Introduction
Middle-aged and elderly people are more vulnerable than are the rest of the population to ischemic optic neuropathy, in particular, anterior ischemic optic neuropathy (AION), which seriously affects visual function. While several methods for treating AION exist, results are not satisfactory (Hayreh, Reference Hayreh1974; Knox et al., Reference Knox, Kerrison and Green2000; Salomon et al., Reference Salomon, Dardik, Steinberg, Kurtz, Rosenberg, Moisseiev and Huna-Baron2000). Further investigations of AION and its treatments are hindered by the lack of suitable animal models (Fazzone et al., Reference Fazzone, Kupersmith and Leibmann2003). Scholars have attempted to establish animal models of optic neuropathy (ON) by pressing and partially amputating the optic nerve, increasing intraocular pressure, injecting vasoconstrictors, and so on. More in-depth investigations to assess the value of treatments for AION are hampered by the lack of suitable and easily accessible animal models (Selles-Navano et al., Reference Selles-Navano, Ellezam, Fajardo, Latour and McKerracher2001). Thus, there is a need to establish an animal model of AION. Our study was aimed at establishing and assessing induced AION (rAION), that is, AION resulting from laser photoactivation of hematoporphyrin derivative (HPD). The aims of this study were to (i) establish an AION rat model; (ii) observe functional changes in the fundus using fundus fluorescein angiography (FFA), optical coherence tomography (OCT), and flash visual-evoked potential (F-VEP); and (iii) confirm the model’s reliability histopathologically.
Materials and methods
The Shanxi Province Institutional Animal Care and Use Committee approved all animal procedures. Normal male Sprague-Dawley rats (140–160 g) were obtained from the Animal Experiment Center of the Fourth Military Medical University of Chinese PLA [SCXK (Military) 2002-005]. Animals were housed in the Xi’an Institute of Ocular Fundus Diseases animal facility and were given food and water. Thirty healthy male rats were maintained in 12-h light (06:00–18:00)/12-h dark (19:00–05:00) cycles, at temperatures ranging from 18 to 29°C, humidity >60%, and background noise <60 dB. Those without eye disease were examined by slit-lamp biomicroscopy and direct ophthalmoscopy after induction of mydriasis.
Rats were randomly divided into naive (n = 5), laser (n = 5), HPD (n = 5), and rAION (n = 15) groups. As right eyes were experimental, and all left eyes were controls. Animal models were established according to the method of Bernstein et al. (Reference Bernstein, Guo, Kelman, Flower and Johnson2003). Prior to induction, animals were anesthetized with an intraperitoneal injection of a narcotics parenteral solution (0.5 ml/kg), SMX. A fundus Volk 78D scope lens (Volk Optical, Inc. Mentor, OH) was placed in front of the laser system to visualize the normal retina. Preoperatively, experimental rAION eyes were given the mydriatic drops three times (30 min each). AION was induced in anesthetized animals using a tail vein injection of HPD (10 mg/kg). After being kept out of strong light for 2 h, the optic nerve was subjected to laser illumination with a multiwavelength krypton red laser (laser beam diameter, 647 nm; 2/3 disc spot laser/120 s/80 mW). After treatment, the left eyes were covered with black rubberized cloth. Those in the HPD group were injected only with HPD via the tail vein. Those in the laser group were illuminated only, as mentioned above. Those in the naive group were untouched. The rAION and HPD groups were kept out of strong light for 2–3 days, except once for FFA and fundus examination. After an FFA examination, three rats each from the rAION, HPD, naive, and laser groups were chosen at random for killing and used for histopathological examination. The remaining 12 rats from the rAION group were examined at 6 and 23 days (FFA), 3 days (VEP), and at 6 and 23 days (OCT). At 23 days, six rats were killed and the eyes prepared for histopathological examination. Six rats were killed 90 days after fundus photography of the optic nerve, and the eyes fixed for histopathological examination.
Visual electrophysiological test
To permit VEP recording, all rats in four groups were anesthetized and secured to a moveable platform, and stainless steel screws were implanted 3 mm lateral to the lambda and 5 mm behind the bregma. Reference electrodes were placed 2 mm lateral to the lambda and 2 mm in front of the bregma. Transient VEPs were evoked by single-flash stimulation (1.3 Hz, 12 ms). VEP signals were recorded by a commercial system (Roland Consult GmbH, Brandenburg, Germany). The number of sweeps per average generally ranged between 40 and 80. Recordings were performed on day 3 secured to a moveable platform after ischemia.
