Development of predictive pharmacokinetic simulation models for drug discovery.

As discovery chemistry produces increased numbers of potential drug compounds, the use of ADME (absorption, distribution, metabolism, and excretion) properties is becoming increasingly important in the drug selection and promotion process. A computer simulation model has been developed and validated to predict ADME outcomes, such as rate of absorption, extent of absorption, etc. using a limited number of in vitro data inputs. The oral bioavailability of ganciclovir in dogs and humans was simulated using a physiologically based model that utilized many biopharmaceutically relevant parameters, such as the concentration of ganciclovir in the duodenum, jejunum, ileum, and colon at various dose levels and solubility values. The simulations were run and compared to dog and human in vivo data. The simulation results demonstrated that the low bioavailability of ganciclovir is limited by compound solubility rather than permeability due to partitioning as previously speculated. This technology provides a breakthrough in in silico prediction of absorption and with its continued development and improvement, will aid drug discovery and development scientists to produce better pharmaceutical products.

Mechanism of dextran transport across rabbit intestinal tissue and a human colon cell-line (CACO-2).

The in vitro permeabilities of 14C labeled dextrans (10, 40, and 70 kD) were calculated from mass transport across Peyer's patches and non-patch tissues derived from rabbit jejunum, and a human colon cell line (Caco-2) grown as a monolayer on polycarbonate filters. Size distribution of dextrans did not change upon transport as judged from size exclusion chromatography. Permeabilities decreased in a size-dependent manner. Ranking of permeabilities for dextran 10 and 40 kD were: Caco-2 > non-patch tissue > Peyer's patches; while dextran 70 kD demonstrated no difference among the barriers. Tissue resistance, expressed as 1/(permeability.tissue thickness) was virtually the same in Peyer's patches and non-patch tissue, suggesting that tissue thickness and not interaction determines the difference in permeability. ATP depletion with ouabain, Na(+)-azide and 2-deoxy-D-glucose, and low temperature (4 degrees C) did not result in reduced permeabilities suggesting passive transport. The results suggest that the investigated intestinal barriers transport dextrans in a similar fashion independent of their source. However, comparison of the ratios dextran 10 kD/mannitol and PEG 900/mannitol between rabbit tissue and Caco-2 monolayers suggests Caco-2 monolayers may serve as a model to study absorption potential of potentially harmful compounds in coeliac disease, gastroenteritis, and colon carcinoma.

Evaluation of the performance of controlled release dosage forms of ticlopidine using in vitro intestinal permeability and computer simulations.

Prototype controlled release formulations of ticlopidine hydrochloride were developed, but when administered to humans, these formulations significantly reduced the bioavailability of intact drug in plasma. In order to examine the intestinal permeability characteristics and gastrointestinal metabolism of 14C-ticlopidine, we employed an in vitro diffusion cell system to directly measure the permeation of ticlopidine across various segments of monkey and rabbit intestine. High pressure liquid chromatography was used to determine the amount of intact ticlopidine on both the mucosal and serosal sides of the intestinal tissue. Simulations based upon the known pharmacokinetics of ticlopidine were conducted using STELLA, a modeling program, to provide insight as to the nature of the decreased bioavailability of these ticlopidine CR dosage forms. These simulations indicate that the absorption of intact ticlopidine is a non-linear phenomena, with inordinately large increases in absorbed intact drug with increases in dose. Conversely, decreases in drug available for immediate absorption, as with the controlled release dosage forms, lead to non-linear decreases in bioavailability. Such a finding is very consistent with the extensive first-pass metabolism suggested from the tissue permeability studies.

A model to predict aqueous humor and plasma pharmacokinetics of ocularly applied drugs.

PURPOSE: To develop methods for constructing a pharmacokinetic model to predict the time course of aqueous humor and plasma drug concentrations after topical dosage in rabbits using the simulation program iThink (formerly STELLA; High Performance Systems, Lyme, NH). METHOD: The model was constructed in experimentally verifiable segments using previously published data on intravenous, nasal, and ocular dosage, and was used to describe the influence of prolonging precorneal retention and varying drug release rate on the ratio of drug absorbed locally to drug absorbed systemically in rabbits. RESULTS: The model developed is comprehensive; it includes precorneal kinetics, nasal absorption kinetics, and plasma kinetics. CONCLUSIONS: Such a model may be useful in designing drug delivery strategies to improve the safety of topical eye medications through minimization of systemic absorption and maximization of drug delivery to ocular tissues.

