Chronic opioid use modulates human enteric microbiota and intestinal barrier integrity

Abstract

Over the past three decades the United States has experienced a devastating opioid epidemic. One of the many debilitating side effects of chronic opioid use is opioid-induced bowel dysfunction. We investigated the impact of methadone maintenance treatment (MMT) on the gut microbiome, the gut bacterial metabolite profile, and intestinal barrier integrity. An imbalance in key bacterial communities required for production of short-chain fatty acids (SCFAs), mucus degradation, and maintenance of barrier integrity was identified. Consistent with dysbiosis, levels of fecal SCFAs were reduced in MMT. We demonstrated that metabolites synthesized by Akkermansia muciniphila modulate intestinal barrier integrity in vitro by strengthening the pore pathway and regulating tight junction protein expression. This study provides essential information about the therapeutic potential of A. muciniphila and warrants development of new clinical strategies that aim to normalize the gut microbiome in individuals affected by chronic opioid use.

Keywords: Akkermansia muciniphila; Intestinal permeability; metabolomics; methadone; short-chain fatty acids; tight junctions.

Figure 1.

β- and α-diversity of the fecal microbiome is modulated by methadone treatment. a, Principal component analysis (PCA) revealed changes in the bacterial composition between the non-opioid users (blue) and methadone treated individuals (red) (PCo2 p < .0001). b, Bacterial biodiversity analysis showed decreased α-diversity in evenness (p = .03) and richness (p = .001) in the methadone cohort (red circles) using the Shannon and Chao1 index, respectively compared to non-opioid users (blue circles). Statistical significance was determined using a two-tailed unpaired t-test (a and b). A violin plot was used to represent the distribution of fecal samples from 28 non-opioid users and 34 methadone-treated individuals (dashed lined indicates the median, dotted line indicates quartiles)

Figure 1.

β- and α-diversity of the fecal microbiome is modulated by methadone treatment. a, Principal component analysis (PCA) revealed changes in the bacterial composition between the non-opioid users (blue) and methadone treated individuals (red) (PCo2 p < .0001). b, Bacterial biodiversity analysis showed decreased α-diversity in evenness (p = .03) and richness (p = .001) in the methadone cohort (red circles) using the Shannon and Chao1 index, respectively compared to non-opioid users (blue circles). Statistical significance was determined using a two-tailed unpaired t-test (a and b). A violin plot was used to represent the distribution of fecal samples from 28 non-opioid users and 34 methadone-treated individuals (dashed lined indicates the median, dotted line indicates quartiles)

Figure 2.

Increased fecal Actinobacteria and decreased Verrucomicrobia abundance with methadone treatment. a, Analysis of bacterial 16S rRNA from fecal samples revealed 13 phyla from the methadone-treated cohort (horizontal red bar) and non-opioid users (horizontal blue bar), in which each column represents an individual donor and the top 8 most abundant bacteria phyla. b, c, f, The relative abundance of the core microbiota, Bacteroidetes, Firmicutes, Proteobacteria, Actinobacteria, and Verrucomicrobia was compared between groups. Bacteroidetes, Firmicutes, and Proteobacteria (b) from methadone-treated individuals (red circles) showed no statistical difference in relative abundance compared to non-opioid users (blue circles). Actinobacteria was significantly increased (c) (p = .005) and Verrucomicrobia significantly decreased (f) (p = .05) in the methadone group (red circles). d, Within the Actinobacteria phylum, eight bacterial families were identified. Only Bifidobacteriaceae (p = .01) showed increased relative abundance in the methadone-treated group (red circles). e, Bifidobacterium bifidum (p = .04) & Bifidobacterium longum (p = .05) were increased in the methadone-treated group (red circles). g,h, Within the Verrucomicrobia phylum, Akkermansiaceae family (p = .05) (g) and the Akkermansia muciniphila species (p = .05) (h) were decreased in methadone-treated individuals (red circles) compared to non-opioid users (blue circles). Statistical significance was determined using a two-tailed unpaired t-test (b-h). Violin plots were used to represent the distribution of fecal samples from 28 non-opioid users (blue) and 34 methadone-treated individuals (red) for the taxonomic classifications (phylum, family, genus, and species) (dashed lined indicates the median, dotted line indicates quartiles)

Figure 2.

