A Rat Model of Chronic Pressure-induced Optic Nerve Damage

Exp. Eye Res. (1997) 64, 85–96

A Rat Model of Chronic Pressure-induced Optic Nerve Damage J O H N C. M O R R I S ONa,c, C. G. M O O R Ea, L I S A M. H. D E P P M E I E Ra, B R U C E G. G O L Db, C H A R L E S K. M E S H U Lc    E L A I N E C. J O H N S O Na a

Kenneth C. Swan Ocular Neurobiology Laboratory, Casey Eye Institute, b The Center for Research in Occupational and Environmental Toxicology and Department of Cell Biology and Developmental Biology, Oregon Health Sciences University, and the c Portland Veterans Affairs Hospital and Medical Center, Departments of Ophthalmology, Medical Psychology and Pathology, Portland, Oregon, U.S.A. (Received Columbia 8 August 1995 and accepted in revised form 25 June 1996) To develop unilateral, chronically elevated intraocular pressure in rats, episcleral veins were injected with hypertonic saline and the intraocular pressure was monitored with a Tono-Pen XL tonometer. Histologic analyses of eyes with differing degrees and durations of intraocular pressure elevation were performed to ascertain the effects of these pressures on the optic nerve. Out of 20 consecutive animals, nine had elevations of intraocular pressure following a single injection, while subsequent injections raised intraocular pressure in seven others. One eye became hypotonous. In the remaining animals, subsequent injections sufficient to raise intraocular pressure were deliberately withheld, to determine the possible direct effects of injections on the optic nerve. Mean sustained pressure elevations ranged from 7 to 28 mm Hg and the retinal vasculature remained perfused in all eyes. Optic nerve cross sections from eyes without intraocular pressure elevation appeared identical to those from uninjected eyes, while nerves from eyes with the greatest intraocular pressure rise demonstrated axonal damage that involved 100 % of the neural area. Eyes with either less severe pressure elevations or shorter durations showed partial damage, ranging from 0±5 % to 10±4 % of the neural area. In 70 % of these nerves, damage was concentrated in the superior temporal region. Within the optic nerve head, often associated with astrocytes, axons contained abnormal accumulations of membrane-bound vesicles and mitochondria. The anterior chamber angles showed sclerosis of the trabecular meshwork with anterior synechiae, but Schlemm’s canal, collector channels and aqueous veins appeared patent. Unilateral sclerosis of the trabecular meshwork produces sustained elevation of intraocular pressure in rats with optic nerve damage that in many ways resembles that seen in human glaucoma. Understanding the mechanism of nerve damage in this model may provide new insights into the pathogenesis of human glaucoma. # 1997 Academic Press Limited Key words : glaucoma ; rat model ; intraocular pressure ; glaucomatous optic neuropathy ; tonometry.

1. Introduction Glaucoma is characterized by progressive optic nerve head cupping, selective loss of retinal ganglion cells and their axons, and loss of visual field. (Glovinsky, Quigley and Dunkelberger, 1991 ; Sommer, et al., 1991 ; Quigley, Dunkelberger and Green, 1988 ; Quigley and Green, 1979 ; Quigley et al., 1983). Elevated intraocular pressure is an important risk factor for glaucoma (Anderson, 1977 ; Sommer, 1989), but little is known about how pressure damages optic nerve fibers. While axonal transport blockade within the optic nerve head has been demonstrated in both human glaucoma specimens and experimental glaucoma models (Anderson and Hendrickson, 1974 ; Gaasterland, Tanishima and Kuwabara, 1978 ;

Reprint requests to : John C. Morrison, Casey Eye Institute, Oregon Health Sciences University, 3375 SW Terwilliger Blvd., Portland, OR 97201, U.S.A. Supported by NIH grants EY10145 (Dr Morrison) and NS19611 (Dr Gold), the Department of Veterans Affairs Merit Review Program (CKM), and unrestricted funds from Research to Prevent Blindness. Dr Morrison is a 1990 RPB Miriam and Benedict Wolfe Scholar. Proprietary interest category : N.

0014–4835}97}010085­12 $25.00}0}EY960184

Quigley and Addicks, 1980 ; Quigley et al., 1981) it has not been demonstrated precisely how increased eye pressure produces this blockade or how this is related to nerve fiber degeneration. Similarly, blockade of axonal transport may not be the sole mechanism underlying glaucomatous nerve damage. Research in this area is limited by the lack of a readily available animal model with which to study, in detail, the effects of elevated intraocular pressure on optic nerve fibers. By improving our understanding of how axons are damaged, such a model may lead to new, effective methods of preventing nerve fiber damage in the face of elevated intraocular pressure. The anatomy of the limbal vasculature in laboratory rats has recently been described (Morrison et al., 1995b). This work suggests that aqueous humor can cross the trabecular meshwork into Schlemm’s canal and then enter a venous plexus through numerous collector channels. This plexus, which encircles the entire limbus, is drained by multiple radial episcleral veins. Based on these relationships, a method to inject mild sclerosants into the aqueous humor outflow pathway has been developed to produce elevated intraocular pressure without affecting the adjacent ciliary body. This report describes the injection method # 1997 Academic Press Limited

