Non-additivity of the functional properties of individual P450 species and its manifestation in the effects of alcohol consumption on the metabolism of ketamine and amitriptyline

Abstract

To explore functional interconnections between multiple P450 enzymes and their manifestation in alcohol-induced changes in drug metabolism, we implemented a high-throughput study of correlations between the composition of the P450 pool and the substrate saturation profiles (SSP) of amitriptyline and ketamine demethylation in a series of 23 individual human liver microsomes preparations from donors with a known history of alcohol consumption. The SSPs were approximated with linear combinations of three Michaelis-Menten equations with globally optimized K_M (substrate affinity) values. This analysis revealed a strong correlation between the rate of ketamine metabolism and alcohol exposure. For both substrates, alcohol consumption caused a significant increase in the role of the low-affinity enzymes. The amplitudes of the kinetic components and the total rate were further analyzed for correlations with the abundance of 11 major P450 enzymes assessed by global proteomics. The maximal rate of metabolism of both substrates correlated with the abundance of CYP3A4, their predicted principal metabolizer. However, except for CYP2D6 and CYP2E1, responsible for the low-affinity metabolism of ketamine and amitriptyline, respectively, none of the other potent metabolizers of the drugs revealed a positive correlation. Instead, in the case of ketamine, we observed negative correlations with the abundances of CYP1A2, CYP2C9, and CYP3A5. For amitriptyline, the data suggest inhibitory effects of CYP1A2 and CYP2A6. Our results demonstrate the importance of functional interactions between multiple P450 species and their decisive role in the effects of alcohol exposure on drug metabolism.

Keywords: Alcohol exposure; Amitriptyline; Drug metabolism; Human liver microsomes; Ketamine; Proportional projection; cytochrome P450.

Fig. 1.. Substrate saturation profiles of N-demethylation of S-ketamine by HLM and their analysis by PCA.

Panel a shows a dataset obtained with a set of 9 pooled HLM preparations. The solid lines represent the approximation of the SSPs by linear combinations of three Michaelis-Menten dependencies with globally optimized K_m values. The results of the application of PCA to this dataset are shown in panel b. The vectors of the first, second, and third principal components are shown in the main panel in blue, black, and red, respectively. The solid and dashed lines in the main panel show their approximations by linear combinations of three Michaelis-Menten dependencies with globally optimized K_m values, which are shown in the insert. The set of SSPs obtained with 23 individual SSP samples along with their approximations is exemplified in the panel c. Each SSP shown in panels a and c represents an average of the results of 3 - 6 individual experiments.

Fig.2.. Correlations between the total maximal rate (V_{max total}, panel a) and the maximal rate of the low-affinity component (V_max3, panel b) of ketamine demethylation with the provisional index of alcohol exposure (PIAE) of liver donors.

The solid line shows the linear approximation of the whole dataset excluding two apparent outliers at very high PIAE shown with open circles. The dashed line shows the linear approximation of the dataset with the points corresponding to homo- and heterozygotes of rare CYP2B6 variants (open stars) excluded.

Fig. 3.. Substrate saturation profiles of N-demethylation of amitriptyline by HLM and their analysis by PCA.

Panel a shows a dataset obtained with a set of nine pooled HLM preparations. The solid lines represent the approximation of the SSPs by linear combinations of three Michaelis-Menten dependencies with globally optimized K_m values. The results of the application of PCA to this dataset are shown in panel b. The vectors of the first, second, and third principal components are shown in the main panel in blue, black, and red, respectively. The solid and dashed lines in the main panel show their approximations by linear combinations of three Michaelis-Menten dependencies with globally optimized K_m values, which are shown in the insert. The set of SSPs obtained with 23 individual SSP samples along with their approximations is exemplified in the panel c. Each SSP shown in panels a and c represents an average of the results of 3 - 6 individual experiments.

Fig. 4.. Correlations of the fractional amplitude of the low affinity (F₃, panel a) and the medium affinity (F₂, panel b) components of amitriptyline demethylation with the provisional index of alcohol exposure (PIAE) of liver donors.

The solid lines show the linear approximation of the whole datasets and the dashed lines represent the approximation of the dataset with the points corresponding to homo- and heterozygotes of rare variants of CYP2B6 (open stars) and CYP2D6 (closed stars) excluded.

Fig. 5.

Correlations between the maximal rate of the high-affinity component of N-demethylation of ketamine (a) and amitriptyline (b) with the depth of inhibition of these reactions by CYP3cide.

Fig. 6.. Analysis of correlations of the total maximal rate of ketamine demethylation with the composition of the cytochrome P450 pool in 23 HLM preparations from individual donors.

Panel a shows the approximation of the dataset of V_{max total} values (black open circles connected with a dashed stepped line) with a linear combination of the vectors of relative fractional content (VFC) of CYP1A2, CYP2C9, CYP3A4, and CYP3A4 taken with the multiplication factors of −9.4,−8.6, +5.9 and −4.1, respectively (red cross points connected with solid dashed line). The X-axis in this panel corresponds to the PIAE of the liver donors. Panel (b) shows the same approximation as a plot of the found VFC combination versus V_{max total}.

Fig. 7.. Correlations of Substrate Saturation Profiles of ketamine (A) and amitriptyline (B) demethylation with the composition of the cytochrome P450 pool in a set of 23 individual and 9 pooled HLM preparations.

The tables illustrate the process of finding the best combinations of four profiles of relative fractional content of 11 major P450 species that approximate the profiles of V_{max total}, V_max1, V_max2, and V_max3. The numbers shown in the table indicate the stage of successive approximations with 1 to 4 VFC combinations. The species with positive correlations are shown with red shadowing. The negative correlation is indicated by a green background. The light red and light green shadowing indicates the species that appeared in the early steps of the analysis but were replaced by other P450s later. The values in the rightmost column are the correlation coefficients of the four-component approximations. In the header rows, the P450 species predicted to play the major role in the metabolism are highlighted in red, orange, yellow, green, and blue colors where the red color designates the predicted primary metabolizer.