Chemoattractant and Degranulation Activities Related to -8 in Vitreous Fluid in and Vitreoretinal Disorders

Joke H. de Boer* C. Erik Hack,f Arthur J. Verhoeven,-\ G. Seerp Baarsma,% Paulus T. V. M. de Jong,% A.JosJ. M. Rademakers,^ Willemien A. E. J. de Vries-Knoppert,^ Aniki Rothova,# and Aize Kijlstra*

Purpose. To investigate whether the interleukin-8 (IL-8), a strong chemoattractant and activator for , is responsible for neutrophil infiltration and degranulation in the eye in uveitis. Methods. IL-8 and elastase were measured with specific enzyme-linked immunoassays in vitre- ous fluid samples obtained from 69 patients with various uveitis entities. Vitreous fluid of nonuveitis patients and eye bank eyes served as controls. The chemotactic activity of vitreous fluid was tested with the Boyden chamber technique. Results. IL-8 was detected in 45% of the vitreous fluid samples from uveitis patients and in 26% of vitreous fluid samples from nonuveitis patients. Vitreous fluid samples with IL-8 levels exceeding 100 pg/ml were chemotactic for neutrophils. This chemotactic activity could be blocked by 41% to 79% with a monoclonal anti-IL-8 antibody. Elastase levels in vitreous fluid of uveitis patients with detectable IL-8 were significantly higher than those in vitreous fluid samples with no detectable IL-8. Conclusion. These results indicate that IL-8 participates in the inflammatory processes in the eye by attracting and degranulating neutrophils. It is suggested that these processes contribute to the pathogenesis of tissue destruction in uveitis. Invest Ophthalmol Vis Sci. 1993; 34: 3376-3385.

Uveitis or intraocular often results in teases, collagenase and gelatinase, that can destroy impairment of vision. Although T cells are generally normal cells and degrade the interstitium.6-7 Accumu- considered to be the key mediators in the pathogenesis lation of neutrophils at sites of inflammation is me- of uveitis, it remains unknown which cells mediate the diated by the release of specific chemotactic factors. tissue damage in this disease.12 In various nonin- The cytokine interleukin-8 (IL-8) appears to be one of fectious inflammatory disorders such as autoimmune the main chemoattractants for neutrophils and also diseases3 and ischemic injuries,4 neutrophils have been for T-lymphocytes.8'9 In addition to its chemoattrac- implicated as mediators of tissue destruction.5 Neutro- tant properties IL-8 can also activate neutrophils re- phils are able to release an array of agents such as the sulting in changes of shape, exocytosis of granules, serine proteinase elastase and the two metallopro- and stimulation of .10 Support for a role of IL-8 in uveitis are animal studies in which intra- vitreal administration of IL-8 has been shown to in-

From "The Netherlands Ophthalmic Research Institute, Amsterdam, the |Central duce ocular inflammation characterized by infiltrating 11 Laboratory of the Netherlands Red Cross Bloodtransfusion Service, Amsterdam, neutrophils. %Eye Hospital, Rotterdam, ^Department of Ophthalmology, Erasmus University, Rotterdam, ^Department of Ophthalmology, Catholic University, Nijmegen, In humans no data are available concerning the ^Department of Ophthalmology, Free University, Amsterdam, #Department of role of IL-8 in clinical uveitis. Here we report a study Ophthalmology, University of Amsterdam, The Netherlands. Submitted for publication January 21, 1992; accepted April 27, 1993. in which we measured IL-8 and elastase levels in vitre- Proprietary interest category: N. ous fluid (VF) of 69 patients with uveitis. VF samples Reprint requests: J.H. de Boer, The Netherlands Ophthalmic Research Institute, Department Ophthalmo-Immunology, P.O. Box 12141, 1100 AC Amsterdam, The from patients with vitreoretinal eye diseases and VF Netherlands. from eye bank eyes served as controls. In addition we

Investigative Ophthalmology & Visual Science, November 1993, Vol. 34, No. 12 3376 Copyright © Association for Research in Vision and Ophthalmology

Downloaded from iovs.arvojournals.org on 09/30/2021 Interleukin-8 Activity in Vitreous Fluid in Uveitis 3377

