Mechanisms of Herb-Drug Interactions Involving Cinnamon and CYP2A6: Focus on Time-Dependent Inhibition by Cinnamaldehyde and 2-Methoxycinnamaldehyde

Abstract

Information is scarce regarding pharmacokinetic-based herb-drug interactions (HDI) with trans-cinnamaldehyde (CA) and 2-methoxycinnamaldehyde (MCA), components of cinnamon. Given the presence of cinnamon in food and herbal treatments for various diseases, HDIs involving the CYP2A6 substrates nicotine and letrozole with MCA (KS = 1.58 µM; Hill slope = 1.16) and CA were investigated. The time-dependent inhibition (TDI) by MCA and CA of CYP2A6-mediated nicotine metabolism is a complex process involving multiple mechanisms. Molecular dynamic simulations showed that CYP2A6's active site accommodates two dynamic ligands. The preferred binding orientations for MCA and CA were consistent with the observed metabolism: epoxidation, O-demethylation, and aromatic hydroxylation of MCA and cinnamic acid formation from CA. The percent remaining activity plots for TDI by MCA and CA were curved, and they were analyzed with a numerical method using models of varying complexity. The best-fit models support multiple inactivator binding, inhibitor depletion, and partial inactivation. Deconvoluted mass spectra indicated that MCA and CA modified CYP2A6 apoprotein with mass additions of 156.79 (142.54-171.04) and 132.67 (123.37-141.98), respectively, and it was unaffected by glutathione. Heme degradation was observed in the presence of MCA (48.5% ± 13.4% loss; detected by liquid chromatography-tandem mass spectrometry). In the absence of clinical data, HDI predictions were made for nicotine and letrozole using inhibition parameters from the best-fit TDI models and parameters scaled from rats. Predicted area under the concentration-time curve fold changes were 4.29 (CA-nicotine), 4.92 (CA-letrozole), 4.35 (MCA-nicotine), and 5.00 (MCA-letrozole). These findings suggest that extensive exposure to cinnamon (corresponding to ≈ 275 mg CA) would lead to noteworthy interactions. SIGNIFICANCE STATEMENT: Human exposure to cinnamon is common because of its presence in food and cinnamon-based herbal treatments. Little is known about the risk for cinnamaldehyde and methoxycinnamaldehyde, two components of cinnamon, to interact with drugs that are eliminated by CYP2A6-mediated metabolism. The interactions with CYP2A6 are complex, involving multiple-ligand binding, time-dependent inhibition of nicotine metabolism, heme degradation, and apoprotein modification. An herb-drug interaction prediction suggests that extensive exposure to cinnamon would lead to noteworthy interactions with nicotine.

Copyright © 2020 by The American Society for Pharmacology and Experimental Therapeutics.

Figures

Fig. 1.

(A) trans-cinnamaldehyde and (B) 2-methoxycinnamaldehyde.

Fig. 2.

Structure-based pharmacophore model generated from the superposition of published X-ray crystal structures of CYP2A6 bound to both reversible and irreversible inhibitors. The majority of the ligands had an H-bond acceptor (part of an aromatic ring) that made H-bond interaction with the amido -NH group of the binding site residue N297. Several phenylalanine residues, including F107, F118, F209, and F480, form an aromatic cage that makes π stacking interactions with the aromatic ring of the ligands. The distal end of ligands is in close proximity to the heme iron atom, coordinating through a lone pair of electrons carried by an electronegative atom such as N.

Fig. 3.

The major binding poses and critical interactions of CA within the binding site of CYP2A6. (A) CA changed its binding pose from the initial docked pose after 2 nanoseconds and assumed an entirely different orientation in which the carbonyl oxygen was engaged in an H-bond with N297. (B) The entry of a water molecule altered CA binding and slightly increased the H-bond length but increased aryl-aryl interactions with several phenylalanine residues.

Fig. 4.

The major binding poses and critical interactions of MCA within the binding site of CYP2A6. (A) MCA changed its binding pose from the initial docked pose after 8 nanoseconds and assumed a slightly different orientation in which the methoxyl oxygen was engaged in an H-bond with N297 through a water molecule. (B) The entry of several water molecules flipped the orientation of MCA to an entirely different binding mode in which the aromatic ring came close to the heme iron atom. The carbonyl oxygen was engaged in a water-mediated H-bond interaction with N297.

