Mechanistic PBPK Modelling for the Locally Acting GIT Drug: Mesalamine

PBPK modelling for locally acting GI drugs has the potential to predict drug concentrations at the site of action for which direct measurements are difficult to perform routinely. Suitably validated PBPK models may assist with optimization of formulations, help to understand local PK/PD relationships, and support development and regulatory decisions within a totality of evidence framework. Such models can also enable comparisons between formulations based on local rather than plasma concentrations (e.g., within a virtual BE framework). Mesalamine is a primary treatment for mild to moderate ulcerative colitis and acts locally in the GI mucosa. Various modified/delayed release formulations target the colon but have low bioavailability (15-30%) with large population variability (PopVar) making development of robust PBPK models for mesalamine challenging. The purpose of this work was to develop and validate PBPK models for mesalamine drug products using the Simcyp population-based PBPK simulator (V24, Certara). A full PBPK model for i.v. bolus dosing in healthy volunteers was developed using literature data to establish disposition parameters. Whole organ metabolic clearances for liver (and subsequently also the gut wall) were optimized to reflect extensive metabolism through N-acetylation. This model was validated against other i.v. bolus and infusion clinical studies. For oral dosing gut lumen degradation of mesalamine in the colon was considered, pH-dependent solubility was collected from various literature, gut regional permeability (Peff) was predicted using the bottom-up Mechanistic Permeability (MechPeff) model without adjustment. The Segregated Transit Time model was selected for orally administered formulations. Dissolution in pH 6.0, 6.8, 7.5 Phosphate Buffer (PB) and FaSSCoF were generated at the University of Florida. In the absence of mechanistic models of dissolution of these formulations in vitro dissolution profiles were directly input to the PBPK model using either the Release or Dissolution profile options for controlled/modified release formulations: (a) Pentasa; (b) Apriso (Dispersible), and (c) Lialda. The release mechanisms and recommended doses for these formulations differ. Apriso and Lialda are enteric-coated thus a “trigger” pH for release was applied which given the known PopVar of lumen pH is a source of PopVar in release. Virtual trial designs matching literature studies were used with ≥10 replicates to consider study power. Mean mesalamine plasma PK profiles were well captured for i.v. bolus, infusion and po dosing. As the aim of formulation development for local GI delivery is to maximise site of action exposure, standard plasma-based assessment may not provide sufficient information about local disposition and thus efficacy which is where PBPK modelling can add insights. It is known that mucosal mesalamine concentration is inversely correlated with disease activity. Thus, after model validation based on systemic concentrations, the model predictions for GIT luminal fluid and colon enterocyte concentrations were compared to available literature values. Reported local concentrations have high variability - between study (linked to methodology), between subject, temporal - but, overall, the simulated concentrations are reasonable subject to caveats which will be presented. The developed mesalamine PBPK model can be considered suitable for applications such as VBE and safespace analysis.