FAAH and away: chemometrics finds novel analgesics

Near or FAAH

Quantitative structure-activity relationship (QSAR) analysis of a series of fatty acid amide hydrolase (FAAH) inhibitors might lead to new drugs for pain relief as FAAH inhibitors interfere with various important metabolic processes involving the sleep-inducing lipid oleamide, the N-acyltaurine calcium channel agonists and the natural cannabis analogue anandamide.

Fatty acid amide hydrolase (FAAH) is membrane enzyme that we cannot do without. A member of the amidase family of serine hydrolases, FAAH terminates the signals from the lipid messengers anandamide (an endocannabinoid, more formally known as N-arachidonoylethanolamine) and oleamide through hydrolysis in the central nervous system and in peripheral tissues. The hydrolysis process deactivates these messengers and so modulates the behaviour with which they are inextricably linked, pain, sleep, feeding and movement.

FAAH inhibitors: pros and cons

There are several classes of FAAH inhibitors available to researchers, including those that reversibly bind with the enzyme, alpha-ketoheterocycles and trifluoromethyl ketones, and those that undergo irreversible binding, the carbamates and fluorophosphonates. Each class has its pros and cons, some are highly selective but activity is short-lived so are less efficacious others are less effective but their effects last longer, which means they work better but have greater side effects. The best approach to finding effective FAAH inhibitors that have manageable side effects is open to debate with pharmaceutical companies, including Pfizer and Amgen, all vying for pole position in the drug discovery process.

Peng Lu, Ruisheng Zhang, Yongna Yuan and Zhiguo Gong of Lanzhou University, Lanzhou 730000, Gansu, P. R. China are well aware that FAAH inhibitors would make useful pharmaceutical leads for treating a wide range of disorders given the variety of processes with which the natural substrates of FAAH are involved. As such, they have used chemometrics methods to carry out a quantitative structure activity relationship (QSAR) analysis on the Ki values of a series of FAAH inhibitors. They looked at six molecular descriptors selected by CODESSA software, calculated from molecular structure alone.

The team explains that the structures of all molecules studied were drawn with the ISIS DRAW 2.3 program and pre-optimized using MM+ molecular mechanics force field and precisely optimized with semi-empirical AM1 method implemented in the Hyperchem software package. A semi-empirical AM1 method in MOPAC6.0 was then used for further geometry refinement. The resulting data were fed to CODESSA and subsequently handled either with a linear heuristic method (HM) or a non-linear support vector machine (SVM), which is a supervised learning method used for classification and regression that was originally developed for pattern recognition problems.

The sets of results from SVM and HM were compared. HM has the advantages of high speed and no software restrictions on the size of the dataset it can handle. However, the team found SVM found to perform the most effectively and the researchers say it offers a novel and effective method for predicting the activity of FAAH inhibitors.

Non-linear approach outshines the linear

"The results indicate that a non-linear relationship can accurately describe the relationship between the structural parameter and the pKi value of the compounds studied," the team says, "and the SVM is a useful tool in the prediction pKi value of FAAH inhibitors. Furthermore, the proposed approach could also be extended to other QSAR investigations."

The team adds that, "The main advantage of SVM is that it adopts the structure risk minimization (SRM) principle, which has been shown to be superior to the traditional empirical risk minimization (ERM) principle, employed by conventional neural networks. Now, with the introduction of an epsilon-insensitive loss function, SVM has been extended to solve nonlinear regression estimation and time-series prediction."