Derivation of preliminary three-dimensional pharmacophores for nonhalogenated volatile anesthetics

Molecular mechanisms of drug action: an emerging view

Volatile anesthetics serve as useful probes of a conserved biological process that is essential to the proper functioning of the central nervous system. A kinetic and thermodynamic analysis of their unusual pharmacological and physiological characteristics has led to a general, predictive theory in which small molecules that adsorb to membranes modulate ion channel function by altering physical properties of membrane bilayers. A kinetic model that is both parsimonious and falsifiable has been developed to test this mechanism. This theory leads to predictions about the structure, function, origin, and evolution of synapses, the etiology of several diseases and disease symptoms affecting the brain, and the mechanism of action of several drugs that are used therapeutically. Neuronal membranes may offer an appealing drug target, given the large number of compounds that adsorb to interfaces and hence membranes.

Comparative molecular field analysis to derive pharmacophore maps for disposition parameters of intravenous anaesthetic agents

BACKGROUND

The present study examines the molecular basis of the disposition kinetics for i.v. hypnotic agents using comparative molecular field analysis (CoMFA).

METHODS

The systemic clearance (Cl(s) litre min(-1)) and apparent volume of distribution at steady state (Vd(ss) litres) for 14 i.v. anaesthetics in human subjects were obtained from the literature, or from unpublished data, and used to form CoMFA models for the two aspects of drug disposition. Molecular alignment was achieved by field-fit minimization with the lead structure for all models eltanolone. The resulting pharmacophore maps were also compared with the pharmacophores for cardiovascular depression (expressed in terms of the drug concentration in 50% patients, associated with a 20% decrease in mean arterial pressure during infusion anaesthesia in the absence of other adjuvant drugs or noxious stimulation), which were taken from the literature.

RESULTS

The CoMFA model for Cl was based on two latent variables, explained 95.2% of the variance in observed activities, and showed good intrinsic predictability (cross-validated q(2) 0.663). The model for Vd(ss) was also based on two latent variables: r(2) 0.986 and q(2) 0.718. Comparison of the pharmacophores for the two disposition parameters showed poor correlation for both electrostatic and steric regions (r=-0.220 and 0.018; both NS). The relative contributions of electrostatic and steric interactions differed between the models (Cl(s) 1.9:1; Vd(ss) 2.5:1). Comparison with the cardiovascular pharmacophores depression models gave r values of 0.551 (P<0.05) and 0.407 (ns) for Cl(s) (for electrostatic and steric models) and -0.225 and -0.448 for Vd(ss) (both ns).

CONCLUSIONS

Comparison of CoMFA models for drug disposition show only small elements of commonality, suggesting different molecular features may be responsible are two properties. There was better similarity for both disposition pharmacophores with the pharmacophores for cardiovascular depression.

Comparative molecular field analysis to derive pharmacophore maps for induction doses of intravenous anaesthetic agents

BACKGROUND

The present study examines the molecular basis of induction of anaesthesia by i.v. hypnotic agents using comparative molecular field analysis (CoMFA).

METHODS

ED(50) induction doses for 14 i.v. anaesthetics in human subjects (expressed as molar dose per kilogram body weight) were obtained from the literature. Immobilizing potency data for the same 14 agents (expressed as the EC(50) plasma free drug concentrations that abolish movement in response to a noxious stimulus in 50% patients) were taken from our previous publication. These data were used to form CoMFA models for the two aspects of anaesthetic activity. Molecular alignment was achieved by field-fit minimization techniques. The lead structure for both models was eltanolone.

RESULTS

The final CoMFA model for the ED(50) induction dose was based on two latent variables, and explained 99.3% of the variance in observed activities. It showed good intrinsic predictability (cross-validated q(2)=0.849). The equivalent model for immobilizing activity was also based on two latent variables, with r(2)=0.988 and q(2)=0.852. Although there was a correlation between -log ED(50) and -log EC(50) (r(2)=0.779), comparison of the pharmacophore maps showed poor correlation for both electrostatic and steric regions when isocontours were constructed by linking lattice grid points, making the greatest 40% contributions; the relative contributions of electrostatic and steric interactions differing between the models (induction dose: 2.5:1; immobilizing activity 1.8:1).

