Computer modeling of bevacizumab drug distribution after intravitreal injection in rabbit and human eyes

Abstract

Age-related macular degeneration (AMD) is a progressive eye disease that causes loss of central vision and has no cure. Wet AMD is the late neovascular form treated with vascular endothelial growth factor (VEGF) inhibitors. VEGF is the critical driver of wet AMD. One common off-label anti-VEGF drug used in AMD treatment is bevacizumab. Experimental efforts have been made to investigate the pharmacokinetic (PK) behavior of bevacizumab in vitreous and aqueous humor. Still, the quantitative effect of elimination routes and drug concentration in the macula are not well understood. In this work, we developed two spatial models representing rabbit and human vitreous to better understand the PK behavior of bevacizumab. This study explores different cases of drug elimination and the effects of injection location on drug concentration profiles. The models are validated by comparing them with experimental data. Our results suggest that anterior elimination is dominant for bevacizumab clearance from rabbit vitreous, whereas both anterior and posterior elimination have similar importance in drug clearance from the human vitreous. Furthermore, results indicate that drug injections closer to the posterior segment of the vitreous help maintain relevant drug concentrations for longer, improving bevacizumab duration of action in the vitreous. The rabbit and human models predict bevacizumab concentration in the vitreous and fovea, enhancing knowledge and understanding of wet AMD treatment.

Keywords: Antibody drug; Distribution; Drug transport; Dynamic simulation; Elimination; In silico modeling; Mathematical model; Mechanistic modeling; Pharmacokinetics.

Fig. 1.

3D geometry of the human eye: a) anatomical region description and b) boundary conditions. The rabbit geometry considers the same anatomical regions and boundary conditions scaled to rabbit dimensions. Dimensions for both the rabbit and human geometries are in Table 1.

Fig. 2.

3D models for vitreous of a) rabbit and b) human. Drug intravitreal injection locations are labeled as I, II, and III for anterior, middle, and posterior locations, respectively.

Fig. 3.

Normalized vitreous concentration in the rabbit and human eye. Injection locations: a) rabbit anterior vitreous, b) rabbit middle vitreous, c) rabbit posterior vitreous, d) human anterior vitreous, e) human middle vitreous, and f) human posterior vitreous. Case 1a: slow convective flow with anterior elimination only, Case 1b: slow convective flow with both anterior and posterior elimination, Case 2a: no convective flow with anterior elimination, and Case 2b: no convective flow with both anterior and posterior elimination. Model results for each case are shown as curves. Experimental data (markers) from Bakri et al., Nomoto et al., Sinapis et al., Ahn et al., Ye et al., Beer et al., and Zhu et al.

Fig. 4.

Drug concentration at the rabbit “fovea” and human fovea. Injection locations: a) rabbit anterior vitreous, b) rabbit middle vitreous, c) rabbit posterior vitreous, d) human anterior vitreous, e) human middle vitreous, and f) human posterior vitreous. Case 1a: slow convective flow with anterior elimination only, Case 1b: slow convective flow with both anterior and posterior elimination, Case 2a: no convective flow with anterior elimination, and Case 2b: no convective flow with both anterior and posterior elimination. C_f: in vitro threshold concentration at the fovea.