Comparison of the permeability characteristics of a human colonic epithelial (Caco-2) cell line to colon of rabbit, monkey, and dog intestine and human drug absorption.

The in vitro permeabilities of Caco-2 monolayers and permeabilities in tissue sections from colon of monkey, rabbit, and dog were compared using a series of compounds. The selected compounds differed in their physicochemical properties, such as octanol/water partition coefficient, water solubility, and molecular weight. Their structure included steroids, carboxylic acids, xanthins, alcohols, and polyethylene glycols. A linear permeability relationship was established between Caco-2 and colon tissue from both rabbit and monkey. The results suggest that Caco-2 is twice as permeable as rabbit and five times as permeable as monkey colon. However, no clear relationship could be established between Caco-2 monolayers and dog colon permeability. A relationship between permeability in Caco-2 monolayers and human absorption was found. The results suggest that within certain limits, permeability of Caco-2 monolayers may be used as a predictive tool to estimate human drug absorption.

Permeability characteristics of various intestinal regions of rabbit, dog, and monkey.

The in vitro permeability of a series of both hydrophilic and lipophilic compounds, as defined by the octanol/water partition coefficient, was measured in four segments of rabbit, monkey, and dog intestine using a side-by-side diffusion cell. A linear relationship was established for tissue resistance to hydrophilic compound diffusion in jejunum and colon among rabbit, monkey, and dog. The results suggest that rabbit jejunum is twice as permeable as monkey and dog jejunum. The colonic tissues of monkey, rabbit, and dog demonstrate similar permeabilities. Measuring the permeabilities of different tissues with compounds of similar physicochemical properties allows comparison of tissue restriction to transport. Thus, in vitro permeability measurements may be used to investigate physiological differences of various intestinal tissue segments that influence tissue permeability. Investigating the permeability of different intestinal segments from various species could allow the identification of an appropriate in vitro intestinal permeability model that will lead to the prediction of intestinal absorption in humans, eliminating the need for extensive and often misleading in vivo animal testing.

Pharmacokinetic basis for nonadditivity of intraocular pressure lowering in timolol combinations.

The authors determined whether the ocular absorption of topically applied timolol in the pigmented rabbit was affected significantly by coadministration with either pilocarpine or epinephrine in the same drop to explain the nonadditivity in intraocular pressure lowering (IOP) seen clinically. They instilled 25 microliters of 0.65% timolol maleate solution (equivalent to 0.5% timolol), both in the presence and absence of 2.6% pilocarpine nitrate or 1% epinephrine bitartrate, into pigmented rabbit eyes. The time course of timolol concentration in the conjunctiva, anterior sclera, corneal epithelium, corneal stroma, aqueous humor, iris-ciliary body, and lens was monitored for 360 min by using reversed-phase high-performance liquid chromatography. The area under the timolol concentration-time curve in all but one of the anterior segment tissues was reduced by 20-50% (mean, 40%) when timolol was coadministered with pilocarpine and by 20-70% (mean, 42%) when timolol was coadministered with epinephrine. Such an effect was not a result of alterations in corneal permeability or aqueous humor turnover rate, nor was it related to the extent of systemic absorption caused by pilocarpine and epinephrine. Rather, the reduction in ocular timolol absorption may have been caused by the accelerated washout of timolol by tears stimulated by the coadministered drugs and, to a lesser extent, by the loss of timolol through binding to the increased amount of tear proteins induced by the coadministered drugs. Thus, the nonadditivity in IOP lowering from timolol-pilocarpine and timolol-epinephrine combinations is probably caused by changes in precorneal timolol clearance.

Improving the ocular to systemic ratio of topical timolol by varying the dosing time.

This study was performed to determine whether the ratio of ocular to systemic absorption of topically applied timolol in the pigmented rabbit can be maximized by varying the time of drop instillation. Twenty-five microliters of 0.65% timolol maleate solutions were instilled into the pigmented rabbit eye at 6 AM, 12 PM, 6 PM, or 12 AM. The time course of timolol concentration in plasma and various eye tissues (conjunctiva, sclera, corneal epithelium, corneal stroma, aqueous humor, iris-ciliary body, and lens) was monitored with the use of reversed phase high-performance liquid chromatography (HPLC). Ocular timolol concentrations were approximately twice as high when the drug was administered at 12 PM than at 6 AM, 6 PM, or 12 AM, whereas timolol concentration in plasma was lowest when the drug was administered at 12 PM. It may, therefore, be possible to maximize the therapeutic index of topically applied timolol by administering the drug at 12 PM. Moreover, the possible influence of dosing time on the extent of ocular and systemic drug absorption must be considered when planning dosing schedules for topically applied ophthalmic drugs.