86

J. C. M O R R I S O N E T A L.

and its mechanism of action, and provides an overview of the effects of elevated intraocular pressure on the rat optic nerve. 2. Materials and Methods Microneedle Assembly The microneedle injection apparatus (Fig. 1) consists of three main components : (1) a glass microneedle approximately 3 mm long, and between 30 and 50 µm in diameter ; (2) a length of polyethylene tubing (PE-50, Clay Adams, Parsippany, NJ, U.S.A.) stretched to a fine taper over a bunsen burner flame ; and (3) a 23 gauge needle with the tip broken off. The glass microneedle segment is made by heating a 10 µl borosilicate glass disposable micropipet (VWR, Seattle, WA, U.S.A.) in a bunsen burner flame and pulling it by hand to a fine taper. All subsequent assembly procedures take place under a dissecting microscope (Wild, Heerbrugg, Switzerland). First, the narrow end of the stretched micropipet is cut with a razor blade to yield a 3 mm segment. Then, the glass segment is inserted into the narrow end of the tapered polyethylene tubing, leaving about 1±5 mm outside, and the 23 G needle is wedged into the opposite, wider end of the tubing. Finally all junctures are sealed with a minimum of hard-drying epoxy cement (Elmer’s, Colombus, OH, U.S.A.). To bevel, the microneedle is held with a pair of forceps at the glass}tubing glue joint, and then the glass tip is gently touched to a fine, rotating abrasive surface for about 4 sec. The forceps is a curved, jeweler-type reverse-action forceps (Dumont style N7), modified with a grove running length-wise along either inner surface, at the tip, similar to a CozeanMcPherson forceps. This groove provides maximum control of the needle by increasing the number of contact points between forceps and needle. A Dremel tool (Dremel, Racine, WI, U.S.A.) equipped with a Microneedle Glue joint

Tapered polyethylene tubing (PE-50) 23g needle

Glue joint

F. 1. Illustration of the microneedle used for retrograde injection into episcleral vessels, demonstrating glass microneedle tip (with bevel), glue joints, tapered polyethylene tubing, 23 gauge hypodermic needle, and 1 cc syringe.

F. 2. Plastic ring (arrow) placed around the equator of the eye to confine saline to limbal veins and aqueous humor outflow pathways. Arrowhead indicates aqueous vein to be injected.

drum-shaped, fine-grit aluminium oxide bit moistened with water, provides the rotating abrasive surface. This microneedle assembly is best complemented by a 1 ml syringe, to provide sufficient fluid pressure at the needle tip. Also, since the assembly tapers to such a small diameter, the shorter the polyethylene tubing section and the shorter the microneedle, the less resistance there will be to the passage of fluid. Generally, an injection rate of 200 µl per minute can be obtained. Fluids used in this system should be relatively non-viscous and must be filtered through a 0±22 µm filter prior to use, to prevent clogging of the needle. Microcannulation Procedures All experiments were performed in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Male brown Norway retired breeder rats (Rattus norvegicus), weighing between 250 and 400 g, were anesthetized by intraperitoneal injection of 1±0 ml kg−" standard rat cocktail, consisting of a solution of 5 ml ketamine (100 mg ml−"), 2±5 ml xylazine (20 mg ml−"), 1 ml acepromazine (10 mg ml−"), and 0±5 ml sterile water. Following a lateral canthotomy, a small polypropylene ring (5±5 mm inner diameter), with a 1 mm groove cut into its inner surface and a 1±0 mm gap cut out of its circumference, was fitted around the globe, straddling the equator (Fig. 2). The gap in the ring was oriented to provide unobstructed passage for one radial aqueous vein in the superior quadrant of the eye, while the ring occluded other aqueous veins and confined the sclerosing agent to the limbus. With an assistant positioning the eye for maximum exposure, the conjunctiva was incised with Vannas scissors to expose the aqueous vein under 25¬ to

R A T P R E S S U R E-I N D U C E D O P T I C N E R V E D A M A G E

40¬ magnification. The microneedle, held by the curved forceps and pointing toward the limbus, was carefully positioned parallel to the vessel wall and then inserted into the vessel lumen, which was stabilized with jewelers forceps. Immediately after cannulation, a volume of 50 µl micro-filtered hypertonic saline solution (NaCl) was injected into the limbal vascular plexus. For injection, a force just sufficient to blanch the limbal artery was used. Injection with a force greater than this tended to result in eyes with excessive intraocular pressure elevation. Following injection, the plastic ring was removed and eyes were examined to assess the degree of vessel blanching and the extent to which the sclerosant traversed the limbal vasculature. Polysporin ophthalmic ointment (Burroughs Wellcome Co., Research Triangle Park, NC, U.S.A.) was applied to the eye, and the animal was allowed to recover from anesthesia. Animals were observed for general activity, weighed, and their intraocular pressure measured as described below. Periodic, detailed examinations of the anterior segment were also performed, with animals anesthetized via inhalational isoflurane (Forane) to assess inflammation or other abnormalities (Moore et al., 1995). If no changes in intraocular pressure were detected after 2 weeks, a second injection of saline was performed as described above, but in an episcleral vein 180 degrees from the first injection site. The optic nerve head and retinal vasculature was assessed by viewing the fundus with a direct ophthalmoscope through pupils dilated with 1±0 % tropicamide (Alcon, Ft. Worth, TX, U.S.A.) and 2±5 % phenylephrine hydrochloride (Bausch and Lomb, Tampa, FL, U.S.A.). Blood vessel appearance was photographed using a Zeiss photo slit lamp through a flat contact lens placed over the cornea using a 2±5 %¬ hydroxypropyl methylcellulose viscous interface (Iolab, Claremont, CA, U.S.A.) to avoid distorting the cornea. Intraocular Pressure Measurements The Tono-Pen XL was calibrated in five cocktailanesthetized brown Norway rats as previously described (Moore, Milne and Morrison, 1993b) using a low displacement pressure transducer (Omega Engineering, Inc., Stamford, CT, U.S.A.) for measuring intraocular pressure and a threaded plunger syringe 87000 (Hamilton Company, Reno, NV, U.S.A.) for adjusting intraocular pressure. Valid individual readings using a factory-calibrated Tono-Pen XL tonometer (Mentor, Norwell, MA, U.S.A.) were defined as those readings resulting from firm but not excessive contact of the tonometer tip with the cornea. Readings resulting from tear-film contact alone, and those occurring just after the tip was removed from the cornea were ignored, as were the periodic, instrumentgenerated averages which we had previously determined to be unreliable.

