Fas Ligand System

Exp. Eye Res. (1997) 65, 575–589

Apoptosis in the Cornea : Further Characterization of Fas/Fas Ligand System R A J I V R. M O H A N, QIANWA LIANG, W O O-JUNG KIM, M A R C O C . H E L E N A, F I O N A B A E R V E L D T,    S T E V E N E. W I L S O N* Eye Institute and Department of Cell Biology/A31, The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH, 44195, U.S.A. (Received Columbia 24 February 1997 and accepted in revised form 1 July 1997) This study was performed to further characterize expression and function of the Fas}Fas ligand system in the cornea. Specifically, these experiments examined (1) the effect of genetic inactivation of Fas or Fas ligand genes on keratocyte apoptosis in response to corneal epithelial wounding, (2) whether cultured human corneal epithelial and endothelial cells are competent to undergo apoptosis in response to Fas activation, (3) expression of membrane bound and soluble Fas and Fas ligand in corneal cells, and (4) the effect of IL-1 on expression of Fas and Fas ligand in corneal fibroblasts. Keratocyte apoptosis in response to corneal epithelial scrape detected by TUNEL assay and transmission electron microscopy was significantly decreased, but not eliminated, in Fas ligand ®}® mice compared with control ­}­ mice. There was also a decrease in Fas ®}® mice that did not reach statistical significance. Thus, while the Fas}Fas ligand system is likely involved in regulating keratocyte apoptosis in response to epithelial wounding, systems with redundant function probably also modulate this response. Activation of the Fas receptor triggered death with ultrastructural changes characteristic of apoptosis in corneal epithelial and endothelial cells in culture. Since these cell types express both Fas and Fas ligand in vivo, systems must be in place to prevent uncontrolled activation via autocrine ligand-receptor interaction. Messenger RNAs coding for both membrane bound and soluble splicing variants of Fas were expressed in each corneal cell type, suggesting that soluble Fas production could be one mechanism to antagonize membrane bound Fas activation. Soluble Fas ligand protein was expressed in wounded ex vivo corneal epithelium, providing a mechanism for Fas ligand from epithelium to mediate keratocyte apoptosis. IL-1, however, also stimulated corneal fibroblasts to express Fas ligand mRNA and protein. Therefore, an alternative mode for epithelial injury to trigger keratocyte apoptosis may be by release of IL-1, subsequent induction of Fas ligand in keratocytes, and apoptosis mediated by autocrine mechanisms. # 1997 Academic Press Limited Key words : cornea ; apoptosis ; programmed cell death ; keratocytes ; fibroblasts ; Fas ; Fas ligand ; interleukin-1.

1. Introduction Recent studies have demonstrated that superficial keratocytes undergo apoptosis in response to corneal epithelial wounding (Wilson et al., 1996a). These studies have suggested injury to the epithelium activates keratocyte apoptosis, rather than loss of trophic influences of the epithelium, exposure to the external environment, or changes in corneal hydration. Apoptosis, in contrast to necrosis, is a form of involutional cell death where demise occurs with limited inflammation and little release of cellular contents, including degradative enzymes that would produce collateral damage to the adjacent tissues and cells (Wyllie, Kerr and Currie, 1980 ; Arends and Wyllie, 1991 ; Bellamy et al., 1995). This rather gentle mechanism for eliminating cells is of obvious significance to the clarity of the cornea. Apoptosis is best identified by characteristic ultrastructural changes which occur in response to stimuli such as specific cytokines (Graham et al., 1994 ; Zou and Niswander, 1996), viral infection (Suarez et al., 1996), and * For correspondence.

0014–4835}97}100575­15 $25.00}0}ey970371

growth factor deprivation (Deckwerth and Johnson, 1993 ; Yao and Cooper, 1996). Apoptosis is critical during development of multicellular organisms, but is also important in the maintenance of tissues and wound healing (Wyllie, Kerr and Currie, 1980 ; Arends and Wyllie, 1991). Molecular signals that have the capacity to trigger apoptosis in cultured keratocytes (corneal fibroblasts) include IL-1 receptor and Fas receptor activation (Wilson et al., 1996a ; 1996b). The role of the Fas}Fas ligand system in mediating keratocyte apoptosis in response to epithelial scrape injury has not been examined in situ, but can now that Fas- and Fas ligand-inactivated mice are available. Even if experiments with Fas- and Fas ligand-inactivated mice confirm a role for this system in keratocyte apoptosis, the mechanisms through which Fas ligand could mediate keratocyte apoptosis in vivo remain enigmatic since Fas ligand has been characterized as a membrane bound protein interacting with Fas receptor on adjacent cells via juxtacrine mechanisms (Griffith et al., 1995 ; Tanaka et al., 1995 ; Mariana et al., 1995 ; Kayagaki et al., 1995 ; Wilson et al., 1996b). If all corneal epithelial Fas ligand remained associated with # 1997 Academic Press Limited

576

the cell membrane it would seem unlikely the cytokine could mediate keratocyte death that occurs to a depth greater than 75 microns after epithelial injury (Wilson et al., 1996a). Fas ligand, however, is produced as a soluble protein by metalloproteinase-mediated release in some cell types (Tanaka et al., 1995 ; Mariana et al., 1995 ; Kayagaki et al., 1995). Soluble Fas ligand could mediate keratocyte apoptosis through paracrine mechanisms. It is not known whether wounded corneal epithelium produces soluble Fas ligand. Another puzzling aspect of the Fas}Fas ligand system in the cornea is the distribution of expression. both the Fas ligand and Fas receptor are expressed by the corneal epithelium and endothelium (Griffith et al., 1995 ; Wilson et al., 1996b). Simultaneous expression of the ligand and receptor in these tissues raises the question as to whether the Fas}Fas ligand system can function in an autocrine fashion to stimulate apoptosis in these cell types. There are many examples of cytokine or growth factor receptors being expressed by specific cell types that do not seem to affect the function of that cell in response to ligand binding. For example, corneal epithelial cells and fibroblasts express hepatocyte growth factor (HGF) receptor mRNA and protein (Wilson et al., 1993, 1994 ; Li et al., 1996). HGF stimulates proliferation and motility of the corneal epithelial cells, but has no effect on these functions in corneal fibroblasts (Wilson et al., 1994). The function of the HGF receptors expressed by fibroblasts are unknown. A type of apoptosis referred to as ‘ anoikis ’ occurs as a normal part of the maturation and death of epidermal tissues such as the corneal epithelium, but there is no evidence Fas-Fas ligand interactions contribute to this process. It would seem especially unlikely the Fas system mediates apoptosis in the normal human corneal endothelium, since endothelial cell density changes very slowly, if at all, over the lifespan of an individual and mitosis is rarely detectable in the human corneal endothelium (Laing et al., 1984). Either the Fas}Fas ligand system is not competent to mediate apoptosis in corneal epithelial and endothelial cells or there are regulatory mechanisms controlling Fas}Fas ligand interaction and}or signal transduction in the absence of other signals. The purpose of this study was to explore these questions by further characterizing the expression and function of the Fas}Fas ligand system in the cornea. We also investigated potential interactions between the IL-1 and Fas systems in regulating keratocyte apoptosis.

