Fas-Fas ligand-mediated apoptosis within aqueous during idiopathic acute anterior uveitis

Fas–Fas Ligand–Mediated Apoptosis within Aqueousduring Idiopathic Acute Anterior Uveitis

Andrew D. Dick, Kirsten Siepmann, Chrisjan Dees, Linda Duncan, Cathryn Broderick,Janet Liversidge, and John V. Forrester

PURPOSE. Despite ocular immune privilege, (auto)immune-mediated acute anterior uveitis (AAU) isrelatively common. However, although relapses of AAU are usually self-limiting, possible regulatorymechanisms remain undefined in humans. Experimentally, Fas–Ligand (FasL)–mediated apoptosisof Fas1 inflammatory cells contributes to the immune privilege within the anterior chamber andprovides an explanation for the success of corneal allograft transplantation. Therefore, whethersuch mechanisms regulate the immune response in AAU was investigated.

METHODS. Aqueous and peripheral blood samples from consecutive patients presenting withidiopathic AAU were obtained with consent. Leukocytic phenotype was analyzed by flow cytom-etry, and apoptosis was determined by both flow cytometry and TdT-dUTP terminal nick-endlabeling analysis. Presence of soluble Fas and FasL was determined by western blot analysis andenzyme-linked immunosorbent assay and compared with control aqueous from patients undergo-ing cataract surgery. The ability of the aqueous to induce apoptosis in a Fas1 Jurkat cell line wasalso determined.

RESULTS. During AAU aqueous-infiltrating Fas1 cells included CD31 T cells and granulocytes,whereas FasL1 cells comprised predominantly of non–CD31 T cells. Higher levels of functionalsoluble FasL were found in aqueous of AAU patients than in normal aqueous, capable of inducingapoptosis in 68.9% 6 7.6% of Fas1 lymphoid cells. Compared with peripheral blood, the CD41 Tcells infiltrate within aqueous showed significantly increased CD69 and CD25(IL-2r) expression.Flow cytometric analysis of aqueous showed that 9.32% 6 1.2% of infiltrating non–granulocyteCD451 cells were apoptotic, confirmed as T cells on subsequent three-color flow cytometricanalysis.

CONCLUSIONS. Taken together with published experimental data, the present study provides evi-dence for FasL-mediated apoptotic cell death contributing to the local immune regulation of ocularinflammatory disease and provides a mechanism to account for the self-limiting clinical course ofAAU. (Invest Ophthalmol Vis Sci. 1999;40:2258–2267)

The eye, like the central nervous system, is regarded as animmune privileged site (principally because of the highsuccess of corneal allograft transplantation), yet intraoc-

ular inflammation remains not an uncommon occurrence. Tra-ditionally, this immune privilege was thought to be due tophysical blood–ocular and blood–retinal barriers, the absenceof lymphatics, and the paucity of antigen-presenting cellswithin ocular tissues. However, it has been documented thatfirst a network of major histocompatibility complex (MHC)class II–positive cells,1,2 some of which behave functionally asdendritic cells existing within the uveal tract of the eye.3

Second, a form of lymphatic drainage exists as antigen-specific

T-cell expansion in the submandibular lymph node occurs afterintraocular antigen administration.4 The phenomenon of ante-rior chamber–associated immune deviation (ACAID),5,6 inwhich foreign antigens and tissues when administered into theanterior chamber fail to elicit an immune response and further-more induce suppression of antigen-specific delayed-type hy-persensitivity (DTH) responses, have prompted investigationinto the cellular and molecular bases of the regulation ofimmune responses within the eye.7 Since the observation thatFas-Ligand (FasL) is constitutively expressed in murine oculartissue such as iris ciliary body and corneal endothelium,8 ex-periments have shown that Fas/FasL-dependent apoptosis isinduced in inflammatory cells entering the eye in response toa viral infection, in which the ability to initiate apoptosis andinterleukin (IL)-10 appears critical for the induction of immuneprivilege.9,10 Furthermore, without Fas/FasL, splenocytesprime for DTH rather than induce ACAID.11 Although themolecular mechanisms of immune deviation are still not fullyunderstood, and despite recent work in autoimmune diabetesthat has cast doubt on the theory of FasL-induced lymphocyteapoptosis as a damage-limiting mechanism,12 present experi-mental evidence within the eye strongly supports an importantrole for Fas-L not only in controlling intraocular inflammationbut also in the induction of immune deviation and the accep-

From the Department of Ophthalmology, University of AberdeenMedical School, Foresterhill, Aberdeen, Scotland, UK.

Supported by the Royal College of Surgeons of Edinburgh, theRoyal Blind School and Asylum (ADD, LD), and the Guide Dogs for theBlind (ADD, CB).

Submitted for publication December 3, 1998; revised April 2,1999; accepted May 4, 1999.

Proprietary interest category: N.Corresponding author: Andrew D. Dick, Department of Ophthal-

mology, University of Aberdeen Medical School, Foresterhill, AberdeenScotland, UK AB25 2ZD.E-mail: [email protected]

Investigative Ophthalmology & Visual Science, September 1999, Vol. 40, No. 102258 Copyright © Association for Research in Vision and Ophthalmology

tance of corneal allografts.13 If Fas-FasL–mediated interactionsare important during immune-mediated ocular inflammatorydisease, such as acute anterior uveitis (AAU), experimental datawould suggest that cell death within the anterior chamberwould require infiltrating mononuclear cells to be Fas1 andresident ocular cells to be FasL110 or that FasL is induced oninfiltrating cells by ocular resident cell products.

