Modified release (MR) formulations are used to enhance the safety and compliance of existing drugs by improving their pharmacokinetics. Predicting the likely success of MR formulations is often difficult before clinical studies. A systematic in vitro approach using mouse and human tissues was adopted to rationalize the in vivo pharmacokinetics of 9- and 15-h MR formulations of an α-adrenoreceptor antagonist, 4-amino-6,7-dimethoxy-2-(5- methanesulfonamido-1,2,3,4 tetrahydroisoquinol-2-yl)-5-(2-pyridyl)quinazoline (UK-338,003). Immediate release UK-338,003 was well absorbed in humans consistent with moderate Caco-2 cell monolayer permeability. In contrast, 9- and 15-h modified release formulations showed marked reductions in Cmax (47.1 and 68.9%) and AUC0-72 (32.6 and 54.0%). Colonic intubation resulted in 81.3 and 73.8% reductions in Cmax and AUC0-72. Mechanistic studies in isolated mouse tissues showed that colonic UK-338,003 permeability (Papp < 0.5 × 10-6 cm/s) was at least 40 times lower than that for ileum with marked asymmetry. UK-338,003 was found to be a substrate for P-glycoprotein (PGP) with a weaker interaction for multidrug resistance-associated protein-type transporters in mouse intestine. PGP inhibition dramatically increased colonic UK-338,003 permeability to the levels observed in ileum. Low UK-338,003 apical to basolateral permeability was also observed in ex vivo human distal intestine, but both the asymmetry and increase in permeability after PGP inhibition were significantly lower. In conclusion, the poor absorption of MR UK-338,003 in humans can be explained by a combination of PGP-dependent efflux and low intrinsic permeability in the lower bowel. Regional permeability studies in ex vivo tissues used during drug development can highlight absorption problems in the distal bowel and assess the feasibility of developing successful MR formulations.