Characterization of the unstirred water layer in Caco-2 cell monolayers using a novel diffusion apparatus.

Caco-2 monolayers grown on Transwell polycarbonate membranes have been characterized as a valuable tool in drug transport studies. Despite the clear advantages of this system, the lack of stirring may create an unstirred water layer (UWL) whose resistance may limit the transcellular transport of lipophilic molecules. The objective of this study was to evaluate a novel diffusion cell where the transport buffer is mixed by gas lift and to determine the mixing flow rate needed to reduce the thickness (h) of the UWL adjacent to cell monolayers. The transport of the leakage marker, mannitol, remained at least 15-fold lower than the flux of testosterone, indicating that the stirring flow rates used did not affect the integrity of the monolayers. The permeability (P) of testosterone (log PC 3.13) across monolayers mounted on this diffusion cell was 4.07, 10.90, and 14.18 x 10(-5) cm/sec at flow rates of 0, 15, and 40 ml/min, respectively, and the apparent UWLs were calculated to be 1966, 733, and 564 microns. P and h in the stagnant Transwell were 3.08 x 10(-5) cm/sec and 2597 microns, respectively. On the other hand, h was significantly smaller in the unstirred, cell-free membranes than in their cell-containing counterparts. P was correlated with lipophilicity and, in the case of the more lipophilic compounds, with the mixing flow rate.

The effects of enprostil and RS-86505-007 on in-vitro intestinal permeability of rabbit and monkey.

Enprostil is a prostaglandin E2 analogue characterized as a racemic mixture of four stereoisomers. Enprostil and a single isomer, RS-86505-007, were evaluated for their effects on the permeability of actively and passively transported compounds in segments of small intestine from rabbits and monkeys. Consistent with human in-vivo studies, which have demonstrated decreases in absorption of D-xylose, both compounds inhibited D-glucose transport. The passively transported compounds mannitol and progesterone were also less permeable in this model in the presence of enprostil or RS-86505-007. In contrast to the concentration-dependent inhibition displayed by ouabain, RS-86505-007 had no effect on purified Na+K(+)-ATPase. It is suggested that an effect of a general nature, possibly an increase in the barrier properties at the intestinal surface, may explain the transport inhibition. Of two other enprostil isomers, RS-86812-007 inhibited D-glucose transport in rabbit small intestine, while RS-86505-008 had no effect. The prostaglandin E1 analogue misoprostol was ineffective in monkey and poorly effective in rabbit. This suggests that the inhibition of D-glucose transport by enprostil and its active stereoisomers is mediated through some structurally specific receptor interaction.

A correlation of permeabilities for passively transported compounds in monkey and rabbit jejunum.

Permeability measurements were conducted for a series of compounds using in vitro tissue sections from monkey and rabbit jejunum. Jejunal segments were stripped of serosal musculature and mounted in a diffusion-cell system, using previously described methods and equipment. Permeability determinations of radiolabeled compounds ranging over two orders of magnitude in molecular weight were conducted. For the compounds examined, the permeability of the rabbit jejunum was approximately twice that of the monkey. This was in contrast to the relationship implied by the stripped tissue thickness measurements of 0.92 and 0.83 mm for rabbit and monkey, respectively. An investigation of the size of the paracellular space in the jejunum was undertaken to account for this apparent discrepancy in tissue permeability. Scanning electron micrographs of intestinal sections revealed a similar packing density of cells between species; however, a difference was noted in the shape and number of villi per unit area. Comparative measurements of the paracellular volume in both species using mannitol and methoxyinulin as extracellular space markers further suggests that the paracellular junctions are similar in size but more numerous per unit area of rabbit jejunum than that of the monkey. In contrast to passively transported compounds, the active transport of D-glucose was greater in monkey jejunum compared to rabbit tissue segments. When active transport was inhibited by blockade of the sodium pump with ouabain, the passive component of D-glucose transport for both rabbit and monkey tissue was in agreement with the relationship demonstrated above for compounds which are solely transported by passive processes.

