Mechanisms of hepatocellular toxicity associated with the components of St. John's Wort extract hypericin and hyperforin in HepG2 and HepaRG cells.

St. John's Wort preparations are used for the treatment of mild to moderate depression. They are usually well tolerated but can cause adverse reactions including liver toxicity in rare cases. To date, the mechanism(s) underlying the hepatotoxicity of St. John's Wort extracts are poorly investigated. We studied the hepatocellular toxicity of hypericin and hyperforin as the two main ingredients of St. John's Wort extracts in HepG2 and HepaRG cells and compared the effects to citalopram (a synthetic serotonin uptake inhibitor) with a special focus on mitochondrial toxicity and oxidative stress. In HepG2 cells, hypericin was membrane-toxic at 100 µM and depleted ATP at 20 µM. In HepaRG cells, ATP depletion started at 5 µM. In comparison, hyperforin and citalopram were not toxic up to 100 µM. In HepG2 cells, hypericin decreased maximal respiration starting at 2 µM and mitochondrial ATP formation starting at 10 µM but did not affect glycolytic ATP production. Hypericin inhibited the activity of complex I, II and IV of the electron transfer system and caused mitochondrial superoxide accumulation in cells. The protein expression of mitochondrial superoxide dismutase 2 (SOD2) and thioredoxin 2 (TRX2) and total and reduced glutathione decreased in cells exposed to hypericin. Finally, hypericin diminished the mitochondrial DNA copy number and caused cell necrosis but not apoptosis. In conclusion, hypericin, but not hyperforin or citalopram, is a mitochondrial toxicant at low micromolar concentrations. This mechanism may contribute to the hepatotoxicity occasionally observed in susceptible patients treated with St. John's Wort preparations.

Effects of Hypericum perforatum (St John's wort) on the pharmacokinetics and pharmacodynamics of rivaroxaban in humans.

AIMS: To investigate the influence of a cytochrome P450 CYP3A4 and efflux transporter P-glycoprotein (P-gp) inducing Hypericum perforatum extract on the pharmacokinetics and pharmacodynamics of rivaroxaban. METHODS: Open-label, nonrandomized, sequential treatment interaction study. Following CYP3A4 and P-gp phenotyping using low-dose midazolam and fexofenadine, 12 healthy volunteers received a single oral dose of 20 mg rivaroxaban and rivaroxaban plasma concentrations and inhibition of the activated coagulation factor X (factor Xa) activity were measured prior to and up to 48 h postdosing. The procedures were repeated after 2 weeks' treatment with the H. perforatum extract. RESULTS: The geometric mean ratios for the area under the concentration-time curve and Cmax of rivaroxaban after/before induction with the H. perforatum extract were 0.76 (90% confidence interval [CI] 0.70, 0.82) and 0.86 (90% CI 0.76, 0.97), respectively. Inhibition of factor Xa activity was reduced with a geometric mean area under the effect-time curve ratio after/before induction of 0.80 (90% CI 0.71, 0.89). No clinically significant differences were found regarding Tmax (median 1.5 vs 1 h, P = .26) and terminal elimination half-life (mean 10.6 vs 10.8 h, P = .93) of rivaroxaban. The H. perforatum extract significantly induced CYP3A4 and P-gp activity, as evidenced by phenotyping. CONCLUSION: The CYP3A4/P-gp inducing H. perforatum extract caused a decrease of rivaroxaban exposure with a proportional decrease of the pharmacodynamic effect. Although the data do not justify a contraindication for the combination or a systematic adjustment of rivaroxaban dosage, avoidance of the combination or laboratory monitoring should be considered in patients taking hyperforin-containing H. perforatum extracts with rivaroxaban.