Increased fecal Actinobacteria and decreased Verrucomicrobia abundance with methadone treatment. a, Analysis of bacterial 16S rRNA from fecal samples revealed 13 phyla from the methadone-treated cohort (horizontal red bar) and non-opioid users (horizontal blue bar), in which each column represents an individual donor and the top 8 most abundant bacteria phyla. b, c, f, The relative abundance of the core microbiota, Bacteroidetes, Firmicutes, Proteobacteria, Actinobacteria, and Verrucomicrobia was compared between groups. Bacteroidetes, Firmicutes, and Proteobacteria (b) from methadone-treated individuals (red circles) showed no statistical difference in relative abundance compared to non-opioid users (blue circles). Actinobacteria was significantly increased (c) (p = .005) and Verrucomicrobia significantly decreased (f) (p = .05) in the methadone group (red circles). d, Within the Actinobacteria phylum, eight bacterial families were identified. Only Bifidobacteriaceae (p = .01) showed increased relative abundance in the methadone-treated group (red circles). e, Bifidobacterium bifidum (p = .04) & Bifidobacterium longum (p = .05) were increased in the methadone-treated group (red circles). g,h, Within the Verrucomicrobia phylum, Akkermansiaceae family (p = .05) (g) and the Akkermansia muciniphila species (p = .05) (h) were decreased in methadone-treated individuals (red circles) compared to non-opioid users (blue circles). Statistical significance was determined using a two-tailed unpaired t-test (b-h). Violin plots were used to represent the distribution of fecal samples from 28 non-opioid users (blue) and 34 methadone-treated individuals (red) for the taxonomic classifications (phylum, family, genus, and species) (dashed lined indicates the median, dotted line indicates quartiles)

Figure 3.

Fecal SCFA profile is altered with methadone use and correlates with core microbiota. a, SCFA content in feces was assessed by gas chromatography-mass spectrometry. Acetate (p = .006), propionate (p = .01), and butyrate (p = .01) levels are decreased in the methadone cohort (red circles). b, Spearman correlation between fecal SCFAs and the relative abundance of the core microbiome from non-opioid users and methadone-treated individuals. Heatmap shows Spearman r coefficient values, and asterisks denotate statistical significance. Acetate correlates positively with Actinobacteria (p = .02) and Firmicutes (p = .03) and negatively with Proteobacteria (p = .0004) in methadone-treated individuals. Butyrate correlates positively with Firmicutes (p = .03) and negatively with Bacteroidetes (p = .05) and Proteobacteria (p = .006) in methadone-treated individuals. p ≤ 0.05 (*), p ≤ 0.005 (**), p ≤ 0.0005 (***). Statistical significance was determined using a two-tailed unpaired t-test (a). A violin plot was used to represent the distribution of fecal samples from 24 non-opioid users and 32 methadone-treated individuals (dashed lined indicates the median, dotted line indicates quartiles)

Figure 3.

Fecal SCFA profile is altered with methadone use and correlates with core microbiota. a, SCFA content in feces was assessed by gas chromatography-mass spectrometry. Acetate (p = .006), propionate (p = .01), and butyrate (p = .01) levels are decreased in the methadone cohort (red circles). b, Spearman correlation between fecal SCFAs and the relative abundance of the core microbiome from non-opioid users and methadone-treated individuals. Heatmap shows Spearman r coefficient values, and asterisks denotate statistical significance. Acetate correlates positively with Actinobacteria (p = .02) and Firmicutes (p = .03) and negatively with Proteobacteria (p = .0004) in methadone-treated individuals. Butyrate correlates positively with Firmicutes (p = .03) and negatively with Bacteroidetes (p = .05) and Proteobacteria (p = .006) in methadone-treated individuals. p ≤ 0.05 (*), p ≤ 0.005 (**), p ≤ 0.0005 (***). Statistical significance was determined using a two-tailed unpaired t-test (a). A violin plot was used to represent the distribution of fecal samples from 24 non-opioid users and 32 methadone-treated individuals (dashed lined indicates the median, dotted line indicates quartiles)

Figure 4.