2. Materials and Methods Effect of Fas or Fas Ligand Inactivation on Keratocyte Apoptosis Following Epithelial Wounding Seven week old Fas ®}® (B6.MRL-Faslpr), Fas ligand ®}® (B6Smn.C3H-Faslgld), and control

R. R. M O H A N E T A L.

(C57BI6) inbred mice (Watson et al., 1992 ; Murphy, Roths and Eicher, 1982 ; Takahashi et al., 1994 ; Jackson Laboratories, Bar Harbor, ME, U.S.A.) were anesthetized with xylazine and ketamine. Central cornea scrape wounds were produced by removing approximately 80 % of the central corneal epithelium with a 64 Beaver blade (Becton-Dickinson, Franklin Lake, NJ, U.S.A.), sparing only approximately 0±5 mm of epithelium adjacent to the limbus. Eyes were enucleated under anesthesia 4 hr after wounding and imbedded in paraffin. Mice were euthenized by intraperitoneal injection of 100 mg kg−" pentobarbitol. Four hours was selected as the time point, since keratocyte TUNEL staining is maximal at this time after epithelial wounding in the mouse, even though apoptosis can be detected by TEM and TUNEL assay immediately after epithelial wounding (Wilson et al., 1996a). It is likely that staining with the TUNEL assay is maximal at 4 hr because DNA fragmentation (detected in the TUNEL assay) is maximal at this time point, even though cell apoptosis begins immediately after wounding. Thus, the sensitivity of detection of keratocytes undergoing apoptosis is likely to be increased by using the 4 hr after wounding time point for TUNEL assay. It is important that the time point after wounding examined was the same in all groups. All procedures in this study were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Six micrometer thick sections extending transversely across the central cornea of mouse eyes were stained using the peroxidase-based TUNEL assay (ApopTag assay, Oncor, Gaithersburg, MD, U.S.A.) according to the manufacturer’s instructions. The TUNEL assay detects fragmented DNA ends characteristic of apotosis (Gavrieli, Sherman and Ben-Sasson, 1992 ; Gold et al., 1994). Photographs were obtained with a Nikon light microscope. The number of keratocyte cells staining with the TUNEL assay in each mouse cornea was determined by summing the total number of cells in ten randomly selected 400¬ magnified fields. The anterior stromal surface bisecting each field during counting. Errors were expressed as the standard errors of the mean. Statistical comparisons were performed with the Student’s t test. A P value less than 0±05 was considered statistically significant. Transmission electron microscopy (Joel, Jem 1200 EX) was performed by fixing whole mouse eyes (controls, 1 hr after corneal epithelial wounding, and 4 hr after corneal epithelial wounding) overnight in 3 % gluteraldehyde, 1 % paraformaldehyde. Corneas were dissected from the eye and electron microscopy was performed as previously described (Wilson et al., 1996a).

F A S/I L-1 A P O P T O S IS

Effect of Fas-stimulating Antibody on First Passage Human Corneal Epithelial and Endothelial Cells Primary human corneal epithelial and endothelial cells were cultured from corneas from donors age term to 2 years old as previously described (Wilson et al., 1993). First passage cells were plated at densities 2±5¬10% cells well−" in standard six-well plates (Corning, NY, U.S.A.). Endothelial cells were maintained in Eagle’s Modified Essential Medium (EMEM) with 10 % fetal bovine serum (FBS) and 5 % fetal calf serum. Epithelial cells were maintained in Keratinocyte Defined Medium (KDM, modified MCDB 153, 5 µg ml−" insulin, 0±5 µg ml−" hydrocortisone, 0±15 m Ca#+, with EGF and bovine pituitary extract, Clonetics, San Diego, CA, U.S.A.). Endothelial cell medium was changed to EMEM with 0±5 % FBS and epithelial cell medium to KDM without EGF or pituitary extract 24 hr after seeding. Either 100 ng ml−" of anti-human Fas mouse monoclonal IgM antibody (Upstate Biotechnology Incorporated, Lake Placid, NY, U.S.A.) that binds the Fas ligand-binding domain and activates the Fas receptor or 100 ng ml−" control mouse IgM (RD Systems, Minneapolis, MN, U.S.A.) that does not bind Fas was added to each well. Photographs were also obtained after 12 to 24 hr with the inverted microscope (Nikon (model TMF), Melville, NY, U.S.A.). First passage corneal epithelial or endothelial cells exposed to anti-Fas monoclonal or control antibody were trypsinized, washed with medium containing 10 % fetal bovine serum, pelleted in a 1±5 ml Eppendorf tube, and fixed in 3 % gluteraldehyde, 1 % paraformaldehyde. Care was taken to insure that cells which spontaneously dissociated from the culture plate were also collected and included in the pellets. Electron microscopy (EM) was performed as previously described (Wilson et al., 1996b) on the fixed cell pellets. EM sections were cut at 70 nm and stained with 3 % uranyl acetate for 15 minutes, followed by 3 minutes in Reynold’s lead citrate. RT-PCR and Southern Blotting to Detect Alternatively Spliced Fas mRNAs in Corneal Cells Alternative mRNA splicing products coded from the genomic sequence for Fas (Cheng et al., 1994 ; Liu, Cheng and Mountz, 1995 ; Cascino et al., 1995 ; Papoff et al., 1996) were detected by reverse transcriptase-polymerase chain reaction (RT-PCR) using previously reported primers (Liu, Cheng and Mountz, 1995). The primers [Fig. 5(C)] recognize five previously reported mRNA splice variants (Cheng et al., 1994 ; Liu, Cheng and Mountz, 1995 ; Cascino et al., 1995 ; Papoff et al., 1996). The sequences from 5« to 3« of the upstream and downstream PCR primers were CACTTCGGAGGATTGCTCAACA and TATGTTGGCTCTTCAGCGCTA, respectively. The PCR reactions generated with the Fas PCR primers allow quantitative comparisons of relative expression of the Fas mRNA

