate tests. However, treating PP mice with the microbiota of MP mice did not produce tolerance to morphine-induced antinociception ( Figures 1D and 1E). 10 mg/kg morphine produced signi cant antinociception in PP mice treated with MP-FMT ( Figures 1D and 1E). These results suggest that dysbiosis is not su cient for producing tolerance to morphine-induced antinociception and that exposure to chronic morphine is required for the induction of tolerance to morphine-induced antinociception.
Exposure to chronic morphine reduced endogenous butyrate concentration in the stool. Butyrate levels are substantially reduced in fecal samples of opioid users 4,7 . Furthermore, the phylum Firmicutes, which comprises butyrate-producing bacteria 22 , was reduced in chronic morphine-treated mice ( Figure S1). Therefore, we analyzed the concentration of endogenous butyrate in fecal samples, colon tissue, and whole blood in chronic morphine-treated mice using GC-MS ( Figure 2). The concentration of butyrate in fecal samples of MP mice was signi cantly reduced compared to that of PP mice (386.1 ± 34.5 vs. 161.0 ± 12.3, respectively). Daily oral supplementation with sodium butyrate restored butyrate levels in the fecal samples of MP mice (Figure 2A). Daily butyrate supplementation in placebo mice did not increase butyrate concentrations in the stool, likely due to quorum sensing 26 . We observed no change in the concentration of butyrate in colon tissue [13.7 ± 2.5 vs. 12.8 ± 3.8, t (df) = 0.186 (8), P = 0.86] or whole blood [8.6 ± 0.7 vs. 7.5 ± 0.4, t (df) = 1.466 (8), P = 0.18] between PP and MP mice ( Figures 2B and 2C, respectively). These data suggest that in chronic morphine-treated mice, butyrate production is signi cantly impaired due to altered intestinal microbiota.
Butyrate administration prevented the development of tolerance to morphine-induced antinociception. We next investigated if butyrate administration altered the development of tolerance to morphine-induced antinociception. Morphine was administered intermittently ( Figure 3A) or continuously through subcutaneously implanted pellets ( Figure 3B). Figure 3A shows that chronic morphine-treated mice were tolerant to morphine-induced antinociception when sodium butyrate was not administered. The 10 mg/kg morphine challenge did not produce signi cant antinociception (32.0 ± 17.6 % MPE vs. 100.0 ± 0.0 % MPE in the saline + 0 M butyrate-treated mice). Oral administration of sodium butyrate prevented the development of tolerance to morphine-induced antinociception in a dose-dependent manner. Signi cant inhibition of tolerance to morphine-induced antinociception was observed starting with the 0.250 M sodium butyrate dose (89.9 ± 5.0 % MPE vs. 32.0 ± 17.6 % MPE in the morphine + 0 M butyrate-treated mice; Figure 3A). A two-way ANOVA analysis comparing % MPE of the 10 mg/kg morphine challenge in saline or chronic morphine-treated animals administered with increasing doses of sodium butyrate revealed a signi cant butyrate dose x morphine treatment interaction [F (4, 40) = 4.015, P = 0.0008] ( Figure 3A). Consistent with results in Figure 3A Figure 4A) or between saline and chronic morphine + sodium butyrate-treated mice ( Figure 4B).
These data demonstrate that the intestinal epithelium remained intact despite chronic morphine treatment.
Butyrate or FMT prevented morphine-induced disruption of the antimicrobial activity of the small intestinal epithelium. AMPs contribute to the innate immune response of the intestinal epithelium by maintaining the intestinal microbiota and preventing colonization of the epithelium by pathogenic strains 21 . It is unclear if the altered antimicrobial activity of the intestinal epithelium contributes to opioidinduced dysbiosis. Therefore, we investigated whether chronic morphine treatment perturbed the antibacterial activity of the ileum and if sodium butyrate or FMT inhibited morphine's effects ( Figure 5A).