Methadone treatment associates with increased plasma IL-6 and TNF-α. Levels of plasma IL-6 (p = .03), TNF-α (p = .006), LBP (p = ns), and as well as fecal LCN2 (p = ns) between non-opioid users (blue) and methadone-treated individuals (red) were measured. Statistical significance was determined using a two-tailed unpaired t-test

Figure 4.

Methadone treatment associates with increased plasma IL-6 and TNF-α. Levels of plasma IL-6 (p = .03), TNF-α (p = .006), LBP (p = ns), and as well as fecal LCN2 (p = ns) between non-opioid users (blue) and methadone-treated individuals (red) were measured. Statistical significance was determined using a two-tailed unpaired t-test

Figure 5.

Components of Akkermansia muciniphila spent media, not short-chain fatty acids (SCFAs) or outer membrane vesicles (OMVs), modulate intestinal barrier integrity. a, TEER of an intestinal epithelial Caco-2 monolayer in a transwell culture was measured after exposure to Akkermansia muciniphila spent media as noted in the model diagram. b, Akkermansia muciniphila spent media (blue) added to the upper chamber of the transwell increased epithelial barrier integrity at 3 and 6 h (p < .0001) compared to control BHI (orange). c, Short-chain fatty acid content in Akkermansia muciniphila spent media (n = 2). d, Akkermansia muciniphila OMV-depleted media (light pink) added to the upper chamber of the transwell increased epithelial barrier integrity at 3 (p = .02) and 6 h (p < .0001) compared to untreated cells (black) and cells exposed to EV-depleted BHI media (light blue). e, Akkermansia muciniphila OMVs (dark pink) and BHI EVs (light yellow) added to the upper chamber of the transwell had no effect on barrier integrity compared to untreated cells (black). f, Percent positive (left) and MFI (right) levels of IgG against Akkermansia muciniphila in non-opioid users and methadone-treated individuals was measured using flow cytometry. Statistical significance was evaluated with a two-way ANOVA with Sidak correction (b), Tukey correction (d & e), and Welch’s t-test (f). Unless otherwise noted, mean and standard deviation from n = 3 biological replicates are shown

Figure 5.

Components of Akkermansia muciniphila spent media, not short-chain fatty acids (SCFAs) or outer membrane vesicles (OMVs), modulate intestinal barrier integrity. a, TEER of an intestinal epithelial Caco-2 monolayer in a transwell culture was measured after exposure to Akkermansia muciniphila spent media as noted in the model diagram. b, Akkermansia muciniphila spent media (blue) added to the upper chamber of the transwell increased epithelial barrier integrity at 3 and 6 h (p < .0001) compared to control BHI (orange). c, Short-chain fatty acid content in Akkermansia muciniphila spent media (n = 2). d, Akkermansia muciniphila OMV-depleted media (light pink) added to the upper chamber of the transwell increased epithelial barrier integrity at 3 (p = .02) and 6 h (p < .0001) compared to untreated cells (black) and cells exposed to EV-depleted BHI media (light blue). e, Akkermansia muciniphila OMVs (dark pink) and BHI EVs (light yellow) added to the upper chamber of the transwell had no effect on barrier integrity compared to untreated cells (black). f, Percent positive (left) and MFI (right) levels of IgG against Akkermansia muciniphila in non-opioid users and methadone-treated individuals was measured using flow cytometry. Statistical significance was evaluated with a two-way ANOVA with Sidak correction (b), Tukey correction (d & e), and Welch’s t-test (f). Unless otherwise noted, mean and standard deviation from n = 3 biological replicates are shown

Figure 6.

