Effect of Aflibercept on Human Corneal Endothelial Cells in Neovascular Age-Related Macular Degeneration

Intravitreal injections of vascular endothelial growth factor (VEGF) inhibitors have being increasingly used in the treatment of neovascular age-related macular degeneration (AMD) in ophthalmic practice.1-4 The most commonly used VEGF inhibitors are bevacizumab (Avastin®, Genentech, San Francisco, California, USA), ranibizumab (Lucentis®, Genentech, San Francisco, California, USA) and aflibercept (Eylea®, Regeneron Pharmaceuticals Inc., Tarrytown, New York, USA) among which aflibercept and ranibizumab were approved by the Food and Drug Administration (FDA) for this indication.

Aflibercept is the most recently developed VEGF inhibitor with a recombinant fusion protein consisting of human VEGF receptor extracellular domains from receptors 1 and 2 (VEGFR1 and VEGFR2) fused to the Fc domain of human IgG. This protein contains all human amino acid sequences, which minimizes the potential for immunogenicity in patients. The prolonged intravitreal half-life of aflibercept compared with ranibizumab can translate to a lower treatment load in terms of injections, monitoring, and medical visits. Several in vitro studies have shown that aflibercept, at the concentration usually used for treating retinal disorders had no toxicity to the ocular cells. However, although both ranibizumab and bevacizumab have been shown not to have harmful effects on corneal endothelium the effect of intravitreal aflibercept on human corneal endothelium has not been reported so far. Considering the functional importance of the corneal endothelium, particularly in aged population, the present study was designed to evaluate the in vivo toxicity of aflibercept on human corneal endothelial cells in patients with neovascular AMD.

Thirty-four eyes of 34 consecutive patients with neovascular AMD (19 male, 15 female; mean age 66.4±3.4 years; age range 57-76 years) were recruited for this observational prospective study. The study protocol was approved by the ethics committee and adhered to the tenets of the Declaration of Helsinki. All participants signed the informed consent before any study-related procedure.

All participants received one monthly intravitreal injection of aflibercept for three consecutive months, and later treatments were applied as needed. The follow-up period was six months.

The procedure for the intravitreal aflibercept injection was performed using standard aseptic techniques. After providing local anesthesia with proparacaine hydrochloride eye drops (Alcaine, Alcon Laboratories Inc, Fort Worth, Texas, USA), the eyelids and the inferior conjunctival fornix were sterilized with 5% povidone iodine. Aflibercept (2.0 mg, 0.05 ml) was injected through the pars plana (4 mm behind the limbus) using a 27-gauge needle.

Noncontact specular microscopy (Tomey EM-3000 Specular Microscope, Tomey Corp, Japan) was performed on the central cornea before the first intravitreal aflibercept injection and 1, 3, 6 months after the injection. A blinded observer (S.D.) obtained the corneal endothelial images. The specular microscope automatically evaluated the endothelial cell density (ECD), the coefficient of variation of the cell size (CoV), an objective measure of polymegathism, and the percentage of the hexagonal cells (Hex%), an index of pleomorphism. Specular microscopy also provided optical pachymetry measurements. The acute toxic of aflibercept on endothelium was evaluated by the the presence of corneal edema and anterior chamber reaction, and intraocular pressure on the first day after the injection. The nontreated eyes served as the control group.

Intravitreal injection of aflibercept did not have any negative effect in human corneal endothelial cells.The signs of acute toxic effect of aflibercept on endothelium, which are corneal edema, anterior chamber reaction, or high intraocular pressure, were not recorded in any of the eyes on the first day after the injection.