Fluorescein angiography
Fluorescein angiography was performed at 30 min and 6 and 23 days after model establishment. Prior to digital fluorography, rats were anesthetized with intramuscular ketamine (10 mg/kg) and xylazine (4 mg/kg), and the pupils dilated. Fluorescein sodium (10%; 75 mg/kg; Fluorescent; Alcon, Alcon Laboratories, Inc. Fort Worth, Texas) was injected into the tail vein of anesthetized rats. Each rat was placed on a stereotactic platform. A contact lens was placed in the front of the cornea in order to sustain corneal clarity throughout the experiment. After injection of sodium fluorescein dye into the tail vein (within 3 s), digital fundus photographs of eyes were taken using a digital fundus camera (Heidelberg Engineering, Heidelberg, Germany), from 5 to 30 s (to record early-phase angiogram) and within 5–8 min (to record late-phase angiograms). Angiograms were retained and reserved for analysis.
Optical coherence tomography
After mydriasis and anesthesia, rats were placed on a detection table in front of the prepositus scope. Using an OCT scanner (Humphrey2000 type; Carl Zeiss Company, Zeiss-Humphrey system, Dublin, CA), cross-scans centered on the optic disc were performed to record morphological changes at 6 and 21 days after model establishment.
Histopathology analysis
For histological analysis, rats were killed, and the eyes immersed in a fixative containing 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 m phosphate buffer (pH 7.2) for 1 h. After being labeled, the anterior portions of the eyes were removed. The posterior portions were fixed for an additional 12 h and then postfixed in 1% osmium tetroxide for 2 h (provided by the Institute of Ophthalmology, Fourth Military Medical University of Chinese PLA). Tissue samples were dehydrated and embedded in paraffin. Serial 6-μm sections were cut along the horizontal meridian through the optic nerve head and stained with methylene blue. The optic disc and surrounding retinal blood vessels were observed under high-power light microscopy. Images of an area 1–1.5 mm from the edge of the optic nerve head were captured and stored.
Statistical analysis
Data were processed using SPSS 13.0 statistical software (SPSS Inc., Chicago, IL) for normality. Student’s t-test was used.
Results
Of the 30 rats, 1 in the laser group and 2 in the AION group died after anesthesia. There were 27 rats used for analysis of results.
Fundus
No abnormality of the optic disc was observed during fundus examination at any time point in the laser (Fig. 1A), HPD, or control groups (data not shown). In the AION group, there was edema in the upper optic disc and an unclear boundary 1 day after model establishment (Fig. 1B). Edema was still observed at 6 days (Fig. 1C), and the upper optic disc was atrophied and appeared pale at 90 days (Fig. 1D).
Fundus phases at different time points. (A) The fundus of a normal control. (B) Upper optic disc was pale and edematous at 1 day in the AION group. (C) Upper optic disc was pale and edematous at 6 day in the AION group. (D) Upper optic disc was pale and atrophied at 90 days in the AION group. (The fundus camera used in panel (A) and (C) was different from the camera used in panel (B) and (D), so the images are different.)
FFA examination
FFA provides the most compelling in vivo evidence of optic disc circulatory impairment in our model. No abnormal changes were observed in the laser (Fig. 2A), HPD, or control groups (data not shown). In the AION group, hyperfluorescence was observed in the upper optic disc postoperatively at 49 s (Fig. 2B) and 3 min (Fig. 2C). Early hypofluorescence (Fig. 2D), middle hyperfluorescence (Fig. 2E), and late hyperfluorescence (Fig. 2F) were observed in the upper optic disc 30 min after model establishment; there was hypofluorescence in the upper optic disc at 21 s (Fig. 2G) and 3 min (Fig. 2H) after 6 days. There was hypofluorescence at all times at 23 days (Fig. 2I).
FFA at different time points. (A) No abnormal changes in the normal control groups. (B) Enhanced fluorescence at 49 s in the HPD group. (C) High fluorescence at 7 min in the HPD group. Low fluorescence at 13 s (D) and 26 s (E), and high fluorescence at 118 s (F) after 1 day of HPD group being made. Low fluorescence at 21 s (G) and 3 min (H) after 6 days of HPD group being made. (I) Low fluorescence at all times after 23 days in the AION group.
OCT changes
No abnormal changes were observed in the laser, HPD, or naive groups. There was a clear boundary between the pigment epithelium layer of the choroid and the nerve fiber layer; the reflective surfaces of the optic disc and retina were smooth (normal optic disc). In the AION group, the reflective surface of a portion of the nerve fiber layer showed creases and was rougher and more elevated than the retinal plane (edema of optic disc) at 6 days after model establishment. The reflective surface of the optic disc was less than that of retina, and the surface was smooth (atrophied) at 23 days.
Changes of visual nerve function
There were no abnormal changes in the laser, HPD, or control groups. However, compared to control eyes, experimental eyes of the AION group had prolonged F-VEP P100 latency [(71.65 ± 8.81), (57.58 ± 8.38) ms, t = 3.148, P = 0.012] and decreased wave amplitude [(4.77 ± 1.90), (10.06 ± 3.66) μv, t = 4.082, P = 0.003] 3 days after model establishment.