CONCLUSIONS

Comparison of two CoMFA activity models shows only small elements of commonality, suggesting that different molecular features may be responsible for these two properties of i.v. anaesthetics.

Development of pharmacophoric maps for cardiovascular depression by intravenous anaesthetic agents: comparison with maps for immobilizing activity

BACKGROUND

The molecular basis of the cardiovascular effects of i.v. anaesthetics was investigated using comparative molecular field analysis (CoMFA).

METHODS

The cardiovascular effects, measured as changes in mean arterial pressure (MAP), compared with awake values of continuous infusions of 13 structurally diverse i.v. anaesthetics were compared at EC(50) plasma concentrations, and by determination of plasma-free drug concentrations associated with a 20% decrease in MAP (dMAP(20)). Data were obtained both from the literature and from unpublished data of the author. The results were fitted to a CoMFA activity model using field-fit minimization techniques to maximize similarities in molecular bulk and electrostatic potential to the lead compound, eltanolone.

RESULTS

The final model for cardiovascular depression based on free drug concentrations associated with dMAP(20) explained 95.8% of the variance in observed activities, with a cross-validated q(2) of 0.824 (n=12). A second model based on change in MAP at EC(50) plasma concentrations explained 98.3% of the variance in arterial pressure, but performed poorly at cross-validation (q(2) 0.526). The comparative model for immobilizing potency had an r(2) value of 0.987 and q(2) 0.823. Comparison of pharmacophoric maps showed several key electrostatic and steric regions common to both models when isocontours were constructed linking lattice grid points, making the greatest 40% contributions (87.57% for electrostatic fields and 86.16% for steric fields).

CONCLUSIONS

Comparison of activity models for cardiovascular depression and immobilizing potency for i.v. anaesthetics shows significant commonality, suggesting that it may not be possible to separate those molecular features associated with each of these effects.

The biological and toxicological activity of gases and vapors

A large amount of data on the biological and toxicological activity of gases and vapors has been collected from the literature. Processes include sensory irritation thresholds, the Alarie mouse test, inhalation anesthesia, etc. It is shown that a single equation using only five descriptors (properties of the gases and vapors) plus a set of indicator variables for the given processes can correlate 643 biological and non-lethal toxicological activities of 'non-reactive' compounds with a standard deviation of 0.36 log unit. The equation is scaled to sensory irritation thresholds obtained by the procedure of Cometto-Muñiz, and Cain provides a general equation for the prediction of sensory irritation thresholds in man. It is suggested that differences in biological/toxicological activity arise primarily from transport from the gas phase to a receptor phase or area, except for odor detection thresholds where interaction with a receptor(s) is important.

What makes a molecule an anaesthetic? Studies on the mechanisms of anaesthesia using a physicochemical approach

Recent studies of mechanisms of anaesthesia have been mainly 'target orientated', investigating the activity of both volatile and i.v. agents at putative sites of action. An alternative approach is one that is 'ligand orientated', focusing on the properties of molecules that define their immobilizing ability and secondly define their potency. The use of conventional descriptors (such as non-polar solubility or the octanol-water partition coefficient [Log P]) are limited in their utility as predictors of potency as they represent three-dimensional molecular properties as a one-dimensional parameter. Using different computer-based molecular modelling methods (molecular similarity studies and comparative molecular field analysis [CoMFA]), we have identified the molecular bases of the activity of structurally diverse anaesthetics, such that they can be described as a single model based on the spatial distribution of molecular bulk and electrostatic potential. The same approach can also be used to model other properties of anaesthetic agents, such as cardiovascular depression. The present data suggest that, for the i.v. agents, it may be difficult to separate immobilizing (anaesthetic) activity and cardiovascular depression within a single molecule.