Aflibercept is a recently approved anti-VEGF offering a new therapy for treatment of neovascular AMD. It is a fusion protein of VEGF receptors 1 and 2. It has higher affinity to VEGF compared with ranibizumab or bevacizumab which indicates longer duration of action for aflibercept. In experimental and clinical trials, the efficacy of aflibercept in treatment of AMD has been shown to be comparable to that of ranibizumab and bevacizumab. In studies comparing aflibercept with bevacizumab or ranibizumab, aflibercept had fewer negative effects on retinal cell lines such as change in cell morphology, apoptosis or permanent decrease in cell viability, cell density or proliferation. Aflibercept has greater binding affinity for VEGF than ranibizumab and bevacizumab and requires less-frequent intravitreal injection than ranibizumab and bevacizumab. The vitreous half-life of aflibercept is shorter than bevacizumab, but longer than ranibizumab. In retina pigment epithelium and retina cultures prepared from pigs' eyes, aflibercept completely inhibited VEGF for six hours at a minimal concentration, and displayed a prolonged VEGF inhibition compared to bevacizumab and ranibizumab. However, this advantage of aflibercept raises concerns about possible side effects of long-term usage.

The corneal endothelium is a barrier to fluid flow from the aqueous humor to the stroma. It is responsible for maintaining corneal transparency by regulating stromal hydration. The endothelial cell density decreases with age, and further damage to corneal endothelium by disease, trauma or drugs, may lead to its loss of function, which cause corneal edema, decreased corneal clarity, and loss of visual acuity. Therefore, keeping corneal endothelium healthy has a vital importance particularly for aged patients. Intravitreal injection causes significant concentration of VEGF inhibitors get into contact with human corneal endothelial cells. Previous studies indicated that VEGF and its receptors are expressed in the corneal endothelium and epithelium, and VEGF inhibitors can be detected in aqueous humour after intravitreal injection, both of which showing that VEGF inhibitors have potential to be cytotoxic to human corneal endothelial cells.The effect of the intravitreal injection of ranibizumab or bevacizumab on corneal endothelium has been studied before. Perez-Rico et al. reported that intravitreal injections of ranibizumab did not have significant effect on endothelial cell density, CoV, and Hex% at seven days or six months after the injection. The repeated intravitreal injections of ranibizumab over six months did not also cause substantial changes in the corneal endothelium. Chiang et al. evaluated human corneal thickness change and corneal endothelial cell density up to six months after intravitreal injection of 2.5 mg bevacizumab and found that intravitreal bevacizumab has no harmful effects on the corneal endothelium. In vitro studies also showed that bevacizumab is not toxic to corneal cells of human origin. Furthermore, although intracameral injection of ranibizumab caused deterioration in endothelial cell morphology in rabbit cornea, intracameral injection of bevacizumab did not affect endothelial cell viability or morphology in the rabbit or human cornea. Bevacizumab did not induce apoptosis or necrosis in human corneal endothelial and fibroblast cells in vitro. In a recent study, it was reported that monthly intravitreal bevacizumab or ranibizumab injections for three consecutive months does not affect corneal morphology and has no harmful effects on the endothelium in patients with diabetic macular edema. The findings of this present study also agree with these studies showing that neither ranibizumab nor bevacizumab has negative effect on corneal endothelial cells. The investigators found that the CCT, ECD, CoV, and Hex% did not change over six months after intravitreal injection of aflibercept in either eyes. Furthermore, the CCT, ECD, CoV, and Hex% did not show statistically significant difference between the eye treated with intravitreal aflibercept and contralateral untreated eye before and during 6-month follow-up after intravitreal injection.

The main limitation of the study was the limited sample size, which precludes us reaching a definitive conclusion on the effect of intravitreal aflibercept on human corneal endothelium. Nevertheless, this pilot study provides first evidence that intravitreal aflibercept injections do not have negative effect on human corneal endothelium.

In conclusion, intravitreal injection of clinicaly effective doses of aflibercept for four times on average during the 6-month period do not induce any harmful effect on human corneal endothelium evaluated by specular microscopy. Further prospective, large-scale, prolonged studies are needed to confirm that intravitreal aflibercept can be used safely without any corneal toxicity to treat neovascular AMD.