Host Cyp450 enzymes and gut microbiota are important factors affecting pharmacokinetics. Diseases such as NASH usually affect the activity of host Cyp450s and the abundance or composition of gut microbiota, thus affecting pharmacokinetic variability. The activity and expression of main host Cyp450s (Cyp2c29, Cyp1a2, Cyp3a11, Cyp2c65, and Cyp2e1) involved in the drug metabolism of omeprazole, phenacetin, midazolam, tolbutamide, chlorzoxazone, and metoprolol were significantly lower in the NASH group than in the control group. Previous reports also support our findings. Li et al. reported that the activity and expression of Cyp450 enzymes, including Cyp1a2, Cyp3a11, and Cyp2c29, decrease in the NASH group induced by MCD diet [40, 41]. By contrast, the number of host Cyp450s involved in those studies is individual, whereas six host Cyp450s were simultaneously investigated in our study, which is almost responsible for the metabolism of over 80% of clinically used drugs. The category and quantity of the gut microbiota in the NASH group significantly reduced, and the composition and abundance were significantly different from those in the control group. Meanwhile, NASH significantly decreased the alpha diversity index (Chao1 and Shannon) of the gut microbiota. Some previous studies support our results. Ye et al. found that MCD diet increases Firmicutes population while decreases Proteobacteria and Bacteroidetes populations [42]. Kai et al. found a decrease in the alpha diversity index (Chao1) in the NASH group [43]. However, our study was more comprehensive and systematic because we considered host CYP450 enzymes and gut microbiota.
Changes in host Cyp450 activity and gut microbiota caused by NASH might cause pharmacokinetic variability. Therefore, the same drug under different administration routes may have different pharmacokinetic variabilities because of the different effects of host Cyp450s and gut microbiota under different drug administration pathways. In general, pharmacokinetics is influenced by the host Cyp450s and the gut microbiota under intragastric administration. By contrast, drugs bypass the intestine and directly enter the bloodstream under intravenous administration, avoiding intestinal absorption. In this case, pharmacokinetics is mainly affected by the host. Based on the above logic, under the intragastric administration, the exposure and pharmacokinetic parameters of omeprazole, phenacetin, midazolam, tolbutamide, and chlorzoxazone were considerably higher in the NASH group than in the control group. For their main metabolites, the pharmacokinetics of 5-OH-omeprazole, acetanomiphen, OH-midazolam, and OH-tolbutamide also significantly changed. However, the pharmacokinetics of metoprolol and α-OH-metoprolol showed no significant change. Under the intravenous administration, the exposure and pharmacokinetic parameters of tolbutamide, OH-toluamide, and OH-midazolam in the NASH group were much higher than those in the control group, whereas those of omeprazole, phenacetin, midazolam, chlorzoxazone, metoprolol, and their metabolites were not obviously different between the two groups. Comparing the pharmacokinetics results between these two administration routes, we found that the pharmacokinetics of omeprazole, phenacetin, midazolam, chlorzoxazone, and their metabolites in the NASH group significantly changed under intragastric administration but not under intravenous administration. The pharmacokinetics of tolbutamide and its metabolite (OH-tolbutamide) significantly changed under the two administration routes. The pharmacokinetics of metoprolol and its metabolite (α-OH-metoprolol) showed no significant changes between the two administration routes. Considering these results, we speculated that the pharmacokinetic variabilities of omeprazole, phenacetin, midazolam, and chlorzoxazone in the NASH group are mainly caused by changes in the gut microbiota, and the pharmacokinetic variability of tolbutamide may be mainly due to changes in host Cyp450 enzymes. Meanwhile, the pharmacokinetic variability of metoprolol was neither affected by host Cyp450s nor by the gut microbiota.
