JOURNAL OF OCULAR PHARMACOLOGYVolume 5. Number 4. 1989Mary Ann Lieben, Inc.. Publishers

Ultrastructural Effects of Topical Betoptic,Betagan, and Timoptic on the Rabbit Corneal

EndotheliumG.S. LIU, G.E. TROPE, and P.K. BASU

Department of Ophthalmology, University of Toronto, Toronto, Ontario, Canada

ABSTRACT

Scanning and transmission electron microscopy was used toevaluate the cytotoxic effect of three commercially available betaadrenergic blockers (Betoptic 0.5%, Betagan 0.5% and Timoptic0.5%), and preservative (benzalkonium chloride 0.01%) on rabbitcorneal endothelium. Evaluation was performed on groups includingintact corneas, de-epithelialized corneas, full-thickness cornealgrafts with intact epithelium, and de-epithelialized cornealgrafts. Both Betoptic and Betagan produced minor damage toendothelial microvilli in the intact epithelial group. The de-epithelialized and the grafted-cornea groups had less damageinduced by Betoptic than Betagan. Timoptic produced moreendothelial damage than Betoptic and Betagan in all groups.Benzalkonium chloride produced minimal endothelial damage. Resultsindicate that Betoptic produces less endothelial toxicity than thetwo non-specific beta blockers when corneal epithelium is damagedand after penetrating keratoplasty.

INTRODUCTION

The ß-adrenergic blocking agents Betoptic (betaxololhydrochloride, a beta-1 specific blocker), Betagan (levobunololhydrochloride, a non specific beta blocker) and Timoptic (timololmaléate, a non specific beta blocker) are used to lower intraocularpressure in patients with various forms of glaucoma(1-4). However,Timoptic has been reported to damage rabbit corneal epithelium andendothelium(5). Previous comparative studies showed that Betopticwas the least toxic of the three drugs to regenerating rabbitcorneal epithelium(6,7). The purpose of this study was to comparethe toxic effects of topical Betoptic, Betagan and Timoptic oncorneal endothelium after both de-epithelialization and penetratingkeratoplasty (P.K.P.).

329

MATERIALS AND METHODS

Eyes from forty adult albino rabbits (New Zealand), eachweighing 3-3.5 kg, were randomly divided into 16 groups of 5 eyeseach, depending on the nature of the treatment they received(Table 1).

TABLE 1:Characteristics of the Sixteen Groups

Group*CornealEpithelium Drug

EpitheliumGroup of Autograft Drug

1

2

3

4

5

6

7

8

Intact

Intact

Intact

Intact

Removed

Removed

Removed

Removed

Placebo**

BetopticBetagan

TimopticPlacebo

BetopticBetagan

Timoptic

9

10

11

12

13

14

15

16

Intact

Intact

Intact

Intact

Removed

Removed

Removed

Removed

Placebo

BetopticBetagan

TimopticPlacebo

BetopticBetagan

Timoptic

* 5 eyes each group

** Placebo: Ophthalmic solution placebo (contains: activeingredient

-

none; preservative - benzalkonium chloride 0.01%) ascontrol

The following three commercially available agents werestudied: Betoptic 0.5% (each ml contains 5.6 mg betaxololhydrochloride and benzalkonium chloride 0.01%, donated by AlconLaboratories Inc., Fort Worth, Texas); Betagan 0.5% (containslevobunolol hydrochloride 0.5% with polyvinyl alcohol 1.4% andbenzalkonium chloride 0.004%, obtained from Allergan Inc, Quebec,Canada): Timoptic 0.5% (timolol maléate eguivalent to timolol 0.5%with benzalkonium chloride 0.01%, obtained from Merck Sharp & DohmeCanada, Merck Frosst Canada Inc., Quebec, Canada). The controlstudy was performed with ophthalmic solution placebo (contains:active ingredient - none; perservative - benzalkonium chloride0.01%, donated by Alcon Laboratories Inc.).

One drop of each drug or placebo was instilled topically inthe eye twice daily for 30 days.

