Inhibition of Caffeine Metabolism by Apiaceous and Rutaceae Families of Plant Products in Humans: In Vivo and In Vitro Studies

Abstract

Daily consumption of caffeinated beverages is considered safe but serious health consequences do happen in some individuals. The Apiaceous and Rutaceae families of plant (ARFP) products are popular foods and medicines in the world. We previously reported significant amounts of furanocoumarin bioactive such as 8-methoxypsoralen, 5-methoxypsoralen, and isopimpinellin in ARFP products. As both caffeine and furanocoumarin bioactive are metabolized by the same hepatic CYP1A1/2 isozyme in humans, caffeine/ARFP product interactions may occur after co-administration. The objectives of the present study were to study in vivo loss of caffeine metabolizing activity by comparing the pharmacokinetics of caffeine in volunteers before and after pre-treatment with an ARFP extract, study the correlation between the decrease in hepatic CYP1A2 activity and the content of furanocoumarin bioactive in ARFP extracts, characterize CYP1A2 inactivation using in vitro incubations containing ¹⁴C-caffeine, a furanocoumarin bioactive, and human liver microsomes (HLMs), and provide a mechanistic explanation for both in vivo and in vitro data using the irreversible inhibition mechanism. The study results showed pre-treatment of volunteers with four ARFP extracts increased the area-under-the-concentration-time-curve (AUC_0-inf) ratio of caffeine in the plasma ranging from 1.3 to 4.3-fold compared to the untreated volunteers indicating significant caffeine metabolism inhibition. The increases in AUC_0-inf ratio also were linearly related to the effect-based doses of the furanocoumarins in the ARFP extracts, a finding which indicated caffeine metabolism inhibition was related to the content of furanocoumarin bioactive in an ARFP product. In vitro incubation studies also showed individual furanocoumarin bioactive were potent inhibitors of caffeine-N-demethylation; the IC₅₀ for 8-methoxypsoralen 5-methoxypsoralen, and isopimpinellin were 0.09, 0.13, and 0.29 µM, respectively. In addition, CYP1A2 inactivation by individual furanocoumarin bioactive was concentration- and time-dependent involving the irreversible inhibition mechanism. The proposed irreversible inhibition mechanism was investigated further using ¹⁴C-labeled 8-methoxypsoralen and HLMs. The formation of ¹⁴C-adducts due to ¹⁴C-8-MOP-derived radioactivity bound to HLMs confirmed the irreversible inhibition of CYP1A2 activity. Thus, furanocoumarin bioactive metabolism in humans would result in reactive metabolite(s) formation inactivating CYP1A2 isozyme and inhibiting caffeine metabolism. Once the CYP1A2 isozyme was deactivated, the enzymic activity could only be regained by isozyme re-synthesis which took a long time. As a result, a single oral dose of ARFP extract administered to the human volunteers 3.0 h before still was able to inhibit caffeine metabolism.

Keywords: P450 cytochrome enzymes; caffeine; chemical mixtures; enzyme inactivation mechanism; furanocoumarin.

FIGURE 1

Chemical structures of caffeine and linear furanocoumarins.

FIGURE 2

The separation of caffeine (1) and benzotriazole (2), the internal standard, in saliva extract using isocratic method by HPLC with UV. The retention time for caffeine and benzotriazole is 10.7 and 14.5 min, respectively.

FIGURE 3

(A) Time course of caffeine concentration in the plasma of volunteers before and after pre-treatment with an extract of A. majus, A. archangelica, C. monnieri, and R. graveolens. Each point is plotted as mean ± SD. Empty squares represent caffeine concentrations in volunteers with no ARFP extract pre-treatment. Filled squares represent caffeine concentrations in volunteers with ARFP extract pre-treatment 3.0 h before receiving the caffeine tablets. (B) Time course of caffeine concentrations in the plasma of volunteers before and after pre-treatment with an extract of P. aniseum, P. crispum, A. pubescens, A. graveolens seeds, and A. graveolens flakes. The point is reported as mean ± SD. Empty squares represent caffeine concentrations in volunteers with no ARFP extract pre-treatment. Filled squares represent caffeine concentrations in volunteers with ARFP extract pre-treatment 3.0 h before receiving the caffeine tablets.

FIGURE 4

Correlation between the area-under-the-concentration-time-curve (AUC) ratio of caffeine and the effect-based dose measure of the furanocoumarin mixture in ARFP extract. The data represent mean ± SD of caffeine AUC ratio with different effect-based dose measures in the ARFP extracts.

FIGURE 5

Effects of furanocoumarin concentration on CYP 1A2 inactivation in vitro. The results are presented as the mean of at least three incubations. The ±SD of the data is omitted for better display and comparison. The IC₅₀ was determined from the concentration-inhibition curve using non-linear regression analysis.

FIGURE 6

(A) Time- and concentration-dependent inactivation of CYP 1A2 activity by 8-MOP (upper), 5-MOP (middle) and ISOP (lower). The results are presented as the mean of at least three independent incubations. The ±SD of the data is omitted for better display and comparison. (B) Maximum inactivation rate of CYP 1A2 activity by 8-MOP (upper), 5-MOP (middle) and ISOP (lower). The results are presented as the mean of at least three independent incubations. The ±SD of the data is omitted for better display and comparison.