Discussion
There have been many reports verifying the involvement of CYP3A4 and CYP2J2 in the metabolism of rivaroxaban; however, the respective metabolic contributions of CYP isoforms were unknown. In this study, the contributions of CYP3A4 and CYP2J2 were systematically evaluated and compared. Firstly, CYP-dependent M1 formation analysis indicated that CYP2J2 showed a strongest catalytic activity. Moreover, in CYP3A subfamily, except CYP3A4, CYPs 3A5 and 3A7 also participated in the rivaroxaban hydroxylation (M1 area, CYP3A4: CYP3A5: CYP3A7 = 25.0: 8.6: 6.3). Secondly, kinetic studies further verified the highest catalytic efficiency of CYP2J2, approximately 39-fold to that of CYP3A4. Finally, CYP-specific inhibition experiments were used to access the contributions of CYPs. Ketoconazole (CYP3A-specific inhibitor) and danazol (CYP2J2-specific inhibitor) inhibited 43.3% and 41.1% rivaroxaban metabolism in HLM. Based on the ratio of M1 area by CYP3A subfamily in CYP screen (M1 area, CYP3A4: CYP3A5: CYP3A7 = 25.0: 8.6: 6.3), the contributions of CYPs 3A4, 3A5 and 3A7 were 27.3%, 9.4% and 6.9% respectively. Therefore, our studies identified the predominated role of CYP2J2 in the rivaroxaban hydroxylation with a contribution of 41.1%, which were much higher than the contribution of 27.3% by CYP3A4.
In the analysis of M1 formation by CYPs, CYP2J2 produced the highest catalytic activity which was nearly 16 times higher than that of CYP3A4 (Figure 3B). Kinetic studies further demonstrated that the intrinsic clearance value of CYP2J2 was far higher than that of CYP3A4, approximately 39-fold (Table 1). Finally, the respective contributions of CYP2J2 and CYP3A4 in HLM were determined by the CYP-specific chemical inhibition study. Even though the inhibitory effect of CYP2J2-specific inhibitor danazol and CYP3A-specific inhibitor ketoconazole were comparable, being 41.1% and 43.3% respectively, there were dramatic differences in the content of each isoenzyme in the HLM. CYP3A subfamily is responsible for the metabolism of approximately 30.2% of clinical drugs, and CYP3A4 as the major isoform of the CYP3A subfamily represents about 14–24% of the microsomal P450 pool, on average [20-23]. In contrast, CYP2J2 is the least abundant P450 isoform of those involved in clinical drug metabolism [23]. CYP3A was the most abundantly-expressed subfamily in the liver, at ~28.8% of the total hepatic CYPs, whereas CYP2J2 abundance was less than 1% [23 24]. Moreover, proteomic analysis demonstrated the protein concentration of CYP3A4 to be about 50-fold that of CYP2J2 in HLM [25]. Therefore, we inferred that the results of inhibitory experiments in HLM, CYP2J2 and CYP3A4-specific inhibitors produced comparable inhibition ratio, were most likely caused by the dramatic differences in content for two isoforms in HLM. With such a low content in HLM, CYP2J2 produced a comparable contribution with that by CYP3A, the most abundant subfamily in liver, it suggested the higher catalytic efficiency of CYP2J2 than that of CYP3A4 in rivaroxaban hydroxylation.
