Effect of Vitrectomy on Aqueous VEGF Concentration and Pharmacokinetics of Bevacizumab in Macaque Monkeys

Abstract

Purpose.: To evaluate the effect of vitrectomy on the concentration of vascular endothelial growth factor (VEGF) and the pharmacokinetics of intravitreally injected bevacizumab in the aqueous humor in cynomolgus macaques.

Methods.: Pars plana lensectomy and a standard three-port vitrectomy were performed in one eye each of six macaques. After a minimal 12-week healing period, the vitrectomized eyes received an intravitreal injection of bevacizumab (1.25 mg/50 μL). Aqueous humor and venous blood samples were obtained from the macaques just before vitrectomy, just before injection of bevacizumab, on days 1, 3, and 7, and during weeks 2, 4, 6, and 8 after the injection. The bevacizumab and VEGF concentrations were measured by using enzyme-linked immunosorbent assay.

Results.: The VEGF concentrations in the aqueous humor ranged from 52.6 to 113.9 pg/mL (mean ± standard deviation [SD], 81.7 ± 27.0 pg/mL) before vitrectomy and 20.7 to 72.7 pg/mL (mean ± SD, 51.4 ± 20.5 pg/mL) 3 months after vitrectomy, a difference that reached significance (P = 0.03). The aqueous VEGF concentrations decreased to less than 9.0 pg/mL, the lower limit of detection, in all eyes between 1 and 7 days after injection of bevacizumab. The mean half-life of 1.25 mg intravitreally injected bevacizumab was 1.5 ± 0.6 days (range, 1.0–2.4 days) in the aqueous humor.

Conclusions.: The VEGF concentration in the aqueous humor decreased and the half-life of the intravitreally injected bevacizumab was shorter in vitrectomized eyes.

Introduction

Bevacizumab (Avastin; Genentech, South San Francisco, CA), a full-length humanized monoclonal antibody to all isoforms of vascular endothelial growth factor (VEGF), received US Food and Drug Administration approval for intravenous treatment of metastatic colorectal cancer in 2004. Intravitreal injection of bevacizumab has been used widely to treat various ocular diseases including age-related macular degeneration, retinal vein occlusion, and proliferative diabetic retinopathy. ^1–8 Although numerous reports on the efficacy of intravitreal injections of bevacizumab have been published, few studies have reported on the pharmacokinetics of the drug. We have previously reported that intravitreal injection of bevacizumab decreases the VEGF concentration in normal eyes for approximately 4 weeks but has no or a minimal effect on the untreated fellow eyes in macaques, and the half-life of 1.25 mg intravitreally injected bevacizumab is 2.8 ± 0.6 days in the aqueous humor. ⁹ To the best of our knowledge, no studies have compared the VEGF concentration between vitrectomized and nonvitrectomized eyes of cynomolgus macaques. To evaluate the effect of vitrectomy on the VEGF concentration and the pharmacokinetics of intravitreally injected bevacizumab in the aqueous humor in cynomolgus macaques, we measured the VEGF and bevacizumab concentrations over time in the aqueous humor of the injected and uninjected eyes before vitrectomy and before and after intravitreal injection of bevacizumab in vitrectomized eyes of cynomolgus macaques. 

Methods

All treatments were performed in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research, and the Animal Experimentation Committee of Shiga University of Medical Science approved the animal research protocol. Six male cynomolgus macaques, aged 7 to 10 years and weighing 4.5 to 8.5 kg, were anesthetized with 5 mg/kg intramuscular ketamine hydrochloride, 1 mg/kg intramuscular xylazine hydrochloride, and 1% isoflurane. The pupils were dilated with 1 drop of 10% phenylephrine hydrochloride ophthalmic solution. The surgical eye was irrigated with 0.02% chlorhexidine gluconate and anesthetized with 1 drop of 4% xylocaine for additional anesthesia. Aqueous humor samples (150 μL) were obtained from the macaques just before vitrectomy. After a limbal-based conjunctival incision was made, three 20-gauge ports were made on the sclera 3.5 mm from the limbus. A standard three-port vitrectomy and pars plana lensectomy was performed in one eye each of six macaques. We did not intend to create posterior vitreous detachment. The monkeys recovered from the postoperative inflammation for a minimum of 12 weeks before entering the pharmacokinetic study. No signs of infections or inflammation were found at 12 weeks after the vitrectomy, and no retinal detachments developed. 

Bevacizumab (1.25 mg/50 μL) was injected with a 29-gauge needle into the vitreous cavity of the vitrectomized eye of each macaque after the recovery period. The fellow eyes did not receive an intravitreal injection and served as controls. Both aqueous humor samples (150 μL) and venous blood samples (2 mL) were obtained from the macaques just before the injection and 1, 3, and 7 days and 2, 4, 6, and 8 weeks after the injection. The samples were stored in a freezer at −80°C until analysis. 