577

splicing variants to be made within an individual cDNA sample. Complementary DNA (cDNA) was generated from ex vivo corneal epithelium and primary cultures of corneal epithelial, fibroblast, and endothelial cells and PCR reactions performed as previously described (Wilson et al., 1993). PCR products were resolved in a 1±2 % agarose gel. One hundred base pair ladder (Gibco BRL, Gaithersburg, MD, U.S.A.) was used to determine sizes of the products. In order to detect splicing variants expressed at lower levels, products were also transferred to Duralon-UV membrane (Stratagene, San Diego, CA, U.S.A.) and probed by Southern blotting (Sambrook, Fritsch and Maniatis, 1989) with a previously reported (Wilson et al., 1996b) 413 base pair Fas sequence extending from exon 3 to exon 7. This probe would recognize each of the previously reported Fas splice variants [Liu, Cheng and Mountz, 1995 ; Fig. 5(C)]. Purified radiolabeled DNA probe with a specific activity of approximately 2¬10* c.p.m. µg−" was prepared with the cloned sequence using a random primer labeling kit ‘ Rediprime ’ and alpha-$#P-dCTP (Amersham Life Science, Arlington Heights, IL, U.S.A.) according to the manufacturer’s protocol. To determine whether IL-1 alpha or IL-1 beta altered the relative expression of the Fas mRNA splicing variants over the time period required to note apoptosis in response to IL-1 in corneal fibroblasts, cDNA was synthesized from first passage human cells exposed to 10 ng ml−" IL-1 alpha, 10 ng ml−" IL-1 beta, or vehicle for 6–10 days (medium, cytokines, and vehicle were replaced at 3 day intervals) until apoptosis was noted in the cytokine, but not vehicle, exposed cultures. RT-PCR and Southern blotting were performed as described above. Northern Blotting to Detect the Effect of IL-1 Alpha on Fas and Fas Ligand Expression in Cultured Corneal Fibroblast Cells First to third passage human corneal fibroblasts exposed to 10 ng ml−" IL-1 alpha or vehicle were used for poly-(A) mRNA isolation. RNA was isolated at 6 to 10 days when cultures showed the first signs of accelerated cell death in response to IL-1 alpha (Wilson et al., 1996a). Total cellular RNA was isolated using TRIzol (BRL, Gaithersburg, MD, U.S.A.). Poly-(A) RNA was isolated using the mRNA Isolation Kit (Boehringer Mannheim, Indianapolis, IN, U.S.A.) according to the manufacturers instructions. RNA size markers and nine µg per lane of poly-(A)­ RNA were resolved on 1±2 % agarose formaldehyde gels. RNA was transferred to Duralon-UV (Stratagene, San Diego, CA, U.S.A.) using a vacuum blotter (Pharmacia, Piscataway, NJ, U.S.A.) and crosslinked to the Duralon-UV with the UV-Stratalinker (Stratagene). Previously described (Wilson et al., 1996b) northern blotting probes for Fas, Fas ligand, and beta actin were generated by RT-PCR and cloned into the PCR II

578

Cloning Vector (Invitrogen, San Diego, CA, U.S.A.). Each probe was sequenced (Sequenase 2.0, United States Biochemical Corp., Cleveland, OH, U.S.A.) according to the manufacturer’s protocols for confirmation. Each sequence was compared to all known sequences in the Genbank and EMBL databases using the Mac Vector 5.0 System (IBI, New Haven, CT, U.S.A.). No other nucleic acid sequences have been reported that would be expected to hybridize to the probes under the conditions utilized for northern blotting. Labeled DNA probes were prepared as described above. Probe for the RNA size markers was prepared by $#P-labeling lambda genomic DNA. A total of 5¬10( c.p.m. of probe was used in each hybridization. After four hours of prehybridization, hybridization was performed at 42°C for 16 hr in hybridization solution containing 5¬ SSPE, 50 % formamide, 5¬ Denhardt’s solution, 0±5 % SDS, and 500 µg ml−" sheared and denatured salmon sperm DNA. Washes were performed with 2 to 0±2¬ SSPE, 0±1 % SDS at 37 to 56°C. Blots were exposed to X-OMAT 5 (Kodak, Rochester, NY, U.S.A.) film and developed after 14 hr to 10 days of exposure at ®85°C. Immunoprecipitation and Western Blotting to Detect Fas Ligand Protein Expression Immunoprecipitation and western blotting were performed to detect Fas ligand protein using ex vivo corneal epithelium removed at the time of excimer laser photorefractive keratectomy or first passage corneal fibroblast cells cultured from donor corneas. This study was approved by the Investigational Review Board of The Cleveland Clinic Foundation. For cultured cells, medium was changed to EMEM with 0±5 % FBS at confluence and cells were exposed to IL-1 alpha 10 ng ml−", IL-1 beta 10 ng ml−", or vehicle until IL-1 exposed cultures began to show accelerated cell death (6–10 days). Medium and cytokines were changed at three day intervals. Reference standard human recombinant IL-1 alpha and IL-1 beta were obtained from the National Cancer Institutes Biological Response Modifiers Program (Bethesda, MD, U.S.A.). The soluble protein fraction used for immunoprecipitation of Fas ligand was isolated from ex vivo corneal epithelium or first passage corneal fibroblast cells using a standard method (Sambrook, Fritsch and Maniatis, 1989). Briefly, 500 µg of total protein from each specimen was used for immunoprecipitation. Thirty µg of protein A-agarose (CalBiochem, Cambridge, MA, U.S.A.) and 1 µg of non-immune IgG (Organon Teknika Corp., Durham, NC, U.S.A.) were added to each protein sample. The mixture was incubated 30 minutes at 4°C and centrifuged 8000 g for three minutes. The supernatant was combined with 1 µg of primary antibody. Twenty-five µg of protein-G plus agarose (CalBiochem) was added and the mixture was incubated for 60 minutes at 4°C. The

R. R. M O H A N E T A L.

mixture was centrifuged at 8000 g for 3 minutes. The pellet was washed four times with 100 m Tris–HCl, pH 7±4, 150 m NaCl, 1 m sodium vanadate with 2 µg ml−" aprotonin and 2 µg ml−" leupeptin. The pellet was resuspended in 2¬ sample buffer (Sambrook, Fritsch and Maniatis, 1989), boiled for 10 minutes, centrifuged briefly at 8000 g, and 25 µl of supernatant loaded on a 4–20 % precast polyacrylamide gel (Novel Experimental Technology, San Diego, CA, U.S.A.). Protein molecular weight standards (Amersham Life Sciences Inc., Arlington Heights, IL, U.S.A.) were run simultaneously on the Western blotting gels. The gel was run at 100 volts. Resolved proteins were transferred to nitrocellulose membrane (BioRad, Richmond, CA, U.S.A.) using a transblot system (Novel). The membrane was blocked for 30 minutes at 37°C and 1 hr at room temperature using blocking buffer (5 % nonfat milk in 10 m Tris–HCl, pH 7±5, 100 m NaCl, 0±1 % Tween 20). The blocking buffer was decanted and the blot washed six times for 5 minutes each with washing buffer (10 m Tris–HCl, pH 7±5, 100 m NaCl, 0±1 % Tween 20). The blot was incubated for 30 minutes at 37°C and 1 hr at room temperature with primary antibody in blocking buffer. The blot was washed six times for five minutes each at room temperature with washing buffer. The protein bands were revealed using the ECL Western Blotting Analysis System (Amersham Life Sciences, Buckinghamshire, England) according to the manufacturers protocol. The anti-Fas ligand antibody was used at a final concentration of 1¬10−% mg ml−" in Western blotting. 3. Results Two mice died immediately after anesthesia and, therefore, there were seven Fas ®}®, eight Fas ligand ®}®, and seven control corneas examined with the TUNEL assay [Fig. 1(A–E)]. The mean number of keratocytes staining with the TUNEL assay at four hours after epithelial scrape injury was 53±3³7±0, 30±7³11±0, and 26±9³7±4 per 10 fields in the control, Fas ®}®, and Fas ligand ®}® groups, respectively [Fig. 1(F)]. The difference between the Fas ligand ®}® and control groups was statistically significant (P ! 0±02). There was a trend towards a difference between the Fas ®}® and control groups, but the difference did not reach statistical significance. There was scatter in the Fas ®}® and Fas ligand ®}® groups with two corneas in each group having keratocyte apoptosis that exceeded the mean in the control group [Fig. 1(F)]. Anterior keratocytes that had cell shrinkage, chromatin condensation and fragmentation, and cellular blebbing consistent with apoptosis were noted by electron microscopy in corneas from the Fas ®}®, Fas ligand ®}®, and control group at four hours after epithelial scrape injury [Fig. 1(G–L)] identical to previous studies (Wilson et al., 1996a). In the control