A comparison of the molecular bases for N-methyl-D-aspartate-receptor inhibition versus immobilizing activities of volatile aromatic anesthetics

BACKGROUND

Aromatic anesthetics exhibit a wide range of N-methyl-d-aspartate (NMDA) receptor inhibitory potencies and immobilizing activities. We sought to characterize the molecular basis of NMDA receptor inhibition using comparative molecular field analysis (CoMFA), and compare the results to those from an equivalent model for immobilizing activity.

METHODS

Published potency data for 14 compounds were supplemented with new values for 2 additional agents. The anesthetics were divided into a training set (n = 12) used to formulate the activity models and a test set (n = 4) used to independently assess the models' predictive capability. The anesthetic structures were geometry optimized using ab initio quantum mechanics and aligned by field-fit minimization to provide the best correlation between the steric and electrostatic fields of the molecules and one or more lead structures. Orientations that yielded CoMFA models with the greatest predictive capability (assessed by leave-one-out cross-validation) were retained.

RESULTS

The final CoMFA model for the inhibition of NR1/NR2B NMDA receptors explained 99.3% of the variance in the observed activities of the 12 training set agents (F(2,)(9) = 661.5, P < 0.0001). The model effectively predicted inhibitory potency for the training set (cross-validated r(2)(CV) = 0.944) and 4 excluded test set compounds (predictive r(2)(Pred) = 0.966). The equivalent model for immobility in response to noxious stimuli explained 98.0% of the variance in the observed activities for the training set (F(2,)(9) = 219.2, P < 0.0001) and exhibited adequate predictive capability for both the training set (r(2)(CV) = 0.872) and test set (r(2)(Pred) = 0.926) agents. Comparison of pharmacophoric maps showed that several key steric and electrostatic regions were common to both activity models, but differences were observed in the relative importance of these key regions with respect to the two aspects of anesthetic activity.

CONCLUSIONS

The similarities in the pharmacophoric maps are consistent with NMDA receptors contributing part of the immobilizing activity of volatile aromatic anesthetics.

A hypothesis on the origin and evolution of the response to inhaled anesthetics

In this article, I present an evolutionary explanation for why organisms respond to inhaled anesthetics. It is conjectured that organisms today respond to inhaled anesthetics owing to the sensitivity of ion channels to inhaled anesthetics, which in turn has arisen by common descent from ancestral, anesthetic-sensitive ion channels in one-celled organisms (i.e., that the response to anesthetics did not arise as an adaptation of the nervous system, but rather of ion channels that preceded the origin of multicellularity). This sensitivity may have been refined by continuing selection at synapses in multicellular organisms. In particular, it is hypothesized that 1) the beneficial trait that was selected for in one-celled organisms was the coordinated response of ion channels to compounds that were present in the environment, which influenced the conformational equilibrium of ion channels; 2) this coordinated response prevented the deleterious consequences of entry of positive charges into the cell, thereby increasing the fitness of the organism; and 3) these compounds (which may have included organic anions, cations, and zwitterions as well as uncharged compounds) mimicked inhaled anesthetics in that they were interfacially active, and modulated ion channel function by altering bilayer properties coupled to channel function. The proposed hypothesis is consistent with known properties of inhaled anesthetics. In addition, it leads to testable experimental predictions of nonvolatile compounds having anesthetic-like modulatory effects on ion channels and in animals, including endogenous compounds that may modulate ion channel function in health and disease. The latter included metabolites that are increased in some types of end-stage organ failure, and genetic metabolic diseases. Several of these predictions have been tested and proved to be correct.

Is a new paradigm needed to explain how inhaled anesthetics produce immobility?

