Single periocular injection of celecoxib-PLGA microparticles inhibits diabetes-induced elevations in retinal PGE2, VEGF, and vascular leakage

Abstract

Purpose: To determine whether celecoxib inhibits VEGF secretion from ARPE-19 cells and to investigate further the safety and effectiveness of periocular celecoxib-poly (lactide-co-glycolide; PLGA) microparticles in inhibiting elevations in retinal PGE(2), VEGF, and blood-tissue barrier leakage at the end of 60 days in a streptozotocin diabetic rat model.

Methods: VEGF mRNA and protein expression in ARPE-19 cells was evaluated in the presence of 0 to 10 microM celecoxib, and cytotoxicity of celecoxib on ARPE-19 and RF6A cells was evaluated over a 0- to 100-microM concentration range. Celecoxib-PLGA microparticles were prepared by a modified solvent evaporation technique, sterilized by 25 kGy of gamma-irradiation, and characterized for size, zeta potential, drug loading, and in vitro release. Normal and streptozotocin-diabetic male Sprague-Dawley rats were divided into five groups: normal, diabetic, diabetic+placebo, normal+celecoxib, and diabetic+celecoxib. Phosphate-buffered saline (PBS) containing celecoxib-PLGA microparticles, placebo PLGA microparticles, or plain PBS in one eye was injected into the posterior subconjunctival (periocular) space in rats under anesthesia. Sixty days after administration, the animals were killed, and retinal PGE2 secretion, VEGF protein, and blood-retinal barrier leakage were estimated. Blood cell counts, blood chemistry and histology were used to assess the safety of the microparticulate system.

Results: Celecoxib (up to 25 microM) did not cause significant cytotoxicity in ARPE-19 or RF6A cells. Nanomolar concentrations of celecoxib reduced VEGF mRNA and VEGF protein secretion. Celecoxib-PLGA microparticles (diameter: 1140 +/- 15 nm), containing 14.93% +/- 0.21% of celecoxib sustained in vitro drug release and in vivo drug levels in the retina for 60 days. Diabetes elevated PGE2 secretion, VEGF protein, the vitreous-plasma protein ratio, and blood-retinal barrier leakage by 3-, 1.7-, 3.1-, and 2.7-fold, and celecoxib-PLGA microparticles significantly reduced these elevations by 40%, 50%, 40%, and 50%, respectively. Neither the placebo-treated eyes nor the contralateral eyes in celecoxib-PLGA microparticle-treated rats showed significant effects. Celecoxib-PLGA or placebo-PLGA particles had no effect on the body weight or blood sugar level of rats. The celecoxib-PLGA microparticles did not cause any changes in blood cell counts or chemistry and caused no histopathological damage to the retina or periocular tissues.

Conclusions: Nanomolar concentrations of celecoxib can inhibit VEGF mRNA and protein expression from ARPE-19 cells. Periocular celecoxib microparticles are useful sustained drug delivery systems for inhibiting diabetes-induced elevations in PGE2, VEGF, and blood-retinal barrier leakage. The periocular celecoxib-PLGA microparticles are safe and do not cause any damage to the retina.

Figure 1

Celecoxib is not toxic to ARPE-19 and RF6A cells up to a concentration of 25 μM. After a 12-hour quiescence period, ARPE-19 cells and RF6A cells were exposed to celecoxib ranging in concentration from 1 nM to 100 μM. Data are expressed as the mean ± SD (n = 8). *Significantly different from the control.

Figure 2

Celecoxib inhibits VEGF mRNA expression in ARPE-19 cells. After a 12-hour quiescence period, ARPE-19 cells were exposed to celecoxib ranging in concentration from 1 nM to 10 μM. (A) Agarose gel electrophoresis of RT-PCR products for VEGF₁₆₅, VEGF₁₂₁, and 18S rRNA. (B) The densitometric band intensities normalized to 18S rRNA. Data are expressed as the mean ± SD of results in three experiments. Significantly different from the control for *VEGF₁₆₅ or †VEGF₁₂₁. (C) Real-time quantitative RT-PCR for VEGF₁₆₅. Data are expressed as the mean ± SD of results in four experiments. ‡Significantly different from the control (P < 0.05).

Figure 3

Celecoxib inhibits VEGF secretion from ARPE-19 cells. After a 12-hour quiescence period, ARPE-19 cells were exposed to celecoxib ranging in concentration from 100 pM to 10 μM. Secreted VEGF was measured at the end of 12 hours of treatment. Data are expressed as the mean ± SD of results in eight experiments. *Significantly different from the control.

Figure 4

In vitro release of celecoxib from drug suspension and sterilized or unsterilized celecoxib-PLGA (85/15) microparticles. Particles equivalent to 20 μg of celecoxib were suspended in a dialysis bag and immersed in 50 mL of PBS. There was no statistically significant difference in celecoxib release at any time point between the sterilized and unsterilized particles. Inset: release of celecoxib from celecoxib suspension (20 μg). Data are expressed as the mean ± SD of three experiments.

Figure 5

Inhibitory effects of celecoxib-PLGA microparticles on diabetes induced elevations in (A) retinal PGE₂ (n = 4); (B) retinal VEGF (n = 8–9); (C) vitreous-plasma protein ratio (n = 4); and (D) blood–retinal barrier leakage (n = 4). The parameters were estimated 60 days after periocular administration of the placebo or celecoxib-PLGA microparticles to rats. Data are expressed as the mean ± SD. Significantly different from the *diabetic group, the †diabetic+placebo group, or the ‡contralateral eye.

Figure 6

Celecoxib-PLGA microparticles are well tolerated and do not have any significant adverse effects in normal rats. Representative histologic sections depicting the health of the retina and the periocular tissues in rats injected with periocular celecoxib-PLGA microparticles. Top: normal rats injected with PBS; bottom: normal rats injected with celecoxib PLGA microparticles. (A, D) Retinal sections stained with hematoxylin and eosin (H&E). (B, E) Periocular tissue (site of administration) stained with H&E. (C, F) Periocular tissue stained with Masson trichrome stain. Magnification: (A, D) ×200; (B, C, E, F) ×100.

Figure 7

Morphometric analysis of the retina in normal rats. The thickness of the different retinal layers was compared between control and celecoxib PLGA microparticle-treated groups. Data are expressed as the mean ± SD of results in four experiments. Inset: representative photomicrographs of the retina showing the different retinal layers. The thickness of the neural retina (double-headed arrow), outer nuclear layer (ONL), and inner nuclear layer (INL) was compared between the two groups. No statistically significant differences were found in the thickness of the layers measured between the two groups. GCL, ganglion cell layer.

Figure 8

Probable pathway for the reduction of diabetes-induced elevation in retinal PGE₂, VEGF, and vascular leakage by periocular celecoxib-PLGA microparticles.