Measurement of VEGF

The VEGF concentrations in the aqueous humor were measured by using a Quantikine Human VEGF Immunoassay (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. The limit of the detectable VEGF concentration was 9.0 pg/mL. VEGF concentration less than the lower limit of detection was considered as 0 in the calculation for mean. 

Measurement of Bevacizumab

The bevacizumab concentration was measured by using a slightly modified enzyme-linked immunosorbent assay as previously described. ¹⁰ Ninety-six well plates were coated with recombinant human VEGF165 (R&D Systems) at a concentration of 1.0 μg/mL overnight at 4°C (100 μL/well). After washing three times with phosphate-buffered saline (PBS) containing 0.05% Tween-20, the wells were blocked with 3% bovine serum albumin (BSA)/PBS overnight at 4°C (200 μL/well). The wells then were washed five times with PBS containing 0.05% Tween-20 and stored dry at 4°C for later use. Aqueous humor or serum diluted in 0.1% BSA/PBS was added to the plates overnight at 4°C (50 μL/well). Bevacizumab was detected by horseradish peroxidase–goat anti-human IgG (H+L) conjugate (Invitrogen Corporation, Carlsbad, CA) with a concentration of 1 μg/mL after a 3-hour incubation period at room temperature. After five washes, color development was performed with 100 μL 3,3′,5,5′-tetramethyl benzidine substrate, and the reaction was stopped by adding 1 M hydrogen chloride (100 μL). The optical density was measured at 450 nm with correction at 570 nm. A standard curve was prepared with bevacizumab ranging from 7.8 to 1000 pg/mL. 

Statistical Analysis

All statistical analyses were carried out by using PASW Statistics 18 (SPSS Japan, Tokyo, Japan). 

Results

The VEGF concentrations in the aqueous humor of the study eyes ranged from 52.6 to 113.9 pg/mL (mean ± standard deviation [SD], 81.7 ± 27.0 pg/mL) before vitrectomy and from 20.7 to 72.7 pg/mL (mean ± SD, 51.4 ± 20.5 pg/mL) 3 months postoperatively (Figs. 1, 2). The aqueous VEGF concentrations decreased significantly (P < 0.03) in the six vitrectomized eyes. One day after the injection of bevacizumab, the VEGF concentrations in the aqueous humor decreased to less than 9.0 pg/mL, the lower limit of detection in all injected eyes. The concentration below the lower limit was maintained for 1 week in all eyes. At 2 weeks, the VEGF concentration remained below the lower limit of detection in two eyes and returned to a detectable level in four eyes. At 4 weeks, the VEGF concentrations in the aqueous humor of the injected eyes ranged from 17.0 to 93.7 pg/mL (mean, 40.4 ± 28.4 pg/mL). The VEGF concentrations in the aqueous humor of the fellow uninjected eyes ranged from 43.4 to 169.5 pg/mL (mean, 104.6 ± 57.3 pg/mL) before the intravitreal injection and did not change after the injection. 

Figure 1. 

 

Pre- and postoperative VEGF concentrations in the aqueous humor of vitrectomized eyes.

Figure 2. 

 

VEGF concentrations in the aqueous humor of vitrectomized eye at various time points after surgery. D, days; W, weeks.

Bevacizumab concentrations in the aqueous humor of the injected eyes peaked at 10.8 ± 6.4 μg/mL (10,800 ± 6400 ng/mL) the day after the injection and declined gradually (Fig. 3). Bevacizumab was not detected in the untreated eyes. The half-life of 1.25 mg intravitreally injected bevacizumab was 1.5 ± 0.6 days (range, 1.0–2.4 days) in the aqueous humor. 

Figure 3. 

 

Bevacizumab concentrations in the aqueous humor of vitrectomized eye and serum at various time points after surgery.

A maximal bevacizumab serum concentration of 42.2 ± 34.0 ng/mL was achieved 1 day after the injection and then declined gradually. However, the reduction rate was much less than that in the aqueous humor of the vitrectomized eyes and higher than that in the aqueous humor in the treated eyes at 4 weeks and thereafter. The bevacizumab concentration in the serum 8 weeks after the injection was 0.25 ± 0.16 ng/mL. The half-life of bevacizumab was 5.9 ± 2.0 days (range, 3.0–8.3 days) in the serum. No complications, such as uveitis or endophthalmitis, developed after bevacizumab was injected. 

