Abstract
Purpose :
Development of drug-releasing silicone oil (SIO) tamponades requires the use of laboratory models, but there is no gold standard. Each model has different advantages, for example studies in in vitro models are easier to perform and quantify, whereas ex vivo globe models represent the geometry of the eye more realistically. Understanding drug release in each model used is very important. We conducted an experimental study to compare the release of a model drug, ibuprofen (Ibu), from SIO in a series of static in vitro and ex vivo models.
Methods :
1 mg/mL Ibu in 1000 cSt SIO (Ibu-SIO) and phosphate buffered saline (PBS) were placed in models at 37°C. In multi-well plates, 1 mL Ibu-SIO was placed on 0.5 mL PBS. Inner walls of spherical [Wetterqvist, Br J Ophthalmol 2004, 88(5)] and eye-on-a-chip (EOC) models [Chan IOVS 2015, 56] were pre-coated with 1% bovine serum albumin (BSA) solution. 3.5 ml Ibu-SIO and 1.5 mL PBS was loaded into the spherical model. 0.6 ml Ibu-SIO and 0.18 mL PBS was loaded into the EOC model. PBS was samples at 24, 48 or 72h. Ibu concentration was measured using UV-vis (λ= 264nm). Freshly enucleated pig eyes were vitrectomised and filled with 2 mL tritiated Ibu-SIO. Eyes were placed in sealed containers containing 20mL PBS. At 72h, oil was removed and the eye dissected. Radiation from different components was measured on a scintillation counter to determine Ibu concentration; this sensititve measurement technique allowed Ibu quantification from tissue.
Results :
Ibu release at 72h in in vitro models varied from 17%-84%. In multiwell plates and the spherical model a burst release was observed, whereas in the EOC the release was linear (R2=0.84). Increasing SIO surface area:volume ratio lead to an increase in drug release that appeared to have a logarithmic relationship, except for the spherical model. Ibu release in 24-well plates was comparable to that in the spherical model. In the ex vivo model, 97.8% Ibu was released at 72h. The majority (~75%) was in the PBS outside the eye.
Conclusions :
Drug release in different models varied, even when normalised for oil surface area:volume and oil:PBS ratios. The ex vivo model has appropriate geometry, but doesn’t replicate the biological processes that affect drug transport within the eye. All the models have some scientific value, but caution must be used when comparing drug release measured in these different, yet complementary models.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.