Steroid-Induced Ocular Hypertension in Normal Cattle

Objectives  To determine whether the bovine eye develops elevation of intraocularpressure (IOP) in response to topical corticosteroid use and to develop areliable model of steroid-induced elevation of IOP in an animal.

Methods  Intraocular pressure was monitored by Perkins applanation tonometryin a group of 12 cows receiving topically administered prednisolone acetatein 1 eye 3 times a day for a period of 49 days after the establishment ofbaseline IOP values. Perkins readings were converted to IOP in mm Hg usingcalibration curves derived from in vitro cannulation manometric experimentsand validated with in vivo manometric measurements. Intraocular pressure wasalso monitored for 50 days after the discontinuation of corticosteroid therapy.

Results  Intraocular pressure began to increase after 3 weeks of treatment in100% of the cow eyes receiving corticosteroid and reached a peak 1 week later.Peak interocular IOP differences between the corticosteroid-treated eye andthe fellow control eye reached up to 15 mm Hg and began to decline after thediscontinuation of treatment but remained significantly elevated for a periodof 3 more weeks.

Conclusions  Bovine eyes exhibit a robust steroid-induced ocular hypertensive response,with 100% occurrence in this trial. The IOP elevation caused by corticosteroidslowly subsides after discontinuation of treatment.

Clinical Relevance  The mechanisms of steroid-induced glaucoma may be related to those involvedin primary open-angle glaucoma and could provide the clues to elucidate thepathogenesis of the latter. The high prevalence of corticosteroid-inducedelevation of IOP in the cow and the large amount of tissue available willpermit studies on the mechanism of this phenomenon not previously possible.

Glucocorticosteroids, administered by a variety of routes, may elevateintraocular pressure (IOP).¹ This induced ocularhypertension,which generally occurs within weeks in susceptible individuals, is usuallyreversible. However, it may produce glaucomatous optic neuropathy if the durationof corticosteroid therapy is lengthy.² Theopen-angle glaucoma so produced resembles primary open-angle glaucoma, includingthe mechanism by which IOP is elevated, a decrease of trabecular meshworkoutflow facility.³ It has been suggested thatthere is a distinct and common relationship between this ocular response tocorticosteroids and primary open-angle glaucoma.⁴ Therefore,understanding the cellular processes that lead to corticosteroid-induced ocularhypertension may illuminate the cause of primary open-angle glaucoma. Unfortunately,these processes remain elusive.

In humans, the elevation of IOP induced by corticosteroids appears tooccur in only about one third of those tested, with an even smaller percentexhibiting marked elevations.⁵ The search forother animal species in which corticosteroids elevate IOP underscores theneed to find animal models in which the investigation of pathogenesis canbe carried out. Corticosteroids will induce ocular hypertension in rabbits,but the results are quite variable, and the amount of corticosteroid necessaryis often lethal.⁶ Moderate elevations of IOPmay be produced in cats by topical administration of corticosteroids.⁷ Corticosteroid-induced ocular hypertension may beinduced in cynomolgus monkeys as well; in one series, 5 of 11 monkeys treatedwith topical dexamethasone showed increased IOP.⁸

Many investigations have been reported that study the effects of glucocorticoidson the trabecular meshwork in an attempt to unravel the family of diseasescalled glaucoma.⁹ These include explorationsin cultured trabecular meshwork cells, organ-cultured eyes, and in vivo models,investigating morphology, gene expression, extracellular matrix, cytoskeleton,and cell adhesion molecules, to name a few. Most often, human trabecular meshworkcells are utilized; sometimes other species' cells are used, rarely the wholeeyes such as those of the rabbit. The cow eye may be ideal for these studiesbecause of similarities with the human eye. For example, recent investigationsreport that the concentration of chloride in the bovine aqueous humor is higherthan that in plasma¹⁰ and that isolated bovineciliary epithelium transports chloride and is inhibited by carbonic anhydraseinhibitors,¹¹ as in humans. Dexamethasone treatmentof cultured bovine trabecular meshwork cells does produce alterations in extracellularmatrix proteins and cell contacts.¹² However,although IOP of normal cattle has been reported,^13-15 itis unknown whether cow eyes are corticosteroid responsive, although they wouldprovide large amounts of tissue material for further studies. Thus, a studywas carried out to explore the IOP response of the bovine eye to the administrationof a topical corticosteroid.

All animal experiments were performed according to the Association forResearch in Vision and Ophthalmology (ARVO) guidelines. Twelve healthy (female)cows between 3 and 5 years of age and weighing 350 to 420 kg were selectedfrom a local ranch in Corrientes, Argentina, for the corticosteroid study.They were of a common type in Argentina named Braford, a cross between Brahmanand Hereford. Two of the cows were pregnant, and 1 lost the fetus during thestudy because of unrelated reasons, without affecting the health of the mother.The cows were tagged for individual identification on their ear lobes. Theywere herded from pasture whenever it was necessary to instill the drops orto measure IOP. They were guided into a funnel corral ending in a loose-fittingyoke (cepo). This allowed movement and holding ofthe head by one person while another instilled the drops (Figure 1). With time, the cows became accustomed to the routine,and drops could be instilled while they were in the open field. To take theIOP, the cows were guided into the funnel corral and then into the neck yoke.This procedure took about 4 minutes per cow. Otherwise, the cows were freeto pasture. The other set of 8 untreated cows of various breeds that underwentIOP measurement were selected at random, and IOPs were measured just priorto sacrifice at the local slaughterhouse.