Using a fine malleable iris spatula (E-706 storz) theepithelium was removed from a 7 mm diameter area in the centralregion of the cornea or corneal graft in groups 5 to 8 and 13 to16. A 7 mm corneal autograft was secured in place with 12interrupted 10-0 nylon sutures (Ethican TG 6-160) and all suturesremoved on the 20th day post operatively. In order to determinethe effects of P.K.P. on re-epithelialization, two extra eyes

330

underwent de-epithelialization before autograft. These two eyeswere not treated with either placebo or drugs.

The rabbit eyes were examined daily by slit-lamp microscopy.All animals were sacrificed on the 31st day post initiation oftreatment with an overdose of intravenous sodium pentobarbitol.Following enucleation, each eye was opened at the limbus. Thecornea was fixed in 2.5% glutaraldehyde for 4 hours at room

temperature and then divided into two equal pieces. For scanningelectron microscopy (SEM) one of the pieces was placed in 1% osmiumtetroxide in phosphate buffer for 1.5 hours at 20°C, and thendehydrated in ascending point drying followed by gold coating.Each specimen was examined under SEM (JSM - 35CF Scanning ElectronMicroscopy, Japan). For transmission electron microscopy (TEM),the central region of the other half of the cornea was cut intosmall pieces and placed in 1% osmium tetroxide in phosphate buffer.The specimen was dehydrated through ascending grades of ethanol andembedded in epoxy resin. Thin sections were stained with uranylacetate and lead citrate. The specimen was examined under H-700Transmission Electron Microscope (Hitachi Ltd., Tokyo, Japan).

RESULTS

Table 2 reviews the damage to corneal endothelium treated withthree beta adrenergic blockers and with preservative.

Pre-experiment Control Group

SEM of the two de-epithelialized grafted corneas treated withneither preservatives nor drugs showed normal endothelial mosaicpattern with fine microvilli. TEM exhibited normal endothelialultrastructure.

Placebo-treated Groups (Nos. 1. 5. 9 and 13)

The endothelium in both group 1 (the corneas with intactepithelium) and group 5 (the de-epithelialized corneas) showednormal mosaic pattern with fine microvilli (Fig. 1 A and B) . TEMexhibited normal endothelial ultrastructure (Fig. 2 A). In thegrafted groups (Nos. 9 and 13), SEM showed normal endothelialmosaic pattern except that some endothelial cells were noted tohave slight decrease in the number of microvilli, especially in thede-epithelialized grafted group (Fig. 1 C and D) . TEM revealednormal endothelial ultrastructure except for a few cells whichshowed minor changes (Fig. 2 B).

Betoptic-treated Groups (Nos. 2. 6. 10 and 14)

Both groups 2 (intact epithelium) and 6 (de-epithelializedcorneas) showed normal endothelial mosaic pattern with finemicrovilli, except for a few cells with minor modifications tomicrovilli (Fig. 3 A and B) . TEM revealed normal endothelialultrastructure except for a few small intracytoplasmic vacuoles(Fig. 4 A). SEM of the grafted corneas (groups 10 and 14), showednormal endothelial mosaic pattern with decreased number ofmicrovilli (Fig. 3 C and D). A few abnormal endothelial cells werenoted at the periphery of the graft (Fig. 3 C, arrows) . TEMrevealed normal endothelial ultrastructure except for some vacuolesand other minor changes (Fig. 4 B).

331

FIGURE 1, A-D: Scanning electron microscopy of cornea treatedwith placebo (contains benzalkonium chloride 0.01%) onedrop twice daily for 30 days. (A) The endothelium fromde-epithelialized cornea showing normal hexagonal mosaicpattern with fine microvilli (Bar = 42.0 ßm). (B)Magnified view of (A) (Bar = 13.6 ßm) . (C) Theendothelium from de-epithelialized graft showing normalappearance, G = Graft. (Bar = 100.0 ßm) . (D) Magnifiedview of (C) showing normal endothelial mosaic patternwith slight decrease in the number of microvilli in somecells (Bar = 10.0 ßm).

332

FIGURE 2, A-B: Transmission electron microscopy of cornea treatedwith placebo one drop twice daily for 30 days. (A)Endothelial cells from de-epithelialized cornea showingnormal ultrastructure (original magnification: X 18,000).(B) Endothelial cells from de-epithelialized graftshowing normal appearance except for minor changes,(original magnification: X 28,920).

