Prodrug approach for alphaIIbbeta3-peptidomimetic antagonists to enhance their transport in monolayers of a human intestinal cell line (Caco-2): comparison of in vitro and in vivo data

Physiological models are good tools to predict rat bioavailability of EF5154 prodrugs from in vitro intestinal parameters

There are only a few reports on the methods that predict in vivo bioavailability from in vitro intestinal parameters. In the present study, we constructed physiological models where we examined if in vivo rat bioavailability was predictable from in vitro intestinal parameters using prodrugs of EF5154, a potent glycoprotein IIb/IIIa receptor antagonist, and other prodrugs. Apparent fraction absorbed (F(a),pred) was calculated using the physiological models that consist of absorption number calculated from Caco-2 cell membrane permeability (P(app)), and Damkohler number calculated from apparent degradation rate constant (K(dapp)) in rat small intestinal fluid. The predicted rat bioavailability that was calculated from F(a),pred corresponded to the observed rat bioavailability, and root mean square error (RMSE) and squared correlation coefficient (r(2)) were 4.58 and 0.904, respectively, suggesting that the physiological models consisting of the membrane permeability and degradation rate constant are good tools for predicting rat bioavailability of EF5154 prodrugs. As for other prodrugs where the chemical structure of their active forms differs from EF5154, the predicted rat bioavailability was not different from fraction absorbed (or rat bioavailability), suggesting the physiological models are generalized to various prodrugs that are not the substrates for active transporters.

Enhanced permeability of the antimicrobial agent 2,5-bis(4-amidinophenyl)furan across Caco-2 cell monolayers via its methylamidoidme prodrug

PURPOSE

DB75 [2.5-bis(4-amidinophenyl)furan] is a promising antimicrobial agent although it has poor oral potency. In contrast, its novel prodrug, 2,5-bis(4-amidinophenyl)furan-bis-O-methylamidoxime (DB289). has excellent oral potency. The mechanisms of transport of DB289 and DB75 across intestinal epithelium have been investigated in these studies to understand differences in their oral potency.

METHODS

Caco-2 cell monolayers were used as an in vitro model to examine the mechanisms of transport of DB289 and DB75. Samples collected from the transport studies were quantified using high-performance liquid chromatography with ultraviolet and fluorescence detection.

RESULTS

A low permeability coefficient (3.8 x 10(-7) cm/s for transport in apical [AP] to basolateral [BL] direction) and high sensitivity to extracellular Ca2+ suggest that AP to BL transport of DB75 across Caco-2 cell monolayers occurs predominantly via a paracellular route. DB289 has an 85-fold higher transport rate (322.0 x 10(-7) cm/s for transport in the AP to BL direction) across Caco-2 monolayers than that of DB75. This, with its insensitivity to extracellular Ca2+ indicates that AP to BL transport of DB289 across Caco-2 cell monolayers occurs predominantly via a transcellular route.

CONCLUSIONS

DB75 is transported across Caco-2 cell monolayers predominantly via paracellular pathways, whereas the prodrug DB289 is transported via transcellular pathways. This could account for the much higher oral activity of DB289 over DB75.

Comparison of human duodenum and Caco-2 gene expression profiles for 12,000 gene sequences tags and correlation with permeability of 26 drugs

PURPOSE

To compare gene expression profiles and drug permeability differences in Caco-2 cell culture and human duodenum.

METHODS

Gene expression profiles in Caco-2 cells and human duodenum were determined by GeneChip analysis. In vivo drug permeability measurements were obtained through single-pass intestinal perfusion in human subjects, and correlated with in vitro Caco-2 transport permeability.

RESULTS

GeneChip analysis determined that 37, 47, and 44 percent of the 12,559 gene sequences were expressed in 4-day andl6-day Caco-2 cells and human duodenum, respectively. Comparing human duodenum with Caco-2 cells, more than 1,000 sequences were determined to have at least a 5-fold difference in expression. There were 26, 38, and 44 percent of the 443 transporters, channels, and metabolizing enzymes detected in 4-day, 16-day Caco-2 cells, and human duodenum, respectively. More than 70 transporters and metabolizing enzymes exhibited at least a 3-fold difference. The overall coefficient of variability of the 10 human duodenal samples for all expressed sequences was 31% (range 3% to 294%) while that of the expressed transporters and metabolizing enzymes was 33% (range 3% to 87%). The in vivo / in vitro drug permeability measurements correlated well for passively absorbed drugs (R2 = 85%). The permeability correlation for carrier-mediated drugs showed 3- 35-fold higher in human above the correlation of passively absorbed drugs. The 2- 595-fold differences in gene expression levels between the Caco-2 cells and human duodenum correlated with the observed 3- 35-fold difference in permeability correlation between carrier-mediated drugs and passively absorbed drugs. CONCLUSIONS; Significant differences in gene expression levels in Caco-2 cells and human duodenum were observed. The observed differences of gene expression levels were consistent with observed differences in carrier mediated drug permeabilities. Gene expression profiling is a valuable new tool for investigating in vitro and in vivo permeability correlation.

Transport of peptidomimetic thrombin inhibitors with a 3-amidino-phenylalanine structure: permeability and efflux mechanism in monolayers of a human intestinal cell line (Caco-2)

PURPOSE

Peptidomimetic thrombin inhibitors derived from Nalpha-(2-naphthylsulfonyl)-3-amidino-phenylalanine with different basic and acidic substituents were investigated with respect to their intestinal transport behavior.

METHODS

Intestinal permeability coefficients were studied using Caco-2 monolayers and a reversed-phase HPLC method for quantitation.

RESULTS

Apparent permeability coefficients Papp of compounds with a free amidino group were in general low (<10 x 10(-8) cm/s) and independent of the structure of the amide part (C-terminus). Polarized efflux, however, was strongly affected by substituents in the amide moiety yielding the following efflux ratios (ER): methylpiperidide (1) (ER 45) > piperidine carboxylic acid methylester (ER 6-11) > piperidine carboxylic acids (ER 1.9-2.9) > piperazide (ER -0.17). Efflux of (1) was temperature-dependent, but independent of the enantiomeric configuration, accompanied by an increase in transepithelial electrical resistance (TEER), and could be reduced by P-gp inhibitors (PSC 833, Cremophor EL) but not by indomethacin. Replacement of the amidine group of (1) by aminomethyl-, amino-, and oxamidine- moieties drastically increased absorptive permeability (46-68 fold) with ER < 3.4. In contrast, the oxamidine with a C-terminal nipecotic acid residue (8) displayed also a temperature dependent efflux- without altering TEER (ER 22). This efflux was sensitive to PSC833/Cremophor EL and indomethacin.

CONCLUSIONS

Basic and acidic residues of amidino-phenylalanine-derived thrombin inhibitors mediate affinity to intestinal efflux pumps. presumably P-gp and MRP. P-gp mediated efflux was related to a net positive charge and accompanied by an increased TEER. Among the methylpiperide (1) promoieties studied the oxamidino group seems to be very promising in overcoming both transport and efflux problems frequently encountered with peptidomimetics containing amidines.