F A S/I L-1 A P O P T O S IS

579

F

F. 1 A–F. For legend see page 580.

580

R. R. M O H A N E T A L.

F. 1. Apoptosis of keratocytes in response to corneal epithelial scrape injury in Fas ligand ®}®, Fas ®}®, and control ­}­ mice. (A) Keratocyte cells stain with the TUNEL assay for fragmented DNA (arrows) consistent with apoptosis four hours after epithelial scrape injury in control ­}­ mice. (B) and (C) Representative sections from Fas ligand ®}® mice stained with the TUNEL assay four hours after corneal epithelial scrape injury. In panel B, two keratocytes (arrows) stain for DNA fragmentation associated with apoptosis. In panel C there were no keratocytes that stained. (D) and (E) Representative sections from Fas ®}® mice stained with the TUNEL assay four hours after corneal epithelial scrape injury. In panel D, four keratocytes (two indicated by arrows) stained for DNA fragmentation associated with apoptosis. In panel E there were no keratocytes that stained. (F) TUNEL positive keratocytes in 10 380¬ microscopic fields in the Fas ®}®, Fas ligand ®}®, and control mice corneas 4 hr after epithelial scrape injury. Means, standard errors of the mean, and totals for individual corneas are provided. P values are for the difference compared with the control group. (G) Transmission electron micrograph of a normal anterior keratocyte (less than 50 µ from the anterior surface) 4 hr after corneal epithelial scrape injury in a Fas ligand ®}® mouse.

F A S/I L-1 A P O P T O S IS

581

F. 2. Effect of Fas-stimulating antibody on first passage human corneal epithelial and endothelial cells. Epithelial (A) and endothelial (C) cells were unchanged after 8 to 24 hr of exposure to control IgM antibody. After 8 to 24 hr of exposure to antiFas agonist antibody, a large proportion of epithelial (B) and endothelial (D) cells had rounded up and dissociated from the culture plate. Magnification 60¬.

group, few, if any, normal keratocytes could be observed within 50 µ of the anterior stromal surface by electron microscopy. In both the Fas ®}® and Fas ligand ®}® groups, however, there appeared to be normal and apoptotic keratocytes in the anterior stroma at four hours after epithelial scrape. Primary cultured corneal epithelial and endothelial cells begin accelerated dying (Fig. 2) within 8 to 12 hr of initiating incubation with anti-Fas agonist antibody in vitro. Examination of pellets of first passage corneal epithelial (Fig. 3) and endothelial (Fig. 4) cells using electron microscopy demonstrated death with classical signs of apoptosis in cells exposed to the anti-Fas agonist, but not control, antibody. Cellular changes noted included cell shrinkage, formation of membrane

bound apoptotic bodies, and chromatin condensation and fragmentation (Figs 3 and 4). Five sizes of alternative Fas mRNA were detected in ex vivo corneal epithelium and primary cultured corneal epithelial, fibroblast, and endothelial cells by RT-PCR and Southern blotting [Fig. 5(A) and 5(B)]. Within an individual PCR reaction a quantitative comparison of the relative levels of expression of the different Fas mRNAs can be made. The full length Fascoding alternative mRNA splicing product (PCR product 1200 base pairs) was expressed at the highest levels in all of the cell types tested. One of the soluble Fas mRNA amplification products (1100 base pairs) was detectable by PCR alone in cultured epithelial, fibroblast, and endothelial cells [Fig. 5(A)]. The other

Note the normal chromatin pattern in the cell nucleus which occupies almost the entire visible portion of the cell. (H) Transmission electron micrograph of an anterior keratocyte located just below the anterior surface of the stroma undergoing apoptosis at 1 hr after epithelial scrape injury in a Fas ligand ®}® mouse. Note chromatin condensation (c) and formation of numerous membrane bound vesicles (arrowheads), with cell shrinkage and almost complete loss of cytoplasm. (I) Transmission electron micrograph of a normal anterior keratocyte 4 hr after epithelial scrape injury in a Fas ®}® mouse. The nucleus is not visible, but note the normal organelles within the cytoplasm. (J) Transmission electron micrograph of an anterior keratocyte that appears to have early chromatin condensation (c) at 1 hr after epithelial scrape injury in a Fas ®}® mouse. Blebs are also beginning to appear at the cell membrane. (K) and (L) Transmission electron micrographs of anterior keratocytes undergoing apoptosis at 1 and 4 hr, respectively, following epithelial scrape injury in control ­}­ mice. Chromatin condensation (c) is seen, along with cell shrinkage with almost complete loss of cytoplasm in both cells. Few, if any, normal keratocytes could be detected within 50 to 75 µ of the anterior stromal surface in control ­}­ mice at 4 hr after epithelial scrape injury. Magnification for all transmission electron micrographs was 3800¬.

582

R. R. M O H A N E T A L.

F. 3. Cultured first passage corneal epithelial cells exposed to anti-Fas antibody or control antibody examined by electron microscopy. Many cells exposed to anti-Fas antibody [(A) and (B) magnification 1000¬] showed cell shrinkage and chromatin condensation and fragmentation (arrows) consistent with apoptosis. Frequent cells exposed to anti-Fas antibody showed cellular blebbing with the formation of membrane bound cell fragments consistent with apoptotic bodies. Large numbers of apoptotic cells were noted at low magnification throughout the cell pellets obtained from corneal epithelial cells treated with anti-Fas antibody. Corneal epithelial cells treated with control IgM antibody [(C) magnification 1000¬] were normal. When cultured first passage corneal epithelial cells exposed to anti-Fas antibody or control antibody were examined by electron microscopy at higher magnification, cells exposed to anti-Fas agonist antibody [(D) and (E) magnification 4000¬] showed cell shrinkage, formation of membrane bound fragments contain organelles (arrows in D) and chromatin condensation and fragmentation consistent with apoptosis. Corneal epithelial cells treated with control IgM antibody [(F) magnification 4000¬] were normal. The arrow in C denotes the nucleolus in the normal nucleus.