87

The means of 15 valid readings obtained at 10 mm Hg increments from 20 to 70 mm Hg were plotted against transducer intraocular pressure to yield a linear regression equation. This equation describes the relationship between Tono-Pen XL readings and actual intraocular pressure. Animals were handled routinely prior to injection to allow reproducible baseline intraocular pressure measurements. Intraocular pressures following episcleral vein injections of hypertonic saline were determined in awake animals using one drop of 0±5 % proparacaine hydrochloride instilled in each eye. Animals were gently restrained with light hand pressure and tonopen measurements obtained using the above criteria. Readings were obtained in groups of 2 to 4 measurements, alternating between eyes to minimize influences of animal activity on either eye. The prominent globes in this species, along with their docile nature, allowed reproducible measurements without resorting to forceful lid manipulation. Following episcleral vein injections, mean intraocular pressures from 15 readings were obtained twice weekly. All animals were maintained in a 12 hr (0600 to 1800) light : dark cycle. Because baseline intraocular pressures varied among animals, but good agreement was always seen between the right and left eyes of normal rats, elevation of intraocular pressure (or intraocular pressure change), was determined for each measurement session by subtracting intraocular pressure of the non-injected eye from that of the injected eye. Mean intraocular pressure changes over the time of pressure elevation were calculated for each eye, ³ the standard error of the mean. All pressure values were expressed as Tono-Pen readings. Tissue Preparation and Histological Analysis Rats were anesthetized with halothane and perfused transcardially with 1 l of either 4 % paraformaldehyde or 5 % glutaraldehyde in 0±1  phosphate buffer (pH 7±2), following intracardiac injection of herparin (1 ml kg−") containing 10 mg ml−" sodium nitroprusside. At the time of perfusion, corneas were punctured with a 23 gauge needle to reduce the eye pressure and improve intraocular delivery of fixative. Eyes were enucleated, the superior cornea marked with a razor blade to preserve orientation and then immersed in fresh fixative for 12 hr. Optic nerve segments 1 mm from the back of the globe, optic nerve heads with adjacent retinas and anterior chamber angles were dissected, washed with fresh buffer, postfixed in 1 % OsO in 25 m potassium ferricyanide % and dehydrated with ascending ethanol and acetone, and embedded in Spurr’s low viscosity resin. For light microscopy, sections (1 µm) were cut on a Reichert Ultracut E microtome and stained with 1 % Toluidine blue or Richardson’s stain (Richardson, Jarret and Finke, 1960). For ultrastructural analysis of the optic

88

3. Results Intraocular Pressure Effects of Hypertonic Saline In normal eyes, Tono-Pen XL readings showed good linear correlation with transducer intraocular pressure from 10 to 70 mm Hg, with a regression line formula of y ¯ 7±4812­0±50834x (r ¯ 0±996). Saline injections of concentrations 2  and higher into episcleral veins produced marked elevations of intraocular pressure in many eyes, with excessive anterior segment inflammation in some. Subsequent

50 45 Average IOP (mm Hg)

nerve heads, tissues were post-fixed in 0±5 % uranyl acetate prior to dehydration, and Embed 812}Spurr’s was used for the embedding resin. Vertical longitudinal thin sections were cut, collected on 200 mesh grids and stained with uranyl acetate and lead citrate. These were viewed on a JEM-1200 EX TEMSCAN transmission electron microscope (JEOL, Tokyo, Japan) for evidence of organelle accumulation within individual nerve fibers. Montages of optic nerve cross sections from 20 eyes, including uninjected control eyes and injection controls, were prepared by photographing the semithin sections at 20¬ on a Zeiss Photomicroscope III and printing to a final magnification of 620¬. The identical sections were then examined twice at 100¬ under oil immersion, marking the location of swollen axons, and axonal and myelin debris on the montages. To check for consistency in identifying affected axons, regions of three of the nerves (two with elevated intraocular pressure and one control) were independently marked by a second investigator unaware of the nerve pressure history (BGG). Degenerating axons, characterized by axonal swellings and myelin debris, were identified and their location marked on the montage. Because it was difficult to determine the exact number of degenerating axons within areas of severe degeneration due to the amount of axonal debris, zones of degeneration were defined as areas where three or more affected axons were separated by no more than 10 µm. These zones were outlined on the montage by connecting the affected axons. The size of each zone was determined by the number and density of individually identified affected axons. For each nerve cross section, a semiquantitative measure of optic nerve injury was then determined by summing these zones, plus the areas of isolated degenerating or swollen axons, using a Bioquant system IV (R&M Biometrics) image analysis system. The total area of lesion was expressed as a percentage of the total optic nerve area. Areas occupied by optic nerve septa and blood vessels were not excluded from the total neural area because their subtraction had an insignificant effect on this percentage. The regional distribution of axonal damage was determined using the inferior location of the central retinal artery as a reference.