Evidence for site-specific absorption of a novel ACE inhibitor.

Moexipril [2-[(1-ethoxycarbonyl)-3-phenylpropyl]amino-1-oxopropyl]-6, 7-dimethoxy-1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid (S,S,S)], an ester prodrug of an ACE inhibitor, was formulated in controlled-release preparations with a range of in vitro release rates, to provide a prolonged input of drug in vivo. However, pharmacokinetic studies with the controlled-release dosage forms in humans produced plasma profiles with the same characteristics and time to peak as an immediate-release capsule. In vitro dissolution data from the controlled-release dosage form, as well as the known characteristics of the polymer used to control drug release from the dosage form, suggest no reason to suspect an abrupt halt to the in vivo release of the drug after 1-2 hr. The lack of sustained blood levels is, therefore, most likely due to failure of the GI tract to absorb the drug beyond some location in the upper small intestine, i.e., site-specific absorption. This theory is supported by a series of computer simulations involving moexipril and the active moiety, moexipril diacid. Possible mechanisms include poor drug permeability, a pH effect whereby the zwitterionic form of the drug is more rapidly absorbed, and esterase cleavage of moexipril to the poorly absorbed moexipril diacid.

In vitro measurement of gastrointestinal tissue permeability using a new diffusion cell.

A new diffusion cell, derived from the Ussing chamber, was developed for the measurement of tissue permeability. This cell incorporates the attributes of using a single material and laminar flow across the tissue surface. In addition, the design allows the cell to be manufactured in a wide range of sizes to allow optimization of surface area to volume for a variety of tissues. The apparatus is applicable to the evaluation of transport of compounds through mucosal/epithelial barriers, i.e., gastrointestinal tissue. Active transport, permeability enhancers, enzymatic degradation, and absorption in various tissue sections can be explored. Preliminary data are consistent with the expected effects of molecular size and partition coefficient of a transported molecule on permeability in epithelial tissue. In addition, active transport of D-glucose and inhibition by phloridzin and ouabain can be demonstrated.

Mechanisms of corneal drug penetration. II: Ultrastructural analysis of potential pathways for drug movement.

Ultrastructure analysis was conducted in an effort to augment the results of classical kinetic studies. Scanning electron microscopy (SEM) allowed visual inspection of cellular junctions on corneal epithelium and endothelium. The addition of calcium-chelating agents to in vivo and in vitro mounted corneas demonstrated a concentration-dependent progressive expansion of the intercellular spaces of epithelium and endothelium, as seen by SEM. The expansion of these cellular junctions correlates with increases in permeability of the cornea to glycerol under similar conditions. The size of the intercellular space was estimated by transmission electron microscopy. Use of lanthanum as a marker of aqueous diffusional pathways demonstrated that the epithelial surface is not a totally occlusive barrier to transport of small hydrophilic compounds. Development of a method whereby an administered drug could be visualized in its actual pathway of movement through the cornea was undertaken, involving precipitation of specific compounds in the tissue with osmium tetroxide vapor. Results suggest that separate pathways of drug movement exist in the cornea for hydrophilic and hydrophobic compounds. Hydrophilic compounds were preferentially located in intercellular spaces, whereas hydrophobic compounds were associated with the lipid structures of the tissue. The results of these studies are consistent with a currently proposed 'pore' model for the penetration of drugs through the cornea.

Mechanisms of corneal drug penetration. III: Modeling of molecular transport.

A model relating the parameters of permeability coefficient in the cornea with partition coefficient and molecular weight of the penetrating species is presented. The development of the model is unique in that it includes the availability of a "pore" pathway with the corresponding kinetic data to substantiate this premise. The "pore" pathway is applied to small hydrophilic compounds and assumes that an aqueous diffusional space is available for transport of these compounds. This is in contrast to an alternate "partitioning" mechanism which is the most probable route of transport for larger or more lipophilic entities. The model is consistent with published data from this and other laboratories.

Mechanisms of corneal drug penetration. I: In vivo and in vitro kinetics.