Akkermansia muciniphila spent media and methadone increase intestinal epithelial barrier integrity in vitro independently with distinct kinetics. a, 100 µM methadone (red) added to the lower chamber of the transwell increased epithelial barrier integrity at 24, 48, and 72 h (p < .0001) compared to an untreated Caco-2 monolayer (black). b, The combination of Akkermansia muciniphila spent media and 100 µM methadone (maroon) added to the upper and lower chambers, respectively, increased barrier integrity with distinct kinetics: 3 h (p = .0004) and 6 h (p < .0001) due to Akkermansia muciniphila spent media and 24 h (p < .0001), 48 h (0.004), and 72 h (p = .004) in response to methadone. c, TEER of the Caco-2 monolayer was measured in the presence of graded concentrations of the OR rescue antagonist naloxone alone or in combination with 100 µM methadone. High dose naloxone enhanced the methadone-stimulated increase in barrier integrity (p < .0001), when compared to an untreated Caco-2 monolayer, naloxone alone, or methadone alone. d, Protein expression and quantification for claudin 1 and claudin 4 after a Caco-2 monolayer was treated with Akkermansia muciniphila spent media (blue), methadone (red), or the combination (maroon) for 0, 6, and 24 h. Densitometric quantification normalized to GAPDH, relative to the untreated sample (black). e, A Caco-2 monolayer was stimulated with A. muciniphila spent media (blue), 100 µM methadone (red), or the combination (maroon) for 6, 9, and 12 h. mRNA expression (fold change measured by RT-qPCR) is relative to untreated sample (black), set to 1. Statistical significance was determined using a two-way ANOVA with Sidak correction (a, b), Tukey correction (c), and one-way ANOVA with Dunnett correction (d-e). Mean and standard deviation from n = 3 biological replicates (a-c, e) and n = 5 biological replicates (d) are shown

Figure 6.

Akkermansia muciniphila spent media and methadone increase intestinal epithelial barrier integrity in vitro independently with distinct kinetics. a, 100 µM methadone (red) added to the lower chamber of the transwell increased epithelial barrier integrity at 24, 48, and 72 h (p < .0001) compared to an untreated Caco-2 monolayer (black). b, The combination of Akkermansia muciniphila spent media and 100 µM methadone (maroon) added to the upper and lower chambers, respectively, increased barrier integrity with distinct kinetics: 3 h (p = .0004) and 6 h (p < .0001) due to Akkermansia muciniphila spent media and 24 h (p < .0001), 48 h (0.004), and 72 h (p = .004) in response to methadone. c, TEER of the Caco-2 monolayer was measured in the presence of graded concentrations of the OR rescue antagonist naloxone alone or in combination with 100 µM methadone. High dose naloxone enhanced the methadone-stimulated increase in barrier integrity (p < .0001), when compared to an untreated Caco-2 monolayer, naloxone alone, or methadone alone. d, Protein expression and quantification for claudin 1 and claudin 4 after a Caco-2 monolayer was treated with Akkermansia muciniphila spent media (blue), methadone (red), or the combination (maroon) for 0, 6, and 24 h. Densitometric quantification normalized to GAPDH, relative to the untreated sample (black). e, A Caco-2 monolayer was stimulated with A. muciniphila spent media (blue), 100 µM methadone (red), or the combination (maroon) for 6, 9, and 12 h. mRNA expression (fold change measured by RT-qPCR) is relative to untreated sample (black), set to 1. Statistical significance was determined using a two-way ANOVA with Sidak correction (a, b), Tukey correction (c), and one-way ANOVA with Dunnett correction (d-e). Mean and standard deviation from n = 3 biological replicates (a-c, e) and n = 5 biological replicates (d) are shown