333

FIGURE 3, A-D: Scanning electron microscopy of the cornea treatedwith Betoptic one drop twice daily for 30 days. (A) Theendothelial cells from de-epithelialized cornea showingnormal hexagonal mosaic pattern with fine microvilliexcept that a few cells had minor modifications ofmicrovilli (Bar = 42.0 ßm) . (B) Magnified view of (A)(Bar = 10.0 ßm) . (C) The endothelium from de-epithelialized graft showing virtually normal except fora few deficient cells at the peripheral region of thegraft (Bar = 100.0 ßm). (D) Magnified view of (C)showing normal endothelial appearance with slightdecrease in the number of microvilli in some cells (Bar =

13.6 ßm) .

334

TABLE 2:Results of the Sixteen Groups

Degree of endo- Degree of endo-Group thelial damage *** Group thelial damage

SEM TEM SEM TEM

1 - - 9 ± ±

2 - 10 + +

3 - - 11 + + to ++

4 - to + ± 12 ++ + to ++

5 - - 13 ± ±

6 - - to + 14 + +

7 - to + - to + 15 + to ++ + to ++

8 + + 16 ++ to +++ + to +++

*** - : Normal endothelial mosaic pattern with fine microvilli oronly a few cells showing minor damage to microvilli(SEM). Endothelial cells with normal ultrastructure andthickness (TEM).

± : Decreased endothelial microvilli or minor damage tocytoplasmic membrane (SEM). Minor endothelialultrastructural change (TEM).

+ : A few mildly edematous endothelial cells. In graftedgroups, only a few damaged cells at the edge of the graft(SEM). Endothelial cells exhibited a fewintracytoplasmic vacuoles and/or other slight butnoticeable ultrastructural damage (TEM).

++ : Abnormal corneal endothelium. In grafted groups, somedamaged cells distributed over the edge and centralregion of the graft (SEM). The endothelial cellsexhibited moderate vacuole formations (TEM).

+++: Severely disrupted endothelial cells at edge and centralregion of the graft (SEM). Moderate to large vacuoleformation and increased cellular thickness (TEM).

335

FIGURE 4, A-B: Transmission electron microscopy of the corneatreated with Betoptic one drop daily for 30 days. (A)Endothelial cells from the de-epithelialized cornea

showing normal ultrastructure with minor modifications(original magnification: X 16,758). (B) Endothelialcells from the de-epithelialized graft showing normalendothelial ultrastructure except for some smallintracytoplasmic vacuoles (original magnification:X 13,680).

Betacran-treated Groups (Nos. 3. 7. 11 and 15)

SEM of the endothelium in both groups 3 (intact epithelium)and 7 (the de-epithelialized corneas) showed normal mosaic patternwith fine microvilli except for a few cells with minormodifications to microvilli (Fig. 5 A). Only a few endothelialcells from one de-epithelialized cornea had slight edema. TEMrevealed normal endothelial ultrastructure except for a few smallvacuoles (Fig. 6 A) . In the grafted corneas (groups 11 and 15) ,

SEM showed normal endothelial mosaic pattern with decrease in thenumber of microvilli and a few damaged cells located in theperipheral region of the graft (Fig. 5 B arrows). One cornea fromthe de-epithelialized grafted group had damaged cells distributedover the edge and the central region of the graft (Fig. 5 C and D).TEM exhibited normal endothelial ultrastructure except for slightto moderate intracytoplasmic vacuole formation (Fig. 6 B).

336

FIGURE 5, A-D: Scanning electron microscopy of the cornea treatedwith Betagan one drop twice daily for 30 days. (A)Endothelial cells from de-epithelialized cornea showingnormal hexagonal mosaic pattern (Bar = 42.0 /¿m). (B) Theendothelium from de-epithelialized graft showing normalappearance; note a few damaged cells at the peripheralregion of the graft (Bar = 150.0 ßm) . (C) Theendothelium from another de-epithelialized graft showingsome damaged cells in the central region of the graft(Bar = 100.0 ßm) . (D) Magnified view of (C) (Bar =

13.6 ßm).

