Exp. Eye Res. (2000) 71, 225±232doi:10.1006/exer.2000.0872, available online at http://www.idealibrary.com on

Caspase Inhibitor z-VAD-FMK Inhibits Keratocyte Apoptosis, butPromotes Keratocyte Necrosis, after Corneal Epithelial Scrape

WOO-JUNG KIMa, RAHUL R. MOHANb, RAJIV R. MOHANb AND STEVEN E. WILSONb*

aDepartment of Ophthalmology, College of Medicine, Sungkyunkwan University, Seoul, Korea andbDepartment of Ophthalmology, University of Washington School of Medicine, Box 356485, Seattle,

WA 98195-6485, U.S.A.

the anter(Wilson e

0014-483

* Author

(Received Lund 8 July 1999 and accepted in revised form 10 April 2000)

The purpose of this study was to determine whether the caspase inhibitor z-VAD-FMK could be appliedtopically prior to epithelial scrape injury to inhibit keratocyte apoptosis. Rabbit corneas were treated withz-VAD-FMK or vehicle alone prior to epithelial scrape injury. Cell fate was analysed at 4 hr after epithelialscrape using quantitative TUNEL assay, propidium iodide staining, and transmission electron microscopy.Less stained anterior stromal keratocytes were detected with the quantitative TUNEL assay in corneas pre-treated with z-VAD-FMK than in corneas pretreated with vehicle at 4 hr after epithelial scrape. Thisdifference appeared to be con®rmed by propidium iodide staining of keratocyte nuclei. It was observed thatfewer nuclei were stained with propidium iodide in the DMSO vehicle treated corneas compared to the z-VAD-FMK treated corneas. Analysis of corneas with transmission electron microscopy, however, indicatedthat many anterior stromal keratocytes in corneas pretreated with z-VAD-FMK, but not vehicle, had cellmorphologic changes more consistent with necrosis. Although pretreatment of corneas with the caspaseinhibitor z-VAD-FMK inhibited keratocyte apoptosis detected with the TUNEL assay, transmission electronmicroscopy revealed that many anterior stromal keratocytes in z-VAD-FMK-treated corneas instead died bynecrosis. Thus, z-VAD-FMK is unlikely to be useful to modulate corneal would healing through inhibition

of keratocyte apoptosis induced by epithelial injury. The TUNEL assay should not be used to monitor cellfate without con®rmation using analyses that also detect necrosis. # 2000 Academic Press

Key words: apoptosis; necrosis; caspases; cornea; wound healing.

1. Introduction

Loss of anterior keratocytes following corneal epi-thelial injury is mediated by apoptosis (Wilson et al.,1996a; Wilson, 1997). Apoptosis is a gentle involu-tional form of controlled cell death associated withcharacteristic ultrastructural changes such as cellshrinkage, chromatin condensation and fragmenta-tion, and budding of apoptotic bodies containing cellcontents such as mitochondria and lysosomes fromthe cell membrane. It has been hypothesized thatkeratocyte apoptosis is an important initiating event inthe wound healing cascade that affects the outcome ofvarious corneal surgeries, including refractive pro-cedures such as photorefractive keratectomy (PRK)and laser in situ keratomileusis (LASIK) (Wilson et al.,1996b; Helena et al., 1998; Wilson and Kim, 1998).Studies have suggested that keratocyte apoptosis ismediated by interleukin-1 alpha (Wilson et al.,1996a), Fas ligand (Wilson et al., 1996b; Mohanet al., 1997), and other cytokines (Mohan et al., 1998)from the injured corneal epithelium or induced bycytokines released from the injured corneal epi-thelium. Following super®cial keratocyte apoptosis

ior stroma is repopulated within a few dayst al., 1996a). Recent studies have suggested

5/00/090225�08 $35.00/0

for correspondence.

that the repopulating cells may be activated bymyo®broblast-like cells that are associated withincreased collagen and glycosaminoglycan depositionand collagen disorganization (Jester et al., 1996;Masur et al., 1996). These repopulating cells areprobably derived through mitosis and migration ofremaining posterior and peripheral keratocytes (ZieskeJD, unpublished data, 1999). The resulting stromaland epithelial changes are associated with haze andregression that may occur following refractive cornealprocedures like PRK (Del Pero et al., 1990; Wu, Starkand Green, 1991; Hanna et al., 1992).