We further analyzed the different measurement values of host Cyp450 activities between these groups to explain the relation of the pharmacokinetic variability of these drugs with the changes in host Cyp450s and gut microbiota from another side. Previous studies reported that the ratio of AUCmetabolite to AUCparent drug of the probe drugs can reflect the activity of the corresponding Cyp450 enzymes in vivo. However, these ratios via intragastric or intravenous administration were inevitably affected by host factors and gut microbiota. By contrast, determination of the host Cyp450 activity by in vitro human liver microsomal incubation was relatively accurate. Therefore, using the results of Cyp450 enzymes in vitro as a reference, we can divide their relationships into three types to compare the differences between the results of Cyp450 enzyme activity determined in vivo. First, if the enzyme activities in vitro and in vivo showed the same trend, the pharmacokinetic variability in NASH mainly caused the change in host Cyp450s. Second, if the enzyme activities were inconsistent between in vitro and in vivo, the pharmacokinetic variability in NASH was mainly caused by the change in gut microbiota. Third, if no difference existed between these enzyme activities, then the pharmacokinetics variability was less affected by both of them. In the present study, the activity of Cyp2c65 in vivo was similar to that in vitro, suggesting that the pharmacokinetic variability of tolbutamide was mainly caused by the changes in host Cyp450s. Although the activity of Cyp2c29 significantly decreased in vivo and in vitro, no significant difference existed in the pharmacokinetics of intravenous administration, suggesting that the pharmacokinetic variability of omeprazole was related to the changes in gut microbiota. The in vivo and in vitro activities of Cyp1a2, Cyp3a11, and Cyp2e1 were inconsistent, indicating that the pharmacokinetic variabilities of phenacetin, midazolam, and chlorzoxazone were mainly related to the changes in gut microbiota. In addition, biotransformation experiments showed that the gut microbiota can metabolize omeprazole, phenacetin, midazolam, and chlorzoxazone. The metabolic rates of the normal control group were higher than those of the NASH group, which supports our inference from other side. The activity of Cyp2d22 did not change, indicating that metoprolol metabolism was not affected by either side.
Regarding the study of pharmacokinetic variability under NASH, only morphine was reported. The Cmax and AUC of morphine-3-glucuronic acid and morphine-6-glucuronic acid significantly increase in patients with NASH, and the changes are related to the increased expression of the transporter MRP3 [37]. Meanwhile, the influence of host Cyp450s and gut microbiota was not considered in the study. For other studies, the analysis of the influencing factors of pharmacokinetic variability was mostly based on single host Cyp450s or gut microbiota, and few studies simultaneously considered these two factors.
The cocktail probe is superior to the single-probe approach in terms of efficiency, saving mice and costs, and in terms of obtaining more information from the same procedure. This study mainly referred to the previously published Cyp450 enzyme probe cocktail method [44–46]. The most common probes for Cyp1a2 are phenacetin and caffeine; for Cyp3a11 are midazolam and dapsone; for Cyp2c65 are tolbutamide and warfarin; for Cyp2c29 are omeprazole and mephenytoin; for Cyp2e1 is chlorzoxazone; for Cyp2d22 are metoprolol and dextromethorphan. According to a group of probe drugs commonly used by researchers [47], phenacetin, midazolam, omeprazole, tolbutamide, chlorzoxazone, and metoprolol were selected as the present probes. Potential interaction between probe drugs is one of the most frequent concerned questions [48]. However, some experiments have already demonstrated that no metabolic interaction occurs when these drugs are administered simultaneously in a relative low dosage [49]. The interactions between the probes can be reduced or controlled by regulating the drug dose. Thus, many previous studies have used the cocktail method to determine the activity of Cyp450 enzymes. Our selection of probe drugs was also based on a combination that had been used frequently in past studies. Thus, the drug metabolic interactions are controllable.
Some limitations need to be noted. Intravenous administration was used rather than gavage in germ-free or pseudo-germ-free mice to rule out the influence of gut microbiota. Germ-free or pseudo-germ-free mice are usually used to study the gut microbiota, but these animals require special breeding conditions or continuous antibiotic administration, respectively. Sterile mice are ideal models but are very expensive. Pseudo-sterile mice might be concerned about drug–drug interactions between probes and antibiotics. Thus, intravenous administration can basically meet the needs of the present research. Intravenous administration is more convenient and quicker and not more expensive than that of sterile mice; pseudo-sterile mouse modeling also causes no drug interactions, nor any effect on the host drug-metabolizing enzymes. Certainly, these phenomena need to be validated progressively in a single drug or further in a clinical disease state to provide better clinical advice.
In summary, we observed the changes in host Cyp450 enzymes and gut microbiota that lead to pharmacokinetic variability in mice with NASH. Results show that the degree of their respective influence varies from drug to drug. NASH is a common type of disease or a risk factor for many metabolic diseases, such as hypertension and diabetes [50]. The medications given to patients with NASH or those with NASH symptoms are often complex in clinical setting. Therefore, the dosage of most drugs needs to be reduced for their high exposure in vivo to avoid their toxicity. The factors leading to the pharmacokinetic variability of individual drugs are often different. Thus, the influence of the host or the gut microbiota according to different medications must be analyzed. The pharmacokinetic variability of individual drugs requires specific analysis and cannot be generalized (Fig. 6). In the future, with the development and popularization of microbial sequencing and high-throughput sequencing technology, the influence of host and gut microbiota should be considered during the individualized treatment of NASH and its complications.