The gut microbiome contributes to somatic morphine withdrawal behavior and implicates a TLR2 mediated mechanism

Abstract

The ongoing opioid epidemic has left millions of people suffering from opioid use disorder due to the over-prescription of highly addictive substances. Chronic opioid exposure leads to dependence, where the absence of the drug results in negative symptoms of withdrawal, often driving patients to continue drug use; however, few therapeutic strategies are currently available to combat the cycle of addiction and the severity of morphine withdrawal. This study investigates the microbiome as a potential therapeutic target for morphine withdrawal, as gut dysbiosis caused by morphine use has been proven to contribute to other aspects of opioid use disorders, such as tolerance. Results show that although the microbiome during morphine withdrawal trends toward recovery from morphine-induced dysbiosis, there continues to be a disruption in the alpha and beta diversity as well as the abundance of gram-positive bacteria that may still contribute to the severity of morphine withdrawal symptoms. Germ-free mice lacking the microbiome did not develop somatic withdrawal symptoms, indicating that the microbiome is necessary for the development of somatic withdrawal behavior. Notably, only TLR2 but not TLR4 whole-body knockout models display less withdrawal severity, implicating that the microbiome, through a gram-positive, TLR2 mediated mechanism, drives opioid-induced somatic withdrawal behavior.

Keywords: Morphine; TLR2; behavior; microbiome; withdrawal.

Figure 1.

Somatic withdrawal symptoms were observed in female and male mice. Schematic of the withdrawal drug treatment paradigm in four groups: morphine-treated females (MORPH_F), morphine-treated males (MORPH_M), placebo treated females (PLCB_F) and placebo treated males (PLCB_M) (a). Total withdrawal (b), as well as top individual symptoms of withdrawal, jumping (c), shaking (d), and grooming (e). All groups had an n = 9–11 animals. Replicates were done with an n = 2–3 per group repeated four times. Statistical analysis for all graphs by two-way ANOVA, Tukey’s multiple comparisons test, **p < .01 between the male and female morphine treated animals. Values plotted are mean with SEM error bars. Schematic of the dependence drug treatment paradigm of two groups: morphine (MORPH) and placebo (PLCB) (f). Mean of total withdrawal scores (g). N = 5 per group, no replicates. Statistical analysis by students t-test, **p < .01.

Figure 2.

Microbiome shows partial recovery during withdrawal. 16s RNA sequencing of stool samples collected during behavior of two groups: morphine (MORPH) and placebo (PLCB). Box plot of alpha diversity measured by the Shannon alpha diversity index with statistical significance labeled with **p < .01, and ***p < .001(A). PcoA plots of beta diversity using the Bray-Curtis metric (b). All differences shown are significant. The phylum density (c) and box plots of relative abundance (d) with T-test and Mann–Whitney test used for statistical analysis and p values numerically represented above graphs. For all graphs in figure: n = 7–10 per group. Replicates were done with an n = 2–3 per group repeated four times.

Figure 3.

Gram-positive bacteria continue to be altered during morphine withdrawal. The relative abundance of the genus was plotted in box plots and separated into groups, recovered gram-negative bacteria (a), recovered gram-positive bacteria (b), altered gram-positive bacteria (c). T-test and Mann–Whitney test used for statistical analysis with numerical p values represented above graph. For all graphs n = 7–10 with four replicates of n = 2–3.

Figure 4.

Antibiotics alter the withdrawal response. Schematic of antibiotic and drug treatment for four groups: antibiotic and morphine (ABX+MORPH), antibiotic and placebo (ABX+PLCB), water and morphine (H2O+MORPH) and water and placebo (H2O+PLCB) (a). The mean total withdrawal score (b) and mean number of jumps (c) were plotted with SEM error bars. N = 9–10 per group with replicates of 2–3 repeated four times. Statistical analysis by two-way ANOVA, Tukey’s multiple comparisons test, with the significance between the ABX+MORPH and H2O+MORPH groups represented with **p < .01, ***p < .001, ****p < .0001. Alternative antibiotic timepoint with the same four treatment groups: antibiotic and morphine (ABX+MORPH), antibiotic and placebo (ABX+PLCB), water and morphine (H2O+MORPH) and water and placebo (H2O+PLCB) (d). The mean total withdrawal was plotted with SEM error bars. Statistical analysis by two-way ANOVA, Tukey’s multiple comparisons test, *p < .05, and ***p < .001 represents the significant differences between the ABX+MORPH and H2O+MORPH groups. N = 9–10 per group with four replicates of 2–3.

Figure 5.

Microbiome after antibiotic treatment. 16s RNA sequencing results of the alpha diversity measured by the Shannon alpha diversity index and plotted in box plots (a). Significance between groups is labeled with *p < .05, **p < .01, ***p < .001. Box plots of the relative abundance of the phylum at 24 h post pellet removal and Wilcox statistical analysis (b). The significant results are indicated by *p < .05, **p < .01, all other differences are not statistically significant. Relative abundance of the genus is displayed in box plot with significant results, of the Wilcox statistical analysis, displayed with *p < .05, **p < .01, ***p < .001, all other differences are p > .05. For all graphs in this figure have an n = 7–9 with four replicates of n = 2–3 per group.

Figure 6.

Germ-Free mice and TLR2 whole body knockout mice display less somatic withdrawal symptoms. Germ-Free mouse paradigm with two groups: germ-free (GF) and specific pathogen free (SPF) (a). The mean total withdrawal (b) and mean number of jumps (c) are plotted with SEM error bars and an n = 6 per group. Statistical analysis by two-way ANOVA, Tukey’s multiple comparisons test, **p < .01, ****p < .0001 compared between the germ-free and SPF mice. TLR4 and TLR2 whole body knockout model paradigm with six total groups: TLR2 knockout with morphine or placebo (TLR2 K/O MORPH and TLR2K/O PLCB) TLR4 knockout with morphine or placebo (TLR4 K/O MORPH and TLR4 K/O PLCB) and wildtype animals with morphine or placebo (WT MORPH and WT PLCB) (d). Mean number of jumps (e) are plotted it with SEM error bars. Statistical analysis by two-way ANOVA, Tukey’s multiple comparisons test, *p < .05 compared between the TLR2 K/O MORPH and WT MORPH groups. There was an n = 9–11 per group. Replicates were run with 1–3 per group repeated five times.