After 1 baseline measurement of IOP (cows 1-6) or 2 such measurements1 week apart (cows 7-12), 1% prednisolone acetate (Falcon Pharmaceuticals,Fort Worth, Tex) or 0.5% prednisolone acetate (Ultracortenol; Novartis Ophthalmics,Hettlingen, Switzerland) was instilled in 1 eye. (The study was initiatedwith the 1% prednisolone solution but was changed to the 0.5% solution forreasons of availability after 1 week of treatment.) As a control, an artificialtear preparation (Alcon Lagrimas II; Alcon Argentina, Tortuguitas, Argentina)was instilled in the contralateral eye. Both control and experimental instillationsconsisted of 2 drops, 3 times daily at 8 AM, 2 PM, and 6:30 PM for the duration of intervention. The plastic bottles containingthe drops were covered with a tape, either red (artificial tears) for thecontrol eyes or green (prednisolone) for the drug-treated eyes, thus maskingthe identity of the agent administered. In addition, the IOP in both eyesof another group of 8 cows was measured to determine the normal cow IOP.

One of the authors taught the cowboys who were in charge of the cowshow to instill the drops, which was done mostly in the field. They were giventhe 2 types of drops as described above from a different investigator withsealed instructions on what "color drop" to instill to right or left eyes,how many times, and when. The cowboys did not know the contents of the bottles.

Once the cow was held in the yoke and a cowboy moved the cow's headto a proper orientation, the ophthalmologist measured the IOP with the Perkinstonometer. Before the IOP measurement, 2 drops of topical 0.5% proparacaine(Alcon Argentina) followed by 2 drops of 0.25% fluorescein were instilled.Two sets of measurements were taken on each eye alternating first one eyeand then the other (Figure 2). Theophthalmologist measuring the IOP was unaware of the treatment of each eye.All IOP measurements were taken between 8:30 AM and 10 AM at least once a week. Although IOPs were taken at various intervals,the drugs were applied 3 times daily during the 49-day treatment period.

A manometric measurement by cannulation was performed in 4 eyes of the12 cows while in the study. The purpose was to determine the actual IOP inthese cows and to use these values to compare with those obtained with thePerkins tonometer in vivo as well as to confirm the calibration of the Perkinstonometer done in isolated cow eyes. With the cows in the yoke, prior to cannulation,3 drops of 0.5% proparacaine were instilled in the eye. About 2 minutes later,with the head held by hand, a 25-gauge butterfly needle was introduced intothe anterior chamber (Figure 3).The butterfly's tubing was connected to a custom-made pressure-recording instrument.This instrument consisted of 3 parts: an inline pressure transducer (Ohmedamodel TNF-R; Ohmeda Ltd, Singapore), a custom-made amplifier, and a high-impedancecustom-made millivolt meter. The pressure transducer was connected to thebutterfly tubing through a valve. The signal of the transducer was fed intothe amplifier. The amplifier was calibrated against a column of water connectedto the transducer so that its output in millivolts corresponded to pressurein mm Hg. The amplifier output was read on the liquid crystal display screenof the voltmeter.

Four cow eyes were obtained from the slaughterhouse. Eyes were transportedto the laboratory on ice immediately after enucleation. All globes were verifiedto be intact with visually clear corneas. The eyes were cannulated with a26-gauge needle at 90° to the visual axis through clear cornea 1 to 2mm anterior to the limbus with the aid of an operating microscope. The absenceof leaks was verified microscopically throughout the experiment. Intraocularpressure was controlled by adjusting the height of a variable column of balancedsalt solution attached to the needle (open stopcock method). Intraocular pressurewas verified and continuously recorded by a pressure transducer (Ohmeda modelTNF-R) connected to a second cannulation needle inserted in the anterior chamberin a similar fashion, 180° away from the previous one. Intraocular pressurewas sequentially adjusted to 15-, 25-, 35-, 45-, and 55-cm water pressures(10 mm Hg = 13.6 cm H₂O pressure). The eyes were supported in asmall cup, and the cornea was applanated after application of fluoresceinsolution. A Perkins handheld applanation tonometer with a clinically usedGoldmann applanation tip was used. The instrument was powered off after eachmeasurement. Five measurements were made at each pressure level, and the meanwas calculated. The readings obtained were plotted against the manometric(true) IOP (after converting to millimeters of mercury values), as shown in Figure 4.