Discussion

In the current study, the VEGF concentrations in the aqueous humor ranged from 52.6 to 113.9 pg/mL (mean ± SD, 81.7 ± 27.0 pg/mL) before the vitrectomy and 20.7 to 72.7 pg/mL (mean ± SD, 51.4 ± 20.5 pg/mL) 3 months after vitrectomy. The aqueous VEGF concentration decreased significantly (P < 0.03) in the six vitrectomized eyes. There are several possible explanations for the reduced VEGF concentration in the vitrectomized eyes. The VEGF concentration in the aqueous humor of normal eyes returned to a detectable level 4 weeks after the bevacizumab injection. ⁹ Therefore, removal of VEGF from the vitreous cannot explain the reduced VEGF concentration 3 months after vitrectomy. Another possible explanation is rapid clearance of VEGF from the vitreous. ^11–13 VEGF produced by retina may be more rapidly cleared from the vitrectomized eyes than from the normal eyes, as intravitreally injected drugs have been reported to be cleared more rapidly in vitrectomized eyes. The concentration of aqueous VEGF in the vitrectomized eyes did not decrease because of temporary removal of VEGF but because of the conditions in the vitreous. 

Pars plana vitrectomy (PPV) reduces the macular thickness of eyes with diabetic macular edema (DME). ^14–16 Yamamoto et al., ¹⁴ who reported the results after PPV was performed in 65 eyes of 63 patients with DME, have found that the postoperative foveal retinal thickness at the final visit is significantly thinner than the preoperative foveal retinal thickness. The authors have suggested that the high oxygen concentrations supplied by the ciliary body cause retinal vasoconstriction and reduce retinal thickness after vitrectomy. Yanyali et al., ¹⁵ who reported the results of PPV with removal of the internal limiting membrane (ILM) in 27 eyes of 27 patients with DME, have found that the mean foveal thickness decreases significantly. They have suggested that removing the ILM might have a beneficial effect in DME by removing tangential traction exerted by the ILM and residual cortical vitreous. Hartly et al., ¹⁶ also have reported that PPV with removal of the ILM results in decreased retinal thickness in patients with DME. In addition to those reports, a decrease in the VEGF concentration in the aqueous humor in vitrectomized eyes might cause a reduction of the foveal thickness in patients with DME. 

In the current study, the VEGF concentration in the aqueous humor of the vitrectomized eyes fell below the lower limit of detection after the bevacizumab injection and remained low for approximately 2 weeks. The VEGF concentration returned to a detectable level after 2 to 4 weeks in vitrectomized eyes. Faster clearance of the bevacizumab in the vitrectomized eye might result in a shorter period of drug effectiveness. 

The current study showed that the half-life of intravitreally injected bevacizumab was 1.5 ± 0.6 days in the vitrectomized eye. We previously have reported that the half-life of intravitreally injected bevacizumab in normal eyes of macaques is 2.8 ± 0.6 days (n = 3; range, 2.3–3.5 days). The half-life of the intravitreally injected bevacizumab in the vitrectomized eyes decreased by 54% compared with the nonvitrectomized normal eyes. Recently, intravitreal bevacizumab injections are being administered more and more often to treat some ophthalmic diseases. However, the precise difference in the intravitreal bevacizumab concentrations between vitrectomized and nonvitrectomized eyes is unknown. Some intravitreal drugs clear more rapidly in vitrectomized eyes than in nonvitrectomized eyes. ^11–13 Doft et al. ¹¹ have studied ocular clearance after 10-μg intravitreal injections of amphotericin B in vitrectomized and nonvitrectomized rabbit models and report half-lives of 1.4 days and 9.1 days, respectively. Chin et al. ¹² have studied the difference in the clearance of 0.3-mg intravitreal injections of triamcinolone acetonide between vitrectomized and nonvitrectomized rabbit models and report half-lives of 1.57 days and 2.89 days, respectively. Lee et al. ¹³ also have studied the differences in clearance of 500 ng intravitreal human VEGF₁₆₅ between vitrectomized and nonvitrectomized rabbit models and report the half-lives of 12.5 minutes and 2.46 hours, respectively. The duration of the decreased concentration of VEGF after intravitreal injection of bevacizumab was shorter in vitrectomized eyes than in nonvitrectomized eyes probably because of the shorter half-life of bevacizumab. Faster clearance of intravitreal bevacizumab should be considered when planning intravitreal injections of bevacizumab in vitrectomized eyes. 

In conclusion, the aqueous VEGF concentration decreased in vitrectomized eyes, and the half-life of intravitreal injections of bevacizumab was shorter in vitrectomized eye than in nonvitrectomized eyes. Careful attention should be paid to the duration of the effectiveness of intravitreal injections of bevacizumab in vitrectomized eyes. 

Footnotes

 Supported in part by a grant from the Ministry of Education, Culture, Sports, Science and Technology of Japan (18591915) and a grant from the Ministry of Health, Labour and Welfare. The authors have no proprietary interest in any aspect of this report.

Footnotes

 Disclosure: M. Kakinoki, None; O. Sawada, None; T. Sawada, None; Y. Saishin, None; H. Kawamura, None; M. Ohji, None