Timoptic Treated Groups (Nos. 4. 8. 12 and 16)

SEM of the endothelium in group 4 (intact epithelium) wasvirtually normal except for a few cells from one cornea with slightdamage to cytoplasmic membranes (Fig. 7 A) . Endothelial TEMconfirmed the overall normal appearance of the cells. SEM of de-epithelialized corneas (group 8) revealed normal endothelial mosaicpattern except for a few endothelial cells with slight edema. Onecornea had abnormal central corneal endothelial cytoplasmicmembranes (Fig. 7 B). TEM revealed normal ultrastructure exceptfor slight alteration to cytoplasmic membranes (Fig. 8 A) . Thegrafted corneas (Nos. 12 and 16) had some damage to endothelialcells at the edge of the graft. Some damage was occasionally seenin the central region of the graft, particularly in the de-

337

FIGURE 6, A-B: Transmission electron microscopy of the corneatreated with Betagan one drop twice daily for 3 0 days.(A) Endothelial cells from the de-epithelialized cornea

showing normal ultrastructure except for a few smallvacuoles (original magnification: X 13,680). (B)Endothelial cells from the de-epithelialized graftshowing normal ultrastructure except for some small and afew moderate intracytoplasmic vacuoles (originalmagnification X 13,680).

epithelialized grafted group (Fig. 7 C and D) . TEM revealed a fewendothelial cells with moderate to larger intracytoplasmicvacuoles, increased thickness, or other intracellular organellechanges (Fig. 8 B).

DISCUSSION

The results of this study indicate that over a 30-day period,topical benzalkonium chloride 0.01% administration caused minimaldamage to rabbit corneal endothelial microvilli. After full-thickness corneal grafts however, preservative caused a slight butnoticeable decrease in the number of endothelial microvilli. Thischange was not seen in the pre-experimental control group with de-epithelialized grafts. Benzalkonium chloride is toxic to cornealendothelium(8,9). Perfusion with benzalkonium chloride 0.01%, evenat very low concentrations (6.5 x 10-4%) , has been reported tocause marked endothelial swelling(lO). However, as indicatedabove, topical application of preservative did not cause severeendothelial damage to the grafted corneas.

338

FIGURE 7, A-D: Scanning electron microscopy of the cornea treatedwith Timoptic one drop twice daily for 30 days. (A)Endothelium from the cornea with intact epitheliumshowing slight change of cytoplasmic membrane in somecells (Bar = 13.6 ßm) . (B) Endothelium from one de-epithelialized cornea showing a few cells with abnormalcytoplasmic membrane at the central region of the cornea

(Bar = 13.6 ßm) . (C) Endothelium from de-epithelializedgraft showing some marked disruptive cells at the centralregion of the cornea (Bar = 150.0 ßm) . (D) Magnifiedview of (C) (Bar =13.6 ßm).

All three beta adrenergic blockers caused some ultrastructuralchanges as shown in Table 2. In both the intact and de-epithelialized groups, Betoptic and Betagan produced minor damageto endothelial microvilli, similar to that seen with benzalkoniumchloride 0.01%. In addition, these drugs also produced slightendothelial intracytoplasmic vacuole formation. One cornea treatedwith Betagan (from the de-epithelialized group) also had a fewslightly edematous endothelial cells. Timoptic, however, producedmore noticeable endothelial toxicity with slight cellular edema andcytoplasmic membrane alteration; one cornea from the de-epithelialized group had very abnormal cytoplasmic membranes.

In the grafted groups, Betoptic caused slight morphologicaldamage to a few endothelial cells at the edge of the graft.Betagan also caused minor but noticeable damage to graftendothelium as evidenced by a few damaged cells at the edge of thegraft and a few intracytoplasmic vacuoles; furthermore some

339

^•¿^ "/"

FIGURE 8, A-B: Transmission electron microscopy of cornea treatedwith Timoptic one drop twice daily for 30 days. (A)Endothelial cells from the de-epithelialized cornea

showing normal ultrastructure except for slight change ofthe cytoplasmic membrane (original magnification: X13,680). (B) Endothelial cells from the de-epithelialized graft showing slight edema, thincytoplasma and vacuole formations, and increased cellularthickness (original magnification: X 13,680).

cellular damage was seen at the central region of the graft in one

de-epithelialized cornea. Timoptic overall produced more damage tocells at both the edge and central region of some grafts comparedwith the other two drugs. Moderate to large intracytoplasmicvacuoles and increased cellular thickness were noted. Althoughhistological evidence of slight endothelial damage was noted inthis and a previous study of ours(5), there was no clinicallysignificant damage to the grafted corneas induced by drug therapy,i.e. all corneas remained clear. It is possible, however, thatlong term treatment with timolol maléate post P.K.P. may predisposeto clinically significant damage. Long term studies on post P.K.P.eyes treated with beta blockers are, therefore, indicated.