There are qualitative and quantitative differences inkeratocyte apoptosis among surgical procedures suchas epithelial-scrape PRK, transepithelial PRK, andLASIK when the procedures are performed in rabbits(Helena et al., 1998; Kim, Shah and Wilson, 1998).The working hypothesis leading to the present studywas that inhibition of keratocyte apoptosis couldreduce the subsequent wound healing processesand lead to more reproducible and safer surgicalprocedures.

Apoptosis is a complex physiological process thatcan be triggered by cell-speci®c extracellular, as well asintracellular, signals. The events leading to apoptosis

are mediated by activation of cascades of cysteineproteases called caspases and other proteins arranged

# 2000 Academic Press

in highly regulated death pathways (Enari, Hug andNagata, 1995; Armstrong et al., 1996; Enari et al.,1996; Glynn, McElligott and Mosier, 1996; Orth et al.,1996; Wilson et al., 1996b). Several caspases havebeen identi®ed as key modulators of apoptosis,including caspase-3 (CPP32), caspase-6, caspase-7,caspase-8, and caspase-9 (Wilson, in press 1999). Thisis in contrast with necrosis, a more random anduncontrolled form of cell death, associated with therelease of cellular contents such as degradativeenzymes that in¯ict damage on surrounding cells andtissues.

Several reports have suggested that inhibitors ofcaspases can partially or completely block apoptosis(Enari et al., 1995; Armstrong et al., 1996; Glynnet al., 1996; Mirzoeva et al., 1996; Rouquet et al.,1996). In this study, the membrane-permeablecaspase inhibitor benzyloxycarbonyl -Val-Ala-Asp-¯uoromethylketone (z-VAD-FMK) was applied topically

226

prior to epithelial scrape in an attempt to inhibit

non-parametric Mann-Whitney U-test. A P value less

keratocyte apoptosis in the rabbit cornea.

2. Materials and Methods

This study adhered to the ARVO Statement for theUse of Animals in Ophthalmic and Vision Research.Adult New Zealand white rabbits (2.5±3.0 kg) wereanesthetized by intramuscular injection of ketaminehydrochloride (30 mg kgÿ1) and xylazine hydrochlo-ride (5 mg kgÿ1). A 20 mM solution of the caspaseinhibitor z-VAD-FMK (Enzyme Systems Products,Livermore, CA, U.S.A.) was prepared in steriledimethyl sulfoxide (Sigma Chemical, St. Louis, MO,U.S.A.) and diluted with an equal amount of balancedsalt solution (Alcon, Ft. Worth, TX, U.S.A.) immedi-ately prior to application to the corneal surface. Thus,the ®nal solution was 10 mM z-VAD-FMK in 50 %dimethyl sulfoxide. Cloning cylinders (Bel-Art Pro-ducts, Pequannock, NJ, U.S.A.) with a 10 mmdiameter were used to ensure suf®cient contact timebetween the solution and the cornea. The animalswere positioned on the procedure table so thatthe cornea was in the frontal plane relative to thetabletop. Cylinders were placed gently overlying thecentral epithelium of the cornea without adhesive.Seventy ml of the 10 mM z-VAD-FMK solution wereapplied to one eye 15 min prior to epithelial scrapeand again 5 min prior to epithelial scrape, with acontact time of 5 min at each application. The eyeswere taped shut between application of z-VAD-FMK.The other eye of each animal was treated in a similarfashion, except with the DMSO-balanced salt solutionvehicle alone. After the second applications of thez-VAD-FMK solution or the control solution, thecornea was lightly marked with a 7 mm diametercorneal marker and the epithelium was scraped witha No. 64 Beaver blade (Becton-Dickinson, Franklin

Lake, NJ, U.S.A.). This method for induction ofkeratocyte apoptosis was selected for this study

because it is the best-characterized in previous studiesand consistently produces apoptosis in the anteriorstromal keratocytes (Wilson et al., 1996a; Helenaet al., 1998). A temporary tarsorrhaphy was placedwith a 5-0 nylon suture (Alcon, Ft. Worth, TX,U.S.A.) following the procedures in each eye. Rabbitswere maintained under anesthesia for 4 hr prior tokilling and excision of the corneoscleral rims. Sometreatment and control corneas were immediatelyembedded in OCT compound (Sakura Finetek,Torrance, CA, U.S.A.), snap frozen in liquid nitrogen,and stored at ÿ808C until analysis. Other treatmentand control corneas were placed immediately into®xative (3 % gluteraldehyde and 1 % paraformalde-hyde) for transmission electron microscopy (TEM).Fixation was continued for 48 hr prior to processingfor TEM. TEM was performed as previously describedusing a Joel (Peabody, MA, U.S.A.) Jem 1200EXelectron microscope (Michel-Salmin et al., 1986).