AAU is pathogenetically distinct from posterior uveitis,14

although animal models show that CD41 T cells are intimatelyinvolved in the immunopathogenesis of both diseases.14,15

Clinically, AAU is an acute self-resolving condition and is fre-quently associated with HLA-B27 MHC class I antigen.16 Mostprevious studies of cellular infiltrate and immune mechanismsin uveitis have been limited because they have included a widespectrum of uveitis conditions, including posterior uveitis as-sociated with systemic disease.17 Despite this, all studies incommon show that CD41 T cells are found within the aqueoushumor in greater numbers than CD81 cells.18–20 Studies havealso shown that apoptosis of infiltrating mononuclear cellswithin the aqueous humor occurs during noninfectious in-traocular inflammation such as Vogt–Koyanagi–Harada (VKH)syndrome.21 In addition to apoptosis, many of the CD41 T cellswithin the aqueous and the cerebral spinal fluid in VKH wereFas1 CD291CD45RA1 T (memory) cells, and regulation of theinflammatory response by Fas-FasL–mediated apoptotic celldeath was therefore implicated. However, no study of humanocular inflammatory disease to date has shown that Fas-FasLinteractions are functionally important in regulating immuneresponses in the eye. The present study further examined thephenotype of cells infiltrating the aqueous humor (particularlywith respect to CD41 T-cell activation, Fas/FasL expression,and levels of soluble Fas/FasL during AAU) and demonstratesthat the aqueous contains soluble FasL (sFasL) capable of in-ducing apoptosis in Fas1 cells.

METHODS

Patients, Diagnosis, and Aqueous Sampling

In this prospective study, patients who presented to the emer-gency rooms of the Eye Department of the Aberdeen RoyalInfirmary (Aberdeen, Scotland) with a clinical diagnosis of AAUwere enrolled after informed consent and local ethical com-mittee approval in accordance with the tenets of the Declara-tion of Helsinki. Patients with AAU had no clinical evidence orlaboratory findings to suggest posterior uveitis or anterior uve-itis associated with systemic disease. At the time of presenta-tion, patients were taking no immunosuppressive agents ortopical dexamethasone therapy for the treatment of their uve-itis. Aqueous sampling was taken as previously described,22

with no complications observed as a result of the paracentesis.Samples (volumes of 0.1–0.2 ml) were placed immediately intoan Eppendorf vial on ice, and after centrifugation (1000 rpmfor 10 minutes) cells were further processed as describedbelow for each assay. Residual aqueous was analyzed for solu-ble Fas (sFas) and sFasL by enzyme-linked immunosorbentassay (ELISA), western blot analysis, and induction of apoptosisin Jurkat cells (see below). Simultaneously, 5 ml of peripheralvenous blood was collected into a sodium heparin evacuatedtube (Vacutainer) for leukocyte flow cytometric phenotypeanalysis and determination of HLA-B27 status. Aqueous andperipheral blood samples of patients undergoing routine cata-

ract surgery (patients without any signs or history of uveitis orother ocular disease) were used as controls.

Flow Cytometric Analysis

Two- and three-color immunophenotyping of peripheral bloodleukocytes were performed using mouse monoclonal antibody(mAb) specific for human cell surface markers, which wereobtained from Becton–Dickinson unless otherwise stated.These included CD45 (leukocyte common antigen), CD14 (LPScell surface receptor on monocytes), CD3 (T cells), CD8, CD4,HLA DR (MHC class II antigen), CD69 (activated T-cell blasts),CD19 (B cells), CD161CD56 (natural killer cells [NK]), CD95(Fas; Calbiochem), NOK-1bio (Fas-L; Pharmingen), and HLAB7/B27 (Serotec, Oxford, UK). Aliquots of 100 ml of peripheralblood were added to round-bottomed polystyrene tubes (Fal-con), and directly conjugated mAb (conjugated to fluoresceinisothiocyanate [FITC], phycoerythrin [PE], or Per chlorophyllprotein [PerCP]) were added at predetermined optimal dilu-tions. After a 30-minute incubation, aliquots were lysed withFACSlyse (Becton–Dickinson) as per manufacturer’s instruc-tions and then washed twice in FACS buffer (phosphate-buff-ered saline [PBS[]/0.2% bovine serum albumin). UnconjugatedmAb was detected with rat absorbed FITC-conjugated sheepF(ab9)2, anti-mouse immunoglobulin (Sigma), and biotinylatedmAb with streptavidin–PE (Caltag). Cells were then fixed in 1%paraformaldehyde and kept at 4°C until analysis. Acquisitionwas performed on FACSCalibur flow cytometer and analyzedusing CellQuest acquisition and analysis software. Leukocytegates and instrument variables were set according to forwardand side scatter characteristics and using appropriate uncon-jugated, biotinylated, or directly conjugated isotype immuno-globulin controls. Analysis of fluorescence was performed afterfurther back gating to exclude dead cells and aggregates. Aque-ous cells (1500–3000 cells per sample) were similarly pre-pared, without lysis, and distributed equally into polystyrenetubes for cell surface labeling.

Estimation of sFas and sFasL

sFas and sFasL were estimated in aqueous of AAU patients andcontrol cataract patients by commercial ELISA. One hundredmicroliters of sample diluent (1:2) was used. sFas (APO-1) wasdetermined using a commercial capture ELISA (Bender Med-systems) and standardized against recombinant sFas. Detectionfor reading at 450 nm was performed with streptavidin–per-oxidase. Similarly, sFasL was determined using a standard com-mercial capture ELISA, using purified capture and peroxidase-conjugated detector antibody pairs (Medical and BiologicalLaboratories, Nagoya, Japan), and standardized against recom-binant sFasL supplied by the manufacturer. For further identi-fication of sFasL, western blot analysis was performed after 8%to 25% sodium dodecyl sulfate–polyacrylamide gel electro-phoresis (SDS–PAGE; Phast system; Pharmacia, Uppsala, Swe-den). In addition to noncellular aqueous, a peptide of sFasL(PP61; Calbiochem Ltd) was run on the gel as a negativecontrol because G247-4 mAb (see below) did not detect pep-tide by ELISA or western blot analysis. Gels then were stainedroutinely with Coomassie blue. After protein separation, gelswere blotted onto a nitrocellulose membrane (0.45 mm) at70°C for 30 minutes for chemiluminescence detection of sFasL.Incubation with primary mAb anti-FasL (G247-4 clone; Pharm-ingen) was performed at room temperature for 1 hour on a

IOVS, September 1999, Vol. 40, No. 10 Apoptosis in Aqueous during Acute Anterior Uveitis 2259

shaker, and after rinsing in Tris-buffered saline, biotinylatedrabbit anti-mouse (Dako) mAb was added for a further 1 hourat room temperature. After further washes, development ofblot was achieved with streptavidin-biotinylated complex(Dako) and placed in development solution (Amersham) for 1to 5 minutes as per manufacturer’s instructions before x-raydetection.