We rst examined antimicrobial activity against the prototypical Gram-positive bacteria, L.reuteri ( Figure  5). Incubation of L.reuteri with conditioned media from the ileum of PP mice resulted in signi cantly reduced colonies on the agar plate compared to when the bacteria were incubated with antibiotic-free nutrient media (6.9 x 10 6 ± 9.4 x 10 5 CFU/mL vs. 2.3 x 10 7 ± 9.2 x 10 5 CFU/mL, respectively; Figure 5B), indicating that the conditioned media exhibited activity against Gram-positive bacteria. In comparison, conditioned media prepared from the ileum of MP mice yielded similar colony numbers to when the bacteria were incubated in antibiotic-free nutrient media (2.1 x 10 7 ± 1.3 x 10 6 CFU/mL vs. 2.0 x 10 7 ± 1.1 x 10 6 CFU/mL, respectively; Figure 5C), suggesting that the antibacterial activity of the conditioned media decreased after chronic morphine treatment. Interestingly, the antibacterial activity of the ileumconditioned media was somewhat restored upon treatment with sodium butyrate ( Figure 5C). The conditioned media from the ileum of mice treated with morphine and sodium butyrate yielded fewer colonies compared to when the bacteria were incubated with antibiotic-free nutrient media (1.5 x 10 7 ± 1.2 x 10 6 CFU/mL vs. 2.0 x 10 7 ± 9.6 x 10 5 CFU/mL, respectively; Figure 5D). A comparison of the normalized data in Figure 5E revealed that chronic morphine treatment signi cantly reduced bactericidal activity of the ileum against Gram-positive bacteria, which was partially restored with oral butyrate treatment (70.1 ± 3.6% in PP mice vs. -3.6 ± 9.9% in MP mice vs. 23.6 ± 8.6% in MP +butyrate-treated mice).
The REG3 family of antimicrobial peptides is highly expressed in the small intestine. Murine REG3γ, and its human analog, REG3α, have been reported to exhibit activity against Gram-positive bacteria 29,30 . Since REG3γ is transcriptionally regulated, we investigated whether morphine altered the expression of the Reg3γ gene in the ileum and if FMT of commensal microbiota comprising butyrate-producing bacteria ( Figure 6A) or direct administration of exogenous butyrate ( Figure 6B) inhibited morphine effects. Reg3γ expression was signi cantly reduced after chronic morphine treatment, i.e., in MP mice compared to PP controls ( Figure 6A) and morphine-injected mice compared to saline controls ( Figure 6B). Interestingly, FMT of from placebo-pelleted mice (PP-FMT) prevented the downregulation of Reg3γ in morphinepelleted mice. However, FMT from morphine-pelleted mice (MP-FMT) produced no effect on Reg3γ expression in placebo-pelleted mice, and Reg3γ expression in the Ileum of these mice remained signi cantly upregulated ( Figure 6A). The absence of the effect of MP-FMT on Reg3γ expression indicated that the antimicrobial response of the epithelial barrier remained intact and that chronic treatment with morphine was necessary for inducing barrier dysfunction. Consistent with the results of PP-FMT treatment on Reg3γ expression in morphine-treated mice, oral administration of sodium butyrate also prevented morphine-induced downregulation of Reg3γ in the ileum ( Figure 6B). Sodium butyrate did not alter the expression of Reg3γ in saline-treated control mice ( Figure 6B). The absence of an effect could be due to quorum sensing by the autochthonous microbiota 26 . The data indicated that the commensal microbiota and/or butyrate regulate Reg3γ transcription to reverse chronic morphine-induced effects.