F A S/I L-1 A P O P T O S IS

583

F. 4. Cultured first passage corneal endothelial cells exposed to anti-Fas antibody or control antibody examined by electron microscopy at higher magnification. Cells exposed to anti-Fas agonist antibody (A) and (B) showed cell shrinkage, formation of membrane bound fragments contain organelles [arrows in (A) and (B)] and chromatin condensation and fragmentation consistent with apoptosis. Corneal endothelial cells treated with control IgM antibody (C) and (D) were normal. Magnification 3360¬.

previously reported (Liu, Cheng and Mountz, 1995) soluble Fas variants appear to be expressed at very low levels in each of the corneal cells tested since they were only detectable when sensitivity of detection was increased by Southern blotting the PCR products [Fig. 5(B)]. Exposure of cultured corneal fibroblasts to IL-1 alpha or beta had no effect on the relative expression of the Fas mRNAs in corneal fibroblast cells (not shown). Fas ligand protein was detected by immunoprecipitation and western blotting in ex vivo human corneal epithelial tissue (Fig. 6). Proteins 37 kDa and 26 kDa in size consistent with the membrane bound and soluble forms (Tanaka et al., 1995 ; Mariana et al., 1995 ; Kayagaki et al., 1995), respectively, of Fas ligand were detected in the corneal epithelial samples. The soluble Fas ligand product appeared to be present at higher levels in many (Fig. 6), but not all [Fig. 7(B)], scraped epithelial specimens [Figs 6 and 7(B)]. Stimulation of second passage corneal fibroblasts with 10 ng ml−" of IL-1 alpha was associated with induction of transcription of Fas ligand mRNA [Fig. 7(A)]. Cells were harvested for mRNA isolation when increased cell death was noted in IL-1 alpha-exposed

cultures. In three separate experiments, Fas ligand mRNA of the expected size of 1±2 kilobases was detectable only in corneal fibroblast cells incubated with IL-1 alpha. Fas ligand protein was detectable in the membrane bound protein fraction isolated from first passage corneal fibroblast cells incubated with either 10 ng ml−" IL-1 alpha or 10 ng ml−" IL-1 beta for 8 days [Fig. 7(B)]. IL-1 alpha appeared to increase the level of the membrane-associated full length 37 kDa, but not soluble 26 kDa, Fas ligand proteins.

4. Discussion Apoptosis (or programmed cell death) is the mechanism for the orderly elimination of unwanted cells during normal development, homeostasis, response to infection, and wound healing in multicellular organisms (Bellamy et al., 1995). Apoptosis, in contrast to necrosis, occurs with minimal collateral damage to surrounding cells and tissues. It is an involutional process accompanied by characteristic ultrastructural changes occurring in cells in response to specific environmental cues which set in motion an orderly-

584

R. R. M O H A N E T A L.

F. 5. Detection of Fas alternative mRNA splicing products expressed in corneal cells. (A) RT-PCR was utilized to detect amplification products that differentiate between Fas mRNAs coding for alternative Fas proteins (Liu, Cheng and Mountz, 1995). Note that the amplification within a particular sample allows a quantitative comparison of the relative expression of each of five previously described Fas variants (full length Fas, 1200 base pair amplification ; 4 soluble alternatives that inhibit Fas-mediated apoptosis (Liu, Cheng and Mountz, 1995 ; Cheng et al., 1994 ; Cascino et al., 1995 ; Papoff et al., 1996), 1100, 900, 800, and 700 base pair amplifications, respectively). Ex vivo corneal epithelium and primary cultured corneal epithelial, fibroblast, and endothelial cells all expressed the full length Fas (1200 base pairs) mRNA at the highest levels. Each sample type also expressed mRNA coding for truncated soluble Fas without a transmembrane domain (1100 base pairs) at lower levels. Other truncated Fas mRNAs expected to yield 900, 800, and 700 base pair amplifications, respectively, were not detected by RT-PCR alone. 1 and 2 represent an independent sample for each cell type. (B) By Southern blotting the Fas PCR products that were less than 1000 base pairs in size with a probe which hybridizes to each variant, soluble Fas mRNAs yielding 900, 800, and 700 base pair amplifications could be detected, but at very low and possibly insignificant levels. (C) Schematic diagram showing the relationship between the mRNAs coding for the full length and truncated Fas proteins. The positions of the exons and the corresponding nucleotides (diagonal numbers) are indicated. Note that PCR products of different length that are recognized by the same probe will be amplified from cDNA with a single set of PCR primers (up and dp). The 413 base pair probe (Wilson et al., 1996b) hybridizes to all five alternative mRNAs. The translated and untranslated regions of the full length and truncated Fas mRNAs are indicated by ovals and lines, respectively. LP, CR, TM, ST, and NR denote leader peptide, cyteinerich subdomains, transmembrane domain, signal transduction domain and negative regulation domain, respectively. AL indicates an altered amino acid sequence coded by the Fas mRNA variants. ∆ indicates Fas mRNA variants lacking TM or portions of the exons listed in parentheses. After Liu, Cheng and Mountz (1995) Biochem J. 310, 957–63 with permission.

F. 5 C. For legend see facing page.

F A S/I L-1 A P O P T O S IS 585

586

F. 6. Examination of ex vivo human corneal epithelial tissue for Fas ligand protein production by immunoprecipitation and western blotting. Two ex vivo corneal epithelial specimens removed at the time of excimer laser photorefractive keratectomy are shown. In each of these samples from different individuals a protein approximately 37 kDa and 26 kDa consistent with the membrane bound and soluble forms, respectively, of Fas ligand were detected. Note that in the first ex vivo epithelial sample there is a doublet band at approximately 37 kDa. The significance of the two proteins at nearly the same size is unknown. Lane C is a control lane containing purified membrane bound Fas ligand antigen (Upstate Biotechnology Incorporated) 37 kDa in size. In specimen 1, and several other samples that are not shown, the soluble 26 kDa band appeared to be darker than the 37 kDa band in ex vivo epithelial samples obtained by scraping.

chain of events terminating in cell death and removal (Wyllie, Kerr and Currie, 1980 ; Arends and Wyllie, 1991 ; Bellamy et al., 1995). Often these cues are delivered or modified by cytokines or growth factors and, depending on the cell type, occur in response to delivery or deprivation of a key modulator(s). Recent studies have demonstrated keratocyte apoptosis occurs in response to corneal epithelial injury (Wilson et al., 1996a ; Nakayasu, 1988 ; Crosson, 1989 ; Campos et al., 1994 ; Wilson et al., 1997), and provided evidence that the IL-1 and Fas}Fas ligand cytoline-receptor systems have a role in regulation of this response (Wilson et al., 1996a, 1996b). The present study demonstrates keratocyte apoptosis in response to corneal epithelial wounding is decreased, but not eliminated, by inactivation of the Fas ligand gene. There was also a trend towards a significant reduction in this response in mice homozygous for the Fas receptor inactivation. The lack of a significant difference in this group is likely attributable to the scatter of the data, since Fas ligand and Fas are the cytokine and receptor, respectively, for the same Fas system. Thus, this quantitative TUNEL method does have sensitivity to detect changes in apoptosis due to defects in modulators mediating apoptosis. In recent studies, we have shown greater than 75 % reduction in apoptosis in Stat 1 ®}® compared with control ­}­ mice, with little scatter of the data, using identical methods (Kim, Kumar, Stark and Wilson, unpublished data). Since keratocyte apoptosis in response to corneal epithelial wounding was reduced, but not eliminated, in Fas ligand ®}® or Fas ®}® animals, we hypothesize other cytokine systems are also involved