Apoptosis Assay

Apoptosis of mononuclear cells from aqueous samples ob-tained from AAU patients was detected by flow cytometry andterminal deoxynucleotidyl transferase (TdT)–dUTP terminalnick-end label (TUNEL) staining of slide preparations. For flowcytometry, cells were fixed for 15 minutes in 1% paraformal-dehyde in PBS (pH 7.4); after resuspending in PBS, the samplewas further centrifuged and pellet resuspended in 70% ethanolat 220°C and stored until analysis. For analysis, ApopTag Plus(Oncor) was used according to manufacturer’s instructions todetect apoptosis by determining the increase in liberated 39OHDNA ends localized in apoptotic bodies. TdT was used tocatalyze the addition of digoxigenin-nucleotide residues toDNA ends generated by fragmentation, which were then de-tected using a FITC-conjugated anti-digoxigenin mAb and pro-pidium iodide. Further flow cytometric analysis was performedby staining aqueous cells with CD3FITC, Annexin V-PE, andViaprobe (Pharmingen) as per manufacturer’s instructions, todetermine the percentage of apoptotic CD31 T cells (seelegend; Fig. 5). In brief, after initial staining with CD3FITC andAnnexin-PE, cells were stained with Viaprobe in calcium–PBS/bovine serum albumin and read on FACSCalibur after voltagesettings and background fluorescence was set with appropriateisotype controls. TUNEL staining23 was also performed onstandard prepared ethanol-fixed slide preparations of cellularaqueous. Before fixation and staining, the slides were treated (1minute) with collagenase (75 mg/ml) and hyaluronidase (5U/ml) to prevent protein precipitation and cross-linking fromthe inhibiting staining procedure. TUNEL procedure was per-formed using a commercial kit (Trevigen), including standardcontrol slides. Any further protein present was digested with20 mg/ml of proteinase K. 39OH end labeling was again per-formed using TdT-digoxigenin labeling, and apoptotic cellswere detected using streptavidin-peroxidase and DAB sub-strate.

Induction of Apoptosis in Jurkat Cells

Jurkat cells (2 3 105/ml) were cultured in triplicate in 96-wellplates (200 ml) with an optimal concentration of mouse IgMmAb to Fas (clone CH-11; Upstate Biotechnology) or recombi-nant human Fas-L in the presence or not of anti-Fas mAb (CloneZB4 [anti-CD95; apo-1]; Immunotech) to block surface Fas onJurkat cells and thus induction of apoptosis. Test noncellularcomponent of aqueous from both AAU and control sampleswas added to cell cultures with and without anti-Fas mAb. Cellswere grown in RPMI containing 10% fetal calf serum for 4 and24 hours, and cells then were processed for apoptosis usingAnnexin-V/propidium iodide (PI) assay24 by flow cytometry(FACSCalibur; Becton–Dickinson). Apoptotic cells are charac-terized by Annexin-V1 PI2 phenotype. To assess any contribu-tion toward induction of apoptosis of tumor necrosis factor(TNF)-a present in aqueous, parallel triplicate cultures were setup with optimal concentration (10 ng/ml) of recombinant

human TNF-a (R1D Systems, Europe Ltd) or test aqueous withand without blocking with mAb anti-human TNF-a at 10 mg/ml(R1D Systems).

RESULTS

Fas-Positive Mononuclear Cells and ActivatedCD41 T Cells in Aqueous Humor during AAU

Cells isolated from aqueous samples of AAU patients wereanalyzed by two- and three-color flow cytometry and comparedwith peripheral blood phenotype. Twenty-seven peripheralblood and 12 aqueous samples were analyzed for phenotype.No patients had evidence of systemic disease or concomitantinfection at the time of sampling or as a possible cause of theiruveitis. Twenty patients were HLA-B27 positive. Figures 1A and1B present data of the phenotype of infiltrating cells within theaqueous compared with that of peripheral blood. Results indi-cated a relative increase in the percentage of T cells in theaqueous compared with that in peripheral blood (23.9% 68.7% versus 15.4% 6 1.5% T cells, respectively), whereas thepercentage of granulocytes in aqueous was reduced from thatseen in peripheral blood (21.25% 6 5.92% versus 61.3% 61.2%, respectively). The CD4-positive T-cell proportion withinthe aqueous was comparable to that in peripheral blood (45.5%6 10.1% and 47.7% 6 9.1% of CD31 gate in aqueous andperipheral blood, respectively). B cells were of a lower per-centage in the aqueous than in peripheral blood (1.2% 6 0.68%and 12% 6 1.2%, respectively). The majority of CD41 T cellswithin the aqueous was activated (as determined by three-color flow cytometric analysis of percentage of CD41CD69 orIL-2R1 expression on gated CD31 T cells), expressing CD69(73.3% 6 6.9%) and IL-2R (54.1% 6 6.4%). There was nodifference in aqueous cell phenotype and T-cell activationbetween HLA-B27–positive or –negative AAU patients (data notshown). Leukocyte subsets do not make up 100% of the cellsanalyzed because of the damage to cells during processing(paracentesis), and these have been excluded from analysis byscatterplot gate, or cells were fragmented as a result of necro-sis/end stage apoptosis. In addition, iris pigment epithelial cellsare liberated during the inflammatory process and do notexpress on their cell surface any leukocytic markers. Figure 1Cshows the percentage Fas and FasL cell surface expression onaqueous-infiltrating CD451 mononuclear cells. The majority ofperipheral T cells expressed Fas (55.6% 6 11.02%), compara-ble with the number of CD451Fas1 cells within the aqueous(68.6% 6 4.2%). Back gating the Fas1 cells within aqueous todetermine scatter characteristics of cells showed that the Fas1

