Intraocular viral infection has become an increasingly serious eye disease, causing visual impairment and even blindness. Viruses can affect any parts of the eyeball and cause various diseases, such as necrotic retinitis, obstinate uveitis, and endotheliitis, resulting in irreversible damage and even blindness. Therefore, a potent antiviral therapy is essential to effectively inhibit virus replication and reduce tissue damage. Although numerous studies have explored treatment of intraocular viral infections, the choice of drugs and administration routes to achieve effective concentrations at the lesion site needs further investigation.
Ganciclovir (GCV), an antiviral medication developed in the 1980s, has a strong antiviral effect against cytomegalovirus (CMV) and other members of the herpes virus family [
GCV is always used to treat or prevent infections of CMV and other members of the herpes virus family. Systemic GCV has been widely used in treatment of CMV-related retinitis, and it has also been recognized as an effective treatment for acute inflammation in CMV endotheliitis [
Topical GCV therapy has been considered as a combination therapy with systemic GCV or as a prophylactic therapy. A topical GCV gel has been approved for treatment of HSV keratitis in more than 30 countries [
In the last few years, intravitreal injection of ganciclovir (IVTG) has been widely used in treatment of CMV retinitis. It has also been introduced in treatment of refractory CMV endotheliitis in a case report [
A total of 46 New Zealand white rabbits (Beijing Vital River Laboratory Animal Technology Co., Ltd.) were used. 34 rabbits were randomly divided into two groups: one group received intravitreal injection of GCV (2 mg/0.1 mL) and the other group received intracameral injection of GCV (2 mg/0.1 mL). Five of each group were used for anterior chamber GCV concentration detection on the 1st, 3rd, and 7th day after injection, and the others were used for histological examinations on the 1st, 7th, 14th, and 28th day after injection. The other 12 rabbits were used as control; one eye received intravitreal injection of an equal volume of balanced salt solution (BSS), and the other eye received intracameral injection of BSS. All experiments were performed in accordance with the ethical standards of Peking University Third Hospital’s Institutional Review Board and with the Association for Research in Vision and Ophthalmology (ARVO) Resolution on the Use of Animals in Research.
Rabbits were anesthetized with intramuscular injections of 20 mg/kg ketamine and 5 mg/kg xylazine. After topical applications of 0.4% oxybuprocaine hydrochloride, intravitreal injections were performed. Under sterile conditions, a 27-gauge needle was introduced through the cornea, and 0.1 mL of aqueous humor was removed from the anterior chamber to ensure adequate hypotony. Then, the drug was injected very slowly with a 25-gauge needle 2 mm posterior to the limbus, with care taken to avoid hitting the lens.
Rabbits were anesthetized as previously described. After anterior chamber, paracentesis was performed to maintain normal intraocular pressure and the drug was injected with a 27-gauge needle through the corneal limbus into the anterior chamber.
Rabbits in both groups were monitored by slit-lamp examination, optical coherence tomography (IVue-100 Fourier-domain optical coherence tomograph, Optovue Inc., Fremont, CA, USA), microscopy for fundus examination, and confocal microscopy (Heidelberg Retina Tomograph 3 with Rostock Cornea Module; Heidelberg Engineering, Heidelberg, Germany) before and 1, 7, 14, and 28 days after intraocular injection of GCV to detect changes (corneal edema, intraocular inflammation, vitreous haze, vitreous hemorrhage, retinal detachment, preretinal membrane formation, or neovascularization).
A pharmacokinetic study was carried out to determine changes in GCV concentration and how long it remained in the aqueous humor. Before aqueous humor collection, animals were anesthetized as previously described. Anterior chamber paracentesis was performed with a 27-gauge needle introduced through the corneal limbus, and 0.1 mL of aqueous humor was removed from the anterior chamber. The samples were frozen and stored at −80°C until analyzed.
The GCV concentration of rabbit aqueous humor was detected by a high-performance liquid chromatography (HPLC) assay with electrochemical detection. Samples were centrifuged, and the supernatant was injected into a hypersil silica column (100 × 4.6 mm) after dilution. Then, we recorded the peak area and calculated the concentration of GCV. The limit of sensitivity of the assay was 1
Rabbits were euthanized at the scheduled time by an intravenous overdose of sodium pentobarbital, and their eyeballs were immediately extracted. Corneal and retinal samples were fixed in fixative.