R. R. M O H A N E T A L.

in mediating this response. This is also suggested by the scatter of the data for Fas ligand [Fig. 1(F)], despite a significant reduction in keratocyte apoptosis. Such redundancy in the regulation of important functions is commonly detected in animals in which a system is inactivated by genetic manipulation. Previous studies have suggested the IL-1 system also modulates keratocyte apoptosis (Wilson et al., 1996a). The role of other systems in this response is being investigated in our laboratory. This study also indicates the importance of tight regulation of the Fas}Fas ligand system lest there be inadvertent destruction of cells important to the structure and function of the cornea. The Fas}Fas ligand system is expressed and apparently competent to mediate apoptosis in many organ systems (Itoh et al., 1991 ; Ni et al., 1994 ; Oishi, Maeda and Sugiyama, 1994 ; Sayama et al., 1994). Fas ligand is expressed in epithelial and endothelial cells, but not keratocytes, in the normal unwounded cornea (Griffith et al., 1995 ; Wilson et al., 1996b). Fas receptor is expressed in epithelial, keratocyte, and endothelial cells in unwounded human cornea and an anti-Fas receptor agonist antibody was previously shown to trigger apoptosis in corneal fibroblast cells (Wilson et al., 1996b). The current study demonstrates antibody activation of Fas also triggers death of corneal epithelial and endothelial cells in vitro. Death of these cells in response to Fas activation is accompanied by classical signs of apoptosis that include cell shrinkage, chromatin condensation, and the formation of apoptotic bodies (Wyllie, Kerr and Currie, 1980 ; Arends and Wyllie, 1991 ; Bellamy et al., 1995). Death of corneal cells in culture in response to Fas activation was delayed to 8 to 12 hr after exposure to Fasactivating antibody. Similar to the delayed apoptosis of cultured corneal fibroblasts in response to Fasactivating antibody compared to the immediate death of anterior keratocytes in vivo after corneal epithelial wounding (Wilson et al., 1996b), the delay in epithelial and endothelial cells could be related to changes that occur to cells when they are cultured. For example, alterations in the expression of antiapoptotic factors such as NF kappa B, BCL-2, BCL-XL could occur in response to serum or growth factors added to the cultures to maintain cell viability. The nature of these changes are poorly characterized, but should be investigated in future studies. Several investigators have demonstrated Fas and Fas ligand interaction resulting in apoptosis can occur by autocrine mechanisms (Dhein et al., 1995 ; Brunner et al., 1995 ; Liles et al., 1996). Simultaneous expression of Fas and Fas ligand by corneal epithelial and endothelial cells, along with competency to undergo apoptosis in response to Fas activation, possess an interesting enigma. How is unwanted activation of Fas and ensuing apoptosis regulated in the cornea ? Clearly there must be regulatory systems in place in the cornea to modulate unpropitious activation of Fas.

F A S/I L-1 A P O P T O S IS

587

The cytoplasmic domain of the membrane associated cognate Fas receptor mediates apoptotic signal transduction pathways leading to cell death. Soluble Fas, similar to numerous other soluble cytokine and growth factor receptors, has been shown to compete with membrane bound Fas for available Fas ligand and protect against apoptosis (Cheng et al., 1994 ; Liu, Cheng and Mountz, 1995 ; Cascino et al., 1995 ; Papoff et al., 1996). Our results demonstrate that corneal cells express mRNAs coding for several soluble variants of Fas. The alternative mRNAs coding for the full length Fas and soluble Fas missing only the transmembrane domain [Liu, Cheng and Mountz, 1995 and Fig. 5(C)] are present at highest levels in vitro [Fig. 5(A)], although mRNAs coding for other soluble Fas proteins can also be detected, albeit at very low levels [Fig. 5(B)]. There are likely to be other mechanisms regulating Fas activation in the cornea. For example, a second signal may be required for Fas activation in vivo. Also, Fas ligand activity can be modulated by binding of Fyn to the cytoplasmic domain of Fas ligand (Hane et al., 1995) and the oncogene abl may serve as a negative regulator of Fasmediated cell death (McGahon et al., 1995). Clearly, the systems modulating Fas activation in the cornea are complex and additional investigation will be needed to delineate the Fas regulatory network and to determine whether pathophysiologic activation of this system contributes to corneal disease. What functions are regulated by the Fas}Fas ligand system in the cornea ? Griffith et al. (1995) demonstrated that Fas ligand produced by ocular cells could stimulate apoptosis of inflammatory cells expressing Fas. They provided strong evidence implicating these interactions in immune privilege within the cornea and other areas of the eye (Griffith et al., 1995). However, it is likely the Fas}Fas ligand system also regulates apoptosis in corneal cells since all three cell types express Fas competent to trigger apoptosis. We previously hypothesized (Wilson et al., 1996b) the Fas}Fas ligand system modulates apoptosis of the underlying keratocytes occurring following corneal epithelial injury (Wilson et al., 1996a). Prior to this study, however, it was not clear how corneal epithelial Fas ligand could mediate keratocyte apoptosis, since

F. 7. IL-1-induced Fas ligand mRNA and protein expression by cultured corneal fibroblast cells detected by northern blotting and immunoprecipitation, respectively. (A) Fas ligand mRNA of the expected size of 1±2 kilobases was detected in second passage corneal fibroblasts cultured for 8 days with 10 ng ml−" IL-1 alpha (IL-1), but not vehicles (CON). Messenger RNA was isolated at the onset of cell death in the corneal fibroblast cultures incubated with IL-1 alpha. The same blot was stripped and reprobed for beta actin mRNA to demonstrate equivalent loading for each sample. Beta actin mRNAs of the expected sizes of 1±9 and 1±3

kilobases were detected in both vehicle (CON) and IL-1 alpha (IL-1) stimulated corneal fibroblast cells. Fas ligand mRNA induction was detected in three identical experiments with IL-1 alpha-stimulated corneal fibroblast cells compared with control cells. (B) Membrane bound Fas ligand protein of the expected size of 37 kDa was detectable in IL-1 beta and IL1 alpha (small arrows), but not vehicle (control), stimulated corneal fibroblasts. The level of expression of the 37 kDa protein in corneal fibroblasts with IL-1 stimulation was nearly the same as that in ex vivo corneal epithelium (small arrow). Soluble Fas ligand protein 27 kDa in size was not detected in the IL-1-stimulated corneal fibroblasts, but was at low levels (hollow arrow) in the ex vivo HCE positive control sample.