Histopathology
Tissue sections of normal rat optic disc showed that the nerve fibers were arranged in order. The pigment epithelium layers of the bilateral choroidal membranes—the internal and external granular layers—had clear stratifications with regular arrangements. Ganglion cells within 2 diameters of the optic disc were dense. No abnormal changes in optic disc tissue were observed in the laser, HPD, or naive groups 1 day after model establishment (Fig. 3A). In the AION group, edema of portions of the nerve fibers and loose tissue were observed in the optic disc; portions of the retina around the optic disc were displaced (Fig. 3B). By day 23, the edema had disappeared and atrophy of the nerve fiber layer had lessened (thinner) (Fig. 3C). Under high power, sections of retina in the control group exhibited a densely arranged ganglion cell layer and normal internal and external granular layers. The typical internal granular layer had 3–6 layers, whereas the external one had 8–11 layers. At 23 days after model establishment, no change in retinal tissue sections was observed in the laser, HPD, or naive groups (Fig. 4A). However, in the AION group, the tissue of retinal clearly reduced, and defects (enlarged space) of nerve could be seen; there was no change in the thicknesses of the internal or external granular layers (Fig. 4B).
Histopathological examination of the optic disc tissue. (A) No abnormal changes in the control group. (B) Optical edema of nerve fibers became more pronounced (thickened) in the ischemic area of the optic disc, edematous tissue was loose, and the surrounding retina was displaced in the AION group after 1 day [hematoxylin and eosin (H&E), ×100]. (C) Optic disc was atrophied and edema was extinct in the AION group after 23 days (H&E, ×150). (D) The occlusion of blood vessels in the AION group in 1 day (H&E, ×150). (E) The higher magnification of optical edema and occlusion of blood vessels in the AION group in 1 day (H&E, ×400).
Histopathological examination of retina after 23 days. (A) Normal retina in the control group [hematoxylin and eosin (H&E), ×200]. (B) The edema disappeared and atrophy of the nerve fibers lessened (thinner) in the AION group (H&E, ×150).
Discussion
In persons older than 50 years of age, AION is the most common acute ON. This is, perhaps, due to infarctions of the anterior portion of the optic disc resulting from inadequate blood supply from the posterior ciliary arteries. Researchers have tried to establish animal models of ON, but this disease differs greatly from AION (Cioffi et al., Reference Cioffi, Orgul, Onda, Bacon and Van Buskirk1995; Johnson et al., Reference Johnson, Deppmeier, Wentzien, Hsu and Morrison2000; Krueger-Nauq et al., Reference Krueger-Nauq, Emsley, Myers, Currie and Clarke2002). The similarity of tissue construction and microcirculation of optical papilla between rat and human eyes are the basis of this study. Recently, Bernstein et al. (Reference Bernstein, Guo, Kelman, Flower and Johnson2003) reported on a new rodent NAION model (rAION), which is based on laser photoactivation of Rose Bengal to generate superoxide radicals. These radicals selectively damage the optic nerve capillary vascular endothelium, producing predictable and reproducible levels of isolated optic nerve ischemia. This research revealed a new method of photodynamically inducing an animal model to establish AION.
The fundamental principle of photodynamics is based on the photochemical action of superoxide anions, not the results of heat injury (Shinji et al., Reference Shinji, Kazuo, Kazunao, Abby and Mitsuyoshi1998). Peak wavelengths of the HPD fluorescence spectrum are 630 and 690 nm. The half-life of HPD in rats is 5–7 h, which is one-fifth that of human beings (25–35 h) (Goldenberg-Cohen et al., Reference Goldenberg-Cohen, Guo, Margolis, Cohen, Miller and Berstein2005). Thus, laser irradiation was initiated ~2 h after injection, a time when the HPD would have been well combined with vascular endothelial cells. The laser beam was aimed directly at the upper optic disc and vascular endothelial cells, which combined with photosensitized active oxygen materials. This, in turn, led to the denaturation and swelling of endothelial cells and exposure of the basilar membrane. Blood platelet white cells also adhered to the endothelial cells or to the exposed intercellular sites and formed mural thrombi, which selectively blocked the microvessels of the supporting optic nerve and resulted in edema of the irradiated areas. The flow rate is rapid in the large blood vessels of the retina, as they are not easy to block. Bernstein et al. (Reference Bernstein, Guo, Kelman, Flower and Johnson2003) used rose bengal and green laser to induce their model. The penetration force of green laser is relatively weak, easily absorbed by protein, and not easily controlled experimentally. But our study used a different photosensitizer, that is, HPD, a blood porphoryrin derivative. The fluorescence wavelength peak is 630–690 nm. The laser wavelength should overlap maximally with the selected absorbed peak of the photosensitizer. We used krypton red light (647 nm), which could blocks and damages the deep blood capillaries according wavelength. This pathological mechanism changes seem to be the same type of lesion in the front of the optic nerve in AION in human beings. Goldenberg-Cohen et al. (Reference Goldenberg-Cohen, Guo, Margolis, Cohen, Miller and Berstein2005) also studied a photodynamically induced rAION model based on Bernstein’s studies. The fundamental parameters were laser wavelength and photosensitizer. The difference in this animal model is in the duration of time of laser illumination. Bernstein suggests that the appropriate illuminating time is 20 s, but the Goldenberg-Cohen group thought that a laser illumination time that exceeded 15 s could cause or contribute to central retinal vein occlusion (CRVO) or central retinal artery occlusion (CRAO). Furthermore, dilation of blood vessels in the retina is also observed, appearing in 10% of eyes within 15 s. The pathology of CRVO is a laser wavelength of 514–535 nm can easily combine with photosensitizers existing in the inner cells of blood vessels and cause blood vessel blockages to occur within the retina. These advantages were confirmed by selecting and using krypton red light (647 nm) as the activating illuminant. We built the rAION model observed by FFA, showing no pathological changes in the optic nerve, and no CRVO or CRAO after HPD.