A portion of corneal and retinal tissue was fixed in 4% formaldehyde solution for 24 h at room temperature. Specimens were dehydrated in a series of ascending concentrations of ethanol, cleared in xylene, and embedded in paraffin wax. Serial sections of the eye were cut at a thickness of 4.0
A portion of corneal and retinal tissue was fixed in 3% glutaraldehyde in PBS for 2 h at room temperature, washed three times with PBS, and then dehydrated through a graded series of ethanol. Samples were critical-point dried, sputter coated with 20 nm gold-palladium, and examined with scanning electron microscopy (SEM; JSM-5600LV; JEOL, Tokyo, Japan).
Samples were fixed with 3% glutaraldehyde for 1 h at room temperature, fixed with fresh 3% glutaraldehyde once, postfixed with 1% osmium tetroxide, washed three times in PBS for 5 minutes each, dehydrated through a graded series of ethanol, and embedded in Epon 812. Ultrathin (80 nm) sections were collected on copper grids and double-stained with uranyl acetate and lead citrate. Then, sections were examined with transmission electron microscopy (TEM; JEM-1230; JEOL).
The data were analyzed with SPSS Statistics 20. Aqueous GCV concentrations and endothelial cell density (ECD) were analyzed separately. To compare the ECD of the cornea at different time points between all these groups, ANOVA was employed. Data are presented as mean ± standard deviation.
Aqueous GCV concentrations were 24.83 ± 6.41 (16.51∼34.02)
Ganciclovir concentration in aqueous humor was detected on the 1st, 3rd, and 7th day after intravitreal/intracameral injection of 2 mg/0.1 mL GCV, separately.
Slit-lamp examination, optical coherence tomography, and microscopy for fundus examination of the rabbit eyes after intravitreal injection of 2 mg/0.1 mL GCV showed no corneal edema, intraocular inflammation, or any other abnormalities (Figures
Representative anterior-segment photographs of rabbit eyes. Photographs of a rabbit eye (a) before injection and (b) 1 day and (c) 7 days after intravitreal injection of 2 mg/0.1 mL GCV. Photographs of a rabbit eye (d) before injection and (e) 1 day and (f) 7 days after intracameral injection of 2 mg/0.1 mL GCV. Photographs of a rabbit eye (g) before injection and (h) 1 day and (i) 7 days after intravitreal injection of 0.1 mL BSS. Photographs of a rabbit eye (j) before injection and (k) 1 day and (l) 7 days after intracameral injection of 0.1 mL BSS. Original magnification, ×16.
Representative optical coherence tomography of rabbit eyes. (a) Optical coherence tomography of a rabbit eye 1 day, 1 week, and 4 weeks after intravitreal injection of 2 mg/0.1 mL GCV. (b) Optical coherence tomography of a rabbit eye 1 day, 1 week, and 4 weeks after intracameral injection of 2 mg/0.1 mL GCV. (c) Optical coherence tomography of a rabbit eye 1 day, 1 week, and 4 weeks after intravitreal injection of 0.1 mL BSS. (d) Optical coherence tomography of a rabbit eye 1 day, 1 week, and 4 weeks after intracameral injection of 0.1 mL BSS.
Endothelial cells were typically hexagonal before injection and 1 day, 1 week, and 2 weeks after intravitreal injection of 2 mg/0.1 mL GCV (Figure
Representative confocal microscopy of the endothelial layers of rabbit eyes. (a) Confocal microscopy of a rabbit eye before and after intravitreal injection of 2 mg/0.1 mL GCV. (b) Confocal microscopy of a rabbit eye before and after intracameral injection of 2 mg/0.1 mL GCV. (c) Confocal microscopy of a rabbit eye before and after intravitreal injection of 0.1 mL BSS. (d) Confocal microscopy of a rabbit eye before and after intracameral injection of 0.1 mL BSS.
We were unable to quantify the ECD of three of all rabbits on the first day after intracameral injection of GCV due to corneal edema (Figure
In the control group, endothelial cells were also typically hexagonal before and after intraocular injection of an equal volume of BSS (Figure
There was no significant difference in ECD before injection (
Light microscopy revealed normal corneal morphology after intravitreal injection of GCV (Figures
Representative light microscopy of the cornea and retina of a rabbit eye after intravitreal injection of 2 mg/0.1 mL GCV. (a, b) Corneal morphology 1 day after intravitreal injection at different magnifications. (c, d) Retinal morphology 1 day after intravitreal injection at different magnifications. (e, f) Retinal morphology 2 weeks after intravitreal injection at different magnifications. Hematoxylin and eosin stain.
