Pharmacomicrobiomics: Investigating the Relationship between the Gut Microbiome and Drug Disposition

Abstract

Pharmacomicrobiomics is an emerging field that aims to understand how variation in the microbiome can impact drug disposition, action, and/or toxicity. In this dissertation, we used a humanized mouse model, multi-omics approaches, in vitro systems, and clinical samples to investigate the relationship between the gut microbiome and drug-metabolizing enzymes and transporters. Specifically, we focused on cytochrome P450 3A4 (CYP3A4), which is responsible for the metabolism of up to 40% of clinically used drugs. Chapter 2 examines how differences in microbial conditions impacted CYP3A4 expression in a pregnane X receptor-constitutive androstane receptor-cytochrome P450 3A4/3A7 (PXR-CAR-CYP3A4/3A7) humanized mouse model. We studied three different groups of male and female mice: conventional (CV), germ-free (GF), and conventionalized (GFCV) mice which were GF mice that were inoculated with sex-matched pooled human feces at 8 weeks (n = 7 per sex per group). As previous studies only investigated how the microbiome impacts the expression of Cyp3a11, the mouse orthologue of CYP3A4, this chapter’s goal was to investigate whether the microbiome affected human CYP3A4. In summary, the presence of microbiome upregulated CYP3A4 expression by 7.6-fold in male mice (p < 0.001) but downregulated CYP3A4 expression by 1.69-fold in female mice (p = 0.012). This highlights that there is both a sex-dependent and microbiome-dependent effect. We found a strong positive correlation between microbial alpha diversity and hepatic CYP3A4 expression (R = 0.66; p = 0.0072), in agreement with previous studies of Cyp3a11. Chapter 3 utilized multi-omics approaches (i.e., metagenomics, transcriptomics, metabolomics, and proteomics) to understand changes in the hepatic environment with alterations in microbial conditions and between sexes in a subset of the PXR-CAR-CYP3A4/3A7 mice in Chapter 2. Male CV mice had a more variable but less diverse microbiome compared to female CV mice. Additionally, principal components analysis (PCA) of the transcriptome and metabolome revealed that male CV mice displayed the greatest interindividual variability within the group, whereas the transcriptome and metabolome of each of the other groups (i.e., male GF, female CV and female GF mice) were more tightly clustered. Lastly, presence of a microbiome had a greater impact on drug metabolizing enzymes and transporters (DMET) gene expression in male mice (43.4% differentially expressed genes) compared to female mice (11% differentially expressed genes). Chapter 4 further expands on the PXR-CAR-CYP3A4/3A7 mouse work through a correlation analysis between CYP3A4 expression and the hepatic metabolome and subsequent in vitro testing of selected microbial metabolites in HepaRG cells. Microbial metabolites, such as short chain fatty acids, secondary bile acids, and tryptophan metabolites (i.e., indole derivatives) have been shown to induce CYP3A4 expression through nuclear receptor binding. Therefore, our goal was to identify other microbial metabolites that can also influence CYP3A4 expression. We identified three microbial metabolites with strong correlations to hepatic CYP3A4 expression: 1 deoxynojirimycin (ρ = 0.72, p = 1 x 10-4), betaine (ρ = 0.66, p = 0.00056), and L-pyroglutamic acid (ρ = 0.62, p = 0.0015). Physiologically-relevant concentrations (up to 150 μM of each microbial metabolite) were incubated in HepaRG cells for 72 hours, however no CYP3A4 induction was observed. In addition, CYP3A4 activity in pooled liver microsomes was not inhibited by any of the metabolites. Therefore, we concluded that the identified metabolites were not regulators of CYP3A4 under the conditions tested. To expand our findings into humans, the aim of Chapter 5 was to explore the relationship between normal variation in the gut microbiome and CYP3A activity. We used clinical samples collected from healthy male volunteers (n = 7) dosed with midazolam semi-simultaneously (2 mg orally at 0 h and 1 mg intravenously at 3 h). Midazolam and its primary metabolites (1’-hydroxymidazolam [1’OH-MDZ] and 4 hydroxymidazolam [4-OH MDZ]) were quantified in plasma samples collected over 12 hours. Metrics related to CYP3A activity (clearance, hepatic and intestinal extraction ratios, and metabolite-to-parent AUC ratios) were calculated and correlated with microbial diversity. Microbial species richness was not correlated with clearance, hepatic extraction ratio or intestinal extraction ratio, however an inverse relationship with observed with the 1’-OH MDZ/MDZ AUC ratio (R = -0.86, p = 0.024) and 4-OH MDZ/MDZ AUC ratio (R = -0.89, p = 0.012). These results suggest that normal variation in the gut microbiome might not influence intestinal or hepatic CYP3A activity but may have an effect on glucuronidation as the parent-to-metabolite AUC ratios account for both formation clearance and metabolite clearance. Following metabolomics analysis of the plasma samples, we identified two tryptophan metabolites strongly correlated with hepatic CYP3A (i.e., hepatic extraction ratio), serotonin (R = 1, p = 0.017) and xanthurenic acid (R = 0.86, p = 0.024). HepaRG cells were treated with these metabolites and we found that 100 nM of serotonin or xanthurenic acid induced CYP3A4 expression by ~2.0-fold (p < 0.01). However, when tested in a PXR reporter assay, these metabolites did not appear to be PXR agonists. In conclusion, this dissertation revealed that differences in microbial conditions more strongly influenced CYP3A4 expression in a humanized mouse model in a sex-dependent manner and the effect on CYP3A metrics in humans was less apparent. We identified two metabolites (i.e., serotonin and xanthurenic acid) as potential inducers of CYP3A4. Findings from this dissertation contribute to further understanding the relationship between the gut microbiome and drug disposition, specifically through CYP3A4 regulation.