measured the chemotactic activity of these VF samples serum, to avoid false-positive reactions) were added in and assessed to what proportion IL-8 contributed to each well and incubated for 2 hours. Then the plates this activity. Our results are consistent with the hy- were developed with streptavidin-horseradish peroxi- pothesis that IL-8 participates in the inflammatory dase (Amersham Buckinghamshire, UK) followed by a process of clinical uveitis as a chemoattractant and ac- signal-amplification step with catalyzed reporter depo- tivator for neutrophils in VF. sition14 and, finally 3,5,3',5'-tetramethylbenzidin (Merck, Darmstadt, Germany) as a substrate. Briefly, purified human recombinant IL-8 (rIL-8) METHODS obtained by transfecting Escherichia coli DH5 with the Patients plasmid pMBLl 1 (British Biotechnology Ltd, Oxford, UK) was used as a standard and results were expressed A diagnosis of uveitis was made according to the crite- 13 ria of the International Uveitis Study Group.12 Paired in picograms per milliliter. All samples were tested in VF samples (n = 69) and serum samples (n = 66) from serial dilutions, starting at a 1:2 dilution in phosphate uveitis patients were collected. From three uveitis pa- buffered saline, pH 7.4, supplemented with 0.02% tients only a VF sample was available. The samples Tween (w/v) and 0.2 % gelatin (w/v). The concentra- were obtained during a therapeutic or diagnostic vi- tion of IL-8 in samples was calculated by comparing trectomy. The indications for a therapeutic vitrectomy the absorption of samples with that of serial dilutions were visual deterioration caused by vitreous opacities of human rIL-8. The detection limit of the assay was or progressive inflammation despite medical treat- 10 pg/ml. To exclude if VF interfered with the ELISA, ment. A diagnostic vitrectomy was performed in a few two different concentrations of rIL-8 (125 pg/ml and patients suspected of having an ocular lymphoma. 31.5 pg/ml) were added to different dilutions of VF Paired VF and serum samples from patients without samples (starting 1:2) of uveitis patients (n = 2) with uveitis undergoing a therapeutic vitrectomy (n = 23) IL-8 levels <10 pg/ml and raised elastase levels (181 and VF from eye bank eyes (n = 28; mean postmortem ng/ml and 111 ng/ml), and VF of eye bank eyes (n time 17 hours) were studied as controls. Undiluted VF = 2). The IL-8 ELISA was not inhibited by the VF samples were collected during a pars plana vitrectomy dilutions tested. before the infusion line was opened. Blood samples were collected the same day the vitrectomy was per- formed, allowed to clot for 1 hour and centrifuged for The chemotactic activity of VF for neutrophils was measured with a modified Boyden chamber tech- 10 minutes at 1500g to obtain the serum. VF and 15 serum samples were stored at — 20°C until use. nique. The two compartments of the chamber were separated by two filters. The upper filter has a pore The diagnosis of the nonuveitis patients included width of 8 ju,m (Sartorius AG, Germany) and the lower proliferative diabetic retinopathy (PDR, n = 9), vitre- filter has a pore width of 0.45 /xm (Millipore, Ireland). ous hemorrhage of unknown cause (n = 4), prolifera- Samples were diluted in complete HEPES medium tive vitreoretinopathy (PVR, n = 5), macular hole (n (containing 132 mM NaCl, 6 mM KC1, 1.2 mM KP;, 1 = 1), occlusion of the retinal vein (n = 1), complicated mM MgSO4, 1 mM CaCl2, 5 mM glucose, supple- cataract surgery (n = 2), vitreal degeneration (n = 1). mented with 5% (w/v) human serum albumin and hep- The exclusion criteria for VF from eye bank eyes were arin 10 IU/ml). Serial sample dilutions, starting undi- eyes from donors seropositive for human immunodefi- luted or 1:2, were placed in the lower compartment of ciency virus, hepatitis C, neuroviral diseases, and the chamber. Neutrophils 105, 95% to 99% pure, iso- sepsis. lated by Percoll (density 1.076, Pharmacia, Uppsala, Sweden) density centrifugation from a healthy do- IL-8 Assay nor,15 were placed in the upper compartment. The IL-8 was assessed with an enzyme-linked immunoassay Boyden chamber was then incubated for 90 minutes at (ELISA) using two monoclonal anti-human IL-8 anti- 37°C in a moist chamber. Thereafter the filters were bodies. The preparation of these monoclonal antibod- removed and the 8-/im filters were fixed in butanol/ ies and the development of the IL-8 ELISA has been ethanol (20/80, v/v) for 10 minutes and stained with described elsewhere.13 Briefly, ELISA plates (Nunc, Weigert solution (1% hematoxylin (v/v) in 95% eth- Roskilde, Denmark) were coated overnight with mono- anol (v/v) and an acidic FeCl3 solution (70 mM) in a clonal antibody anti-IL-8-1 (Central Laboratory of the volume ratio of 1:1). The filters were dehydrated with Netherlands Red Cross blood transfusion service ethanol and made transparent in xylene. The filters (CLB, Amsterdam). After washing, 100 /ul of the sam- were fixed upside down in malinol. The number of cells per field that passed the filter were determined by ple dilutions together with 10 /zl of biotinylated mono- 8 9 clonal anti-IL-8-4 (CLB; diluted in buffer containing light microscopy. Human rIL-8 (10~ M and 10~ M) 10% [v/v] of a 1:1 mixture of normal bovine and sheep was used as a positive control. Controls included ran-

Downloaded from iovs.arvojournals.org on 09/30/2021 3378 Investigative Ophthalmology & Visual Science, November 1993, Vol. 34, No. 12