J. C. M O R R I S O N E T A L.

40 35 30 25 20 15 10 5 0

20 40 60 80 100 120 140 160 180 200 220 Day number

F. 3. Example of long term intraocular pressure course in an eye (OD) with prolonged, mildly elevated intraocular pressure, produced by episcleral injection with 1±85  hypertonic saline. Pressures are expressed as mean tonopen readings, without conversion to actual intraocular pressure. E, OD IOP ; D, OS IOP.

injections of 1±80  to 1±85  hypertonic saline without regard to the force of the injection, produced more moderate pressure rise in several eyes that persisted for prolonged periods of time. A representative intraocular pressure course in one of these eyes appears in Fig. 3. However, many other eyes had variable pressure responses and often demonstrated increased inflammation. Therefore, injection pressure was reduced to a force just sufficient to blanch the limbal artery. A total of 28 animals were injected with reduced force using either normal saline, 0±5 , 1±0 , 1±5 , 1±65 , or 1±75  hypertonic saline to establish a ‘ dose-response ’ relationship. Normal saline, 0±5 , 1±0  and 1±5  saline (two animals each) failed to produce a sustained elevation of intraocular pressure after 4 weeks, in spite of using a second injection through an inferior vein after the first 2 weeks. Table I summarizes our experience with injecting either 1±65  or 1±75  saline into episcleral veins with reduced force in the remaining 20 animals. In nine animals, a single injection produced elevated intraocular pressures. In seven animals, subsequent injections of either 1±65  or 1±75  saline resulted in intraocular pressure elevations. In three of these animals (R75, R78 and R81), elevated intraocular pressure readings were sporadic, resulting in the lowest mean elevations. One animal developed hypotony after a double injection with 1±75  saline. None of the remaining three animals demonstrated an elevation in intraocular pressure following either one or two injections. Further injections sufficient to produce an intraocular pressure rise were deliberately witheld in these animals to allow assessment of the effects of episcleral vein injection alone on the optic nerve. All animals with sustained intraocular pressure elevation were killed at predetermined times after

R A T P R E S S U R E-I N D U C E D O P T I C N E R V E D A M A G E

89

T I Intraocular pressure response to reduced-force hypertonic saline injection : summary of 20 consecutive eyes Animal no. R65 R66 R67 R68 R69 R70 R71 R72 R73 R74 R75 R76 R77 R78 R79 R80 R81 R82 R83 R84

First injection 1±75 M 1±75 1±75 1±65 1±65 1±65 1±65 1±75 1±75 1±75 1±75 1±75 1±75 1±75 1±75 1±75 1±75 1±75 1±75 1±75

Subsequent injection 1±75 M(¬2) — 1±75 1±65, 1±75 1±65, 1±75 — 1±75 — — — — — — — — 1±70 1±70 1±70 1±75 —

∆IOP³... (mm Hg)

Duration (days)*

18³7 28³12 28³13 7³4 20³5 9³7 24³8 8³3 9³6 14³5 3³3** No change No change 1³5** 12³6 No change 1³7** 7³2 ®5³4 15³5

36 32 7 7 35 10 12 34 11 7 21 — — 28 14 — 30 18 — 14

* Duration indicates period of intraocular pressure elevation prior to killing. ** Pressure elevations limited to one or more isolated readings.

T II Correlation of mean intraocular pressure change with percent area of optic nerve degeneration

F. 4. Fundus photograph of eye with mean IOP elevation of 10 mm Hg showing well perfused retinal blood vessels radiating out from the optic nerve head.

durations ranging from 7 to 38 days to allow us to document the effects of pressure elevation on the optic nerve. Tissue Effects of Sustained Intraocular Pressure Elevation All eyes demonstrated slight conjunctival inflammation and mild corneal haze for the first several days after injection. Conjunctival wounds healed rapidly and no infections were encountered. Once intraocular pressure elevation occurred, a slight enlargement of

Animal no.

∆IOP³... (mm Hg)

Duration (days)

Degeneration (percent area)

R72 OS* R82 OS* R76** R80** R82 OD R68 R72 OD R73 R79 R74 R84 R69 R71 R67

— — — — 7³2 7³4 8³3 9³6 12³6 14³5 15³5 20³5 24³8 28³13

— — — — 18 7 34 11 14 7 14 35 12 7

0±023 0±045 0±063 0±073 3±4 1±3 0±53 10±4 1±2 5±7 7±1 100 100 100

* Uninjected fellow eyes. ** Eyes without IOP elevation despite hypertonic saline injection.

the globe was seen in some eyes but there was no change in animal activity or alteration in normal weight gain pattern. Periodic, direct ophthalmoscopic inspection showed perfusion of the retinal vasculature in all eyes with elevated IOP (Fig. 4). In histologic examinations of optic nerve cross sections, control eyes and injected eyes without any measurable pressure elevations showed degeneration of only sporadic axons, measuring less than 0±1 % of the total neural area (Table II). Nerves from many

90

J. C. M O R R I S O N E T A L. S

N T

I

(A)

S

F. 5. Region of affected optic nerve 1 mm posterior to the globe shows light microscopic evidence of damaged axons, with axonal swelling (arrows) and myelin debris (arrowheads). ¬683.