Next, we tested the activity of the ileum supernatants against the prototypical Gram-negative bacteria, E.coli (Figure 7). Bacterial colony counts were signi cantly diminished when E.coli were incubated with conditioned media from the ileum of PP mice instead of antibiotic-free nutrient media (5.9 x 10 7 ± 2.7 x 10 6 CFU/mL vs. 8.1 x 10 6 ± 4.0 x 10 6 CFU/mL, respectively; Figure 7A), indicating that the ileum-derived conditioned media exerted substantial antibacterial activity against Gram-negative bacteria. Sodium butyrate did not affect the antibacterial activity of the ileum-derived conditioned media. E.coli colony counts were substantially reduced when incubated with the conditioned media from the ileum of PP mice treated with sodium butyrate instead of antibiotic-free DMEM/F12 media (5.0 x 10 7 ± 2.3 x 10 6 CFU/mL vs. 6.2 x 10 6 ± 1.9 x 10 6 CFU/mL, respectively Figure 7B). Figure 7C shows the reduced antibacterial activity of ileum supernatants obtained from chronic morphine-treated mice. Incubation of E.coli with conditioned media from the ileum of MP mice resulted in a non-signi cant reduction in colony numbers compared to when the bacteria were incubated with antibiotic-free nutrient media (3.8 x 10 7 ± 2.9 x 10 6 CFU/mL vs. 4.8 x 10 7 ± 3.4 x 10 6 CFU/mL, respectively). Conditioned media derived from the ileum of chronic morphine-pelleted mice treated with sodium butyrate exerted signi cant antibacterial activity ( Figure 7D). E.coli colony counts were substantially larger when antibiotic-free nutrient media was used instead of conditioned media from the ileum of sodium butyrate-treated chronic morphine-pelleted mice (5.9 x 10 7 ± 3.4 x 10 6 CFU/mL vs. 1.3 x 10 7 ± 3.4 x 10 6 CFU/mL, respectively; Figure 7D). A comparison of the normalized data in Figure 7E revealed that the % Bactericidal activity of morphinetreated ileum media was signi cantly reduced compared to that from PP mice (17.5 ± 10.0% vs. 87.1 ± 6.0%, respectively). While oral administration of sodium butyrate did not alter the antibacterial activity of the ileum from PP mice, it signi cantly increased the bactericidal activity of the ileum of MP mice (87.9 ± 3.9% and 77.3 ± 6.4%, respectively; Figure 7E). These data indicated that chronic treatment with morphine inhibited the homeostatic activity of the ileum from ghting against Gram-positive and Gram-negative bacteria and that oral administration of sodium butyrate can prevent these effects.
Oral butyrate administration altered the fecal bacterial composition in morphine-treated mice. Analysis of the Chao1 Index of alpha diversity revealed that the bacterial abundance in butyrate-treated MP mice was diminished compared to saline-treated MP mice and was comparable to saline-treated PP animals. Oral butyrate administration did not alter the Chao1 diversity of the fecal bacteria of PP mice ( Figure S2A). Evaluation of the beta diversity using the Bray-Curtis Index revealed distinct clustering of butyrate-treated MP mice and saline-treated MP mice (Table S1; Figure S2B). PERMANOVA analysis indicated a signi cant difference in the beta diversity of butyrate-treated MP mice and saline-treated PP mice (Table   S1; Figure S2C). PERMANOVA analysis did not reveal differences in the beta diversity of saline-treated PP mice and butyrate-treated PP mice (Table S1; Figure S2D). Altogether, these data indicated that butyrate altered the fecal microbial composition of chronic morphine treated mice.
Naloxone antagonized the morphine-induced decrease in the antimicrobial activity of the ileum. We next tested if the inhibitory effect of morphine on the antibacterial activity of the ileum was opioid receptormediated. Morphine treatment reduced the activity of the ileum supernatants against L.reuteri (52.5 ± 2.1% in morphine-treated vs. 73.8 ± 1.7% in saline-treated; Figure 8A) and against E.coli (36.1 ± 14.5% in morphine-treated vs. 93.3 ± 2.8% in saline-treated) ( Figure 8B). Concomitant treatment with naloxone, a non-selective opioid receptor antagonist, prevented morphine effects (L.reuteri: 69.9 ± 2.6%; E.coli: 83.4 ± 4.8%; Figure 8), thus indicating that the effect of chronic morphine exposure on the antibacterial activity of the ileum is opioid receptor-mediated.