dom neutrophil migration in the presence of medium was used as a substrate and the reaction was stopped alone and VF samples of eye bank eyes (n = 2). VF with 100 fi\ of 2 M H2SO4. The absorbency of the samples of uveitis patients with detectable IL-8 (n = 9) reaction product was measured in a Titertek Multiskan and with IL-8 levels less than 10 pg/ml (n = 6) were Plus (Biomedicals, Costa Mesa, CA) at 450 nm. Elas- tested in duplicate at different dilutions starting undi- tase levels were calculated by comparison with a stan- luted or 1:2 in medium. The same diluted experimen- dard curve of purified human neutrophil elastase (Fig. tal samples were also incubated with anti-IL-8-1 (CLB) 1). Controls included coating with nonspecific rabbit antibody at a concentration of 2 Mg/ml immunoglobu- immunoglobulin G. The detection limit of this assay lin G, a concentration that inhibits 80% of the chemo- was 2.5 ng/ml elastase. tactic activity of 10~8 M human rIL-8. IL-6 Measurements Elastase Assay IL-6 in VF was measured using an ELISA performed 17 Elastase in VF was measured with a capture ELISA as described by Helle et al. A monoclonal anti-hu- using polyclonal rabbit anti-human elastase antibod- man IL-6 antibody (mAB CLB IL-6/16, donated by ies. These antibodies were prepared by immunization Dr. LA. Aarden from the Central Laboratory of the with purified human neutrophilic elastase (Elastin Netherlands Red Cross blood transfusion Service) was Products Co. Inc., Pacific, MO).16 ELISA plates were used as a coating capture antibody. Polyclonal biotiny- coated overnight with 100 jul of anti-elastase antibod- lated sheep anti-human IL-6 immunoglobulin G ies at a concentration of 1.5 fig/m\ in phosphate- (CLB) was used as a detection antibody. The detection buffered saline. One hundred microliters of serial di- limit of this assay was 10 pg/ml. Results with experi- lutions of samples in 0.02% Tween (w/v) and 0.2/% mental samples were related to a standard curve of gelatin (w/v) were added to the wells and incubated for recombinant IL-6, and expressed in picograms per one hour at 37 °C. After washing, biotinylated polyclo- milliliter. nal rabbit anti-human elastase antibodies were added at the optimal concentration in the presence of 1% Miscellaneous (v/v) of a 9:1 mixture of normal rat and rabbit serum Albumin levels in serum and VF were measured by (to prevent nonspecific binding) and incubated for 1 radial immunodiffusion (LP-Partigen-Albumin, Behr- hour. After washing, streptavidin-horseradish peroxi- ing [Marburg, Germany]). The concentrations were dase was added diluted 1:1000 in 0.02% Tween (w/v) calculated from a standard curve using a human refer- and 0.2% gelatin (w/v). 3,5,3',5'-tetramethylbenzidin ence serum (HOO-05, CLB). For calculation of the

2500

elastase (ng/ml)

FIGURE 1. Dose-response curve of human neutrophilic elastase in the ELISA with polyclonal rabbit anti-human elastase antibodies. Data show the mean and SD (bar) of eight experi- ments.

Downloaded from iovs.arvojournals.org on 09/30/2021 Interleukin-8 Activity in Vitreous Fluid in Uveitis 3379

relative concentration ratio (RCR)18 the following (P < 0.0001) and tended to be higher than that in VF equation was used: fluid from patients with vitreoretinal disorders al- though this latter difference was not significant (Fig. vitreous IL-8 vitreous albumin 2). Circulating IL-8 was raised in 20 of 66 serum sam- (l)RCR serum IL-8 serum albumin ples of uveitis patients (30%; median of 20 samples = 42 pg/ml; range 10 to 1329 pg/ml) and in one of 23 Statistical Analysis samples of patients with noninflammatory eye diseases (4%; 17 pg/ml) (P< 0.01). Statistical analysis of the data was performed by using the SPSS statistical software package (SPSS Inc., Chi- To investigate VF IL-8 levels among various uvei- cago, IL). The differences among multiple patient tis entities, the patients were subdivided into sub- groups were assessed with the nonparametric Kruskal- groups according to the localization of inflammation Wallis test. Differences between two patients groups in the eye and to cause (Table 1). Raised IL-8 levels were estimated with the nonparametric Mann-Whit- were observed in various groups. There was no - ney test or Chi-squared test using the Bonferroni tionship between localization of the inflammation and method to correct for multiple comparison. Correla- the presence of IL-8 in VF, but there were signifi- tion coefficients were determined with the nonpara- cantly more patients with IL-8 in serum in the total metric Spearman test on log transformed data. group with anterior or intermediate uveitis compared to posterior uveitis (P < 0.006; P < 0.006, respec- tively) or panuveitis (P < 0.0024; P < 0.012, respec- tively). In six often patients with acute retinal necrosis RESULTS syndrome, IL-8 was detectable in the VF but could not be detected in the serum. Serum IL-8 was detected in IL-8 Levels 11 of 20 samples from patients with intermediate uvei- IL-8 levels (10 pg/ml) were detected in 31 of 69 (45%) tis (P < 0.015 versus patients with acute retinal necro- VF samples of uveitis patients (median of 31 samples sis). IL-8 could be detected in sera of all three patients = 33 pg/ml; range 10 to 1830 pg/ml, Fig. 2). In VF positive for HLA B27, but only one of these patients samples of patients with noninflammatory eye dis- had increased levels in VF. IL-8 levels were raised both in serum and VF in eight of the uveitis cases (Table 2). eases, IL-8 levels were elevated in six of 23 samples (26 18 %; median of 6 samples = 25 pg/ml; range 10 to 116 For these patients samples, the RCR was calculated pg/ml). Only one of 28 (4%) VF fluid samples obtained using paired albumin levels as a reference. The RCR from eye bank eyes had an increased IL-8 value (20 values ranged between 3 and 216. In two patients with pg/ml). IL-8 in VF fluid from uveitis patients was sig- severe albumin diffusion in the eye (10% and 50% of nificantly higher than that in VF fluid of eye bank eyes the serum value, respectively) and raised IL-8 levels in

10000 1 vitreous fluid sera p<0.0001 -1*0.01

1000 -

100 -

•*

10 4

uveitis non-uveitis eyebank eyes uveitis non-uveitis FIGURE 2. 11-8 levels in VF and sera from uveitis patients and nonuveitis patients. The broken line represents the detection limit of the assay.