N

T

I (B)

S

T N

F. 6. Optic nerve of animal with 5±7 % damage shows focal damage to axons in the superior temporal region *. ¬172.

injected eyes with elevated pressure readings showed only partial optic nerve damage. Mean intraocular pressure change of less than 10 mm Hg for 2 to 34 days produced focal optic nerve lesions (Figs 5 and 6). Similar lesions were observed in nerves from eyes with intraocular pressure change of 10 to 20 mm Hg for less than 3 weeks. Intraocular pressure change of 10

(C)

I

F. 7. Diagrams of optic nerve cross sections showing zones of damage and location of isolated abnormal axons in (A) a control eye with normal intraocular pressure (0±045 % damage), (B) R68, with pressure elevation of 7 mm Hg for 7 days (1±3 % damage) and (C) R74, with pressure 14 mm Hg above the fellow eye for 7 days (5±7 % damage). T, S, N and I designate temporal, superior, nasal and inferior regions of the nerve.

R A T P R E S S U R E-I N D U C E D O P T I C N E R V E D A M A G E

91

to 20 mm Hg for more than 3 weeks, or greater than 20 mm Hg for over 1 week, resulted in total involvement of the optic nerve, with occasional axons that appeared morphologically normal. Digitization of damaged areas in partially damaged nerve revealed lesions comprising 0±5 % to 10±4 % of the total neural area (Table II) and subsequent analysis of other nerves outside this series have shown more extensive partial damage. In 70 % of partially damaged nerves, lesions were focused in the superior temporal quadrant (Figs 6 and 7). Within the optic nerve head, eyes with early damage showed axonal swellings and increased cellularity that was greatest at the level of the lamina cribrosa (Fig. 8). Swollen axons were characterized by an accumulation of organelles, dense bodies and swollen mitochondria, indicative of blocked axonal transport [Fig. 8(C), and Fig. 9]. Eyes with partial nerve damage showed relatively little damage to the retina. However, in eyes with high pressures, the overall retinal thickness was reduced, due primarily to loss of the entire ganglion cell and the nerve fiber layers (Fig. 10). Histologic analysis of eyes with elevated intraocular pressure showed anterior synechiae and loss of the normal trabecular meshwork architecture (Fig. 11). Schlemm’s canal, the collector channels, and the veins of the limbal vascular plexus were patent, indicating that the main impediment to aqueous outflow was at the level of the trabecular meshwork. The ciliary process epithelium appeared normal.

4. Discussion Current understanding of the chronic pathology of glaucomatous optic nerve damage stems from the study of human glaucoma eyes and experimental nonhuman primate models (Gaasterland et al., 1978 ; Quigley et al., 1981 ; Quigley et al., 1983 ; Quigley and Green, 1979). This work has shown astrocyte proliferation, axon loss and accumulation of organelles, vesicles and mitochondria within remaining axons at the level of the lamina cribrosa, which is frequently compressed and posteriorly displaced. Since nearly all of these eyes have chronic damage, the precise relationships between these changes and the mechanisms of axon damage may be obscured by secondary effects due to loss of nerve fibers and other unknown factors.

F. 8. Light and electron micrographs of optic nerve head from R74. (A) Focal area of damage appears within the transition zone (outlined), demonstrating disorganization of nerve fiber bundles and increased cellularity. Anterior region of the ONH * appears relatively normal. (B) High power view of outlined region shows focal axonal swellings (arrows). (C) Swollen axon from same region appears packed with vesicular inclusions. A ¯¬334 ; B ¯¬582 ; C ¯¬6365.

92

More acute studies in primates have established that fast axonal transport, both anterograde and retrograde, can be blocked by elevated intraocular pressure (Anderson et al., 1974 ; Gaasterland et al., 1978 ; Quigley and Addicks, 1980). However, establishing the precise role of this and other phenomena in the eventual death of ganglion cells will require extensive correlations between eyes with acute and chronic exposure to elevated intraocular pressure using numerous animals. The expense of performing such studies on primates justifies the search for a more readily available laboratory animal. Histologically, the rat anterior optic nerve has a bottleneck configuration, and can be divided into three regions (Hildebrand, Remahl and Waxman 1985). The neck is most anterior. It is bordered by the choroid and sclera and is devoid of myelinated axons. A conical transition zone begins near the posterior edge of the sclera and gradually expands with increased myelination of the axons. Posteriorly, the optic nerve proper begins where the nerve is widest and fully myelinated. Despite its structural simplicity, the rat optic nerve head possesses several anatomic similarities to primates. Hildebrand described transverse vessels in the neck and transition zone of the rat nerve head (Hildebrand, Remahl and Waxman 1985). Ultrastructurally, these vessels have wide, collagen-filled perivascular spaces with astrocytic borders. The presence of these structures has been confirmed, and it has been found that they contain small amounts of elastin with abundant collagen I, III, and VI, and both chondroitin and dermatan sulfate-containing proteoglycans in the core, with laminin and collagen type IV along their borders (Morrison et al., 1995a). Based on their ultrastructure, and their nearly identical composition to primate laminar beams (Hernandez et al., 1987 ; Morrison et al., 1988 ; Morrison et al., 1989), it is hypothesized that these collagenous structures comprise the rat lamina cribrosa. As in the primate, astrocytes at the level of the rat lamina cribrosa appear specialized to support axons. In both species, astrocyte processes are packed with glial filaments and are oriented transversely, perpendicular to the axons (Anderson, 1969 ; ffrenchConstant et al., 1988). Using electron microscopy and freeze fracture techniques, Black has also shown that astrocyte processes at this level enter the nerve bundles where they intimately contact the unmyelinated axons (Black, Waxman and Hildebrand 1985). In our initial report, forceful injections of 2  hypertonic saline into episcleral vessels increased intraocular pressure in 60 % of eyes (Moore, Milne and Morrison 1993a). Some of these eyes had sustained elevations for as long as 6 months. However, many eyes had excessive inflammation and pressures were often much higher than desired. This prompted the change to a more dilute saline and less forceful injection pressure. These modifications have allowed

