Morphine induces changes in the gut microbiome and metabolome in a morphine dependence model

Abstract

Opioid analgesics are frequently prescribed in the United States and worldwide. However, serious comorbidities, such as dependence, tolerance, immunosuppression and gastrointestinal disorders limit their long-term use. In the current study, a morphine-murine model was used to investigate the role of the gut microbiome and metabolome as a potential mechanism contributing to the negative consequences associated with opioid use. Results reveal a significant shift in the gut microbiome and metabolome within one day following morphine treatment compared to that observed after placebo. Morphine-induced gut microbial dysbiosis exhibited distinct characteristic signatures, including significant increase in communities associated with pathogenic function, decrease in communities associated with stress tolerance and significant impairment in bile acids and morphine-3-glucuronide/morphine biotransformation in the gut. Moreover, expansion of Enterococcus faecalis was strongly correlated with gut dysbiosis following morphine treatment, and alterations in deoxycholic acid (DCA) and phosphatidylethanolamines (PEs) were associated with opioid-induced metabolomic changes. Collectively, these results indicate that morphine induced distinct alterations in the gut microbiome and metabolome, contributing to negative consequences associated with opioid use. Therapeutics directed at maintaining microbiome homeostasis during opioid use may reduce the comorbidities associated with opioid use for pain management.

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1

Morphine treatment leads to temporal modulation of the gut microbiome in 3 days post-treatment. Beta diversity measures of the gut microbiome following treatment with placebo, morphine, naltrexone, or morphine plus naltrexone. Wild-type mice were implanted with placebo (P), 25 mg morphine (M), 30 mg naltrexone (N), or morphine and naltrexone (MN) pellets subcutaneously (n = 4 in each group). Fecal samples were taken for analysis at the following time points: day 0, day 1, day 2, and day 3 post-treatment. Principal coordinates analysis of samples from day 0 (A), day 1 (B), day 2 (C) and day 3 (D) using the UniFrac metric at the OTU level. In Fig. 1A, F value = 4.766, total degree of freedom (DF) = 27, DF(treatment, between columns) = 2, DF(residual, within columns) = 25. In Fig. 1B, F = 1.789, DF (total) = 27, DF(treatment, between columns) = 2, DF(residual, within columns) = 25. In Fig. 1C, F = 17.25, DF (total) = 27, DF(treatment, between columns) = 2, DF(residual, within columns) = 25. In Fig. 1D, F = 89.39, DF (total) = 27, DF(treatment, between columns) = 2, DF(residual, within columns) = 25. Mantel test is run over these distance classes versus the ecological distance matrix. Parametric p-value (Bonferroni-corrected) < 0.01.

Figure 2

Morphine treatment leads to sustained modulation of the gut microbiome for 6 days post-treatment. Beta diversity measures of the gut microbiome following treatment with placebo, morphine, naltrexone, or morphine plus naltrexone. Wild-type mice were implanted with placebo (P), 25 mg morphine (M), 30 mg naltrexone (N), or morphine and naltrexone (MN) pellets subcutaneously (n = 4 in each group). Fecal samples were taken for analysis at the following time points: day 0, day 4, day 5, and day 6 post-treatment. Principal coordinates analysis of samples from day 0, day 4, day 5, and day 6 post-treatment using the UniFrac metric at the OTU level. F value = 7.304. Total degree of freedom (DF) is 495, including DF of treatment (between columns), which is 35, and DF of residuals (within columns), which is 460. A Mantel test is run over these distance classes versus the ecological distance matrix. Parametric p-value (Bonferroni-corrected) 

Figure 3

Morphine treatment induces distinct changes in the gut microbiome. Wild-type mice were implanted with placebo or 25 mg morphine pellets subcutaneously. Fecal samples were taken for analysis at day 3 post-treatment. (A) Alpha diversity was assessed using the chao1 index. Morphine treatment (n = 8) results in decreased alpha diversity compared to that of controls (n = 8) measured by using the chao1 index. The OTU tables were rarefied at the cutoff value of 31000 sequences per sample. (B) t-test was conducted on the chao1 index. **Indicates a significant difference, P value = 0.0030. (C) Principal coordinates analysis (PCoA) of samples using the UniFrac metric at the OTU level. (D) UniFrac distance significant tests were performed using QIIME. The tests of significance were performed using a two-sided student’s t-test. *Parametric p-value (Bonferroni-corrected) < 0.05, **Parametric p-value (Bonferroni-corrected) < 0.01. ANOVA test indicated P < 0.0001, F value = 63.29. Total degree of freedom (DF) is 119, including DF of treatment (between columns), which is 2 and DF of residuals (within columns), which is 117. (E) Morphine treatment results in a significant increase in pathogenic bacteria. Multiple hypothesis test with the given threshold (FDR = 0.05) demonstrates that the relative abundance of potential pathogenic bacteria (genus level) increases significantly at day 3 post-treatment with morphine compared to that following placebo treatment. Increased (red color) representative pathogenic bacteria include Flavobacterium, Enterococcus, Fusobacterium, Sutterella, Clostridium.

Figure 4

Enterococcus faecalis is a biomarker of morphine-induced alteration of the gut microbiome. Real-time PCR expression profiling of species-specific 16S-rRNA gene in gut microbiota. (A) The expression of species-specific 16S-rRNA gene was profiled in stool samples from mice s.c. implanted with placebo (control) or morphine using a qPCR assay (n = 8 in each group). The heat map was generated by a log transformation of the real-time PCR data presented as ΔCT (CT_species – CT_universal_16SrRNA). Red color indicates increased levels of amplification. (B) E. Faecalis 16S-rRNA gene amplification fold change due to treatments on day 3 post-treatment (n = 4 in each group). (C) E. Faecalis 16S rRNA genes amplification fold change due to treatments in a short-term study (n = 8 in each group).