Seven mm sections extending transversely acrossthe central cornea were prepared with a Reichert-Jung cryostat (Leica, Deer®eld, IL, U.S.A.) and placedon Superfrost plus slides (Fisher Scienti®c, Pittsburgh,PA, U.S.A.). Fresh frozen sections were stained with0.05 % propidium iodide and examined with ¯uor-escent microscopy with an Optiphot-2 microscope(Nikon, Melville, NY, U.S.A.) to detect keratocytes. Theperoxidase-based terminal deoxyribonucleotidyltransferase-mediated dUTP-digoxigenin nick andlabeling (TUNEL) assay was performed according tothe manufacturer's instructions (Intergen, Gaithers-burg, MD, U.S.A.) to detect fragmentation of DNAassociated with apoptosis in the fresh frozen cornealsections (Gavrieli, Sherman and Ben-Sasson, 1992;Gold et al., 1994; Mori et al., 1994). Counter stainwas not used in the TUNEL assay to allow moresensitive detection of the staining keratocytes. Thetotal number of apoptotic cells in ten randomlyselected non-overlapping, full thickness columnsfrom the central cornea were counted for each corneaas previously described (Helena et al., 1998). Thediameter of each column was the 400� Optiphot-2microscopic ®eld.

Variations were expressed as standard errors ofthe mean (S.E.(M.)). Statistical comparison between thecontrol and treated groups were performed with the

W.-J. KIM ET AL.

than 0.05 was considered statistically signi®cant.

3. Results

Dose response experiments were performed toevaluate the effect of 0.5, 1, 5, 10 and 25 mM ®nalconcentrations of z-VAD-FMK in 50 % DMSO onkeratocyte apoptosis when applied prior to epithelialscrape injury. The four highest concentrations werefound to inhibit the keratocyte apoptosis response as

detected by the TUNEL assay. There were nosigni®cant differences between these concentrations

SIS

(data not shown). The 0.5 mM concentration of z-VAD-FMK was no different from the vehicle alone.Thus, the 10 mM ®nal concentration was within theeffective concentration range evaluated in thesedosing experiments. Cell morphology detected withtransmission electron microscopy was similar witheach concentration of z-VAD from 1±25 mM. At eachconcentration, many keratocyte cells had morphologythat was not consistent with apoptosis, but consistentwith necrosis.

Seven rabbits were included in the 10 mM z-VAD-FMK/DMSO- and DMSO vehicle-treated groupsincluded in the ®nal experiment. Only one eye ofeach animal was treated. Five in each group were freshfrozen in OCT for TUNEL assay and other histologicanalyses. Two from each group were processed fortransmission electron microscopy. Two additionalrabbits in the z-VAD-FMK/DMSO- and DMSO vehicle-treated groups were treated with the drugs, but did nothave epithelial scrape injury.

Propidium iodide staining was used to detect cell

CASPASE INHIBITORS AND KERATOCYTE APOPTO

nuclei. There was no difference in keratocyte densitymonitored by propidium iodide staining in the anterior

FIG. 1. Propidium iodide staining of cell nuclei in corneas treatecorneal epithelial (e) scrape. (A) Cornea treated with DMSO vehdensity and distribution were normal. (B)±(D) Four hr after epithedecrease in keratocyte density in the anterior stroma. Arrowheahowever, a much larger number of smaller structures (arrows)(E) Keratocyte distribution and density was normal after treatm(F)±(H) Four hr after epithelial scrape in eyes pre-treated withkeratocyte nuclei (arrows) in the anterior stroma. However, manyH). Also the density of staining nuclei appeared to be somewhat rAll magni®cations 400�, except (A) 100�.

