Transport CoMFA

The same assumptions apply to CoMFA as to ordinary Hansch analysis. These are additivity of effects and the availability of structurally similar (congeneric) molecules. The method does not account for pharmacokinetic effects, such as distribution, elimination, transport and metabolization. A prospective drug may appear to bind well to the receptor or enzyme, but may not reach the target site due to undesirable pharmacokinetic properties [8]. 

Wellsow, J., Machulla, H.-J. and Kovar, K.-A. (2002) 3D QSAR of serotonin transporter ligands CoMFA and CoMSIA studies. Quant. Struct. -Act. Relat., 21, 577-589. 

Structural studies of PepTl substrates were extended using comparative molecular field analysis (CoMFA) of 10 known substrates for the peptide transporter with data derived from an in situ rat model (222). The CoMFA approach required manual alignment of relatively rigid molecules however, it allowed explanation of the variation in the permeability with respect to steric and electrostatic interaction energies of the molecules (222). This 3D-QSAR produced by our laboratory provided a valuable starting point for future prediction of... 

However, despite the simplicity of the analyses and the good correlations obtained in these studies, a ligand interaction-based model like the CoMFA method should not be used to model nonlinear effects arising from transport and distribution no reasonable results can be expected for sets of compounds which are no homologous series. Better and theoretically more consistent alternatives would be the addition of suitably weighted log P values to the CoMFA table, the use of lipophilicity similarity matrices (chapter 9.4), or the correlation with log P values in the classical manner, applying either the parabolic or the bilinear model. 

While such lipophilidty similarity matrices do not consider the 3D structures of the molecules, they seem to be appropriate for the incorporation of nonlinear lipophilicity-activity relationships in 3D QSAR analyses, e.g. in CoMFA studies. At least from a theoretical point of view lipophiUcity similarity matrices should be preferred when the nonlinear lipophilicity-activity relationship does not result from binding but from transport and distribution of the drugs in the biological system, which most often is the case. 

T. I. Oprea, unpublished work (1996). Because CoMFA fields are derived from a three-dimensional grid placed in Cartesian space, one can empirically estimate that CoMFA models should exhibit three PLS components. A requirement for higher complexity to fit the data set may indicate that mechanisms in addition to ligand fit (e.g., transport, difhision) are included in the target property. 

---