DISCUSSION
In this paper we show that the antimicrobial activity of the gastrointestinal tract is markedly reduced in a morphine-dependent mouse model. Chronic morphine inhibits the expression of the antimicrobial peptide, Reg3γ, that is prevented by fecal transplant from morphine-naïve mice or by the short chain fatty acid, butyrate. Supplementing the enteric microbiome of morphine-treated mice with butyrate or the fecal microbiota of morphine-naïve mice attenuated tolerance to morphine-induced antinociception.
Altogether, the results indicate that morphine-induced disruption of the antimicrobial activity of the intestinal epithelium contributes to enteric dysbiosis and impacts the development of antinociceptive tolerance.
The intestinal epithelium restricts pathogenic strains from colonizing intestinal tissue and entering systemic circulation through multiple contingencies that maintain spatial segregation between the intestinal epithelium and the gut bacteria. For example, tight junction proteins between adjacent intestinal epithelial cells limit paracellular transport between the apical and basolateral membranes 31 20 . Similar effects were observed in the present study in mice exposed to chronic morphine. Here, we observed that repeated exposure to morphine, whether through intermittent injections or subcutaneously implanted continuous-release pellets, reduced the antimicrobial activity of the intestinal epithelium. Activity against both Gram-positive and Gram-negative bacteria was reduced. Speci cally, the antimicrobial c-type lectin, Reg3γ, was downregulated in chronic morphinetreated mice. Reg3γ promotes host-bacteria segregation, and consequently, Reg3γ -/mice exhibit altered mucus distribution, increased colonization of the intestinal epithelium by mucosa-associated bacteria, bacterial translocation, and elevated in ammatory responses in the intestine 32,33 . Naloxone, a non-selective opioid receptor antagonist, prevented morphine effects on the antimicrobial activity of the ileum against Gram-positive and Gram-negative bacteria, indicating that this effect was opioid receptormediated. In addition to the perturbed antimicrobial activity of the intestinal epithelium, chronic morphine exposure altered the composition of the gut bacteria. Thus, these data suggest that impaired ability to neutralize bacteria contributes to opioid-induced intestinal dysbiosis and predisposes to systemic translocation of luminal bacteria via disruption of the epithelial barrier.
μ-opioid receptors are not detected in intestinal epithelial cells; they are primarily expressed in neurons 1,34 .
Thus, altered activity of neurons might form the basis for opioid-induced intestinal barrier dysfunction and the resultant dysbiosis. There is increasing evidence that enteric neurons can modulate the response of the intestinal epithelial barrier to gut microbiota. Chemogenetic activation of cholinergic enteric neurons altered the intestinal transcriptome, including the expression of genes responsible for mucosal immunity and antimicrobial responses, and produced effects on the gut microbiome and metabolome 35 . Approximately half of the μ-opioid receptor-expressing neurons in the enteric nervous system are cholinergic 25,34 . A subset of μ-opioid receptor-containing neurons also expresses the inhibitory neurotransmitter, vasoactive intestinal peptide (VIP) 25,34 . VIPergic neurons have been shown to regulate the activity of intestinal epithelial cells via crosstalk with innate lymphoid cells, ILC2, and ILC3 [36][37][38] . Morphine is known to modulate the excitability of enteric neurons by altering the activity of voltage-gated ion channels [39][40][41][42] . Morphine can also alter the expression of IL-18 in enteric neurons via opioid receptors 25 . IL-18 from enteric neurons, but not immune or epithelial cells, is necessary to produce AMPs, which protect against bacterial in ltration of the intestinal epithelium 43 . Thus, disrupting the homeostatic AMP production through effects on enteric neurons might be a potential mechanism by which morphine inhibits the intestinal barrier function. The intestinal mucosa also receives innervation from extrinsic nerve bers of the vagus, spinal afferents, and sympathetic neurons 44 ; these neurons also play an essential role in maintaining the intestinal epithelial barrier. Gut-innervating nociceptive dorsal root ganglia neurons can direct goblet cells to produce mucus and maintain the epithelial barrier 45 . These neurons can also regulate the density of specialized antigen-sampling microfold cells in the gutassociated lymphoid tissue of the Peyer's patches. Breakdown of this crosstalk predisposes the intestinal mucosa to invasion by pathogenic microbes 46 . Like with enteric neurons, chronic exposure to morphine signi cantly alters the excitability of nociceptive dorsal root ganglia neurons 11,12 . Studies are required to investigate the contribution of gut-innervating extrinsic neurons versus enteric neurons in the mechanism of morphine-induced disruption of the intestinal barrier.
