Physico-chemical Modelling of Anaesthetic Agents

 

Jason C. Sewell and John W. Sear

 

Nuffield Dept. of Anaesthetics, University of Oxford, John Radcliffe Hospital, Headington, Oxford OX3 9DU.

 

Over 150 years after the first use of modern anaesthetic agents, we still do not know the mechanism of action of these drugs. Recent studies of mechanisms have been predominantly target-orientated, focusing on the interactions of general anaesthetics at putative sites of action. For example, several studies have investigated the molecular specificity of anaesthetic interactions with ligand-gated ion channels in vitro [1, 2], yet the relative importance of these targets in achieving general anaesthesia remains unresolved.

 

An alternative approach to investigating the molecular basis of activity is to focus on the anaesthetic molecules themselves. The aims of such ligand-orientated studies are to identify the physico-chemical properties of the molecules that determine activity, and to formulate a model that correlates the magnitude of these properties with potency. The property conventionally considered for general anaesthetics is lipophilicity (the Meyer-Overton correlation), represented by the octanol/water partition coefficients (log P) of the compounds. However, log P has been shown to be a poor predictor of anaesthetic activity for both structurally homologous [3] and chemically-diverse [4] intravenous general anaesthetics. It is also unable to predict the potency order for the enantiomers of chiral anaesthetics, such as ketamine [4].

 

Three-Dimensional Molecular Fields

The fundamental problem with conventional descriptors such as log P is that they represent three-dimensional molecular properties as one-dimensional parameters. We have been developing a modelling approach [4-7] to address this shortcoming. Our approach considers both the magnitude and the spatial distribution of the physico-chemical properties of the anaesthetics by calculating molecular fields; an example is shown in fig 1. This diagram shows the atomic structure of thiopental overlaid with two physico-chemical properties of the anaesthetic molecule: the geometric shape and the electrostatic potential. The latter describes the uneven distribution of charge around the structure due to the different electro-negativities of the anaesthetic’s substituent atoms. Regions of positive potential (e.g. associated with hydrogen atoms) are shown in purple, and regions of negative potential (e.g. sulphur and oxygen atoms) are shown in red. It is our working hypothesis that general anaesthetic activity is determined by the magnitude of such physico-chemical properties at specific, key regions around the molecules; and that pharmacophoric maps can be derived which describe the spatial arrangement of these key regions.

 

Comparative Molecular Field Analysis Studies

The relationship between the spatial distribution of the physico-chemical properties and the activities of the anaesthetics can be determined using Comparative Molecular Field Analysis (CoMFA), a computer-aided drug design methodology frequently used where the detailed structure of a receptor is unknown [8]. In CoMFA, the anaesthetic structures are aligned and placed in a grid consisting of regularly spaced lattice points. A charged probe atom is placed at each lattice point, and the steric and electrostatic interaction energies between the probe and the anaesthetics calculated. The values of these interaction energies at each point are correlated with anaesthetic activity using partial least squares regression. By identifying which lattice points make the greatest contribution to the activity model, three-dimensional pharmacophoric maps of the key regions where steric and electrostatic interactions are important in determining activity can be derived. To date, we have derived CoMFA models and pharmacophoric maps for two different aspects of intravenous general anaesthetic activity:

 

(1) Abolition of movement in response to noxious stimuli. Potency data (expressed as EC₅₀ub - the plasma free drug concentration of anaesthetic that abolishes movement in response to noxious stimuli in 50% of subjects) were obtained from the literature for 14 chemically diverse intravenous general anaesthetics. The compounds were randomly divided into a training set of 10 agents (eltanolone, minaxolone, ORG 21465, thiamylal, thiopental, methohexital, R-ketamine, R-etomidate, ORG 25435 [an a-amino acid phenolic ester derivative] and chlormethiazole) used to derive the CoMFA model; and a test set of 4 agents (Althesin as alphaxalone, S-ketamine, pentobarbital and propofol) used to assess the model’s predictive capability. The CoMFA model explained 94.0% of the observed variance in the activities of the training set compounds (F_1,8 = 125.843, P < 0.0001). It was a good predictor of activity for both the training set (assessed using leave one out cross-validation, q² = 0.768) and test set anaesthetics (r² = 0.799), correctly predicting the potency order of the ketamine enantiomers [5]. Comparable pharmacophores have also been produced for 107 inhalational general anaesthetics [6, 7].

 

(2) Cardiovascular depression. We have some preliminary data which compares CoMFA models for in vivo anaesthetic potency (EC₅₀ub) and cardiovascular depression as measured by a 20% decrease in mean arterial blood pressure (dMAP₂₀). The CoMFA model for blood pressure depression, based on 8 of the 14 i.v. agents, explained 87.0% of the variance in the observed dMAP₂₀ values (F_1,6 = 39.293, P = 0.0008). However, the model currently has limited predictive capability under cross-validation, with a q² of 0.477.

 

Both models show that molecular shape and electrostatic potential are important determinants of intravenous general anaesthetic activity. Further investigations are in progress to determine whether the pharmacophores derived from the in vivo potency and cardiovascular depression models indicate a common molecular basis for these two different aspects of activity. 

 

References

1. Krasowski MD, Harrison NL. General anaesthetic actions on ligand-gated ion channels. Cell Mol Life Sci 1999; 55: 1278-1303.

2.  Yamakura T et al. Anesthetics and ion channels: Molecular models and sites of action. Annu Rev Pharmacol Toxicol 2001; 41: 23-51.

3.  Krasowski MD et al. 4D-QSAR Analysis of propofol analogues: Mapping binding sites for an anesthetic phenol on the GABA_A receptor. J Med Chem 2002; 45: 3210-3221.

4.  Sewell JC, Sear JW. Can molecular similarity-activity models for intravenous general anaesthetics help explain their mechanism of action. Br J Anaesth 2002; 88: 166-174.

5.  Sewell JC, Sear JW. Derivation of preliminary three-dimensional pharmacophoric maps for chemically diverse intravenous general anaesthetics. Br J Anaesth 2004; 92: 45-53.

6.  Sewell JC, Sear JW. Derivation of preliminary three-dimensional pharmacophores for nonhalogenated volatile anesthetics. Anesth Analg 2004; 99: 744-751.

7.  Sewell JC, Sear JW. Determinants of volatile general anesthetic potency: A preliminary three-dimensional pharmacophore for halogenated agents. Anesth Analg In press.

8.  Kubinyi H. QSAR and 3D-QSAR in drug design part 1. Drug Discov Today 1997; 2: 457-467.