stroma between z-VAD-FMK/DMSO and DMSO treatedcontrol corneas in which there was no corneal scrape[Fig. 1(A) and (E)]. Thus, there was no effect of thepharmacological agents on keratocyte density in theabsence of scrape injury. There were less propidiumiodide-stained super®cial keratocytes detected in thecentral corneas treated with DMSO alone [Fig. 1(B)±(D)] compared with corneas treated with z-VAD-FMK/DMSO [Fig. 1(F)±(H)] at 4 hr after epithelial scrape.Some of the staining in the anterior stroma of DMSO-treated corneas appeared to be very small bodies thatmay have not been intact cells [Fig. 1(B)±(D)].However, this could not be determined with certaintyat the light microscopic level. In preliminary exper-iments, we evaluated the effect of application of 50 %DMSO after the corneal scrape injury on keratocyteapoptosis. There was no difference in keratocyteapoptosis evaluated with the TUNEL assay or electronmicroscopy at 4 hr after scrape injury in rabbitcorneas that were treated with 50 % DMSO alone(applied immediately after epithelial scrape, with a

227

contact time of 5 min) compared with no treatmentwith DMSO (data not shown). No keratocyte necrosis

d with DMSO vehicle or z-VAD-FMK/DMSO with or withouticle without scrape. At 4 hr after treatment the keratocytelial scrape in eyes pre-treated with DMSO vehicle there was ads indicate nuclei similar to those in (A). There were also,that may be remnants of cells that underwent apoptosis.ent with z-VAD-FMK/DMSO without epithelial (e) scrape.

z-VAD-FMK/DMSO there appeared to be relatively normalof the nuclei-like structures appeared to be small (arrow in

educed compared with the control that was not scraped (E).

(�1 %) in corneas treated with DMSO vehicle prior to

was noted in corneas that were treated with 50 %DMSO after epithelial scrape injury.

There were no TUNEL-stained keratocytes detectedin either z-VAD-FMK/DMSO- or DMSO-treated corneasif there was no epithelial scrape (Fig. 2(A) and (D)].Many more TUNEL-stained keratocytes were detectedin DMSO-treated corneas [Fig. 2(B) and (C)] than inz-VAD-FMK/DMSO-treated corneas [Fig. 2(E) and (F)]at 4 hr after epithelial scrape. TUNEL-positive cellswere counted in sections from ®ve scraped corneasfrom the z-VAD-FMK/DMSO and ®ve scraped corneasfrom the control DMSO alone group. The number ofTUNEL-positive keratocytes per ®eld in ten 400� ®eldsof central stroma was 0.5 + 0.4 in the z-VAD-FMK/DMSO-treated corneas (Fig. 3). This compared with2.9 + 2.3 TUNEL-positive cells per ®eld in the DMSO-treated control corneas. The difference was statisticallysigni®cant (P � 0.04).

Transmission electron microscopy (TEM) providedimportant results not obtainable at the light micro-scopic level with either propidium iodide or TUNELstaining. In control corneas that were not scraped aftertreatment with DMSO or z-VAD-FMK/DMSO therewere only normal keratocytes at the TEM level (notshown). The anterior stroma to approximately 20±

228

30 % of corneal depth had numerous cells withmorphologic features suggestive of apoptosis in

FIG. 2. TUNEL assay to detect keratocyte apoptosis in responsevehicle (DMSO) or z-VAD-FMK/DMSO (z-VAD). (A) Little keratocvehicle without epithelial scrape. e indicates epithelium, s indicatemembrane. (B), (C) A high level of keratocyte apoptosis (arrowheawas applied prior to epithelial scrape. Both panels were stainedMagni®cation 200�. At higher magni®cation (not shown) indivseen, but this lower power view provides a better analysis of the dFMK/DMSO was applied to the cornea without scraping no keraindicates epithelium, s indicates stroma, and * indicates the levelVAD-FMK/DMSO was applied to the cornea prior to scrape there wassay. Corneas shown were analysed at 4 hr after epithelial scrapestained (arrowhead) is seen in (F). All magni®cations 200�.