340

This study confirms that Timoptic produces ultramicroscopyevidence of rabbit corneal endothelial damage that is more

pronounced than seen with Betagan and Betoptic. Betoptic producedthe least toxicity of the three drugs. Why this should be is notclear but factors have been discussed elsewhere in relation tocorneal epithelial damage which may apply to endothelium as

well(7) .

In conclusion, the results of this study indicate that topicaltreatment with Betoptic for one month produces minimal toxicity torabbit corneal endothelium. Betoptic may be the beta blocker ofchoice for patient use after penetrating keratoplasty.

ACKNOWLEDGEMENTS

This work was supported by grants from the Medical ResearchCouncil of Canada, The Lions' International District A-16Ophthalmological Research Foundation, Independent Order of OddFellows and Rebekahs of Ontario, Canadian National Institute forthe Blind out of the E.A. Baker Foundation for the Prevention ofBlindness, Toronto, Canada, and Alcon Laboratories Inc., Texas,U.S.A.

We thank Mrs. Elena Walker, Mr. Stan Niewojt, and Ms. JoanMacneil, Department of Ophthalmology, and Mr. Battista Calvieri,Electron Microscopy Laboratory, Faculty of Medicine, University ofToronto, for their valuable technical assistance, and Dr. T.A.McDonald from Alcon Laboratories Inc. Texas (U.S.A.) for the supplyof drugs and placebo.

REFERENCES

Berrospi, A.R. and Leibowitz, H.M. Betaxolol - A new beta-adrenergic blocking agent for the treatment of glaucoma.Arch. Ophthalmol., 100:943-946, 1982.

Stewart, R.H., Kimborough, R.L., Ward, R.L. Betaxolol vstimolol: a six-month double blind comparison. Arch.Ophthalmol., 104:46-49, 1986.

Berson, F.G., Cohen, H.B., Foerster, R.J., Lass, J.H., Novack,G.D., Duzman, E. Levobunolol compared with Timolol for thelong-term control of elevated intraocular pressure. Arch.Ophthalmol., 103:379-382, 1985.

Zimmerman, T., Kaufman, H.E. Timolol, a beta adrenergicblocking agent for the treatment of glaucoma. Arch.Ophthalmol., 95:601-604, 1977.

Liu, G.S., Basu, P.K., Trope, G.E. Ultrastructural changes ofthe rabbit corneal epithelium and endothelium after timoptictreatment. Graefe's Arch. Ophthalmol., 225:325-330, 1987.

Liu, G.S., Trope, G.E., Basu, P.K. A comparison of topicalBetoptic and Timoptic on corneal re-epithelialization inrabbit. J.Toxicol. Cut. & Oc. Toxicol., 6(4):335-343, 1987.

Trope, G.E., Liu, G.S., Basu, P.K. Toxic effects of topicallyadministered Betagan, Betoptic and Timoptic on regeneratingcorneal epithelium. J.Oc. Pharmacol., 4(4): 359-366, 1988.

Gasset, A.R., Yasuo, I., Kaufman, H.E., Miller, T.

341

Cytotoxicity of ophthalmic preservatives. Am. J.Ophthalmol.,78(1):98-105, 1974.

9. Lemp, M.A. and Zimmerman, L.E. Toxic endothelial degenerationin ocular surface disease treated with topical medicationscontaining benzalkonium chloride. Am.J. Ophthalmol.,105(6):670-673, 1988.

10. Hull, D.S. Effects of epinephrine, benzalkonium chloride andintraocular miotics on corneal endothelium. South. Med. J.,72(11):1380-1381, 1979.

Received: August 16, 1989Accepted for Publication: October 16, 1989

Reprint Requests: Dr. P.K. BasuDepartment of OphthalmologyUniversity of Toronto1 Spadina CrescentToronto, Ontario M5S 2K6CANADA

342