Replacing the dysbiotic gut ora with "normal" ora by FMT is clinically utilized for treating several intestinal pathologies [47][48][49][50] . Banerjee et al. 14 reported that FMT from control mice prevented morphineinduced dysbiosis and disruption of the intestinal epithelial barrier. Consistent with these ndings, in the present study, PP-FMT prevented morphine-induced epithelial permeability and downregulation of Reg3γ, indicating improved epithelial barrier function. Metabolomic analysis of the fecal samples of PP and MP mice revealed reduced butyrate levels in chronic morphine-treated mice. Similar results have been noted in the stool samples of human patients exposed to opioids 7 . Consistent with prior ndings 11 , the abundance of the phylum Firmicutes, which comprises butyrate-producing bacteria 22 , was diminished in chronic morphine-treated mice. The depletion of Firmicutes might explain low fecal butyrate levels in chronic morphine-treated mice.
Butyrate is well-known for its positive effects on host-gut microbiota interactions as it engages pathways that regulate epithelial permeability, in ammation, and immune function 51 . Butyrate administration protected against the morphine-induced increase in epithelial permeability. This effect of butyrate could be due to its actions on epithelial tight junction proteins, such as occludin and zonula occluden protein-1 51 , which are downregulated in the ileum of chronic morphine-treated mice 27 [59][60][61][62] . It is important to note that G-protein-coupled receptors for SCFAs are found not only on epithelial cells but also expressed by enteric neurons, and autonomic and somatosensory ganglia [63][64][65] . Hence, the effects of butyrate on the epithelial barrier could also be modulated, in part, by direct actions on neurons.
Oral butyrate administration reduced the alpha diversity of the fecal microbiota to placebo levels.
However, the beta diversity remained signi cantly different. The intestinal epithelium maintains the composition of the enteric bacteria through multiple mechanisms, including secretion of AMPs by Paneth cells and IgA by lamina propria immune cells 19 . Morphine exposure has been previously reported to reduce IgA production by the gut-associated lymphoid tissue through the downregulation of TGF-β 66 . Further studies are required to investigate the effect of butyrate on IgA production in morphine-treated mice.