J. C. M O R R I S O N E T A L.

us to sclerose the trabecular meshwork to produce more moderate elevations in intraocular pressure, while simultaneously avoiding excessive inflammation and ciliary body damage. Sixteen of the 20 consecutive eyes presented here achieved intraocular pressure elevations in a range that produces identifiable nerve damage. Of the remaining four eyes, one became hypotonos. The others were enucleated, without further saline injections. None of the animals without documented pressure rise following injection showed any evidence of axonal degeneration greater than that seen in the normal controls. Because repeated injections do not diminish the usefulness of the model, success rates closer to 100 % could be anticipated by ‘ titrating ’ the response using multiple injections of 1±75  saline. While most eyes with elevated intraocular pressure had pressures greater than 7 mm Hg above that of the fellow eyes, three (R75, R78, and R81) showed mean intraocular pressure change of 3 or less. Unlike the others, the elevated pressures in these eyes were isolated readings, and were not sustained over the entire observation period. It is likely that a further saline injection would have produced a more consistent pressure elevation in these eyes as well. In the eyes with sustained pressure rise, the mean intraocular pressure change ranged from 7 to 28 mm Hg, with the majority between 7 and 20 mm Hg. Because this model relies upon scarring the trabecular meshwork and impeding aqueous humor outflow, the resulting elevation of intraocular pressure is dependent on many factors, including the degree of obstruction as well as the rate in which the eye produces aqueous humor. For these reasons, variability in intraocular pressure elevation is to be anticipated in our model, and, for that matter, is a recognized phenomenon in human glaucoma eyes (Drance, 1960 ; Katavisto, 1964). Preliminary analysis of selected aspects of the effects of elevated intraocular pressure on the rat optic nerve are included in this report, with more complete studies to follow. Several features common to primate models and human glaucoma warrant specific comment here. It has recently been determined that prolonged elevation of intraocular pressure in this model causes abnormal deposition of extracellular matrix (ECM) components within the nerve fiber bundles at the level of the nerve head (Johnson et al.). Similar depositions have previously been described in human glaucoma eyes and in the eyes of non-human primates with increased intraocular pressure (Fukuchi et al., 1991 ; Hernandez, Andrzejewska and Neufeld 1990 ; Morrison et al., 1990). They are not seen in eyes with optic nerve transection and appear to represent a pressure-related response (Morrison et al., 1990) unique to the optic nerve head. Detailed compliance studies of the optic nerve head in such eyes would need to be carried out to determine if these depositions correlate with altered physical behavior, as has been

R A T P R E S S U R E-I N D U C E D O P T I C N E R V E D A M A G E

93

F. 9. Electron micrographs of optic nerve head from R68. (A) Enlarged axons with marked accumulation of mitochondria and membrane-bound organelles appear in association with astrocytes (As). (B) Similar region from the normal fellow eye. ¬10 000.

shown to occur in humans and primates (Zeimer and Ogura, 1989). Such studies may provide a means to understand better the nature of this response and its role in glaucomatous optic nerve damage.

A full study of the ultrastructural pathology of optic nerve heads in these eyes will be presented in a subsequent report. However, preliminary electron micrographs show that axons within these nerve

94

J. C. M O R R I S O N E T A L.

F. 10. Retinal cross sections located approximately 200 microns from the optic nerve head in R65. Ganglion cells (arrows) and nerve fiber layer * are identifiable in the normal fellow eye (A), but absent in the experimental eye (B), with mean pressure change of 18³7 mm Hg for 36 days. Attrition of cells in outer layers may be due to periodic high pressures experienced by this eye (¬307).

heads are frequently swollen, with accumulation of vesicles, dense bodies and swollen mitochondria. These findings parallel similar observations in both human glaucoma eyes and chronic experimental primate models (Gaasterland et al., 1978 ; Lampert, Vogel and Zimmerman, 1968 ; Quigley and Addicks, 1980 ; Quigley and Addicks, 1981) and support the concept that chronically elevated intraocular pressure obstructs axonal transport in rats as well. Previous reports on the association of axonal transport and intraocular pressure in rat eyes are limited to acute experiments, where pressures of 35 and 50 mm Hg have been shown to obstruct retrograde transport in a reversible fashion (Johansson, 1986, 1988). The mechanism of this effect is yet to be determined. However, a mechanical compression due

F. 11. Anterior chamber angle from (A) a normal, uninjected eye illustrating trabecular meshwork * and Schlemm’s canal (arrow). (B) Angle from an experimental eye shows iris adhesions and loss of trabecular meshwork architecture. Schlemm’s canal (arrow) remains patent. (C) Another eye, at lower power shows marked thickening of the trabecular meshwork while Schlemm’s canal (arrow) and collector channels (arrowheads) remain patent. A and B, ¬526 ; C, ¬307.