588

Fas ligand has been best characterized as a membrane bound protein. Fas ligand, however, is converted to a soluble protein by metalloproteinase-mediated release in some cell types (Tanaka et al., 1995 ; Mariana et al., 1995 ; Kayagaki et al., 1995). Our results demonstrate that human corneal cells, including injured ex vivo human corneal epithelium, produce soluble Fas ligand. In fact, in many ex vivo corneal epithelial specimens (Fig. 6, sample 1, for example) the soluble Fas ligand protein is detected at higher levels than the membrane bound protein. Since the soluble form is derived from the membrane bound form, it would be expected that the level of the latter would fall as the level of the soluble form increases, unless there was rapid induction of new Fas ligand protein production in response to epithelial scraping. Thus, soluble Fas ligand released from injured corneal epithelial cells may bind to anterior stromal keratocytes and trigger apoptosis. Other cytokine systems may also participate in keratocyte apoptosis in response to epithelial injury. We previously demonstrated that IL-1 alpha and beta (IL-1) can trigger apoptosis of corneal fibroblasts in vitro (Wilson et al., 1996a). Similarly, IL-1 can mediate keratocyte apoptosis in vivo when microinjected into the corneal stroma (Wilson et al., 1996a). One puzzling aspect of the previous study, however, was the significant delay (6–10 days) before IL-1 triggered corneal fibroblast apoptosis (Wilson et al., 1996a), whereas Fas agonist antibody stimulated rapid apoptosis (hours). The present study suggests the effect of IL-1 in stimulating corneal fibroblast apoptosis may be indirectly mediated via the Fas}Fas ligand system. Thus, IL-1 stimulates Fas ligand mRNA and protein production in corneal fibroblasts. Fas ligand protein expression is not detected in keratocytes in the unwounded cornea (Griffith et al., 1995 ; Wilson et al., 1996b) or in cultured corneal fibroblasts in the absence of IL-1 (Fig. 7). Thus, IL-1 may trigger corneal fibroblast apoptosis by induction of Fas ligand mRNA and protein expression. We did not examine the effect of IL-1 beta on Fas ligand mRNA expression in corneal fibroblasts in these studies. It is likely, however, that the effect of IL-1 beta on Fas ligand mRNA expression in corneal fibroblasts would be the same as that for IL1 alpha since (1) the two cytokines bind the same receptors and the effects of IL-1 alpha and IL-1 beta are similar in other systems (Dinarello, 1994) and (2) IL-1 beta induced expression of Fas ligand protein in corneal fibroblasts [Fig. 7(B)]. It appears only the expression of the membrane bound 37 kDa form, and not the soluble 26 kDa form, of Fas ligand is increased in corneal fibroblasts by IL-1 alpha [Fig. 7(B)]. Thus, this study suggests there may be two mechanisms through which epithelial injury could trigger keratocyte apoptosis. First, soluble Fas ligand released from the injured epithelium may bind keratocyte Fas. Alternatively, IL-1 released from the injured epithelium could induce Fas ligand expression in kerato-

R. R. M O H A N E T A L.

cyte cells leading to autocrine modulation of death (Dhein et al., 1995 ; Brunner et al., 1995 ; Liles et al., 1996). These alternatives are not mutually exclusive. Other yet to be identified systems may also mediate this response. Regulation of apoptosis in the cornea is complex and involves a number of mediators that are likely to interact in maintaining the balance between proliferation, differentiation, and apoptosis. This study provides some insights into these mediators and potential interactions between them. A more thorough understanding of apoptosis should lead to a better understanding of normal physiology, pathophysiology, and response to wounding of the cornea and, possibly, to strategies for controlling these processes. Acknowledgements Supported by US Public Health Service grants EY10056 from the National Institutes of Health, Bethesda, Maryland and The Cleveland Clinic Foundation. Presented in part at the Keystone Symposium Ocular Cell and Molecular Biology, Tammaron, CO, U.S.A. in January, 1997, and Association for Research in Vision and Ophthalmology, Ft Launderdale, FL, U.S.A. in May, 1997. We would like to acknowledge John Gabrovsek of The Cleveland Clinic Foundation for assistance with electron microscopy.

References Arends, M. J. and Wyllie, A. H. (1991). Apoptosis : mechanisms and roles in pathology. Int. Rev. Exp. Pathol. 32, 223–54. Bellamy, C. O., Malcomson, R. D., Harrison, D. J. and Wyllie, A. H. (1995). Cell death in health and disease : the biology and regulation of apoptosis. Sem. Cancer Biol. 6, 3–16. Brunner, T., Mogil, R. J., LaFace, D., Yoo, N. J., Mahboubi, A., Echeverri, F., Martin, S. J., Force, W. R., Lynch, D. H. and Ware, C. F. (1995). Cell-autonomous Fas (CD95)}Fas-ligand interaction mediates activation-induced apoptosis in T-cell hybridomas. Nature 373, 441–4. Campos, M., Szerenyi, K., Lee, M., McDonnell, J. M. and McDonnell, P. J. (1994). Keratocyte loss after corneal deepithelialization in primates and rabbits. Arch Ophthalmol. 112, 254–60. Cascino, I., Fiucci, G., Papoff, G. and Ruberti, G. (1995). Three functional soluble forms of the human apoptosisinducing Fas molecule are produced by alternative splicing. J. Immunol. 154, 2706–13. Cheng, J., Zhou, T., Liu, C., Shapiro, J. P., Brauer, M. J., Kiefer, M. C., Barr, P. J. and Mountz, J. D. (1994). Protection from Fas-mediated apoptosis by a soluble form of the Fas molecule. Science 263, 1759–62. Crosson, C. E. (1989). Cellular changes following epithelial abrasion. In Healing processes in the cornea, (Eds Beuerman, R. W., Crosson, C. E. and Kaufman, H. E.). Pp. 3–14. Gulf Publishing Co : Houston, TX, U.S.A. Deckwerth, T. L. and Johnson, E. M. Jr. (1993). Temporal analysis of events associated with programmed cell death (apoptosis) of sympathetic neurons deprived of nerve growth factor. J. Cell Biol. 123, 1207–22. Dhein, J., Walczak, H., Baumler, C., Debatin, K. M. and Krammer, P. H. (1995). Autocrine T-cell suicide mediated by APO-1}(Fas}CD95). Nature 373, 438–41.

