Discussion

The aim of the current study was to find out which CYPs are involved in the metabolism of 4-MAA and whether inhibition of the CYPs involved leads to clinically relevant drug interactions. The in vitroexperiments with supersomes revealed that CYP1A2 is the most important enzyme for 4-MAA metabolism and that CYP2C19 and CYP2D6 contribute to 4-MAA N-demethylation. Our in vivo data confirmed these findings and showed that the effects of ciprofloxacin and fluconazole regarding inhibition of the formation of 4-AA and 4-FAA are additive.
Both our in vitro and in vivo experiments showed that CYP1A2 is the dominant enzyme for the demethylation of 4-MAA to 4-AA but also for the conversion of 4-MAA to 4-FAA. So far, it had been demonstrated that both reactions are catalyzed by hepatic microsomes [22-24] but the CYPs involved had not been clearly identified and verified in a clinical study. The results of the current study are in agreement with those in a recent study where we investigated the effect of metamizole on the activity of different CYPs [30]. In this study, we found that metamizole inhibits the conversion of caffeine to paraxanthine, which is catalyzed by CYP1A2, most likely in a competitive fashion. The results of the current study support this interpretation. However, the findings in the current study disagree with the reports of Geisslinger et al. [23] and of Bachmann et al. [24], which were both performed with human microsomes as the enzyme source. Geisslinger et al. showed that the demethylation of 4-MAA to 4-AA by human hepatic microsomes could partially be inhibited by ketoconazole, suggesting a major contribution of CYP3A4 [23]. Geisslinger et al. used a ketoconazole concentration of 10 µM in their study, a concentration far below the Ki of ketoconazole for CYP1A2 and 2C19 [31], excluding inhibition of CYP1A2 and CYPC19 as an explanation for their findings. Bachmann et al. investigated the conversion of 4-MAA to 4-AA by human liver microsomes using specific CYP inhibitors and identified CYP2B6, 2C8, 2C9 and 3A4 as the most important contributors to 4-MAA demethylation [24]. The discrepancies between the current study and the studies of Geisslinger et al. and Bachmann et al. may be due to the different enzyme sources used, human recombinant CYPs expressed in supersomes versus human liver microsomes. The discrepancy is relevant, indicating that results obtained in in vitro studies should be confirmed in another in vitro system or, preferably, by a clinical study. The fact that we obtained almost identical results for the in vitro and in vivo investigations in the current study supports the notion that CYP1A2 is the most important CYP for 4-MAA N-demethylation and conversion to 4-FAA.
The FDA defines strong, moderate and weak enzyme inhibitors as drugs that increase the AUC of a sensitive substrate by ≥5-fold, ≥2 to <5-fold, and ≥1.25 to <2-fold, respectively. [32]. The FDA lists ciprofloxacin as a strong inhibitor of CYP1A2 and fluconazole as a strong inhibitor of CYP2C19 and a moderate inhibitor of CYP2C9 and CYP3A4. In the current study, ciprofloxacin increased the AUC of 4-MAA time-dependently by a factor of 1.31, 1.56 and 1.51 and fluconazole by a factor of 1.10, 1.32 and 1.17 at 6, 8 and 12 h, respectively (suppl. Table 1). According to the FDA, the inhibition of 4-MAA metabolism by ciprofloxacin and fluconazole was therefore weak. The reason for an only weak inhibition in the presence of strong inhibitors could be the existence of alternative metabolic pathways. Therefore, we also assessed the combined application of ciprofloxacin and fluconazole, which increased the effect on the AUC of 4-MAA, resulting in AUC ratios of 1.54, 1.90 and 1.92 at 6, 8 and 12 h, respectively. The combined application of ciprofloxacin and fluconazole showed that the effects of ciprofloxacin and fluconazole on the AUCs of the metamizole metabolites are additive, excluding a mutual compensation between CYP1A2 and CYP2C19. However, Volz and Kellner have investigated the pharmacokinetics of orally administered14C-labelled metamizole in healthy volunteers [12]. They detected 7 metabolites in serum and could identify four of them as 4-MAA, 4-AA, 4-FAA and 4-AAA. Forty-eight hours after administration, 90% of the radioactivity had been excreted renally, but the four metabolites accounted only for approximately 60% of the excreted radioactivity. The study therefore indicated the existence of additional metabolic pathways that may compensate for the inhibition of CYP1A2 and CYP2C19.
Based on the results in the supersomes, compensation by CYP2D6 is an obvious possibility, which we didn’t study in our clinical trial. However, the activity of CYP2D6 in supersomes was not higher than for CYP2C19, and, in contrast to CYP1A2 and CYP2C19, the CYP2D6 genotype showed no correlation with the metabolic activity, rendering a major contribution of CYP2D6 to the metabolism of 4-MAA unlikely.
An additional possibility is the contribution of myeloperoxidase in granulocytes, as proposed in the report of Bachmann et al. [24]. The results of the current study do not exclude this possibility but suggest that the contribution of this pathway would most likely be less than the contribution by the described hepatic metabolism. If the extrahepatic metabolism were dominant, inhibition of the (in that case minor) hepatic pathway would not be expected to increase the AUC of 4-MAA and to decrease the formation of 4-AA. The inhibition of the (minor) hepatic pathway should be compensated by the dominant extrahepatic pathway. In the current study, pretreatment with ciprofloxacin/fluconazole increased the AUC0-12h of 4-MAA by 92% and decreased the AUC0-12h of 4-AA by 24%, rendering the existence of a dominant extrahepatic pathway unlikely.
Even if the interaction with CYP1A2 and CYP2C19 inhibitors is quantitatively small, this does not mean that this interaction is clinically negligible. Important adverse reactions of metamizole are hypotensive reactions, skin eruptions, myelotoxicity possibly leading to agranulocytosis and hepatic injury [33]. While hepatic injuries and skin toxicities are mainly dose-independent, immunological reactions [33, 34], hypotensive reactions and myelotoxicity appear to be dose-dependent. Regarding myelotoxicity, a recent genome-wide association study failed to reveal an HLA association and suggested that impaired antioxidative defense mechanisms in granulocytes or granulocyte precursors could be a risk factor for neutropenia and agranulocytosis [35]. In vitro studies using HL60 cells support toxicological mechanisms associated with the formation of reactive metabolites, which is dose-dependent [36, 37]. An increase in the plasma concentration of 4-MAA by drug interactions and/or enzyme polymorphisms could therefore enhance the risk for certain adverse reactions associated with metamizole.
In conclusion, we provide evidence that CYP1A2 is the major CYP for the conversion of 4-MAA to 4-AA and 4-FAA. CYP1A2 inhibition increases the 4-MAA exposure by a factor of approximately 1.5, which could be relevant for dose-dependent adverse reactions.