cells were present within characteristic lymphocyte scatterprofile, although Fas expression on NK cells or the smallnumber of B cells present within the infiltrate cannot beexcluded. Compared with peripheral leukocytes (0.78% 60.5%), a high percentage (58% 6 20%) of leukocytes within theaqueous expressed FasL, and on back gating this populationdemonstrated a characteristic scatter profile consistent withnon–T-cell population. We subsequently performed three-color flow cytometric analysis for CD3, Fas, and FasL expres-sion, which confirmed our interpretation of the scatterplotsand showed that 98% of infiltrating CD31 T cells expressed Fasof which only 15% were also FasL positive.

2260 Dick et al. IOVS, September 1999, Vol. 40, No. 10

FIGURE 1. Comparison of cell sur-face phenotype and activation mark-ers between leukocytes infiltratingaqueous and peripheral blood. (A)Percentages of lymphocytes, mono-cytes, and granulocytes were calcu-lated according to CD45/CD14 ex-pression, NK cells according toCD561CD161 expression, and Bcells according to CD19 expression(10,000 events and 1,500–3,000events were collected from periph-eral blood and aqueous samples, re-spectively). (B) CD41 T-cell activa-tion was calculated by three-coloranalysis by gating on CD31 cells andthen subsequently calculating CD69or IL-2R expression on CD41 cells.(C) Fas and FasL was expressed as apercentage of infiltrating CD451

cells. *Denotes statistically significantdifferences of P , 0.02 (Mann–Whit-ney U test).


Increased Levels of sFasL and MononuclearApoptotic Cell Death Present in Aqueous Humorduring AAU

ELISA estimation of Fas and sFasL levels in the supernatant ofaqueous is shown in Figure 2. The protein concentration (Fig.2), not unexpectedly because of the concurrent breakdown ofthe blood-ocular barrier during ocular inflammation, was in-creased in AAU patients. Samples of vitreous were also in-cluded in patients with noninfectious autoimmune posterioruveitis as positive controls to compare protein concentrationsin other intraocular inflammatory conditions. In 8 of 16 pa-tients with AAU sFasL was detected within the aqueous byELISA (25.96 6 13.9 pg/ml; P , 0.039; Fig. 2), compared withundetectable levels in control patients with cataract. In addi-tion, there was no difference in protein concentration betweenFasL1 and FasL2 AAU patients, which may account for sFasL2

status. sFas was detected in only 4 AAU patients, whereas nosFas was detected by ELISA in controls. Increased levels ofsFasL were also seen in vitreous samples of patients withidiopathic noninfectious posterior uveitis compared with con-trols, although there was no significant difference in sFasLlevels between aqueous (AAU) and vitreous (posterior uveitis)samples. Noncellular aqueous was also analyzed by SDS–PAGEand western blot analysis. Electrophoresis of the aqueous hu-mor from AAU and control cataract patients showed a detect-able band of molecular weight (MW) similar to that of sFasL(approximately 26 kDa; Fig. 3A). Western blot analysis con-firmed that sFasL was present in the aqueous from AAU pa-tients (Fig. 3B). SDS–PAGE protein bands were also of MWcomparable to that of TNF-a in aqueous and cataract controlpatients, with the presence of TNF-a was confirmed by west-ern blot analysis (data not shown). Levels of apoptotic celldeath, as detected by ApopTag Plus (see the Methods section)in the aqueous humor of AAU patients (n 5 6), was 9.32% 61.25% of infiltrating lymphocytes (as defined by cell scatterprofile; Fig. 4A). Less than 2% of granulocyte scatter wasapoptotic. In a separate specimen, three-color flow cytometricanalysis confirmed that 55.6% of CD31 T cells were apoptotic(Annexin V1 Viaprobe2; Fig. 4C). Histochemical confirmationof apoptosis was obtained with cytospin TUNEL preparations(data not shown).

sFasL in Aqueous of AAU Patients and ApoptoticCell Death in Fas1 Jurkat Cells

Jurkat cells are greater than 90% Fas1 (data not shown), andrecombinant human sFasL induces apoptosis in 40.5% 6 2.92%of Jurkat cells after only a 4-hour incubation compared with13.54% 6 3.4% in medium alone (Fig. 5). Also we confirmedthe active induction of apoptosis in Jurkat cells with CH-11mAb and confirmed the inhibition of CH-11–induced apoptosiswith ZB4 clone, which was therefore used throughout theremainder of the experiments (CH-11–induced apoptosis in21.25% 6 5% of cells in 4 hours and was inhibited by priorincubation with ZB4 clone to basal levels of apoptosis of 12.9%6 1% of cells). Although the increase in the percentage ofapoptotic cell death after 4 hours’ incubation of Fas1 Jurkatcells with aqueous from AAU patients was not significant, by24 hours a significant increase in apoptotic cell death wasrecorded (69.8% 6 7.6%; P , 0.001). Apoptosis was effectivelyblocked with prior incubation with anti-Fas mAb (25.9% 67.6%; P , 0.01). Jurkat cells do not express quantifiable levels

FIGURE 2. ELISA quantification of sFasL and sFas in aqueous.Comparison of protein concentration (top), sFasL (middle), andsFas (bottom) in aqueous of AAU patients (AAU) and cataractpatients (controls) or vitreous from patients with idiopathic poste-rior uveitis undergoing vitrectomy. There were statistically signifi-cantly increased levels of sFasL in aqueous of AAU patients com-pared with that in controls (P , 0.039). Levels of sFasL withinaqueous were not accounted for by differences in protein concen-tration.


of membrane tumor necrosis factor receptor (TNFr) (data notshown), and therefore, not surprisingly, the addition of recom-binant human TNF-a did not induce apoptosis in Jurkat cellsafter 24 hours of incubation.