DMSO-treated corneas at 4 hr after scrape[Fig. 4(A)±(D)]. No normal keratocytes were noted inthe anterior stroma of DMSO-treated corneas. Normalkeratocytes, however, were detected in corneas treatedwith z-VAD-FMK/DMSO at 4 hr after scrape[Fig. 4(E)]. There were occasional keratocytes withchanges suggestive of apoptosis detected in theanterior stroma of scraped corneas that had beentreated with z-VAD-FMK/DMSO [Fig. 4(F) and (G)], butthese cells were less than 1±2 % of all cells. In contrast,cells with morphology consistent with apoptosis werepredominant in the anterior stroma of corneas treatedwith DMSO alone prior to epithelial scrape. Unexpect-edly, however, we also detected a large number ofkeratocytes (approximately 50±75 % of cells) withdisrupted morphology that was more consistent withnecrosis than apoptosis in the anterior stroma ofcorneas treated with z-VAD-FMK/DMSO [Fig. 4(H)±(J)]. Necrotic appearing keratocytes were nearly absent

W.-J. KIM ET AL.

scrape injury.

4. Conclusions

The results of this study demonstrate that there is

signi®cant inhibition of epithelial injury-inducedkeratocyte apoptosis by the topical application of

to corneal epithelial scrape in corneas treated with DMSOyte apoptosis was detected in corneas treated with DMSOs stroma, and * indicates the level of the epithelial basementds) was detected in the anterior stroma when DMSO vehicle

4 hr after epithelial scrape. * indicates the stromal surface.idual cells staining with the TUNEL assay are more clearlyistribution of the cells within the stroma. (D) When z-VAD-

tocytes that stained with the TUNEL assay were detected. eof the epithelial basement membrane. (E) and (F). When z-as a marked reduction in cells that stained with the TUNEL

injury. No stained cells are seen in (E). Only a single cell that

FIG. 3. Quantitation of TUNEL-positive cells at 4 hr aftercorneal epithelial scrape in rabbit corneas pretreated withz-VAD-FMK/DMSO (zVAD) or DMSO vehicle (DMSO control).Signi®cantly less cells stained per 400� ®eld in corneastreated with z-VAD-FMK/DMSO than DMSO vehicle. The

CASPASE INHIBITORS AND KERATOCYTE APOPTOSIS

z-VAD-FMK prior to epithelial scrape injury. In anotherexperiment in which z-VAD-FMK or vehicle wasapplied following epithelial scrape injury, the level ofkeratocyte apoptosis after scrape injury was notdifferent between the two groups (Kim and Wilson,unpublished data, 1998). Thus, the inhibitory agentmust be applied prior to epithelial injury to have anyeffect on cell death. Although the sample size of thisstudy was limited due to the expense of z-VAD-FMK,there was statistically signi®cant inhibition of kerato-cyte apoptosis in the z-VAD-FMK group compared tothe vehicle group detected using the TUNEL assay(Fig. 3). Surprisingly, however, transmission electronmicroscopy detected many keratocytes in the anteriorstroma of corneas treated with z-VAD-FMK prior toscrape injury that had morphology consistent withnecrosis rather than apoptosis (Fig. 4). The necrotic-appearing keratocytes in the z-VAD-FMK-treatedcorneas may or may not have been detected as dyingcells by the TUNEL assay that recognizes fragmentedDNA ends since this assay is relatively, but notabsolutely, speci®c for apoptosis (Gold et al., 1994;Darzynkiewicz et al., 1998). Many necrotic kerato-cytes were probably not detected by the TUNEL assaysince the number of cells that stained was markedlyreduced in the z-VAD-FMK/DMSO-treated compared toDMSO-treated corneas. If all necrotic cells weredetected by the TUNEL assay, there most likely wouldhave been no distinction between corneas treated withz-VAD-FMK/DMSO and DMSO alone.

It is possible that there was still a reduction in thetotal number of cells that died (apoptosis plus necrosis)with the z-VAD-FMK treatment since there were more

difference was statistically signi®cant.

anterior stromal cells that stained with propidiumiodide in this group compared to the DMSO group

(Fig. 2). This could, however, be attributable to cellsthat are in the process of undergoing necrosis that isdetectable by transmission electron microscopy at 4 hrafter scrape. Many of these cells were probably stilldetectable by propidium iodide staining of nucleic acidmaterial since they had not completely disintegrated.This study emphasizes the critical importance of inde-pendent con®rmation of TUNEL assay results usingtechniques such as transmission electron microscopyof detection of activated components of the apoptosispathway such as apopain. Use of an apoptosis-speci®canalysis such as the apopain assay would not detectnecrotic cells. The TUNEL assay remains useful oncethe response detected in a particular experiment witha speci®c tissue has been con®rmed to be apoptosisusing other techniques. TUNEL assay ®ndings alonemay be misleading and result in erroneous con-clusions regarding the fate of cells.