Corneal penetration studies were conducted in unanesthetized albino rabbits using various organic compounds representing both polar and nonpolar species. Very low molecular weight compounds demonstrate rapid uptake into the aqueous humor despite the lipid-like barrier imposed by the corneal epithelium. Evidence that these compounds may have access to a diffusional channel for corneal transport is presented. In vitro permeability studies were also conducted in an effort to quantitate the corneal diffusion of compounds covering a range of molecular weights and partition coefficients; the results corresponded well with the results of in vivo experiments. Calculations of energies of activation, taken from Arrhenius plots, indicate that the diffusion of drug across the cornea may be by two different mechanisms that depend on the physical-chemical characteristics of the perfusant. One mechanism appears similar to drug movement in an aqueous environment and is characterized by an activation energy similar to that for diffusion in water. The other relates to the expected partitioning of a compound across cellular membranes represented by a relatively high activation energy for diffusion. For hyrdophilic compounds, the epithelium appears to be rate limiting to drug movement, whereas for hydrophobic compounds, the stroma is rate limiting. In the presence of calcium-chelating agents, glycerol demonstrated an increase in corneal penetration in vivo. This effect appears to be reversible at specific concentrations of chelator. In contrast, divalent cations reduced corneal penetration of glycerol. The known calcium chelator EDTA was shown to penetrate the cornea, conjunctiva, and iris/ciliary body from a topically applied dose. The implications of this observation pertain to toxicity effects when EDTA is incorporated into ocular drug products for stability purposes, or novel stratagems for improving ocular bioavailability of topically applied drugs are employed. The addition of calcium-chelating agents to in vivo mounted corneas demonstrated increases in permeability of the cornea to glycerol which were directly related to the concentration of chelating agent used. These results paralleled the findings of similar in vivo studies. The results of these studies are consistent with a currently proposed 'pore' model for the penetration of drugs through the cornea which demonstrates both a partition coefficient and molecular weight dependency on the permeability of the cornea to transported compounds.

Effects of calcium chelating agents on corneal permeability.

Corneal penetration studies have been conducted in unanesthetized albino rabbits using various organic compounds representing both polar and nonpolar species. In the presence of calcium chelating agents, polar compounds generally demonstrate an increase in corneal penetration. Evidence that this corneal effect is reversible is presented. Concomitant with an increase in both corneal and aqueous humor drug levels was a decrease in drug concentration in both iris and conjunctival tissues tentatively attributed to chelation effects on vascular permeability of these tissues. EDTA, a known calcium chelator, was shown to penetrate the cornea, conjunctiva, and iris/ciliary body from a topically applied dose. The implications of this observation pertain to both toxicity effects, when EDTA is incorporated into ocular drug products for stability purposes, and novel strategems for improving ocular bioavailability of topically applied drugs.

Relationship of chemical structure to corneal penetration and influence of low-viscosity solution on ocular bioavailability.

Current understanding of the mechanism of corneal penetration by organic molecules proposes the epithelial layer as the rate-limiting membrane for water-soluble compounds and the stromal layer as rate limiting for lipid-soluble compounds. This suggests that the relationship between corneal permeability and the logarithm of oil/water partition coefficients, for a series of drugs, should not be the typical, single, continuous, parabolic-shaped curve. Corneal penetration studies have been conducted in unanesthetized albino rabbits using various organic compounds, representing five orders of magnitude in partition coefficient, at a constant concentration of 4 X 10(-5) M dispensed in either a 1- or 90-centipoise (cps) solution. It has been shown that for non-ionizable compounds, a pair of bell-shaped curves were generated, one for lipid-soluble and one for water-soluble compounds. Small water-soluble species demonstrate very high apparent permeabilities, which may relate to the presence of aqueous pores or other paracellular drug movement. Penetration of ionizable compounds does not appear to correlate well with the structural relationships invoked for un-ionized compounds. Consistent with the proposed mechanisms of corneal penetration, oil-soluble drug substances show no improvement in drug bioavailability when dosed from a 90-cps solution, and water-soluble drugs show a modest improvement in ocular drug bioavailability. Small water-soluble substances demonstrate no improvement due to their already high bioavailability. The importance of nonproductive absorption and precorneal drainage on bioavailability is addressed.

Ocular delivery of pilocarpine from erodible matrices.

The present study examined the feasibility of sustaining the release of a water-soluble drug, pilocarpine, to the tear film. Both gels and dried films were utilized as drug delivery systems. In vitro studies demonstrated significant prolongation of drug release from these systems as compared with simple aqueous or viscous solutions. The in vitro results were supported by in vivo miosis studies in albino rabbits.