R A T P R E S S U R E-I N D U C E D O P T I C N E R V E D A M A G E

to movement of laminar beams would appear to be unlikely, since the collagenous lamina in the rat is much less prominent than in the primate. The close association of swollen, vesicle-filled axons with astrocytes observed suggests that glial cells may play a role in this process. Although this has been proposed previously by Anderson and others (Anderson, 1969 ; Elkington et al., 1990), it is unknown if astrocytic changes are primary, or simply secondary responses to axonal injury. The eyes examined in this report all experienced relatively short durations of pressure elevation. Consequently, eyes with extensive optic nerve damage were those with higher degrees of IOP elevation, which may explain why retinal damage in these eyes was predominant in, but not completely limited to the ganglion cell and nerve fiber layers. Further analyses of eyes with less severe elevation of IOP over more prolonged periods would prove very useful in understanding further the mechanism of ganglion cell death, as well as the effects of pressure on the other retinal layers. Because the Tono-Pen XL was used in these studies, this instrument had to be calibrated for the rat eye. Although the relationship was linear with a good correlation (r ¯ 0±996), the slope of the regression line was less than what was reported with the Tono-Pen 2 (Moore et al., 1993a). This difference between instruments may reflect factory changes in design with the newer generation instrument and reinforces the need for re-calibration when beginning to use it on the rat eye. Given this correction, animals with intraocular pressure change greater than 20 mm Hg were likely experiencing actual mean intraocular pressures above 60 mm Hg. The almost complete nerve damage seen in these animals is likely to be equivalent to Quigley’s findings of extensive nerve damage in monkeys with intraocular pressure over 50 mm Hg, even for durations as little as 10 days (Quigley and Addicks, 1980). At the other end of the spectrum, pressures in the animals with the lower mean elevations convert to actual intraocular pressure of approximately 35 mm Hg. All of these nerves showed at least some identifiable, focal axonal damage. This compares with the work of Gaasterland, who noted that, of four monkey eyes with mean pressures of 26 to 32 mm Hg for 19 to 28 days, all had identifiable focal nerve fiber damage (Gaasterland et al., 1978). The preliminary correlations of IOP and nerve damage (Table II) show that this damage is, in general, influenced both by the degree and the duration of pressure elevation, although individual variation in pressure susceptibility may be present. More extensive analysis with greater numbers of eyes would be required to provide an accurate assessment of the relationship between the degree of intraocular pressure elevation and the amount of axonal pathology.

95

Finally, it is of extreme interest that all eyes with less than 100 % damage showed focal axonal degeneration, and that this was primarily located in the superior temporal region of the nerve. Regional susceptibility of optic nerve fibers is common in human glaucoma optic nerves, with a predilection for their superior and inferior poles (Quigley and Green, 1979). This has been correlated with relatively sparse laminar beams in these regions (Quigley and Addicks, 1981 ; Radius, 1981), suggesting less support or protection of nerve fibers. Recent work indicates that many of the laminar beams in the rat are primarily oriented vertically when viewed in cross section (Morrison et al., 1995a). It is possible that, in the rat as well as in the human, regional susceptibility to pressure-induced nerve damage may be due to the underlying anatomy of the nerve head. More detailed examination of the distribution and orientation of laminar beams and astrocytes in the rat optic nerve head, and their association with axon damage in this model will help us understand this phenomenon better and advance our understanding of how elevated intraocular pressure damages optic nerve fibers. Acknowledgement We thank Cindy Allen for assistance with tissue preparation for electron microscopy.

References Anderson, D. (1969). Ultrastructure of human and monkey lamina cribrosa and optic nerve head. Arch. Ophthalmol. 82, 800–14. Anderson, D. (1977). The management of elevated intraocular pressure with normal optic discs and visual fields, I, therapeutic approach based on high risk factors. Surv. Ophthalmol. 21, 479. Anderson, D. and Hendrickson, A. (1974). Effect of intraocular pressure on rapid axoplasmic transport in monkey optic nerve. Invest. Ophthalmol. Vis. Sci. 13, 771–83. Black, J., Waxman, S. and Hildebrand, D. (1985). Axo-glial relations in the retina-optic nerve junction of the adult rat : freeze-fracture observations on axon membrane structure. J. Neurocytol. 14, 887–907. Drance, S. (1960). Diurnal variation of intraocular pressure in treated glaucoma : significance in patients with chronic simple glaucoma. Arch. Ophthalmol. 70, 302–11. Elkington, A., Inman, C., Steart, P. and Weller, R. (1990). The structure of the lamina cribrosa on the human eye : an immunocytochemical and electron microscopical study. Eye 4, 42–57. ffrench-Constant, C., Miller, R., Burne, J. and Raff, M. (1988). Evidence that migratory oligodendrocyte-type2 astrocyte (O-2A) progenitor cells are kept out of the rat retina by a barrier at the eye-end of the optic nerve. J. Neurocytol. 17, 13–25. Fukuchi, T., Sawaguchi, S., Yue, B., Iwata, K., Hara, H., Shirakashi, M. and Kaiya, T. (1991). Sulfated proteoglycans in the lamina cribrosa of the optic nerve head in normal and in laser-induced glaucomatous monkey eyes. Invest. Ophthalmol. Vis. Sci. 32, 874. Gaasterland, D., Tanishima, T. and Kuwabara, T. (1978).

