Permeability In The Gastrointestinal Tract And At The Bloodbrain Barrier

Measured permeability (especially when combined with solubility and charge state) can be viewed as a surrogate property for predicting oral (gastrointestinal) absorption of preclinical drug candidate molecules. This chapter considers the transport of molecules by passive diffusion through phospholipid bilayers. The emphasis is on (1) the current state-of-the-art measurement of permeabilities by the so-called PAMPA method and (2) the theoretical physicochemical models that attempt to rationalize the observed transport properties. Such models are expected to lead to new insights into the in vivo absorption processes. In oral absorption predictions, the established in vitro assay to assess the permeability coefficients is based on Caco-2 cultured-cell confluent monolayers [48,510-515]. We refer to this topic in various places, drawing on the biophysical aspects of the work reported in the literature. We also consider some physicochemical properties of the blood-brain barrier (BBB), insofar as they contrast to those of the gastrointestinal tract (GIT). Our main focus, however, is on results derived from simpler in vitro systems based on artificial membranes.

In order to rationalize membrane permeability and oral absorption in terms of physicochemical drug properties, good experimental data and sound theoretical

Absorption and Drug Development: Solubility, Permeability, and Charge State. By Alex Avdeef ISBN 0-471-423653. Copyright © 2003 John Wiley & Sons, Inc.

models are needed. Since lipophilicity is such an important concept in ADME (absorption, distribution, metabolism, excretion) predictions, models that address the permeability-lipophilicity relationships are expected to provide important insights. In the simplest models, permeability is linearly related to the membrane-water partition coefficient [Eq. (2.3)], but in practice, linearity is not generally observed over a wide range of lipophilicities. To explain this, different theoretical models for passive membrane diffusion have been described in the literature.

In assays based on synthetic membranes, the nonlinearity may be caused by (1) unstirred water layer; (2) aqueous pores in oily membranes; (3) membrane retention of lipophilic solute; (4) excessive lipophilicity (non-steady-state conditions, long acceptor-side solute desorption times); (5) transmembrane pH gradients; (6) effects of buffers (in the unstirred water layer); (7) precipitation of solute in the donor side; (8) aggregation of solutes in the donor side (slowing diffusion); (9) specific hydrogen bonding, electrostatic, and hydrophobic/lipophilic interactions with membrane constituents; (10) solute charge state (pKa effects) and membrane surface charge (Gouy-Chapman effects); and (11) the use of inappropriate permeability equations (e.g., neglecting membrane retention of lipophilic drugs).

In vitro systems based on cultured cells are subject to all the above mentioned nonlinear effects, plus those based on the biological nature of the cells. The apical and basolateral membranes have different lipid components, different surface charge domains, and different membrane-bound proteins. Active transporters abound. Some enhance permeability of drugs, others retard it. A very important efflux system, P-gp (where ''P'' denotes permeability), prevents many potentially useful drugs from passing into the cells. P-gp is particularly strongly expressed in the BBB and in cancer cells. The junctions between barrier cells can allow small molecules to permeate through aqueous channels. The tightness of the junctions varies in different parts of the GIT. The junctions are particularly tight in the endothelial cells of the BBB. The GIT naturally has a pH gradient between the apical and basolateral sides of the epithelial cell barrier. Metabolism plays a critical role in limiting bioavailability of drugs.

In penetrating biological barriers, drugs may have simultaneous access to several different mechanisms of transport. To develop an integrated model for the biological processes related to oral absorption is a daunting challenge, since many of these processes are not entirely understood. Most practical efforts have been directed to deriving sufficiently general core models for passive membrane transport (both transcellular and paracellular), addressing many of the effects observed in artificial membrane studies, as listed above. Components of the active transport processes, derived from more complex in vitro cultured-cell models, can then be layered on top of the core passive models.

In the bulk of this chapter we will focus on the rapidly emerging new in vitro technology based on the use of immobilized artificial membranes, constructed of phospholipid bilayers supported on lipophilic filters. One objective is to complete the coverage of the components of the transport model explored in Chapter 2, by considering the method for determining the top curve (horizontal line) in the plots in Fig. 2.2 (i.e., intrinsic permeabilities P0 of drugs). Also, a new model for gastrointestinal (oral) absorption based on permeability measurements using artificial membranes will be presented.

Approximately 1400 measurements of permeability are presented in tables and figures in this chapter. Most of the data are original, not published previously. Unless otherwise noted, the permeability and membrane retention data are from pION's laboratories, based on the permeation cell design developed at pION. Cells of different designs, employing different filter and phospholipid membrane materials, produce different permeability values for reasons discussed below. Although the analysis of the measurements is the basis of the presentation in this chapter, much of the data can be further mined for useful quantitative structure-property information, and the reader is encouraged to do so. First-person references in this chapter, such as ''our laboratory," refer specifically to pION's laboratory, and ''our results" are those of several colleagues who have contributed to the effort, covering a period of >4 years, as cited in the acknowledgment section. Where possible, comparisons to published permeability results from other laboratories will be made.

The survey of over 50 artificial lipid membrane models (pION) in this chapter reveals a new and very promising in vitro GIT model, based on the use of levels of lecithin membrane components higher than those previously reported, the use of negatively charged phospholipid membrane components, pH gradients, and artificial sink conditions. Also, a novel direction is suggested in the search for an ideal in vitro BBB model, based on the salient differences between the properties of the GIT and the BBB.

We return to using the Kp and Kd symbols to represent the partition coefficient and the apparent partition (distribution) coefficient, respectively. The effective, apparent, membrane, and intrinsic permeability coefficients are denoted Pe, Pa, Pm, and P0, respectively, and D refers to the diffusivity of molecules.

The coverage of permeability in this book is more comprehensive than that of solubility, lipophilicity, and ionization. This decision was made because permeability is not as thoroughly treated in the pharmaceutical literature as the other topics, and also because much emphasis is placed on the PAMPA in this book, which is indeed a very new technique [547] in need of elaboration.

0 0

Post a comment