z-VAD-FMK inhibits apoptosis by interfering withthe function of mediators such as caspase-3 andcaspase-8 (Mirzoeva et al., 1996; Wilson, in press1999). This study shows that, similar to the recent®ndings in several other systems (Hirsch et al., 1997;Lemaire et al., 1998), inhibition of the keratocyteapoptosis after epithelial injury with z-VAD-FMKappears to inhibit the orderly, programmed cell deathof keratocytes, but not the subsequent death of manyof the anterior keratocytes by a process more akin tonecrosis. Keratocytes that die by necrosis wouldrelease intracellular components such as collagenasesthat could in¯ict damage on surrounding cells andtissues (Wilson and Kim, 1998). Similar to the re-sponse to keratocyte apoptosis, these necrotic anteriorstromal keratocytes would most likely be replaced overa period of days through proliferation and migration ofremaining posterior and peripheral keratocytes thathave the myo®broblast phenotype (Jester et al., 1996;Masur et al., 1996; Wilson and Kim, 1998). Thus,despite inhibition of keratocyte apoptosis, the sub-sequent wound healing response is unlikely to beaffected by the application of z-VAD-FMK.

Recent studies have shown that apoptosis occurs bystimulus-speci®c and cell type-speci®c cascades indifferent organs (Wilson, in press 1999). Some stimulitrigger more than one apical caspase, such ascaspase-8 or caspase-9, which can independentlytrigger effector caspase activation and apoptosis. Inorder to design pharmacological strategies to inhibitkeratocyte apoptosis it may be important to identifythe different cascades activated and to utilize broad-acting caspase inhibitors that simultaneously blockeach of the relevant cascades. It will probably beimportant that the inhibitors block apical steps in thecascades to avoid transition to necrosis (Wilson, inpress 1999). Recently, apoptosis inhibitors that actearlier in the cascade have been identi®ed (Wanget al., 1998). In addition, stimulation of transcription

229

factors such as nuclear factor kappa B activation hasbeen shown to effectively inhibit corneal ®broblast

FIG. 4. Transmission electron microscopic analysis at 4 hr after corneal epithelial scrape in rabbit corneas pre-treated withDMSO vehicle or z-VAD-FMK/DMSO. (A), (B) After treatment with DMSO vehicle a large number of keratocyte cells in theanterior stroma had morphologic changes consistent with apoptosis including cell shrinkage, vacuolization, chromatincondensation, and chromatin fragmentation. Cells with vacuolization and shrinkage of the cytoplasm along with chromatincondensation (arrowheads) can be noted (6000�). The arrow in (A) indicates what appear to be membrane-bound bodiesconsistent with apoptotic bodies. In corneas treated with z-VAD-FMK/DMSO there were relatively few normal anterior stromalcells. An example of a normal keratocyte detected in the anterior stroma of one z-VAD-FMK/DMSO-treated cornea is shown in(C) (4000�). The arrowhead indicates the normal chromatin pattern. Some cells with morphologic changes suggestive ofapoptosis were detected in the anterior stroma of corneas treated with z-VAD-FMK/DMSO as shown in (D) and (E) (4000�).Arrowheads indicate chromatic condensation. However, in corneas treated with z-VAD-FMK/DMSO prior to epithelial scrape itwas noted that the majority of cells in the anterior stroma appeared to be disintegrating without signs of chomatincondensation or typical fragmentation (F) (4000�), (G) (15 000�), and (H) (15 000�). Many of these cells appeared asthough they had exploded, with most of the cellular contents not contained within membrane bound bodies. Note free cellulardebris in the extracellular spaces in (G) and (H) not contained in membrane bound structures (arrowheads). No evidence of thecell membrane (arrows) can be seen in many areas of the cells shown in (G) and (H). In every case, the nucleus was completelydisrupted and could not be detected as condensed chromatin or another remnant. The morphology of these cells is consistentwith necrosis rather than apoptosis. No cells with morphology consistent with apoptosis of necrosis were found in corneas thatdid not have epithelial scrape whether they were treated with z-VAD-FMK/DMSO or DMSO vehicle (not shown).