Several studies have reported CNS effects of microbial-derived butyrate or orally administered butyrate, including in substance use disorders 67,68 . In the present study, oral administration of PP-FMT or sodium butyrate to chronic morphine-treated mice attenuated the development of tolerance to morphine-induced antinociception. The inhibition of tolerance was dependent on the dose of sodium butyrate. Although SCFAs can cross the blood-brain barrier [76][77][78][79] , approximately 95% of orally administered butyrate is absorbed locally in the intestine and extremely low concentrations of oral butyrate are detected in systemic circulation or brain [69][70][71] . Thus, the effect of orally administered sodium butyrate on the development of tolerance to morphine-induced antinociception is most likely through peripheral processes mediating the crosstalk between the gut and the brain. Accumulating evidence indicates that gut microbes, through their products, such as SCFAs, can modulate neural circuits that relay information between the PNS and the CNS [72][73][74] . A recent study showed that eliminating gut bacteria increased the expression of c-Fos, a marker of neuronal activity, in gut-innervating neurons that connect the brainstem sensory nuclei to the sympathetic ganglia, but colonization of mice with butyrate-producing bacteria or feeding butyrate reduced c-Fos expression 72 . Furthermore, eliminating enteric bacteria with antibiotics prevented the development of morphine tolerance in nociceptive dorsal root ganglia neurons 11,12 . Nociceptive dorsal root ganglia neurons transduce noxious stimuli from the periphery to the CNS and have been implicated in the induction of tolerance to morphine-induced antinociception 75,76 . Treating naïve dorsal root ganglia neurons with gut-derived mediators from chronic morphine-treated mice induced morphine tolerance 12 , which was attenuated when cells were incubated with gut-derived mediators from chronic morphine-treated mice exposed to antibiotics 12 .
In the present study, FMT from placebo-pelleted mice attenuated tolerance in chronic morphine-treated mice. The reverse was not observed when placebo-pelleted mice received FMT from morphine-pelleted mice. Placebo-pelleted mice treated with FMT from morphine-pelleted mice did not develop tolerance to morphine. These results were inconsistent with a previous report from Zhang et al. 16 that noted the induction of tolerance to morphine-induced antinociception in germ-free mice treated with FMT from chronic morphine-treated animals 16 . Germ-free mice exhibit signi cant developmental de cits, especially in the gut-associated immune system 77 ; consequently, the microbiota could disrupt the epithelial barrier without resistance and activate downstream events that result in the development of tolerance. Speci c pathogen-free mice have an intact barrier function that maintains spatial segregation between the intestinal epithelium and the bacteria contained within the FMT of morphine-pelleted mice. In the present study, placebo-pelleted mice treated with FMT from morphine-pelleted mice exhibited reduced epithelial permeability and intact Reg3γ expression comparable to that in sham-treated placebo mice. Placebopelleted mice treated with FMT from morphine-treated animals also exhibited reduced IL-17 levels compared to morphine-pelleted mice treated with FMT from morphine-treated animals 14 . Elevated IL-17 levels have been previously reported in chronic morphine-treated mice, and neutralization of IL-17 protected epithelial barrier function and prevented opioid-induced dysbiosis 15 . Therefore, our results that MP-FMT does not disrupt epithelial barrier function and induce tolerance to morphine-induced antinociception in placebo-pelleted mice indicate that dysbiosis alone is insu cient for downstream events and that morphine-induced disruption of the homeostatic functions of the intestinal epithelium is essential for the development of tolerance.

LIMITATIONS OF THE PRESENT STUDY
In the present study, there was a signi cant reduction in the antimicrobial activity of the intestinal epithelium of chronic morphine-treated mice against Gram-negative bacteria. However, the identity of speci c AMPs active against Gram-negative bacteria was not determined. One possible candidate could be Reg3β, an isoform of Reg3γ, which is inducible and exhibits activity against Gram-negative bacteria 78 .
A second limitation of the current study is that only male mice were investigated. Studies in humans and rodents have reported sex as a biological variable in uencing opioid-induced antinociception 82 .
Additionally, sexual dimorphism in the composition of gut bacteria has been noted in laboratory animals and humans, and studies have shown that sex hormones actively in uence the gut microbiome 83 .

CONCLUSION
The intestinal epithelium plays a critical role in controlling the composition of the intestinal microbiota and helps maintain a symbiotic relationship with commensal bacteria. Here, we report that chronic treatment with morphine reduced the antimicrobial activity of the intestinal epithelium via opioid receptors, implicating a potential mechanism underlying opioid-induced dysbiosis. Preventing morphineinduced disruption of the epithelial barrier function with FMT or sodium butyrate inhibited the development of tolerance to antinociception. Finally, dysbiosis alone is insu cient for inducing tolerance to morphine-induced antinociception and disruption of the intestinal epithelium is required for the development of tolerance. In conclusion, our results implicate a mechanism by which morphine disrupts homeostasis in the microbiota-gut-brain axis.