After intracameral injection of GCV, the corneal stroma layer was obviously thickened. The endothelium was separated from Descemet’s membrane during tissue processing (Figures
Representative light microscopy of the cornea and retina of a rabbit eye 1 day after intracameral injection of 2 mg/0.1 mL GCV. (a, b) Corneal morphology at different magnifications. The endothelium was separated from Descemet’s membrane (black arrows). (c, d) Retinal morphology at different magnifications.
The morphology of corneal and retinal tissue was normal after intravitreal/intracameral injection of BSS (Figure
Representative light microscopy of the cornea and retina of rabbit eyes 1 day after intraocular injection of 0.1 mL BSS at different magnifications. (a, b) Corneal and retinal morphology after intravitreal injection. (c, d) Corneal and retinal morphology after intracameral injection.
SEM showed no corneal endothelial abnormalities after intravitreal injection of GCV. Corneal endothelial cells exhibited normal cytoarchitecture. The boundary of hexagonal cells was clear, and intercellular connections were tight. Endothelial cells intersected with each other. Microvilli on the cell surface appeared normal (Figure
Representative scanning electron microscopy of the endothelial layers of rabbit eyes after intravitreal injection of GCV (a), intracameral injection of GCV (b, c), and intracameral injection of BSS (d).
In contrast, the boundary of corneal endothelial cells was unclear after intracameral injection of GCV. Intercellular connections were destroyed, and microvilli on the cell surface were shortened and abnormal (Figure
Corneal endothelial cells exhibited normal cytoarchitecture after intravitreal injection of BSS (results not shown) and intraocular injection of BSS (Figure
TEM showed normal architecture of all the corneal layers after intravitreal injection of 2 mg/0.1 mL GCV. The corneal endothelial cells exhibited normal cytoarchitecture with tight intercellular junctions (Figure
Transmission electron microscopy of the corneal endothelial layer of a rabbit eye at different magnifications on the 1st day (a, b), 7th day (c, d), and 28th day (e, f) after intravitreal injection of 2 mg/0.1 mL GCV.
Transmission electron microscopy of the retina of a rabbit eye on the first day after intravitreal injection of 2 mg/0.1 mL GCV. (a, b) All retinal layers exhibited significant edema at different magnifications. (c) Bubble-like structures were found in the retinal nerve fiber layer. (d, e) The intercellular space was dilated in the inner/outer nuclear layer. (f) Photoreceptors were swollen and loosely arranged, and photoreceptor outer segments were decreased in number.
Transmission electron microscopy of the retina of a rabbit eye on the 7th day after intravitreal injection of 2 mg/0.1 mL GCV. (a) Retinal edema remitted on the 7th day after injection. (b) Bubble-like structures remained in the retinal nerve fiber layer but were significantly reduced. (c, d) The intercellular space was smaller in the inner/outer nuclear layers. (e, f) The photoreceptors were less swollen at different magnifications.
Transmission electron microscopy of the retina of a rabbit eye 4 weeks after intravitreal injection of 2 mg/0.1 mL GCV. (a) Retinal architecture was almost normal except for remaining fluid retention in the retinal nerve fiber layer (b). The inner plexiform layer (c), inner nuclear layer (d), outer nuclear layer (e), and photoreceptor layer (f) all exhibited normal architecture.
Endothelial cells were swollen and even broken on the first day after intracameral injection (Figure
Transmission electron microscopy of the corneal endothelium and retina of a rabbit eye after intracameral injection of 2 mg/0.1 mL GCV. The cytoarchitecture of corneal endothelial cells on the 1st day (a, b) and 7th day (c, d) after injection at different magnifications. (e, f) Fluid retention was still observed in the retinal nerve fiber layer 4 weeks after intracameral injection of 2 mg/0.1 mL GCV.