DISCUSSION

Several pathways of apoptotic regulation of inflammation andtolerance have been proposed,25 but the relative contributionto and importance of these in human disease remains unclear.In the present study we attempted to evaluate characteristics,both functional and phenotypical, of the inflammatory infil-trate within the immune privileged anterior chamber of the eyeby investigating aqueous samples from a cohort of patientswith idiopathic AAU. AAU is characterized by a brisk, unpre-dictable, self-resolving yet recurrent inflammation of the irisand ciliary body within the anterior chamber of the eye, ofwhich 50% of cases are HLA-B271.16 Recent experimentalevidence suggests that within the anterior chamber of the eyeFas/FasL-induced death is the mechanism by which cells arekilled, and furthermore it is the apoptotic signal that remainscritical to induce immunoregulation.5–7,9 We therefore hypoth-esized that regulation of the inflammatory response in AAU ismediated via Fas/FasL apoptotic cell death of infiltrating mono-nuclear cells.

Previous studies have shown that the aqueous leukocyticinfiltrate during a range of uveitis conditions comprises, inaddition to monocytes and CD81 T cells, CD451 CD291

CD45RO1 Fas1 CD41 T cells.20,21,26 Furthermore, in a cohortof patients with AAU CD41 and CD81 T cells are found to beassociated with high levels of IL-12 and IL-10 as well as inter-feron-g within the aqueous, although the precise cell origin ofcytokine production remains undefined.19 Despite obvious lim-itations to any study of leukocytes in the aqueous because ofthe small sample size routinely obtained, we have still con-firmed experimental data that Fas-FasL interactions are activeduring AAU. In addition, the data are confirmatory and showthat the aqueous leukocytic infiltrate consists of an increasedproportion of T cells compared with blood and that, althoughmonocyte percentage was less than T cells within the aqueous,they occurred in greater proportion than in peripheral blood(Fig. 1). Infiltrating CD41 T cells are activated, expressing highlevels of CD69 and IL-2R.27,28 Whether CD41 T cells are anti-gen-specific or not is not known, because putative autoanti-gens in AAU have not been confirmed to test this. Extrapola-tion from animal models suggests that AAU has an integralT-cell component and that no one cytokine is requisite to thepathogenesis of AAU.29 During AAU, the aqueous containsleukocytes, particularly T cells expressing Fas (CD95) and thenon–T-cell population expressing its ligand (CD95L); andmoreover, sFasL and Fas are detected in greater quantities than

FIGURE 3. SDS–PAGE electrophoresis and western blot analysis of sFasL in aqueous. Aqueous supernatant was obtained and run on SDS–PAGE gel(A) where a predominant band was noted with MW of approximately 30 kDa, corresponding to soluble ligands FasL and TNF-a after proteolyticprocessing of membrane proteins (accounting for the variance of MW between samples according to degree of glycosylation and breakdownpatterns). Chemiluminescence development (1 minute) of western blot analysis (B) to detect FasL with mAb (G272-4; Pharmingen) demonstratedsFasL in the aqueous of AAU patients and not control patients with cataract. However, prolonged development (5 minutes) showed traces of sFasLin the aqueous of control patients, on western blots (data not shown). PP61 peptide of sFasL which is not recognized by mAb used for detection(as determined by ELISA) was run as negative control. Bands visualized represent contaminating proteins within peptide preparation. Lanes 1through 4, aqueous of AAU patients; Lanes 5 and 6, aqueous of cataract patients; Lane 7, peptide of FasL (PP61: residues 261 to 267).


in control aqueous. Later experiments showed that the aque-ous could induce apoptotic FasL-dependent cell death in Fas1

Jurkat cells (Fig. 5). Normal aqueous also contains functionalsFasL and induces apoptosis of Fas1-dependent cell line, albeitat a lower level, suggesting a constitutive role for Fas/FasLsignaling in the anterior chamber.

FasL is a type II integral membrane protein homologouswith TNF.30,31 Membrane-bound FasL (mFasL) is released as a26-kDa soluble form, like TNF,32 by matrix metalloproteinases(MMPs),33 resulting in an equally functional active form. Re-cent data suggest that with certain cell lines and under certainconditions, in which mFasL is cleaved, the resultant sFasL doesnot induce apoptosis and therefore is downregulatory34,35 (seebelow). MMPs have been recorded in the normal aqueoushumor as well as raised levels in the aqueous during ocular

inflammation,36 which may account for the increase in func-tional sFasL in AAU samples. Although TNF-a is also found inincreasing amounts early in the course of endotoxin-induceduveitis,37,38 the neutralization of TNF-a activity resulted indisease exacerbation.39 Our data do not support or deny a rolefor TNF-mediated activation–induced apoptosis40 in regulationof the inflammatory response. In AAU TNF-a may play a dualrole. First, during the initial inflammatory response as a proin-flammatory cytokine, but secondly as a result of the chronicproduction of low levels of TNF-a,41 suppressing T-cell activa-tion,42 and thus contributing to the immunoregulatory envi-ronment of the anterior chamber. TNF-a may also act byinducing inactivation and apoptosis of the CD691 CD161

CD561 NK cells,43 which we observed infiltrating the anteriorchamber during AAU (data not shown).

FIGURE 4. Apoptosis of infiltrating mononuclear cells during AAU. (A) Flow cytometric analysis of apoptosis shows a range of positiveFITC-digoxigenin expression. Percentage of apoptosis was calculated from expression within T-cell scatter gates, because granulocyte gate showedless than 2% apoptosis. Regions of high FITC expression were taken to calculate percentage apoptotic cells, thus excluding staining by necroticcells. (B) Flow cytometric 3-color analysis confirmed CD31 T-cell apoptosis. (C) Possible necrotic cells were excluded with Viaprobe (7-aminoactinomycin D; Pharmingen), and remaining cells were analyzed for dual expression of CD3 and Annexin V53 (arrow) to confirm T-cell apoptosis.