DISCLOSURES
The authors declare no competing interests.

DATA AVAILABILITY
The data supporting these ndings have been deposited with NCBI BioProject       Effect of morphine and sodium butyrate on the activity of the ileum against L.reuteri. (A) Schematic of the antibacterial activity assay protocol. Each agar plate is divided into quadrants and in quadrant I ileum-derived conditioned media incubated with bacteria was smeared. In quadrant II ileum conditioned media containing only nutrient broth and no bacteria was applied; in quadrant III antibiotic-free DMEM/F12 media incubated with bacteria was smeared; and in quadrant IV antibiotic-containing DMEM/F12 media incubated with bacteria was applied. Conditioned media obtained from the ileum of (B) PP mice, (C) MP mice or (D) MP mice treated with 0.250 M sodium butyrate (MP + Butyrate) was used to evaluate the antibacterial activity of the ileum against the Gram-positive bacteria, L.reuteri. Representative images of L.reuteri growing on agar plates have been shown in B-D. Figure E is the normalized data of the %bactericidal activity of the ileum supernatants from PP, MP and MP + butyrate mice. Data in B-D were analyzed by repeated-measures one-way ANOVA and Tukey's multiple comparisons test was used for post hoc analysis. Data in E was analyzed by one-way ANOVA and Holm-Sidak's multiple comparisons test was used for post hoc analysis. P values for relevant comparisons have been indicated. All data are mean ± SEM and scatter represent data from samples from individual mice. Sample sizes are 6, 7, and 6 per group for B, C, and D, respectively. Sample sizes for E are: 6 (PP), 7 (MP) and 6 (MP + butyrate).  Effect of morphine and sodium butyrate on the activity of the ileum against E.coli. Conditioned media obtained from the ileum of (A) PP mice, (B) PP mice treated with 0.250 M sodium butyrate (PP + butyrate), (C) MP mice or (D) MP mice treated with 0.250 M sodium butyrate (MP + Butyrate) was used to evaluate the antibacterial activity of the ileum against the Gram-negative bacteria, E.coli. Representative images of E.coli growing on agar plates have been shown in A-D. Each agar plate is divided into quadrants and in quadrant I ileum-derived conditioned media incubated with bacteria was smeared. In quadrant II ileum conditioned media containing only nutrient broth and no bacteria was applied; in quadrant III antibiotic-free DMEM/F12 media incubated with bacteria was smeared; and in quadrant IV antibiotic-containing DMEM/F12 media incubated with bacteria was applied. Figure E is the normalized data of the %bactericidal activity of the ileum supernatants from PP, PP + butyrate, MP and MP + butyrate mice. Data in A-D were analyzed by repeated-measures one-way ANOVA and Tukey's multiple comparisons test was used for post hoc analysis. Data in E was analyzed by two-way ANOVA and Tukey's multiple comparisons test was used for post hoc analysis. P values for relevant comparisons have been indicated. All data are mean ± SEM and scatter represents data from samples from individual mice. Sample sizes are 5 per group for A-D. Sample sizes for E are: 5/group. Effect of naloxone on morphine-induced decrease in antibacterial activity of the ileum. The effect of the non-selective opioid receptor antagonist, naloxone, was evaluated on the inhibitory effect of morphine on the antibacterial activity of the ileum against (A) L.reuteri, and (B) E.coli. Data are normalized %bactericidal activity and presented as mean ± SEM. Scatter represents data from ileum-derived conditioned media from individual mice. Data were analyzed by one-way ANOVA and Tukey's multiple comparisons test was used for post hoc analysis. P values for relevant comparisons have been indicated. Sample sizes in A and B are 6, 6, and 7 for saline, morphine and morphine + naloxone groups, respectively.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.