96

Axoplasmic flow during chronic experimental glaucoma. 1. Light and electron microscope studies of the monkey optic nerve head during development of glaucomatous cupping. Invest. Ophthalmol. Vis. Sci. 17, 838–46. Glovinsky, Y., Quigley, H. and Dunkelberger, G. (1991). Retinal ganglion cell loss is size dependent in experimental glaucoma. Invest. Ophthalmol. Vis. Sci. 32, 484–91. Hernandez, M., Andrzejewska, W. and Neufeld, A. (1990). Changes in the extracellular matrix of the human optic nerve head in primary open-angle glaucoma. Am. J. Ophthalmol. 109, 180–8. Hernandez, M., Luo, X., Igoe, F. and Neufeld, A. (1987). Extracellular matrix of the human lamina cribrosa. Am. J. Ophthalmol. 104, 567–76. Hildebrand, C., Remahl, S. and Waxman, G. (1985). Axoglial relations in the retina-optic nerve junction of the adult rat : electron-microscopic observations. J. Neurocytol. 14, 597–617. Johansson, J. (1986). Retrograde axoplasmic transport in rat optic nerve in vivo. What causes blockage at increased intraocular pressure ? Exp. Eye Res. 43, 653–60. Johansson, J. (1988). Inhibition and recovery of retrograde axoplasmic transport in rat optic nerve during and after elevated intraocular pressure in vitro. Exp. Eye Res. 46, 223–7. Johnson, E., Morrison, J., Farrell, S., Deppmeier, L., Moore, C. and McGinty, M. (1996). The effect of chronically elevated intraocular pressure on the rat optic nerve head extracellular matrix. Exp. Eye Res. 62, 663–74. Katavisto, M. (1964). The diurnal variations of ocular tension in glaucoma. Acta Ophthalmol. 78, 1–130. Lampert, P., Vogel, M. and Zimmerman, L. (1968). Pathology of the optic nerve in experimental acute glaucoma. Invest. Ophthalmol. 7, 199–213. Moore, C., Milne, S. and Morrison, J. (1993a). A rat model of pressure-induced optic nerve damage. Invest. Ophthalmol. Vis. Sci. 34, 1141. Moore, C., Milne, S. and Morrison, J. C. (1993b). Noninvasive measurement of rat intraocular pressure with the Tono-Pen. Invest. Ophthalmol. Vis. Sci. 34, 363–9. Moore, C., Epley, D., Milne, S. and Morrison, J. (1995). Longterm non-invasive measurement of intraocular pressure in the rat eye. Curr. Eye Res. 14, 711–7. Morrison, J., Dorman-Pease, M., Dunkelberger, G. and Quigley, H. (1990). Optic nerve head extracellular matrix in primary optic atrophy and experimental glaucoma. Arch. Ophthalmol. 108, 1020–4. Morrison, J., Farrell, S., Johnson, E., Deppmeier, L., Moore, C.

J. C. M O R R I S O N E T A L.

and Grossmann, E. (1995a). Structure and composition of the rodent lamina cribrosa. Exp. Eye Res. 60, 127–35. Morrison, J., Fraunfelder, F., Milne, S. and Moore, C. (1995b). Limbal microvasculature of the rat eye. Invet. Ophthalmol. Vis. Sci. 36, 751–6. Morrison, J., Jerdan, J., L’Hernault, N. and Quigley, H. (1988). The extracellular matrix composition of the monkey optic nerve head. Invest. Ophthalmol. Vis. Sci. 29, 1141–50. Morrison, J., L’Hernault, N., Jerdan, J. and Quigley, H. (1989). Ultrastructural location of extracellular matrix components in the optic nerve head. Arch. Ophthalmol. 107, 123–9. Quigley, H. and Green, W. (1979). The histology of human glaucoma cupping and optic nerve damage : clinicopathologic correlation in 21 eyes. Ophthalmology 86, 1803–27. Quigley, H. and Addicks, E. (1980). Chronic experimental glaucoma in primates. II. Effect of extended intraocular pressure elevation on optic nerve head and axonal transport. Invest. Ophthalmol. Vis. Sci. 19, 137–52. Quigley, H. and Addicks, E. (1981). Regional differences in the structure of the lamina cribrosa and their relation to glaucomatous optic nerve damage. Arch. Ophthalmol. 99, 137–43. Quigley, H., Adicks, E., Green, W. and Maumenee, A. (1981). Optic nerve damage in human glaucoma. II. The site of injury and susceptibility to damage. Arch. Ophthalmol. 99, 635–49. Quigley, H., Hohman, R., Addicks, E., Massof, R. and Green, W. (1983). Morphologic changes in the lamina cribrosa correlated with neural loss in open-angle glaucoma. Am. J. Ophthalmol. 95, 673–91. Quigley, H., Dunkelberger, G. and Green, W. (1988). Chronic glaucoma causing selectively greater loss of large optic nerve fibers. Ophthalmology 95, 357–63. Radius, R. (1981). Regional specificity in anatomy at the lamina cribrosa. Arch. Ophthalmol. 99, 478–80. Richardson, K., Jarret, L. and Finke, E. (1960). Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol. 35, 313. Sommer, A. (1989). Intraocular pressure and glaucoma. Am. J. Ophthalmol. 107, 186–8. Sommer, A., Katz, J., Quigley, H., Miller, N., Robin, A., Richter, R. and Witt, K. (1991). Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss. Arch. Ophthalmol. 109, 77–83. Zeimer, R. and Ogura, Y. (1989). The relation between glaucomatous damage and optic nerve head mechanical compliance. Arch. Ophthalmol. 107, 1232–4.