Optimization of laser-induced choroidal neovascularization in African green monkeys.

We developed and validated a new nonhuman primate model of laser-induced choroidal neovascularization (CNV) that addresses study design limitations prevalent in laser-induced CNV-based efficacy studies. Laser-induced Bruch's membrane disruption triggers CNV and has been widely utilized in animals to model neovascular ("wet") age-related macular degeneration (AMD). Despite widespread use of the approach, detailed assessment of experimental parameters and their influence on pathophysiological endpoints critical for disease modeling has been extremely limited and largely based on anecdotal observations. We evaluated laser power parameters and endpoint measures to optimize methods for CNV formation and quantification to facilitate drug efficacy screening in African green monkeys. Six laser spots of 350, 550, 750, 950 or 1500 mW laser power were positioned bilaterally 1.5 disc diameters from the fovea, within the macula. Fluorescein angiograms were collected 3-5 weeks later and scored by trained masked investigators using graded (I-IV) and densitometric methods. Histopathology assessments were also performed, including determination of CNV area. Test system sensitivity to angiogenesis inhibition was subsequently assessed by evaluating the effect of intravitreal bevacizumab (Avastin) pretreatment (one day prior to laser photocoagulation) on incidence of CNV. Grade III and grade IV lesions were considered clinically relevant, demonstrating early hyperfluorescence and late leakage within or beyond the lesion borders. By 4 weeks post-laser all treatment groups demonstrated evidence of grade III lesions with greatest incidence observed in lesions induced by 750 and 950 mW laser power (72.9% and 69.4% respectively). Grade IV lesions were confined to eyes receiving 550 mW laser power or higher, with highest incidence of grade IV lesions observed in eyes receiving 950 (19.4%) and 1500 mW (31%) laser spots, incidence peaking 4 weeks post-laser photocoagulation. Densitometric analyses of angiograms corroborated visual scoring. Bevacizumab completely abolished grade IV lesion development and significantly lowered lesion fluorescein signal intensity (P < 0.0001) and CNV area (P = 0.038) compared to vehicle-treated controls. Our studies demonstrate that laser power of 950-1500 mW and angiography analysis 4 weeks post-laser are optimal parameters to evaluate treatment effects on CNV induction following laser photocoagulation. Bevacizumab significantly attenuated CNV development, as determined by fluorescein angiography and histopathology assessments in this model, supporting the application of African green monkeys in preclinical modeling of CNV. Laser parameters and time points for therapeutic dosing and angiography endpoints are critical factors to the laser-induced CNV model and must be validated for robust assessment of efficacy. The newly optimized nonhuman primate model described will facilitate preclinical efficacy assessments of novel therapeutics for CNV.