In vitro and in vivo evaluation of an anti-VEGF controlled release drug delivery system

Abstract

Purpose: Current therapies for posterior segment diseases often require monthly bolus injections of anti-vascular endothelial growth factors (anti-VEGFs). The purpose of this study was to validate a drug delivery system (DDS) capable of bioactive anti-VEGF release for over 6 months.

Methods: The DDS consists of poly(lactic-co-glycolic acid) microspheres suspended in a thermoresponsive poly(N-isopropylacrylamide)-based hydrogel. Microspheres were loaded with ranibizumab or aflibercept. The MTS assay was used to determine bioactivity of release samples on human umbilical vascular endothelial cells (HUVECs) under VEGF-induced proliferation. A rat model of choroidal neovascularization (CNV) was used whereby CNV was induced in male Long-Evans rats using an Ar-green laser. At 1, 2 and 4wk post-CNV induction/treatment, lesion areas were measured using a multi-Otsu threshold technique. Dark-adapted electroretinogram (ERG) intensity series were elicited with full-field Ganzfeld stimulation (max flash intensity: 307 scotopic cd×s×m^-2; duration: 2ms). Rats treated with our DDS were compared to bolus anti-VEGF injection counterparts, non-treated and negative control (DDS without anti-VEGF) rats.

Results: Our DDS is capable of releasing ranibizumab and aflibercept for ~200 days with an initial burst (first 24hr) of 21.0±2.0% and 20.1±0.8%, respectively, followed by controlled release of 0.2μg/day and 0.07μg/day, respectively. No toxicity was seen in HUVECs at any time for any treatment group; negative control showed no inhibitory effect on HUVEC proliferation. Both anti-VEGFs remained bioactive throughout release. In vivo results confirmed our DDS is non-toxic with no significant changes in ERG maximal response (R_max) or half-saturation (σ) compared to control measurements (A_max: -472±20μV; B_max: 770±42μV; σ_a: 2.8±0.7cd×s×m^-2; σ_b: 0.003±0.0005cd×s×m^-2). Negative control lesion areas (0.05±0.001mm²) were not significantly different than non-treated at any time point (0.05±0.002mm²). At 4wk, rats treated with our DDS had CNV lesion areas that were 37% and 32% smaller than their bolus-injection counterparts for ranibizumab and aflibercept, respectively (p<0.05).

Conclusions: The results indicate our DDS can deliver bioactive anti-VEGFs with significant impacts in both in vitro and in vivo models. Our DDS may provide a significant advantage over current bolus injection therapies in the treatment of posterior segment diseases.