F A S/I L-1 A P O P T O S IS

Dinarello, C. A. (1994). The interleukin-1 family : 10 years of discovery. FASEB J. 8, 1314–25. Gavrieli, Y., Sherman, Y. and Ben-Sasson, S. A. (1992). Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J. Cell Biol. 119, 493–501. Gold, R., Schmied, M., Giegerich, G., Breitschopf, H., Hartung, H. P., Toyka, K. V. and Lassmann, H. (1994). Differentiation between cellular apoptosis and necrosis by the combined use of in situ tailing and nick translation techniques. Lab. Invest. 71, 219–25. Graham, A., Francis-West, P., Bickell, P. and Lumsden, A. (1994). The signalling molecule BMP4 mediates apoptosis in the rhombencephalic neural crest. Nature 372, 684–6. Griffith, T. S., Brunner, T., Fletcher, S. M., Green, D. R. and Ferguson, T. A. (1995). Fas ligand-induced apoptosis as a mechanism of immune privilege. Science 270, 1189– 92. Hane, M., Lowin, B., Peitsch, M., Becker, K. and Tschopp, J. (1995). Interaction of peptides derived from the Fas ligand with the Fyn-SH3 domain. FEBS Lett. 373, 265–8. Itoh, N., Yonehara, S., Ishii, A., Yonehara, M., Mizushima, S., Sameshima, M., Hase, A., Seto, Y. and Nagata, S. (1991). The polypeptide encoded by the cDNA for human cell surface antigen Fas can mediate apoptosis. Cell 66, 233–43. Kayagaki, N., Kawasaki, A., Ebata, T., Ohimoto, H., Ikeda, S., Inoue, S., Yoshino, K., Okumura, K. and Yagita, H. (1995). Metalloproteinase-mediated release of human Fas ligand. J. Exp. Med. 182, 1777–83. Laing, R. A., Neubauer, L., Oak, S. S., Kayne, H. L. and Leibowitz, H. M. (1984). Evidence for mitosis in adult corneal endothelium. Ophthalmology 91, 1129–34. Li, Q., Weng, J., Mohan, R. R., Bennett, G. L., Schwall, R., Wang, Z.-F., Tabor, T., Kim, J., Hargrave, S., Cuevas, K. H. and Wilson, S. E. (1996). Hepatocyte Growth Factor (HGF) and HGF receptor protein in lacrimal gland, tears, and cornea. Invest. Ophthalmol. Vis. Sci. 37, 727–39. Liu, C., Cheng, J. and Mountz, J. D. (1995). Differential expression of human Fas mRNA species upon peripheral blood mononuclear cell activation. Biochem. J. 310, 957–63. Liles, W. C., Kiener, P. A., Ledbetter, J. A., Aruffo, A. and Klebanoff, S. J. (1996). Differential expression of Fas (CD95) and Fas ligand on normal human phagocytes : implications for the regulation of apoptosis in neutrophils. J. Exp. Med. 184, 429–40. Mariana, S. M., Matiba, B., Ba$ umler, C. and Krammer, P. H. (1995). Regulation of cell surface (APO-1}Fas (CD95) ligand expression by metalloproteases. Eur. J. Immunology 25, 2303–7. McGahon, A. J., Nishioka, W. K., Martin, S. J., Mahboubi, A., Cotter, T. G. and Green, D. R. (1995). Regulation of the Fas apoptotic cell death pathway by Abl. J. Biol. Chem. 270, 22625–31. Murphy, E. D., Roths, J. B. and Eicher, E. (1982). Generalized lymphoproliferative disease (gld). Mouse News Lett 67, 20–1. Nakayasu, K. (1988). Stromal changes following removal of epithelium in rat cornea. Jpn. J. Ophthalmol. 32, 113–25. Ni, R., Tomita, Y., Matsuda, K., Ichihara, A., Ishimura, K., Ogasawara, J. and Nagata, S. (1994). Fas-mediated apoptosis in primary cultured mouse hepatocytes. Exp. Cell Res. 215, 332–7. Oishi, M., Maeda, K. and Sugiyama, S. (1994). Distribution of apoptosis-mediating Fas antigen in human skin and

589

effects of anti-Fas monoclonal antibody on human epidermal keratinocyte and squamous cell carcinoma cell lines. Arch Dermatol. Res. 286, 396–407. Papoff, G., Cascino, I., Eramo, A., Starace, G., Lynch, D. H. and Ruberti, G. (1996). An N-terminal domain shared by Fas}Apo-1 (CD95) soluble variants prevents cell death in vitro. J. Immunol. 156, 4622–30. Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989). Molecular cloning. A laboratory manual. Second edition, Pp. 7.37–7.50, 8±60–8±74, 18.30–18.75. Cold Spring Harbor : Cold Spring Harbor Laboratory Press. Sayama, K., Yonehara, S., Watanabe, Y. and Miki, Y. (1994). Expression of Fas antigen on keratinocytes in vivo and induction of apoptosis in cultured keratinocytes. J. Invest. Dermatol. 103, 330–6. Suarez, P., Diaz-Guerra, M., Prieto, C., Esteban, M., Castro, J. M., Nieto, A. and Ortin, J. (1996). Open reading frame 5 of porcine reproductive and respiratory syndrome virus as a cause of virus-induced apoptosis. J. Virol. 70, 2876–82. Takahashi, T., Tanaka, M., Brannan, C. I., Jenkins, N. A., Copeland, N. G., Suda, T. and Nagata, S. (1994). Generalized lymphoproliferative disease in mice, caused by a point mutation in the Fas ligand. Cell 76, 969–76. Tanaka, M., Suda, T., Takahashi, T. and Nagata, S. (1995). Expression of the functional soluble form of human Fas ligand in activated lymphocytes. EMBO J. 14, 1129–35. Watson, M. L., Rao, J. K., Gilkeson, G. S., Ruiz, P., Eicher, E. M., Pisetsky, D. S., Matsuzawa, A., Rochelle, J. M. and Seldin, M. F. (1992). Genetic analysis of MRL-lpr mice : relationship of the Fas apoptosis gene to disease manifestations and renal disease-modifying loci. J. Exp. Med. 176, 1645–56. Wilson, S. E., Walker, J. W., Chwang, E. L. and He, Y.-G. (1993). Hepatocyte Growth Factor (HGF), Keratinocyte Growth Factor (KGF), their Receptors, FGF Receptor-2, and the Cells of the Cornea. Invest. Ophthalmol. Vis. Sci. 34, 2544–61. Wilson, S. E., He, Y.-G., Weng, J., Zeiske, J. D., Jester, J. V. and Schultz, G. S. (1994). Effect of epidermal growth factor, hepatocyte growth factor, and keratinocyte growth factor, on proliferation, motility, and differentiation of human corneal epithelial cells. Exp. Eye Res. 59, 665–78. Wilson, S. E., He, Y.-G., Weng, J., Li, Q., McDowall, A. W., Vital, M. and Chwang, E. L. (1996a). Epithelial injury induces keratocyte apoptosis : Hypothesized role for the interleukin-1 system in the modulation of corneal tissue organization and wound healing. Exp. Eye Res. 62, 325–38. Wilson, S. E., Li, Q., Weng, J., Barry-Lane, P. A., Jester, J. V., Liang, Q. and Wordinger, R. J. (1996b). The Fas}Fas Ligand System and Other Modulators of Apoptosis in the Cornea. Invest. Ophthalmol. Vis. Sci. 37, 1582–92. Wilson, S. E., Pedroza, L., Beuerman, R. and Hill, J. M. (1997). Herpes simplex virus type-1 infection of corneal epithelial cells induces apoptosis of the underlying keratocytes. Exp. Eye Res. 64, 775–9. Wyllie, A. H., Kerr, J. F. and Currie, A. R. (1980). Cell death : the significance of apoptosis. Int. Rev. Cytol. 68, 251–306. Yao, R. and Cooper, G. M. (1996). Growth factor-dependent survival of rodent fibroblasts requires phosphatidylinositol 3-kinase but is independent of pp70S6K activity. Oncogene 13, 343–51. Zou, H. and Niswander, L. (1996). Requirement for BMP signaling in interdigital apoptosis and scale formation. Science 272, 738–41.