The IOP measurements in both eyes of 8 normal cows were used to determinethe baseline values of IOP in these animals (Table 1). The corresponding Perkins tonometer reading and the equivalentIOP as determined from the Perkins calibration curve (Figure 4) indicated a normal IOP of between 16 and 17 mm Hg.

The Perkins IOP measurements during the course of the experiment (correctedfrom the calibration curve in Figure 4),as depicted in Figure 5 in meanabsolute values and in Figure 6 asIOP differences (ΔIOP) between the corticosteroid-treated eye and thecontralateral eye of each animal, showed substantial elevations of IOP inthe eyes treated with corticosteroid. The ΔIOP was significant betweendays 28 and 77 (21 to 70 days after the onset of steroid administration) (P<.001, analysis of variance; P<.05,Bonferroni test). The IOP remained elevated at the same level between days35 and 56 (days 28-49 on steroid) (P>.05, Bonferronitest).

To confirm the accuracy of the elevated IOP measurements obtained withthe Perkins tonometer in the experimental cows, IOP was determined manometricallyin vivo in 4 eyes (the eyes of cows 4 and 7 receiving corticosteroid dropsand cows 1 and 8 receiving artificial tear drops) on 2 occasions (occasion1, day 67 for cows 1 and 4 and day 60 for cows 7 and 8; occasion 2, day 84for cows 1 and 4 and day 77 for cows 7 and 8) immediately after measurementwith the Perkins tonometer (Table 2).The IOP measured by cannulation manometrically was not significantly differentfrom the IOP determined by calibrated Perkins applanation tonometry (P>.11, Wilcoxon signed rank test). In addition, the manometricIOP correlated very well (P<.001; R² = 0.88) with the IOP as determined by Perkins tonometrycorrected from the calibration curve.

Intraocular pressure in the cow has been previously reported. In onestudy,¹⁵ using the MacKay-Marg tonometer, meanIOP values of 27.5 and 28.2 mm Hg were obtained. Mean IOP with the TonoPenwas 26.9 mm Hg¹⁵; however, with both instruments,there was a large range of values (12-42 mm Hg). Another study¹⁴ reporteda mean IOP of 23.4 mm Hg with the TonoPen. However, both studies did not correlateIOP measurements with these instruments with actual IOP as measured by cannulation.Using Perkins handheld applanation tonometry, the readings in the normal coweye are numerically very low. Using a calibration curve derived from in vitromanometric measurements, the control, presumably normal, mean of IOP in thecow is 16 to 17 mm Hg. The reliability of these calibration curves is confirmedby the in vivo cannulations performed on some of the experimental eyes. Corticosteroiddrops produce a substantial elevation of IOP in cow eyes. This occurs rapidlyover a few days during the third to fourth weeks of treatment and is subsequentlymaintained during corticosteroid administration. Peak IOPs of 30 to 35 mmHg are typical in all cow eyes so treated. The IOP slowly reverts to normalover 4 to 5 weeks after discontinuation of corticosteroid administration.A small rise of IOP occurs in the fellow eyes of these cows, the cause ofwhich will require additional study. The intensity and maintenance of thissteroid response in all the cows treated with corticosteroid drops are unlikethose in any other species previously tested. Moreover, the large volume ofocular tissue that can be made available from the cow represents a remarkableresource for further studies of this phenomenon.

Further studies of cytoskeleton, extracellular matrix, receptors, integrins,and growth factors certainly can be carried out by comparing the corticosteroid-treatedand untreated fellow eyes.⁹ The availabilityof this model indicates that tissue-cultured cells may no longer be necessaryfor these investigations. Limitations on prior studies using whole tissue,such as those indicating that there are changes in the aqueous pathway glycosaminoglycansin rabbit eyes treated with dexamethasone,¹⁶ willbe obviated by this model.

This study establishes the bovine eye as a reliable and reproduciblein vivo model for steroid-induced glaucoma. We have defined a dosage level,dosage frequency, and rate of onset to reach a doubling of IOP, as well asthe rate of return to normal IOP when treatment is discontinued. This willenable other investigators to use this model for further studies. Such studiescould include experiments in bovine eyes with increased IOP to determine thedynamics of aqueous humor flow. In addition, the large amounts of tissue availablecan allow investigations of gene and protein expression^17,18 thatmay shed light on the mechanisms involved in steroid-induced glaucoma and,potentially, primary open-angle glaucoma.

Correspondence: Steven M. Podos, MD, Mount Sinai School of Medicine,One Gustave L. Levy Place, Box 1183, New York, NY 10029 (steven.podos@mssm.edu).

Submitted for publication September 3, 2003; final revision receivedJanuary 29, 2004; accepted February 2, 2004.

This study was supported in part by an unrestricted grant, a Sybil S.Harrington Scholar Award (Dr Danias), and a research sabbatical grant (DrCandia) from Research to Prevent Blindness Inc, New York, NY, and grants EY01867,EY11649, R03 EY13732, and K08 EY00390 from the National Institutes of Health,Bethesda, Md.