Downloaded from iovs.arvojournals.org on 09/30/2021 3380 Investigative Ophthalmology & Visual Science, November 1993, Vol. 34, No. 12

TABLE l. IL-8 Levels in Vitreous Fluid and Serum in Different Uveitis Entities Vitreous Fluid Samples Serum Samples With Increased IL-8 With Increased IL-8

No. Median Range No. Median Range Anatomical Diagnosis Specific Diagnosis of Total (pg/ml) (pg/ml) of Total (Pg/ml) (pg/ml) Anterior uveitis HLA B27 related* 1/3 44 3/3 48 11-455 Anterior uveitist 0/2 — — 1/2 944 — Intermediate uveitis Intermediate uveitisf 7/21 21 10-170 1I/20J 27 10-1329 Posterior uveitis Toxoplasniosis 4/5 11 10-25 0/5 — — Cytomegalovirus retinitis 0/2 — — 0/2 — — Toxocariasis 1/2 57 — 1/2 11 — 0/1 — — NP — — Posterior uveitist 2/3 173 21-325 1/3 10 — Masquerade syndrome§ 3/4 22 12-35 0/4 — — Pan uveitis Acute retinal necrosis (ARN) 6/10 115 17-233 o/iot — Fuchs' heterot hromic cytiitis 0/2 — — 1/2 35 — Mycolic 1/1 272 — 0/1 — — Bechets disease 0/1 — — 0/1 — — Crohn's disease 1/1 10 — 0/1 — — Panuveitisf 5/11 33 12-1830 2/10 67 56-78 Total 31/69 33 10-1830 20/66 42 10-1329 NP = not performed. * One patient with Reiter's disease. t Unknown origin. X Difference in serum 11.-8 levels between patients with intermediate uveitis and acute retinal necrosis is significant (P < 0.015). § Specific diagnosis: lymphoma (n = 3) and teratoid medullar epithelioma (n = 1).

serum (455 pg/ml and 944 pg/ml, respectively) no IL- serum (17 pg/ml) from a patient with PDR with no 8 was detected within the VF. raised IL-8 in the VF. In the nonuveitis group, IL-8 was detected in the No relationship was observed between IL-8 in VF VF of three of nine PDR patients (median 13 pg/ml; or serum and age or sex of the patients. There was no range 10 to 76 pg/ml), in one of four patients with evidence that IL-8 was related to duration of the dis- vitreous hemorrhage of unknown cause (23 pg/ml), in ease or medical treatment. Most of the patients under- one patient with a retinal vein occlusion (27 pg/ml), went surgery during a chronic stage of the disease ex- and in one patient with PVR after perforating ocular cept seven patients with acute retinal necrosis who un- trauma (116 pg/ml). IL-8 was slightly raised in one derwent surgery within 4 months of disease onset. Six

TABLE 2. Albumin, IL-8 Levels, and Relative Concentration Ratio (RCR) in VF and Serum of Patients With Detectable IL-8 in Both Fluids VF Serum VF Serum IL-8 IL-8 Albumin Albumin Diagnosis (pg/ml) (pg/ml) (mg/ml) (mg/ml) RCR* HLA B27 associated uveitis. Reiter's disease 44 48 6.90 48.2 6 Intermediate uveiiist 170 110 2.00 47.8 37 Intermediate uveitist 37 75 2.16 46.0 11 Intermediate uveitist 21 10 0.36 37.1 216 Intermediate uveitist 15 172 1.00 39.4 3 Intermediate uveitist 12 273 0.62 52.4 4 Posterior uveitist 21 10 3.90 41.6 22 Panuveitist 58 56 7.00 47.0 7 * The RCR value was calculated as follows: (vitreous IL-8/serum IL-8):(vitreous albumin/serum albu- min). t Unknown origin.

Downloaded from iovs.arvojournals.org on 09/30/2021 Interleukin-8 Activity in Vitreous Fluid in Uveitis 3381

of them had IL-8 in VF. Fourteen patients were 100 treated with systemic steroids at the time of vitrectomy and two of them had IL-8 in serum (455 pg/ml and 90- H no anti-IL-8 944 pg/ml). Half of the patients with raised IL-8 in VF H anti-IL-8 were treated with corticosteroids systemically (n = 7), 80- subconjunctivally (n = 3), or in eyedrop form (n = 6) at

the time of vitrectomy. eld 70 H= k_ (D Chemotactic Activity of VF Samples Q. 60 in To verify that the immunochemically detected IL-8 in