TEM also showed that the corneal and retinal tissue of rabbit eyes exhibited normal cytoarchitecture after intravitreal/intracameral injection of 0.1 mL BSS (Figure
Transmission electron microscopy of the corneal endothelium and retina of rabbit eyes after intraocular injection of 0.1 mL BSS. (a, b) The cytoarchitecture of corneal endothelial cells on the 1st day after intracameral injection. The retinal nerve fiber layer (c), inner nuclear layer (d), outer nuclear layer (e), and photoreceptor layer (f) all exhibited normal architecture after intravitreal injection.
In the last few years, intravitreal injection of GCV has been widely used in treatment of CMV retinitis. Intravitreal injection of GCV has also been introduced in treatment of corneal endotheliitis. This method involves injection of GCV directly into the vitreous body, bypassing all the external barriers. Despite its invasive nature and other complications associated with intravitreal injections, this method of administration remains expedient, provides fast delivery of high intraocular drug concentrations, and has fewer systemic side effects.
In our study, GCV concentrations in the aqueous humor of rabbits after intravitreal injection were evaluated. Aqueous GCV concentrations were 16.51∼34.02 (24.83 ± 6.41)
Schulman also studied the clearance of GCV in aqueous humor following intravitreal injection of 400
The concentration of GCV is significantly lower after topical administration on the ocular surface than after intravitreal injection. The peak concentrations of GCV in the aqueous humor after administration of 0.2% GCV in situ gelling eye drops and common GCV eye drops were 4.79
In our study, we also evaluated GCV concentrations in the aqueous humor of rabbits after intracameral injection. Notably, GCV could not be detected on the first day after intracameral injection, which suggests that GCV cannot remain in the anterior chamber for a long time. Gunda et al. once placed a well on the cornea of rabbits with linear probes implanted in the aqueous humor and allowed 200
Currently, intravitreal injection of GCV is widely used in the treatment of CMV retinitis. When taking the severity of retinal lesions in CMV retinitis into account, intravitreal injection of GCV constitutes a safe and efficient treatment for this disease. Recently, GCV was used to treat endotheliitis in a case report. However, whether intravitreal administration of GCV is safe for intraocular tissues remains controversial.
Several clinical studies have evaluated the safety of retina after intravitreal injection of GCV. The reported dose varies from 0.2 to 5 mg of GCV [
In our study, the retina showed significant tissue edema and swollen cytoarchitecture 1 day after intravitreal injection of GCV. Fortunately, retinal edema gradually receded over time. However, TEM showed that fluid retention remained in the retinal nerve fiber layer 4 weeks after injection. The results indicated that intravitreal injection of 2 mg GCV had a toxic effect on retina; however, the damage to the retina receded over time and may be reversible. Fortunately, there were no abnormal changes in the cornea. As the damage caused by GCV receded over time, it might explain why there was no significant effect on visual acuity in clinical practice in patients with CMV retinitis. This finding also suggested that intravitreal injection of GCV should be used with great caution for treatment of endotheliitis.
Pulido investigated the retina toxicity of GCV intravitreal injection in the rabbit. Intravitreal doses of 400
Nevertheless, Moschos revealed that intravitreal injection of GCV doses of 300–600
In our study, we found corneal edema on the first day after intracameral injection of GCV, and we were unable to quantify the ECD of rabbits due to corneal edema. The cytoarchitecture of endothelial cells was nearly destroyed. The cornea recovered clarity within one week because the corneal endothelial cell of rabbits could regenerate with the support of Descemet’s membrane [
In conclusion, intravitreal injection of GCV is optional for treatment of refractory viral corneal endotheliitis. This method offers effective drug concentrations in the anterior chamber and forms a drug depot that releases the drug in a sustained manner. Notably, intravitreal injection of GCV has a toxic effect on the retina. Although the damage to the retina receded over time, intravitreal injection should be used with great caution for treatment of endotheliitis. However, intracameral injection of GCV results in rapid elimination of drugs from the aqueous humor and can severely damage endothelial cells; moreover, GCV can induce endothelial dysfunction in humans. Therefore, intracameral injection of GCV is not recommended.
The data used to support the findings of this study are included within the article.
The authors declare that they have no conflicts of interest.
Drs. Sun and Peng contributed equally to the work.
This work was supported by the National Natural Science Foundation of China (grant number 81650027). The authors thank Yu-xin Guo, Yuan Qiu, and Ting Yu for their assistance with the experiments.