A paradox arises from present information concerning the mechanism(s) by which inhaled anesthetics produce immobility in the face of noxious stimulation. Several findings, such as additivity, suggest a common site at which inhaled anesthetics act to produce immobility. However, two decades of focused investigation have not identified a ligand- or voltage-gated channel that alone is sufficient to mediate immobility. Indeed, most putative targets provide minimal or no mediation. For example, opioid, 5-HT3, gamma-aminobutyric acid type A and glutamate receptors, and potassium and calcium channels appear to be irrelevant or play only minor roles. Furthermore, no combination of actions on ligand- or voltage-gated channels seems sufficient. A few plausible targets (e.g., sodium channels) merit further study, but there remains the possibility that immobilization results from a nonspecific mechanism.

Effect of anesthetic structure on inhalation anesthesia: implications for the mechanism

Many previous attempts (e.g., the Meyer-Overton hypothesis) to provide a single set of physical or chemical characteristics that accurately predict anesthetic potency have failed. A finding of a general predictive correlation would support the notion of a unitary theory of narcosis. Using the Abraham solvation parameter model, the minimum alveolar concentration, MAC, of 148 varied anesthetic agents can be fitted to a linear equation in log (1/MAC) with R(2) = 0.985 and a standard deviation, SD = 0.192 log units. Division of the 148 compounds into a training set and a test set shows that log (1/MAC) values can be predicted with no bias and with SD = 0.20 log units. The two main factors that determine MAC values are compound size and compound hydrogen bond acidity, both of which increase anesthetic activity. Shape has little or no effect on anesthetic activity. Our observations support a unitary theory of narcosis by inhalation anesthetics. A two-stage mechanism for inhalation anesthesia accounts for the observed structural effects of anesthetics. In this mechanism, the first main step is transfer of the anesthetic to the site of action, and the second step is interaction of the anesthetic with a receptor(s).

QSAR Studies on the Anesthetic Action of Some Polyhalogenated Ethers

There has been an on-going debate about the mode of action of general anesthetics and until now, many sites have been postulated as target site for action of these compounds. Here, some quantum chemical-based quantitative structure-activity relationship (QSAR) models were developed for a set of polyhalogenated ethers in order to investigate the aspects of their anesthetic action, which is not completely defined yet, although some hypotheses have been suggested. A data set including 25 polyhalogenated methyl ethyl ethers were selected, and different descriptors were calculated for each molecule using density functional theory calculations, and subsequently some multilinear QSAR models were built by using different sets of the calculated molecular descriptors. The result showed that polar (polarizability) and non-polar (log P) parameters have mixed role on the anesthetic activity i.e. models with high statistical quality were obtained in combination with these two parameters. Also a good model between anesthetic action and electrostatic potentials was obtained, which may imply the important role of electronic interactions in the anesthetic activity of the compounds. Finally, a four-parametric QSAR model containing log P, molecular polarizability, most positive charge and an electrostatic potential parameters was obtained, which indicated that the anesthetic action of the polyhaloganted ethers may be proceeded through lipophilic, steric and columbic interactions.

Determinants of volatile general anesthetic potency: a preliminary three-dimensional pharmacophore for halogenated anesthetics

We investigated the molecular basis for the immobilizing activity of halogenated volatile anesthetics using comparative molecular field analysis. In vivo potency data (expressed as minimum alveolar concentrations) for 69 structurally diverse anesthetics were obtained from the literature. The drugs were randomly divided into a training set (n = 52) used to derive the activity model and a test set (n = 17) used to independently assess the model's predictive power. The anesthetic structures were aligned so as to maximize their similarity in molecular shape and electrostatic potential to the most potent drug in the group, CF2H-(CF2)3-CH2OH. The conformers and alignments of the anesthetics with maximum similarity (calculated as Carbo indices) were retained and used to derive the comparative molecular field analysis models. The final model explained 94.2% of the variance in the observed activities of the training set compounds. The model showed good predictive capability for both the training set (cross-validated r2 = 0.705) and randomly excluded test set anesthetics (r2 = 0.837). Three-dimensional pharmacophoric maps were derived to identify the spatial distribution of key areas where steric and electrostatic interactions are important in determining immobilizing activity of the halogenated drugs and were compared with our previously published maps obtained for nonhalogenated volatile anesthetics.