Mechanistic Monte Carlo assessment of colocalized permeation enhancement and efflux modulation for oral peptide delivery

The oral absolute bioavailability of 2- to 4-kDa peptides remains near 1%. We constructed a mechanistic Monte Carlo model of a colocalized, reversible permeation-enhancer and efflux-modulator architecture (Dual-Lock) with explicit mass balances and evidence-graded, correlated priors and asked when it would increase bioavailability to at least 4% with predicted reversible epithelial exposure. The arms shared random draws so that the ablations and negative controls isolated each component. Global sensitivity, benefit-risk and falsification analyses, spatial and uptake-constrained transport models, reversibility-endpoint scenarios, observation-conditioned analysis and an out-of-sample evaluation of an optimized design tested robustness. With the reference calibrated to 1.2%, Dual-Lock reached a median of 5.0% at half the enhancer dose and met the target in 45% of draws against 15% for the reference. Confinement caused most of the benefit: a matrix-only control gained 8-fold over peptide alone, from thinning the niche (3.3-fold) and the protease shield, and the gain survived stagnant-layer and solubility limits. In the uptake-constrained model success was 54% (reference 9%); it was 21% to 34% if the reference performs like oral semaglutide, and the upper tail stayed beyond published values. The advantage survived damage at the efficacious concentration but not below it. A credible release-rate optimum added 8 to 12 points out of sample. The transport structure is not independently validated; failure boundaries define the first experiments. Results are conditional on author-assigned priors.