In the inhibition study, the total inhibitory activity in HLMs was over 100% at 130.27%, which may have been due to the poor selectivity of the high-concentration inhibitors. Additionally, the concentration of specific inhibitors had considerable impact on their inhibition selectivity. For example, although ketoconazole is known as a specific inhibitor of the CYP3A subfamily, it has also shown inhibitory activity in the CYP1A1-mediated metabolism of 7-ethoxycoumarin, with anIC 50 value of 0.33 ± 0.03 μM [26]. In the present study, the concentration of ketoconazole used for inhibiting CYP3A was set as 1 μM―which was about 10 x Ki ―to ensure absolute inhibition. At this concentration, ketoconazole may also inhibit CYP1A1, which was also found to be involved in the metabolism of rivaroxaban (Figure 3); this could explain why high concentrations of ketoconazole affected high total inhibition activity. A similar outcome was observed with quinidine, which was used to specifically inhibit the activity of CYP2D6. Quinidine displayed significant inhibition of CYP1A1-mediated 7-ethoxyresorufin O-deethylation, with anIC 50 value of 1.1 μM [27]. However, the concentration of quinidine in the present study was set at 10 μM, which probably also inhibited the CYP1A1-mediated metabolism of rivaroxaban and led to a much higher total inhibition activity. Hence, poorly selective inhibitory effects of high-concentration inhibitors resulted in a more than 100% inhibition ratio in the present study.
In addition to poor selectivity of some inhibitors, the higher expression level of CYP2C9 in the liver may also partly account for its high inhibition. In the CYP-specific inhibition study, the inhibition ratio of CYP2C9 was the third highest after CYP2J2 and CYP3A4, while the M1 peak area produced by CYP2C9 in the CYP screen was less than 10, and approximately 0.38% of that produced by CYP2J2. The abundant content of CYP2C9 in the liver played an important role in this difference between the two results. First, CYP2C9 is one of the most highly expressed members of all P450 isoforms, with a similar or lower protein level to CYP3A4 [22 28]. More importantly, the CYP2C9 protein level was higher than that of CYP3A4 in HLM, and much higher than that of CYP2J2 [25]. Based on this rationale, the inhibition ratio achieved with CYP2C9-specific sulfaphenazole was much higher than its actual efficacy.
The difference in the protein contents of CYP2J2 and CYP3A4 in the liver resulted in their different status in clinical drug-drug interaction research. The drug-drug interactions of rivaroxaban have been extensively assessed in combination with many drugs, including CYP3A4 or P-gp substrates, inhibitors and inducers. However, results have demonstrated that the combination of rivaroxaban with agents that are strong inhibitors of both CYP3A4 and P-gp can increase rivaroxaban plasma concentrations in vivo, prompting caution regarding its co-administration [15]. In contrast, owing to its low expression levels in liver microsomes, CYP2J2 is usually not considered in routine drug-drug interactions, thus, the potential for CYP2J2 regulators to change rivaroxaban clearance has not been widely evaluated to date. However, it is expressed at an extremely high level in the cardiovascular system: an evaluation of the mRNA levels of P450 isoforms in the heart showed CYP2J2 mRNA levels largely exceeded those of other detected isozymes by 3 million to 62 times [29 30]. And, in the aorta and coronary artery, the expression of CYP2J2 ranks second only to CYP2C9 [31]. These distribution characteristics of CYP2J2 are consistent with the function of transforming arachidonic acid into epoxyeicosatrienoic acids, which plays a vital role in cardiovascular homeostasis and regulating vascular tone [32 33]. Importantly, a few rivaroxaban drug-drug interaction studies indicated that it may interact with other cardiovascular drugs that target CYP2J2, such as amiodarone and dronedarone [34]. Therefore, the possibility of the interaction between rivaroxaban and other drugs, especially with drugs to treat cardiovascular diseases, can be more exactly evaluated if the heart is set as the target organ and the content of CYP2J2 is considered.
In vitro drug-drug interaction (DDI) studies have been one of the major methods for evaluating the efficiency and safety of drugs. For rivaroxaban, the evaluations targeting CYP3A4 and P-gp are the main DDI research direction so far, which was due to the rich content of CYP3A4 in the liver and the key role of P-gp in transporter-mediated DDIs. Our results, showing the dominant role of CYP2J2 in the metabolism of rivaroxaban, fill a gap in the basic metabolism studies of rivaroxaban and also give a new insight into DDI studies involving rivaroxaban.
In summary, multiple CYP isoforms were found to be involved in the hydroxylation of rivaroxaban, with CYP2J2 identified as the predominant isoenzyme involved.