230 W.-J . KIM ET AL.

SIS

apoptosis in response to cytokines such as bonemorphogenic protein 2 (Mohan et al., 1998) andtumor necrosis factor alpha (Mohan, Mohan, Kim andWilson, unpublished data, 1998). Thus, methods toactivate the anti-apoptotic effects of NF-kB might beeffective in preventing apoptosis without promotingcell necrosis. Future investigations of these inter-

CASPASE INHIBITORS AND KERATOCYTE APOPTO

ventions should include methods that distinguish

apoptosis and necrosis.

Acknowledgements

Supported in part by the U.S. Public Health Service grantEY10056 from the National Eye Institute, National Insti-tutes of Health, Bethesda, Maryland, U.S.A., The Universityof Washington School of Medicine, Seattle WA, U.S.A. and

an Unrestricted Grant from Research to Prevent Blindness, New York, NY, U.S.A.

References

Armstrong, R. C., Aja, T., Xiang, J., Gaur, S., Krebs, J. F., Bal,X., Korsmeyer, S. J., Karanewsky, D. S., Fritz, L. C. andTomaselli, K. J. (1996). Fas-induced activation of thecell death-related protease CPP32 is inhibited by Bcl-2and by ICE family protease inhibitors. J. Biol. Chem.271, 16850±5.

Darzynkiewicz, Z., Bedner, E., Traganos, F. and Murakami,T. (1998). Critical aspects in the analysis of apoptosisand necrosis. Hum. Cell 11, 3±12.

Del Pero, R. A., Gigstad, J. E., Roberts, A. D., Klintworth,G. K., Martin, C. A., L'Esperance, F. A., Jr. and Talor,D. M. (1990). A refractive and histopathologic study ofexcimer laser keratectomy in primates. Am. J. Ophthal-mol. 109, 419±29.

Enari, M., Hug, H. and Nagata, S. (1995). Involvement of anICE-like protease in Fas-mediated apoptosis. Nature 375,78±81.

Enari, M., Talanian, R. V., Wong, W. W. and Nagata, S.(1996). Sequential activation of ICE-like and CPP32-like proteases during Fas-mediated apoptosis. Nature380, 723±6.

Gavrieli, Y., Sherman, Y. and Ben-Sasson, S. A. (1992).Identi®cation of programmed cell death in situ viaspeci®c labeling of nuclear DNA fragmentation. J. CellBiol. 119, 493±501.

Glynn, J. M., McElligott, D. L. and Mosier, D. E. (1996).Apoptosis induced by HIV infection in H9 T cells isblocked by ICE-family protease inhibition but not by aFas (CD95) antagonist. J. Immunol. 157, 2754±8.

Gold, R., Schmied, M., Giegerich, G., Breitschopf, H.,Hartung, H. P., Toyka, K. W. and Lassmann, H.(1994). Differentiation between cellular apoptosis andnecrosis by the combined use of in situ tailing and nicktranslation techniques. Lab. Invest. 71, 219±25.

Hanna, K. D., Pouliquen, Y. M., Waring, G. O., 3rd,Savoldelli, M. and Thompson, K. P. (1992). Cornealwound healing in monkeys after repeated excimer laserphotorefractive keratectomy. Arch. Ophthalmol. 110,1286±91.

Helena, M. C., Baerveldt, F., Kim, W.-J. and Wilson, S. E.(1998). Keratocyte apoptosis following corneal surgery.Invest. Ophthalmol. Vis. Sci. 39, 276±83.

Hirsch, T., Marchetti, P., Susin, S. A., Dallaporta, B.,Zamzami, N., Marzo, I., Geuskens, M. and Kroemer, G.

(1997). The apoptosis-necrosis paradox. Apoptogenicproteases activated after mitochondrial permeability

transition determine the mode of cell death. Oncogene15, 1573±81.

Jester, J. V., Barry-Lane, P. A., Cavanagh, H. D. and Petroll,W. M. (1996). Induction of alpha-smooth muscleactin expression and myo®broblast transformation incultured corneal keratocytes. Cornea 15, 505±16.