(cell? 50 VF was also biologically active, chemotactic activity of V) VF for neutrophils was investigated using the Boyden

taxi 40 chamber technique. The mean random neutrophil mi- o E gration induced by HEPES medium was nine cells per a) 30 field (range 5 to 13 cells per field). The chemotactic JuZ activity of the two VF samples of eye bank eyes was 11 20- and 14 cells per field. In Figure 3 chemotactic activity of different dilutions of one VF sample with an IL-8 10- concentration of 325 pg/ml is shown. The dilution factor with the optimal chemotactic activity of VF sam- ples of 15 patients is shown in Figure 4. The chemo- VF1:2 VF1:6 VF 1:20 tactic activity of two VF samples increased after ten- FIGURE 3. Chemotactic activity of different dilutions of one fold dilution of the samples. VF samples with IL-8 lev- uveitis VF sample with an IL-8 level of 325 pg/ml. The bro- els exceeding 100 pg/ml (n = 5) were all chemotactic ken line indicates the mean random neutrophil migration for neutrophils whereas VF samples with IL-8 levels induced by HEPES medium. less than 100 pg/ml (n = 4) and VF samples with IL-8 levels less than 10 pg/ml (n = 6) did not attract neutro- phils. Neutralization of the chemotactic activity of VF compared with IL-6 levels. Both cytokine levels were for neutrophil by incubating them with a monoclonal measured in an immunochemical assay in 61 VF sam- anti-IL-8 antibody varied between the fiveV F samples ples (Fig. 6). In the VF of uveitis patients, raised IL-8 and ranged from 41% to 79%. levels were always accompanied by raised IL-6 levels but not vice versa. IL-6 levels were more frequently Elastase raised as compared to IL-8. Instead of studying the cytology of VF during intraocu- lar inflammation, a neutrophil degranulation product, elastase, was measured as a marker for neutrophil acti- DISCUSSION vation (Table 3). An elastase level of 22 ng/ml (mean + This study shows that IL-8 levels in VF are increased in 2 SD of elastase levels in VF of eye bank eyes) was used 45% of the patients with uveitis and in 26% of the as the upper normal level for VF. Elastase levels were patients with noninflammatory eye diseases. To con- elevated in 28 of 66 VF samples from uveitis patients firm that the immunochemically detected IL-8 in VF (42%) and in 7 of 21 VF samples of nonuveitis patients was biologically active, we performed chemotactic as- (33%). Elastase levels were significantly higher in VF of says. VF samples of uveitis patients with an IL-8 level uveitis patients and nonuveitis patients than in that of exceeding 100 pg/ml were chemotactic for neutro- eye bank eyes (P < 0.0005; P < 0.0015, respectively). phils, an activity that could be blocked by a monoclo- Raised elastase levels in VF samples with detectable nal anti-human IL-8 antibody. Only 10% of the uveitis IL-8 were observed in 18 of 28 VF samples of uveitis VF samples and one VF sample from a patient with patients (64%), in 3 of 6 VF samples of the nonuveitis PVR had IL-8 levels in this range. Experimental ani- group, and one VF sample of eye bank eyes (Fig. 5). mal models in the rat have shown that intravitreal in- Elastase levels in VF of uveitis patients with detectable jection of 100 pg to 103 ng of human rIL-8 induced IL-8 were significantly higher compared to elastase cell infiltration into the eye, with an optimal effect levels in VF samples without detectable IL-8 (P seen at 10 ng.11 However, the concentration of IL-8 < 0.001). that affects the human eye might be different from that inducing ocular inflammation in the rat model. IL-8 Versus IL-6 Levels Furthermore, it is possible that the active in vivo level To investigate the possible interaction between differ- of IL-8 is significantly less than that required for a ent in VF, the intraocular IL-8 levels were positive reaction in the Boyden chamber technique.

Downloaded from iovs.arvojournals.org on 09/30/2021 3382 Investigative Ophthalmology & Visual Science, November 1993, Vol. 34, No. 12

100

90- no anti-IL-8 anti-IL-8 _ 80-

^ 70 H a> Q. 2 60- ^ 50- "S 40-

30-

20-

10-

reclL-810"M IL-8 > 100 pg/ml IL-8<100pg/ml IL-8 < 10 pg/ml

FIGURE 4. Chemotactic activity (cells per field) of rec-IL-8 10 8 M and VF samples obtained from 15 uveitis patients. The broken line indicates the mean random migration of neutro- phils induced by HEPES medium. The dilution factor with optimal chemotactic activity of VF with IL-8 levels >100 pg/ml (n = 5; dilution factor in the Boyden chamber: undiluted, undiluted, 1:2, 1:10, 1:10 [left to right], respectively), IL-8 levels < 100 pg/ml (n = 4; tested undiluted), and IL-8 < 10 pg/ml (n = 6; tested undiluted) are shown, and the chemotactic activity after blocking with a monoclonal anti-IL-8-1 antibody (2 /ig/ml).

Low IL-8 levels in VF might participate in processes ready been demonstrated in VF of patients with PVR within the eye other than neutrophil recruitment, be- and PDR.19"21 It was also recently reported that IL-8 cause it is known that the various biologic effects of can induce corneal neovascularization in rabbits.22 We IL-8 levels are concentration dependent, whereby low found IL-8 in the VF of three patients with PDR and in levels induce T cell chemotaxis, whereas higher levels one patient with vitreous hemorrhage. Whether the are needed for neutrophil attraction.9 angiogenic properties of IL-8 might play a role in reti- Different explanations are possible for the pres- nal neovascularization deserves further investigation. ence of IL-8 in VF in vitreoretinal disorders. PVR Unfortunately, no information was available seems to be an exaggerated wound healing response about the number of infiltrating cells in VF, clinically after retinal detachment surgery or perforating or counted at vitrectomy. Instead of studying the cytol- trauma and neutrophil recruitment can participate in ogy of the VF during intraocular inflammation, we the elimination of necrotic tissue.19 Other inflamma- measured a neutrophil degranulation product, elas- tory mediators than IL-8 and growth factors have al- tase, as a marker for neutrophil activation in vivo.