Although the data strongly suggest that the majority ofFas1 cells are T cells (either CD41 or CD81), some cells withinmonocyte and macrophage scatter profiles also expressed Fas.It is conceivable, therefore, that Fas1 monocytes and neutro-phils also undergo apoptosis. However, flow cytometric anal-ysis showed that apoptotic events occurred predominantlywithin the scatter profile of T cells, which was confirmed byidentifying the CD31Annexin V1 apoptotic T-cell population(Fig. 4B). The role of Fas-FasL is still controversial. Althoughthere is compelling evidence that Fas-FasL is protective withinimmunoprivileged sites, particularly the anterior chamber ofthe eye,6 Fas-FasL interactions may result in exacerbation ofinflammation and solid organ graft rejection,12,44,45 via possi-ble FasL-mediated activation of Fas1 granulocytes, althoughinterestingly not Fas-dependent T-cell cytotoxicity.45 In addi-tion, as intimated earlier, MMP cleavage, environment (i.e.,presence of other cytotoxic agents such as TNF), and cell type(unactivated T cells are not responsive to sFasL-mediated apo-ptosis34) are integral as to whether sFasL remains proapop-totic.34,35 Evidence suggests that cleaved trimeric sFasL re-quires aggregation for proapoptotic action. Therefore, as aprotective response, sFasL is downregulatory mainly so thatcirculating sFasL does not possess devastating systemic conse-quences. Interestingly, in some samples we have observed andconfirmed on western blot analysis, MW of sFasL at around 70kDa, representing either aggregated sFasL or mFasL (data notshown). What makes sFasL proapoptotic in aqueous samples is

likely to be secondary to both a function of local MMP activityas well as yet undefined immune factors generated duringsequestered inflammatory response in addition to the underly-ing immunoregulatory environment, all of which may altersignaling on receptor engagement (similar to differential TNFeffects42). It is also possible that within the aqueous otherproteolytic activity as a result of granulocyte infiltration andactivation results in nonspecific cell death and thus the releaseof active mFasL from killed cells, such as iris epithelium. Thisin turn would represent a local protective response by gener-ating FasL to kill infiltrating activated Fas1 T cells. Potentially,therefore, one may postulate that during AAU a critical bal-ance, as a result of constitutive FasL expression within theanterior chamber, exists between Fas-dependent “activated”T-cell apoptosis attempting to suppress the inflammatory re-sponse and a proinflammatory nonspecific Fas-dependent gran-ulocyte toxicity. These results in part strongly support exper-imental evidence of immune regulation within the anteriorchamber of the eye during ocular inflammation, particularlywith respect to the observation, experimentally at least, thatFasL-induced apoptosis is necessary for corneal graft survival.13

Apoptosis has been proposed as a regulatory mechanismpivotal to generation of ACAID.8 In ACAID, TGF-b can preferen-tially induce antigen presenting cells to secrete IL-10, which itselfis central to the induction of ACAID.46 Interleukin-10 is present ingreater quantities than interferon-g within the aqueous duringAAU19 and therefore IL-10 may suppress DTH reactivity and anti-gen-specific responsiveness47 and direct the inflammatory re-sponse toward Th2 as has been shown experimentally.48 Addi-tionally, antigen presenting cells may traffic from the eye andinduce Th2 responses when antigen presentation occurs withinthe local drainage lymph nodes.4 Alternatively, other mechanismsthat preserve Th2 responses may be secondary to the increasedsusceptibility of Th1 cells to FasL-mediated apoptosis and, thus,their preferential deletion when entering the eye.49 Although wehave been unable to identify which subset of T cells is undergoingapoptosis (because of the low numbers of cells for analysis), wepostulated that it is predominantly CD41 T cells because thesecells are highly activated and express Fas (Fig. 1). Moreover,CD81 T cells are less susceptible to FasL-mediated apoptosis40

and may indeed themselves contribute to immune regulationwithin the anterior chamber first by cytotoxic killing of Th1 cellsand second as regulatory cells via the production of IL-10.50

Conversely, however, as we have mentioned, Fas-FasL interac-tions may be proinflammatory such as via activation of Fas1

granulocytes. During AAU, IL-8 levels within the aqueous areincreased51 contributing toward neutrophil recruitment. In addi-tion, neutrophil apoptosis is central to the resolution of acuteinflammatory responses, and IL-8 impairs proapoptotic functionof Fas-FasL.52 In this study, the majority of apoptotic events withinthe aqueous was noted within T-cell populations, and the per-centage of apoptosis within granulocytes was low.

Although previous studies have investigated the ocular in-flammatory infiltrate during uveitis, none have studied either apure cohort of noninfectious anterior uveitis or potential regula-tory mechanisms within the immunoprivileged anterior chamberenvironment. The role of Fas-FasL interactions is diverse, not onlyacting as a signal for apoptosis and thus regulating the immuneresponse but also proinflammatory, principally by inducing Fas1

cytotoxicity. These data report on the role of FasL-mediated apo-ptosis of lymphocyte populations during AAU and provide apossible explanation for the self-limiting course of the disease and

FIGURE 5. sFasL in aqueous induces apoptotic cell death in Fas1

Jurkat cells. Apoptosis was induced with sFasL in a 4-hour incubationand was significantly inhibited when Jurkat cells were blocked withanti-Fas mAb (clone ZB4). Increased Jurkat cell apoptosis was observedafter incubation over 24 hours with aqueous from both AAU patientsand controls. Apoptosis was blocked by anti-Fas mAb. TNF-a did notinduce apoptosis in Jurkat cells. *Denotes statistically significant apo-ptosis (P , 0.02; Mann–Whitney U test).


the maintenance of ocular immune privilege as inferred by theability of sFasL within the control aqueous to also induce apopto-sis. Knowledge of he nature of sFasL and MMP activity is requiredto elucidate why sFasL is proapoptotic and not blocking as pre-viously shown under certain conditions.34,35 Further investigationis also required to investigate the role of Fas-FasL and TNF-ainteraction on other cell types (e.g., neutrophils and NK cells andtheir contribution to the inflammatory response within the ante-rior chamber).