No abnormal manifestations by fundus or FFA examinations were observed in the laser, HPD, or control groups. In the rAION group, morphological changes observed under the funduscope 30 min after model establishment indicated microcirculatory disorders and capillary endothelial damage in the hyperfluorescent areas of the optic nerve. As the damage became more pronounced, typical papilledema was observed at 1 day. FFA showed early hypofluorescence in the area of optic nerve ischemic damage. The compensatory dilation of near capillaries and those on the optic disc surface led to the fluorescent phases of middle and late periods being greater than the hyperfluorescence in the damaged area of the optic nerve. FFA showed persistent hypofluorescence of the optic disc after 23 days. Areas of gray atrophy were observed in the upper optic disc after 90 days. Furthermore, no abnormal fluorescent changes in retinal blood vessels were seen. These fluorescent changes are similar to those seen in AION in human beings. Goldenberg-Cohen’s first attempt has carried out FFA verification in an in vivo animal in an extraordinary manner. We were guided by animal model building based on retinal ophthalmoscopy. At appropriate times, FFA was carried out, and we observed and recorded the natural course of rAION progress.
Histopathological examination showed a series of pathological changes, including denaturation or necrosis of a portion of optic disc nerve fibers, edema of local and surrounding tissues, and loose tissue. The volume increased, extended to the surrounding deeper layers toward the corpus vitreum, and blocked the tough madrepenic plate, which also pressed in on the retina around part of the optic disc, displacing it on day 1. At 23 days, optic disc tissue sections showed disappearance of papilledema and a lessening of nerve fiber atrophy. Observation of retinal tissue sections suggests that this method of model establishment does not destroy structures underlying the retina. OCT, carried out on an in vivo animal model at 6 days, showed that the creases on the reflective surface of portions of the optic nerve fiber layers were rough and more elevated than in the retinal plane (papilledema). This was clearly different from optic disc and retinal reflective surfaces before model establishment or those after 23 days. OCT and FFA results at acute and atrophic periods are in confirmation. These results histopathologically explain and confirm the manifestations and characteristics of optic nerve ischemia. rAION resembles the natural pathophysiology of human AION. Flicker VEP latency period and amplitude reflect the state of optic nerve function in myelin and axonal sheaths, respectively. Prolonged latency period and amplitude values are reduced in the rAION model eye at 3 days. These abnormal changes were more significant than those of the other groups.
The main pathophysiological process in AION primarily involves the prelaminar and laminar regions of the optic nerve. The major vascular supply to the prelaminar and laminar regions of the optic nerve head is from end arteries of the circle of Haller and Zinn. This circle is formed by branches of the short posterior ciliary arteries, is located intrasclerally, and may be constricted in eyes with a small scleral outlet. Discs with a small or absent physiological cup have been defined as one of the most important risk factors in the development of AION. When blood perfusion is decreased because of reduced vessel caliber, systemic hypotension, or increased intraocular pressure, the mentioned regions of the optic nerve suffer ischemia. Following ischemia, unmyelinated axons in the prelaminar region are often edematous, with the border between ischemic and viable tissue being at the lamina cribrosa. The resultant edema further increases constriction and compromises perfusion, causing a vicious cycle. Several other risk factors, for example, ischemic heart disease, hypercholesterolemia, and diabetes mellitus also play significant roles. The mechanism of rAION in our experimental models is different from that of human beings, particularly the photodynamically induced in vivo isolated optic nerve axonal ischemia model. In the rAION model, the prelaminar region is approached directly, and the adjacent tissues (retina and choroid) are not involved. This rAION model may be used to investigate effective or neuroprotective drugs and various proposed medical treatments for the treatment of AION.