TABLE 3. Vitreous Fluid Elastase Levels in Uveitis, Noninflammatory Eye Diseases, and Eye Bank Eyes No. of Samples With

Elastase >22 ng/ml IL-8 > JO pg/ml Median elastase per No. of Samples per No. of Samples With Samples (ng/ml) Range (ng/ml) Investigated (%) Elastase >22 ng/ml (%)

Uveitis 15.5* <2.5-10,124 28/66 (42%) 18/28 (64%) Nonuveitis 14.0* <2.5-86 7/21 (33%) 3/7 (43%) Eye bank eyes 4.0* <2.5-31 2/28 (7%) 1/2 * Difference in elastase levels in VF of uveitis and nonuveitis patients compared with VF from eye bank eyes is significant (P < 0.0005 and P < 0.0015, respectively).

Downloaded from iovs.arvojournals.org on 09/30/2021 Interleukin-8 Activity in Vitreous Fluid in Uveitis 3383

Elastase is a proteolytic enzyme considered to play a role as a mediator of tissue destruction.5 Comparable • elastase levels were observed in VF of patients with uveitis and vitreoretinal disorders, which indicates

that neutrophils participate in both eye diseases. The 1000- observation that elastase levels in VF with detectable 1 . " * IL-8 were significantly higher than in VF samples without IL-8 suggested that IL-8 is a chemoattractant and activator for neutrophils in VF. In 10 of the 28 uveitis samples (36%) and in 4 of the 7 nonuveitis sam- 100- :

ples with raised elastase levels no IL-8 was detectable. i One explanation for this might be that IL-8 may have t i • already been cleared from VF. Conversely, this raised t

elastase might be due to the presence of other neutro- 10- phil chemotactic factors, such as C5a or leukotriene < 10- k 23 24 B4, known to be released into the eye. - Other elas- <10 10 100 1000 10000 tase containing cells, for instance, monocytes and mac- IL-8 (pQ/ml) rophages, that infiltrate in the eye during uveitis can theoretically be a source of elastase in VF. Monocytes FIGURE 6. Relationship between IL-8 (in pg/ml) and IL-6 (in only contain 6% of the amount of elastase of neutro- pg/ml) in vitreous fluid (r = 0.53; P < 0.000015). phils and the proteolytic activity of is largely due to cysteine proteases.25'26 also contain elastase but these cells are not commonly in- 27 29 (Table 2). For these patients we calculated the RCR to volved in uveitis. " Therefore, neutrophils may be correct for changes in permeability of the blood vitre- regarded as the main source of elastase in VF. In the ous barrier. The detection limit of the IL-8 assay was majority of patients with detectable IL-8 in VF, serum 10 pg/ml, which prevented us from calculating normal levels of IL-8 were normal, suggesting that IL-8 in VF RCR values for IL-8 in VF and serum. It is therefore of these patients reflected local production. Our data difficult to accurately estimate which RCR indicates do not allow conclusions regarding the source of IL-8. intraocular production of IL-8. Furthermore, it is not The cytokine may have originated from resident ocu- clear whether IL-8 behaves in the same way as albumin lar cells, such as retinal pigment epithelial cells, which in its diffusion behavior across the blood-retinal have been shown to produce IL-8 in vitro after stimula- 30 barrier. Two patients with an increased permeability tion with IL-1 or TNF. It is also possible that IL-8 of the blood-retinal barrier had raised IL-8 in serum was produced by infiltrating cells. Eight patients had but no detectable IL-8 in their eye. More studies con- detectable IL-8 both in the eye and in the circulation cerning intraocular kinetics of cytokines, are warranted to clarify this issue. Previous studies have already demonstrated that 100000 uveitis cytokines participate in the process of ocular inflam- non-uveitis mation.31 The cytokines TNF-a, IL-1 a//?, IL-6, IL-8,

10000 - and -7 have uveitogenic properties when injected intravitreally.11'32"35 Furthermore, the pres- ence of the cytokines IL-2, IL-6, and IFN-7 has been

& 1000 - demonstrated in ocular material obtained during clin- ical uveitis.36'2237 In vitro studies have demonstrated that human retinal pigment epithelial cells produce

100- IL-8 as well as IL-6 on stimulation with IL-1 or TNF- a 30.38,39 jo investigate if a similar mechanism was operating in the in vivo situation, we compared IL-6 !> 10 and IL-8 levels in human VF. Raised IL-8 levels were ;! always accompanied by raised IL-6 levels but not vice 2-5 ,J <2.5 versa (Fig. 5) and IL-6 levels were more frequently raised than IL-8 levels. This latter observation might 10 100 1000 10000 IL-8 (pg/ml) be caused by differences in the regulatory processes of production and elimination of these cytokines during FIGURE 5. Relationship between IL-8 (in pg/ml) and elastase inflammation. (in ng/ml) in vitreous fluid (r = 0.48; P < 0.000001). Raised IL-8 levels in VF and sera were detected in

