New in vitro prototypes (PK-Eye™) were tested with and without eye movement to understand diffusion and convection effects on intraocular clearance. Port placement in front ((i) ciliary inflow model) and behind the model lens ((ii) posterior inflow model) was used to study bevacizumab (1.25 mg/50 µL) and dexamethasone (0.1 mg/100 µL) in phosphate-buffered saline (PBS, pH 7.4) and simulated vitreal fluid (SVF). Dexamethasone was studied in a (iii) retinal-choroid-sclera (RCS) outflow model (with ciliary inflow and two outflow pathways). Ciliary vs. posterior inflow placement did not affect the half-life for dexamethasone at 2.0 µL/min using PBS (4.7 days vs. 4.8 days) and SVF (4.9 days with ciliary inflow), but it did decrease the half-life for bevacizumab in PBS (20.4 days vs. 2.4 days) and SVF (19.2 days vs. 10.8 days). Eye movement only affected the half-life of dexamethasone in both media. Dexamethasone in the RCS model showed approximately 20% and 75% clearance from the RCS and anterior outflows, respectively. The half-life of the protein was comparable to human data in the posterior inflow model. Shorter half-life values for a protein in a ciliary inflow model can be achieved with other eye movements. The RCS flow model with eye movement was comparable to human half-life data for dexamethasone.
Real-time measurement is important in modern dissolution testing to aid in parallel drug characterisation and quality control (QC). The development of a real-time monitoring platform (microfluidic system, a novel eye movement platform with temperature sensors and accelerometers and a concentration probe setup) in conjunction with an in vitro model of the human eye (PK-Eye™) is reported. The importance of surface membrane permeability when modelling the PK-Eye™ was determined with a “pursing model” (a simplified setup of the hyaloid membrane). Parallel microfluidic control of PK-Eye™ models from a single source of pressure was performed with a ratio of 1:6 (pressure source:models) demonstrating scalability and reproducibility of pressure-flow data. Pore size and exposed surface area helped obtain a physiological range of intraocular pressure (IOP) within the models, demonstrating the need to reproduce in vitro dimensions as closely as possible to the real eye. Variation of aqueous humour flow rate throughout the day was demonstrated with a developed circadian rhythm program. Capabilities of different eye movements were programmed and achieved with an in-house eye movement platform. A concentration probe recorded the real-time concentration monitoring of injected albumin-conjugated Alexa Fluor 488 (Alexa albumin), which displayed constant release profiles. These results demonstrate the possibility of real-time monitoring of a pharmaceutical model for preclinical testing of ocular formulations.