Theoretical Calculation and Prediction of Drug Transport Processes Using Simple Parameters and Partial Least Squares Projections to Latent Structures (PLS) Statistics. The Use of Electrotopological State Indices ULF NORINDER, 1 THOMAS O È STERBERG 2 1 AstraZeneca R&D So È derta Èlje, SE-151 85 So È derta Èlje, Sweden 2 Pharmacia Corporation, Pharmaceutical Sciences, SE-112 87 Stockholm, Sweden Received 28 February 2000; revised 1 November 2000; accepted 12 January 2001 ABSTRACT: A method of modeling and predicting drug transport processes using simple, theoretically computed molecular descriptors and multivariate statistics has been investigated in four data sets related to Caco-2 cell permeability, human intestinal absorption, brain±blood partitioning, and immobilized arti®cial membrane (IAM) chromatography. The program Molconn-Z was used to compute theoretical molecular descriptors related to electrotopological state indices. Additional parameters related to size and lipophilicity [i.e., calculated molar refraction (CMR) and octanol±water partition coef®cient (CLOGP)] were also used in the statistical modeling. Good statistical models were derived (r 2 and Q 2 values ranged from 0.75 to 0.93 and 0.70 to 0.89, respectively) that permit fast computational (electronic) screening and prioritization of virtual libraries. ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 90:1076±1085, 2001 Keywords: Molconn-Z; partial least squares projections to latent structures (PLS); Sybyl; human intestinal absorption; Caco-2 cell permeability; brain±blood partitioning; IAM chromatography; electrotopological state indices INTRODUCTION A need for fast computational (electronic) screen- ing methods for virtual compound collections (e.g., virtual libraries) has become apparent as a result of the introduction of combinatorial chemistry and high throughput screening methods. The compu- tational methods are required for prioritization and strategic considerations when planning new campaigns. They may also be used for the evalua- tion of commercial compound libraries. Thus, there is a high demand for rapid and ef®cient computa- tional methods capable of differentiating drugs with acceptable drug transport properties (e.g., intestinal absorption, permeability across cell membranes, as well as membrane partitioning) at an early stage in the drug discovery process. Furthermore, even though experimental proce- dures may evolve that operate in short time frames, theoretical analysis of transport processes is still important. These investigations may give valuable feedback and insight into the possible mechanisms of drug transport as well as identify important molecular parameters for the property in question that is being investigated. We recently reported a new method of modeling various drug transport properties by MolSurf technology 1 in conjunction with multivariate statistics. 2±4 In this work our aim is to focus attention on simple and rapidly computed mole- cular descriptors derived from two-dimensional (2-D) graphs that are expected to encompass the information needed [i.e., hydrogen bonding, lipo- philicity (octanol±water partition coef®cient; CLOGP), and calculated molar refraction (CMR), for the development of multivariate statistical 1076 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 8, AUGUST 2001 Correspondence to: Ulf Norinder (Telephone: 46 8 553 25057; E-mail: ulf.norinder@astrazeneca.com) Journal of Pharmaceutical Sciences, Vol. 90, 1076±1085 (2001) ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association