Downloaded from iovs.arvojournals.org on 09/30/2021 3384 Investigative Ophthalmology 8c Visual Science, November 1993, Vol. 34, No. 12

different intraocular diseases, which suggest that IL-8 of the Netherlands Red Cross blood transfusion Service for is a common inflammatory mediator within the eye their technical assistance during the experiments and Dr. A. (Table 2). In 60% of patients with acute retinal necro- F. de Vos from the Netherlands Ophthalmic Research Insti- sis, IL-8 was detectable in VF but not in serum. Acute tute for his valuable advice. retinal necrosis is characterized by a severe necrotiz- ing inflammation of the retina. It is caused by reactiva- References tion of latent neurotropic viruses of the herpes group 40 1. Lightman S, Chan CC. Immune mechanisms in and often leads to blindness. The observation that chorio-retinal inflammation in man. Eye. 1990;4:345- IL-8 is not systemically detectable in these patients 353. supports the opinion that acute retinal necrosis is a 2. Forrester JV. Uveitis: pathogenesis. Lancet. 1991; 41 local inflammatory process. IL-8 was detectable in 338:1498-1501. 55% of the serum samples of patients with interme- 3. Weissmann G, Korchak H. Rheumatoid arthritis: the diate uveitis of unknown origin. Other investigators role of neutrophil activation. Inflammation. 1984;8 have detected high levels of a new circulating (suppl):3-14. in these patients.42 These observations may suggest a 4. Rossen RD, Swain JL, Michael LH, Giannini E, Ent- systemic involvement in this type of inflammatory eye man ML. Selective accumulation of the first compo- disease. Because some of the patients with interme- nent of complement and leucocytes in ischaemic ca- diate uveitis are known to develop a systemic disease nine heart muscle: a possible initiator of an extra myo- later on, it would be interesting to investigate whether cardial mechanism of ischemic injury. Circ Res. this is related to the occurrence of circulating cyto- 1985;57:119-130. kines. Circulating IL-8 has been detected in various 5. Malech HL, Gallin JI. Neutrophils in human diseases. NEnglJMed. 1987; 317:687-694. systemic diseases. The range of serum IL-8 levels is 6. Weiss SJ. Tissue destruction by neutrophils. N EnglJ very wide between different diseases but also within Med. 1989; 6:365-376. one disease. Compared to serum IL-8 levels in uveitis, 7. Weissmann G, Smolen JE, Korchak H. Release of in- low serum IL-8 levels were observed in patients with flammatory mediators from stimulated neutrophils. N Crohn's disease (maximum 21 pg/ml), comparable lev- EnglJ Med. 1980; 303:27-34. els were observed for patients with acute pancreatitis 8. Thelen M, Peveri P, Kernen P, vonTscharner V, Walz (maximum 1388 pg/ml), and, as suspected, high levels A, Baggiolini M. Mechanism of neutrophil activation were detected in patients with septic shock (maximum by NAF, a novel monocyte-derived peptide agonist. 66,000 pg/ml).13-43-44 FASEBJ. 1988; 2:2702-2706. 9. Larsen CG, Anderson AO, Appella E, Oppenheim JJ, This study shows that IL-8 participates in the Matshushima K. The neutrophil-activating protein pathogenesis of clinical uveitis and that neutrophil at- (NAP-1) is also chemotactic for T lymphocytes. traction and activation may be involved in the process Science. 1989; 243:1464-1466. of tissue destruction in ocular inflammation. Cytologic 10. Baggiolini M, Walz A, Kunkel SL. Neutrophil-activat- and histologic studies have repeatedly demonstrated ing peptide-1/, a novel cytokine that acti- the presence of neutrophils in the eye during uveitis. vates neutrophils./ Clin Invest. 1989;84:1045-1049. For example, Hogan observed that next to lympho- 11. Ferrick MR, Thurau SR, Oppenheim MH, et al. Ocu- lar inflammation stimulated by intravitreal interleu- cytes and monocytes, neutrophils were often present kin-8 and interleukin-1. Invest Ophthalmol Vis Sci. in the aqueous humor of uveitis patients.29 In the ret- 1991;32:1534-1539. ina of patients with Behcets disease, neutrophils infil- 45 12. Bloch-Michel E, Nussenblatt RB. International Uvei- trate the area around obstructed blood vessels. In tis Study Group recommendations for the evaluation other diseases such as rheumatoid arthritis, the pres- of intraocular inflammatory disease. Am] Ophthalmol. ence of IL-8 in synovial fluid has been reported and it 1987; 103:234-235. is known that neutrophils play a role in tissue destruc- 13. Hack CE, Hart M, Strack van Schijndel RJM, et al. tion in this disease.3'46 Studies concerning the interac- Interleukin-8 in sepsis: Relation to shock and inflam- tion between different chemoattractants and infiltrat- matory mediators. Infect Immun. 1992; 60:2835-2842. ing cells may yield important information about the 14. Bobrow MN, Harris TD, Shaughnessey KJ, Litt GJ. inflammatory processes within the eye and may offer Catalyzed reporter deposition, a novel method of sig- new prospects for the treatment of clinical uveitis. nal amplification. Application to immunoassays. J Im- munol Methods. 1989; 125:279-285. Key Words 15. Bruijnzeel PLB, Warringa RAJ, Kok PTM. Inhibition of -activating factor- and zymosan-activated interleukin-8, elastase, chemotaxis, uveitis, vitreous fluid serum-induced chemotaxis of human neutrophils by Acknowledgments nedocromil sodium, BN 52021 and sodium cromogly- cate. Br J Pharmacol. 1989; 97:1251-1257. The authors thank Mrs. A. J. M. Eerenberg, Mrs. Y. T. P. 16. Nuijens JH, Abbink JJ, Wachtfogel YT, et al. Plasma Lubbers, and Mr. F. P. J. Mul from the Central Laboratory elastase al-antitrypsin and lactoferrin in sepsis: evi-