Fig. 5.

The major binding poses and critical interactions of two CA molecules bound to CYP2A6 at the same time. (A) CA-1 was stabilized in its initial docked pose for 60 nanoseconds, after which it changed its orientation in which the carbonyl oxygen that was initially in close proximity to the heme iron shifted to near N297 and formed an H-bond. (B) The second CA molecule moved toward the heme and positioned its aromatic ring in π stacking with F209, which is also in π-π interaction with CA-1. The carbonyl oxygen atom of CA-2 made an H-bond interaction with T308, although the interaction was dynamic in nature (see Supplemental Fig. 2C).

Fig. 6.

The major binding poses and critical interactions of two MCAs bound to CYP2A6 at the same time. (A) MCA-1 was stabilized in its initial docked pose in which the methoxyl oxygen group was engaged in an H-bond with N297. (B) After 100 nanoseconds, the second MCA molecule moved toward the heme and pushed the carbonyl oxygen away from the heme iron. The carbonyl oxygen atom of MCA-2 reoriented its position toward the aromatic cage (see Supplemental Fig. 3, C and D).

Fig. 7.

(A) Representative binding difference spectra of purified rCYP2A6 with increasing concentrations of MCA. (B) A representative curve generated from fitting changes in absorbance (A386 nm–A418 nm) to model using specific binding with Hill slope, as a function of MCA concentration.

Fig. 8.

Kinetic schemes and model fits for CYP2A6 inactivation by CA (1000–0 μM) in HLM. (A) Kinetic scheme for EII-MIC-M-IL model 1. (B) Kinetic scheme for EII-MIC-M-IL model 2. (C) Kinetic scheme for PI-M-IL model. (D) Experimental (points) and EII-MIC-M-IL model fitted (solid lines) PRA plots. (E) Experimental (points) and EII-MIC-M-IL model 2 fitted (solid lines) PRA plots. (F) Experimental (points) and PI-M-IL model fitted (solid lines) PRA plots. The colors indicate different inactivator concentrations: 1000 (red), 750 (blue), 500 (orange), 250 (green), 200 (violet), 175 (cyan), 150 (magenta), 100 (brown), 50 (black), 25 (gray), 12.5 µM (pink), and solvent control (light gray).

Fig. 9.

Kinetic schemes and model fits for CYP2A6 inactivation by MCA (250–0 μM) in HLM. (A) Kinetic scheme for MIC-M-IL model. (B) Experimental (points) and MIC-M-IL model fitted (solid lines) PRA plots. (C) Kinetic scheme for PI-M-IL model. (D) Experimental (points) and PI-M-IL model fitted (solid lines) PRA plots. (E) Kinetic scheme for EII-PI-M-IL model. (F) Experimental (points) and EII-PI-M-IL model fitted (solid lines) PRA plots. The colors indicate different inactivator concentrations: 250 (red), 200 (blue), 175 (orange), 150 (green), 125 (violet), 100 (cyan), 80 (magenta), 40 (brown), 20 (black), 10 (gray), 5 µM (pink), and solvent control (light gray).

Fig. 10.

Representative LC-MS/MS traces showing detection of heme from incubations of rCYP2A6 with NADPH (top panel) and rCYP2A6 with MCA and NADPH (bottom panel). Incubations with MCA and NADPH exhibited a 48.5% heme loss relative to enzyme-plus-NADPH controls (S.D. = 13.4%; 95% CI = 34.4%–62.6%; P = 0.0082; N = 6).

Fig. 11.

Masses of CYP2A6 from incubations containing MCA (top left), 8-MOP (bottom left), CA (top right), or no inhibitor (bottom right). Main peaks indicate the unmodified CYP2A6 mass, and minor peaks indicate possible adducts. Deconvolution was achieved by extracting the spectra around the vertex of the peak at the retention time of CYP2A6 and processed using the Biotoolkit add-on in the PeakView software (AB Sciex).