Kim, W.-J., Shah, S. and Wilson, S. E. (1998). Differences inkeratocyte apoptosis following transepithelial and laser-scrape photorefractive keratectomy. J. Refract. Surg. 14,526±33.

Lemaire, C., Andreau, K., Souvannavong, V. and Adam, A.(1998). Inhibition of caspase activity induces a switchfrom apoptosis to necrosis. FEBS Lett. 425, 266±70.

Masur, S. K., Dewal, H. S., Dinh, T. T., Erenburg, I. andPetridou, S. (1996). Myo®broblasts differentiate from®broblasts when plated at low density. Proc. Natl. Acad.Sci. USA 93, 4219±23.

Michel-Salmin, L. A., Tosi-Couture, E., Gautier, A.,McDowall, A. W. and Dubochet, J. (1986). Electronmicroscopy of the chromosomes in Dino¯agellatePorocentrum micans: Con®rmation of BouligandõÂssliquid crystal hypothesis. J. Ultrastruct. Molec. Struct.Res. 97, 10±30.

Mirzoeva, O. K., Yaqoob, P., Knox, K. A. and Calder, P. C.(1996). Inhibition of ICE family cysteine proteasesrescues murine lymphocytes from lipoxygenaseinhibitor-induced apoptosis. FEBS 396, 266±70.

Mohan, R. R., Kim, W. J., Mohan, R. R., Chen, L. andWilson, S. E. (1998). Bone morphogenic proteins 2and 4 and their receptors in the adult human cornea.Invest. Ophthalmol. Vis. Sci. 39, 2626±36.

Mohan, R. R., Liang, Q., Kim, W.-J., Helena, M. C.,Baerveldt, F. and Wilson, S. E. (1997). Apoptosis inthe cornea: further characterization of Fas/Fas ligandsystem. Exp. Eye Res. 65, 575±89.

Mori, C., Nakamura, N., Okamoto, Y., Osawa, M. andShiota, K. (1994). Cytochemical identi®cation of pro-grammed cell death in the fusing fetal mouse palate byspeci®c labeling of DNA fragmentation. Anat. Embryol.190, 21±8.

Orth, K., O'Rourke, K., Salvesen, G. S. and Dixit, V. M.(1996). Molecular ordering of apoptotic mammalianCED-3/ICE-like proteases. J. Biol. Chem. 271,20977±80.

Rouquet, M., Pages, J. C., Molina, T., Briand, P. and Joulin,V. (1996). ICE inhibitor zVADcmk is a potent thera-peutic agent against in vivo liver apoptosis. Curr. Biol. 6,1192±5.

Wang, C.-Y., Mayo, M. W., Korneluk, R. G., Goeddel, D. V.and Baldwin, A. S. (1998). Nf-kB antiapoptosis:induction of TRAF1 and TRAF2 and c-IAP1 andc-IAP2 to suppress caspase-8 activation. Science 281,1680±3.

Wilson, S. E. (1997). Programmed cell death, woundhealing, and laser refractive surgical procedures:molecular biology for the corneal surgeon. J. Refract.Surg. 13, 171±5.

Wilson, S. E. (1999). Stimulus-speci®c and cell type-speci®ccascades: emerging principles relating to control ofapoptosis. Exp. Eye Res., in press.

Wilson, S. E., He, Y.-G., Weng, J., Li, Q., Vital, M. andChwang, E. L. (1996a). Epithelial injury induceskeratocyte apoptosis: hypothesized role for the inter-leukin-1 system in the modulation of corneal tissueorganization. Exp. Eye Res. 62, 325±38.

Wilson, S. E. and Kim, W.-J. (1998). Keratocyte apoptosis:implications on corneal wound healing, tissue organiza-

231

tion, and disease. Invest. Ophthalmol. Vis. Sci. 39,220±6.

Wilson, S. E., Li, Q., Weng, J., Barry-Lane, P. A., Jester, J. V.,Liang, Q. and Wordinger, R. J. (1996b). The Fas/Fas

232

ligand system and other modulators of apoptosis in thecornea. Invest. Ophthalmol. Vis. Sci. 37, 1582±92.

Wu, W. C., Stark, W. J. and Green, W. R. (1991). Cornealwound healing after 193 nm excimer laser keratect-

W.-J. KIM ET AL.

omy. Arch. Ophthalmol. 109, 1426±32.