Figure 5

Enterococcus faecalis infection accelerates morphine induced analgesic tolerance. The mice were administrated with 5 g/L of streptomycin sulphate in the drinking water for 2 days and switch to normal drinking water for 24 hours before E. faecalis (EF) inoculation by oral gavage. The spectinomycin sulphate selective E. faecalis were diluted up to the concentration of 2 × 1010/mL in phosphate buffered saline (PBS). Each mouse was administered with 200ul spectinomycin solution by oral gavage daily. After 48 hours post gavage, the mice were treated with 250 mg/L spectinomycin sulfate (to prevent overgrowth of pathogenic gram-negative bacteria) in the drinking water during the behavior study. To maintain the population of E. faecalis in the mouse gut, mice were administered the same dose of E. faecalis and the same dose of spectinomycin sulphate by oral gavage daily during the behavior experiment for 8 days. (A) Analgesic effectiveness of morphine was evaluated by mouse reaction to heat. Analgesic tolerance, interpreted as percentage of maximum possible effect (MPE%), was determined by tail flick analgesic test. Mice were intraperitoneally injected with saline or 15 mg/kg morphine twice daily for 8 days with 12 hours interval. Behavioral assessment was performed before and 30 min after saline or morphine administration in the morning. All groups had a minimum of 10 mice/group. For positive controls, a new group of 10 mice treated with Placebo + PBS and a new group of 10 mice treated with Placebo + E. faecalis were administered with 15 mg/kg morphine at each time point to test analgesic reaction of these naive mice to morphine treatment. T-test analyses were used to compare MPE% in each group daily. P value of 0.05 or less was considered significant. (B) Daily nociceptive behavior and morphine antinociception throughout an 8-day chronic morphine schedule (15 mg/kg, intraperitoneally, twice daily): antinociceptive behavior by tail flick. Voltage to the light source was adjusted to achieve baseline latency between 2–3 seconds. The cut-off time is 10 seconds to avoid tissue damage. The mice were put on the tail flick assay for 5 min of habituation everyday for two days before the behavior test. The averages of each two measurements before and after morphine injection were recorded as baseline and response to morphine antinocipective effect at each time point daily during experimental period. The tests of significance were performed using a two-sided student’s t-test. *: morphine compared to morphine + E. faecalis 30 min after morphine injection, #: morphine compared to morphine + E. faecalis baseline. *p < 0.05, **p < 0.01 and ***p < 0.001, #### and ****p < 0.0001.

Figure 6

Metabolomic analysis of fecal matter and identification of significant shifts of gut metabolome following morphine treatment. (A) Mice were treated with 25 mg morphine or placebo pellet subcutaneously (n = 8 in each group). Fecal samples were taken for analysis at 3 days post-treatment. Scores scatter plot of the partial least square discriminant analysis (PLS-DA) model of fecal samples from wild-type mice (C57B6/J) with morphine (□) or placebo (Δ) treatment. The t[1] and t[2] values represent scores of each sample in principal components 1 and 2, respectively. (B) Loadings plot of the principal components (n = 8 in each group). Metabolites contributing to the differences in fecal samples from mice following morphine and placebo treatment were labeled, and their chemical identities were confirmed. (C) In a short-term study, fecal samples were taken from mice at the following time points: day 0, day 1, day 2, and day 3 post-morphine. Scores scatter plot of the partial least square discriminant analysis (PLS-DA) model of fecal samples from wild-type mice (C57B6/J) with morphine (□) or placebo (Δ) treatment at day 0, day 1, day 2, and day 3 post-treatment (n = 4 in each group). The t[1] and t[2] values represent scores of each sample in principal components 1 and 2, respectively. (D) Heat map plot of significant associations with morphine treatments and the loading of indicator metabolites in fecal matter from mice at day 0, day 1, day 2, and day 3 post-treatment (n = 4 in each group). All relative abundances are row z-score-normalized for visualization. (E) Metabolomic analysis of fecal samples and measurement of effects of naltrexone on morphine-induced gut metabolomic shifts at day 3 post-treatments. Mice were treated with placebo, 25 mg morphine, 30 mg naltrexone, or morphine + naltrexone pellets subcutaneously. (F) Relative abundance analysis of metabolites reveals naltrexone–an opioid receptor antagonist–reversed the effect of morphine on loading of deoxycholic acid (DCA), a secondary bile acid, and phosphatidylethanolamines (PEs), a class of phospholipids found in biological membranes. The tests of significance were performed using a two-sided student’s t-test (n = 4 in each group at day 3 post-treatments).

Figure 7

The M3G/MS concentration ratio in serum and fecal matter increases post-morphine treatment. LC-MS analysis identified MS and M3G concentrations. The M3G/MS concentration ratio increases following morphine treatment in both mouse serum and intestinal matter. Statistical significance tests were performed using a two-sided student’s t test (P 

Figure 8

Model of metabolism and biotransformation of morphine and bile acids. In liver, cholesterol is transformed to primary bile acids, and morphine is conjugated to M3G. In the gut, intestinal bacteria transform primary bile acids and M3G into secondary bile acids and morphine, respectively. Bile acids and morphine are reabsorbed and recycled through enterohepatic circulation.