References

1. Forrester JV, McMenamin PG, Holthouse I, Lumsden L, LiversidgeJ. Localization and characterization of major histocompatibilitycomplex class II-positive cells in the posterior segment of the eye:implications for induction of autoimmune uveoretinitis. InvestOphthalmol Vis Sci. 1994;35:64–77.

2. McMenamin PG, Crewe J, Morrison S, Holt PG. Immunomorpho-logic studies of macrophages and MHC class II1 dendritic cells inthe iris and ciliary body of the rat, mouse and human eye. InvestOphthalmol Vis Sci. 1994;35:3234–3250.

3. Steptoe RJ, Holt PG, McMenamin PG. Functional studies of majorhistocompatibility class II-positive dendritic cells and resident tis-sue macrophages isolated from the rat iris. Immunology. 1995;85:630–637.

4. Egan RM, Yorkey C, Black R, Loh WK, Stevens JL, Woodward JG.Peptide-specific T cell clonal expansion in vivo following immuni-zation in the eye, an immunoprivileged site. J Immunol. 1996;157:2262–2271.

5. Streilein JW. Molecular basis of ACAID. Ocul Immunol Inflamm.1997;5:217–218.

6. Ferguson TA. The molecular basis of anterior associated immunedeviation (ACAID). Ocul Immunol Inflamm. 1997;5:213–215.

7. Griffith TS, Ferguson TA. The role of FasL-induced apoptosis inimmune privilege. Immunol Today. 1997;18:240–244.

8. Griffith TS, Brunner T, Fletcher SM, Green DR, Ferguson TA. Fasligand-induced apoptosis as a mechanism of immune privilege.Science. 1995;270:1189–1192.

9. Ferguson TA, Gao Y, Herndon JM, Griffith TS. Apoptotic cell deathand IL-10 production in the eye are essential for the induction ofimmune deviation [ARVO Abstract]. Invest Ophthalmol Vis Sci.1997;38(4):S492. Abstract nr 2270.

10. Griffith TS, Yu X, Herndon JM, Green DR, Ferguson TA. CD95-induced apoptosis of lymphocytes in an immune privileged siteinduces immunological tolerance. Immunity. 1996;5:7–16.

11. Kawashima H, Yamagami S, Tsuru T, Gregerson DS. Anteriorchamber inoculation of splenocytes without Fas/Fas-ligand inter-action primes for a delayed-type hypersensitivity response ratherthan inducing anterior chamber-associated immune deviation. EurJ Immunol. 1997;27:2490–2494.

12. Kang SM, Schneider DB, Lin Z, et al. Fas ligand expression in isletsof Langerhans does not confer immune privilege and instead tar-gets them for rapid destruction. Nat Med. 1997;3:738–743.

13. Stuart PM, Griffith TS, Usui N, Pepose J, Yu X, Ferguson TA. CD95ligand (FasL)-induced apoptosis is necessary for corneal allograftsurvival. J Clin Invest. 1997;99:396–402.

14. Forrester JV. Uveitis: pathogenesis. Lancet. 1992;388:1498–1501.15. Forrester JV, Liversidge J, Dua HS, Towler H, McMenamin PG.

Comparison of clinical and experimental uveitis. Curr Eye Res.1990;9:75–84.

16. Feltkemp TE, Ringrose JH. Acute anterior uveitis and spondyloar-thropathies. Curr Opin Rheumatol. 1998;10:314–318.

17. Forrester JV. Duke Elder lecture: new concepts on the role ofautoimmunity in the pathogenesis of uveitis. Eye. 1992;5:433–447.

18. Deschennes J, Freeman WR, Char DH, Garavoy MR. Lymphocytesubpopulations in uveitis. Arch Ophthalmol. 1986;104:233–236.

19. Muhaya M, Calder V, Towler HM, Shaer B, McLaughlan M, Light-man S. Characterisation of T cells and cytokines in the aqueoushumour (AH) in patients with Fuch’s heterochromic cyclitis (FHC)and idiopathic anterior uveitis. Clin Exp Immunol. 1998;111:123–128.

20. Ohta K, Norose K, Wang XC, Ito S, Yoshimura N. Abnormal naiveand memory T lymphocyte subsets in the peripheral blood ofpatients with uveitis. Curr Eye Res. 1997;16:650–655.

21. Ohta K, Norose K, Wang XC, Susuma I, Yano A, Segawa K.Apoptosis-related Fas antigen on memory T cells in aqueous humorof uveitic patients. Curr Eye Res. 1996;15:299–306.

22. Van der Lelij A, Rothova A. Diagnostic anterior chamber paracen-tesis in uveitis: a safe procedure? Br J Ophthalmol. 1997;81:976–979.

23. Gavriele Y, Sherman Y, Ben–Sasson SA. Identification of pro-grammed cell death in situ via specific labeling of nuclear DNAfragmentation. J Cell Biol. 1992;119:493–501.

24. Vermes I, Haansen H, Steffens–Nakken H, Reutelingsperger C. Anovel assay for apoptosis: flow cytometric detection of phosphati-dylerine expression on early apoptotic cells using fluoresceinlabelled Annexin V. J Immunol Methods. 1995;184:39–51.

25. Lenardo MJ. Fas and the art of lymphocytic maintenance. J ExpMed. 1996;183:721–724.

26. Ohta K, Yoshimura N. Expression of Fas antigen on helper Tlymphocytes in Vogt-Koyanagi-Harada disease. Graefes Arch ClinExp Ophthalmol. 1998;236:434–439.

27. Ziegler SF, Ramsdell F, Alderson MR. The activation antigen CD69.Stem Cells. 1994;12:456–465.

28. Biselli R, Matricardi PM, D’Amelio R, Fattorossi A. Multiparametricflow cytometric analysis of the kinetics of surface molecule ex-pression after polyclonal activation of human peripheral blood Tlymphocytes. Scand J Immunol. 1992;35:439–447.

29. Smith JR, Hart PH, Coster DJ, Williams KA. Mice deficient in tumornecrosis factor receptors p55 and p75, interleukin-4, or induciblenitric oxide synthase are susceptible to endotoxin-induced uveitis.Invest Ophthalmol Vis Sci. 1998;39:658–661.

30. Tanaka M, Suda T, Takahashi T, Nagata S. Expression of thefunctional form of human Fas ligand in activated lymphocytes.EMBO J. 1995;14:1129–1135.

31. Suda T, Takahashi T, Golstein P, Nagata S. Molecular cloning andexpression of the Fas ligand: a novel member of the tumournecrosis factor family. Cell. 1993;75:1169–1187.

32. Gearing AJH, Beckett P, Christodoulou M, et al. Processing oftumour necrosis factor-a precursor by metalloproteinases. Nature.1994;370:555–557.

33. Kayagaki N, Kawasaki A, Ebata T, et al. Metalloproteinase-mediatedrelease of human Fas ligand. J Exp Med. 1995;182:1777–1783.

34. Suda T, Hashimoto H, Tanaka M, Ochi T, Nagata S. Membrane Fasligand kills human peripheral blood T lymphocytes, and solubleFas ligand blocks the killing. J Exp Med. 1997;186:2045–2050.

35. Schneider P, Holler N, Bodmer JL, et al. Conversion of membrane-bound Fas (CD95) ligand to its soluble form is associated withdownregulation of its proapoptotic activity and loss of liver toxic-ity. J Exp Med. 1998;187:1205–1213.

36. Di Girolamo N, Verma MJ, McCluskey PJ, Lloyd A, Wakefield D.Increased matrix metalloproteinases in the aqueous humour ofpatients and experimental animals with uveitis. Curr Eye Res.1996;15:1060–1068.

37. De Vos AF, Van Haren MAC, Verhagen C, Hoekzema R, Kijlstra A.Kinetics of intraocular tumor necrosis factor and interleukin-6 inendotoxin induced uveitis in the rat. Invest Ophthalmol Vis Sci.1994;35:1100–1106.

38. Yoshida M, Yoshimura N, Hangai M, Tanihara H, Honda Y. Inter-leukin-1a, interleukin-1b and tumor necrosis factor gene expres-sion in endotoxin induced uveitis. Invest Ophthalmol Vis Sci.1994;35:1107–1113.

39. Kasner L, Chan CC, Whitcup SM, Gery I. The paradoxical effect oftumor necrosis factor alpha (TNF-a) in endotoxin induced uveitis.Invest Ophthalmol Vis Sci. 1993;34:2911–2917.

40. Osborne GR. Apoptosis and the maintenance of homeostasis in theimmune system. Curr Opin Immunol. 1996;8:245–254.

41. Bakunowicz–Lazaraczyk A, Moniuszko T, Sredzinska–Kita D,Chwiecko J. Levels of selected cytokines in aqueous humor ofpatients with cataract extraction. Klin Oczna. 1997;99:235–237.

42. Cope AP, Liblau RS, Yang XD, et al. Chronic tumor necrosis factoralters T cell responses by attenuating T cell receptor signaling. JExp Med. 1997;185:1573–1584.


43. Jewett A, Cavalcanti M, Bonavida B. Pivotal role of endogenousTNF-a in the induction of functional inactivation and apoptosis inNK cells. J Immunol. 1997;159:4815–4822.

44. Kabelitz D. Apoptosis, graft rejection and transplant tolerance.Transplantation. 1998;65:869–875.

45. Wever PC, Boonstra JG, Laterveer JC, et al. Mechanisms of lym-phocyte-mediated cytotoxicity in acute renal allograft rejection.Transplantation. 1998;66:259–264.

46. D’Orazio TJ, Niederkorn JY. A novel role for TGF-b and IL-10 in theinduction of immune privilege. J Immunol. 1998;160:2089–2098.

47. Groux H, Bigler M, de Vries JE, Roncarolo MG. Interleukin-10induces a long term antigen-specific anergic state in human CD41

T cells. J Exp Med. 1996;184:19–29.48. Gao Y, Herndon JM, Zhang H, Griffith TS, Ferguson TA. Antiin-

flammatory effects of CD95 Ligand (FasL)-induced apoptosis. J ExpMed. 1998;188:887–896.

49. Ramsdell F, Seaman MS, Miller RE, Picha KS, Kennedy MK, LynchDH. Differential ability of Th1 and Th2 cells to undergo activation-induced cell death. Int Immunol. 1994;6:1545–1553.

50. Seder RA, LeGros GG. The functional role of CD81 type 2 cells. JExp Med. 1995;181:5–7.

51. Verma MJ, Lloyd A, Rager H, et al. Chemokines in acute anterioruveitis. Curr Eye Res. 1997;16:1202–1208.

52. Leuenroth S, Lee C, Grutkoski P, Keeping H, Simms HH. Interleu-kin-8-induced suppression of polymorphonuclear leukocyte apo-ptosis is mediated by suppressing CD95 (Fas/Apo-1) Fas-1 interac-tions. Surgery. 1998;124:409–417.

53. Vermes I, Haansan H, Steffens–Nakken H, Reutelingsperger C. Anovel assay for apoptosis: flow cytometric detection of phos-phatidylserine expression on early apoptotic cells using fluo-rescein labelled Annexin V. J Immunol Methods. 1995;184:39 –51.


Documents

Fas-Fas ligand-mediated apoptosis within aqueous during idiopathic acute anterior uveitis