Downloaded from iovs.arvojournals.org on 09/30/2021 Interleukin-8 Activity in Vitreous Fluid in Uveitis 3385

dence for neutrophils as mediators in fatal sepsis. J 32. Fleischer LN, Ferrell JB, McGahan MC. Ocular in- Lab Clin Med. 1992; 119:159-168. flammatory effects of intravitreally injected tumour 17. Helle M, Boeije L, de Groot E, de Vos AF, Aarden LA. necrosis factor-alpha and endotoxin. Inflammation. Sensitive ELISA for interleukin-6, detection of IL-6 in 1990; 14:325-335. biological fluids: synovial fluids and sera. J Immunol 33. Rosenbaum JT, Samples JR, Hefeneider SH, Howes Meth. 1991; 138:47-56. EL. Ocular inflammatory effects of intravitreal Inter- 18. Dernouchamps JP. of the aqueous humour. leukin 1. Arch Ophthalmol. 1987; 105:1117-1120. Doc Ophthalmol. 1982;53:193-248. 34. Hoekzema R, Murray PI, Kijlstra A. Cytokines and 19. Wiedemann P. Growth factors in retinal diseases: pro- intraocular inflammation. Curr Eye Res. 1990;9:207- liferative vitreoretinopathy, proliferative diabetic reti- 211. nopathy, and retinal degeneration. Surv Ophthalmol. 35. Hamel CP, Detrick B, Hooks JJ. Evaluation of la ex- 1992;36:373-384. pression in rat ocular tissues following inoculation 20. Conner TB, Roberts AB, Sporn MB et al. Correlation with interferon-gamma. Exp Eye Res. 1990; 50:173- of fibrosis and transforming -/? type 2 182. levels in the eye./ Clin Invest. 1989;83:1661-1666. 36. Murray PI, Hoekzema R, van Haren MAC, de Hon 21. de Boer JH, van Haren MAC, de Vries-Knoppert FD, Kijlstra A. Aqueous humour levels WAEJ, et al. Analysis of IL-6 levels in human vitreous in uveitis. Invest Ophthalmol Vis Sci. 1990; 31: fluid obtained from uveitis patients, patients with pro- 917-920. liferative intraocular disorders and eye bank eyes. 37. Hooks JJ, Chan CC, Detrick B. Identification of the Curr Eye Res. 1992; ll(suppl):181-186. , interferon-gamma and interleukin-2, in 22. Elner VM, Strieter RM, Kunkel SL, et al. Interleukin- inflammatory eye diseases.Invest Ophthalmol Vis Sci. 8: a corneal factor that induces neovascularisation. 1988;29:1444-1451. ARVO Abstracts. Invest Ophthalmol Vis Sci. 1992;33 38. Planck SR, Dang TT, Graves D, Tara D, Ansel JC, Ro- (suppl):1465. senbaum JT. Retinal pigment epithelial cells secrete 23. Shimada K, Yaoita H, Shikano S. Chemotactic activity Interleukin-6 in response to Interleukin-1. Invest Oph- in aqueous humour of patients with Behcets disease. thalmol Vis Sci. 1992;33:78-82. Jpn J Ophthalmol. 1972; 16:84-92. 39. Elner VM, Scales W, Elner SG, Danforth J, Kunkel SL, 24. Parker JA, Goetzl EJ, Friedlaender MH. Leukotrienes Strieter RM. Interleukin-6 (IL-6) expression and in aqueous humour of patients with uveitis. Arch Oph- secretion by cytokine-stimulated human retinal epithe- thalmol. 1986; 104:722-724. lial cells. Exp Eye Res. 1992;54:361-368. 25. Campbell EJ, Silverman EK, Campbell MA. Elastase 40. Culbertson WW, Blumenkranz MS, Haines H, Gass and cathepsin G of human monocytes. Quantification JDM, Mitchell KB, Norton EWD. The acute retinal of cellular content, release in response to stimuli, and necrosis syndrome, part 2: histopathology and etiol- heterogeneity in elastase-mediated proteolytic activ- ogy. Ophthalmology. 1982;89:1317-1324. ity. J Immunol. 1989; 143:2961-2968. 41. Urayama A, Yamada N, Sasaki T, et al. Unilateral 26. Chapman HA, Stone L. Comparison of live human acute uveitis with periarteritis and detachment. JpnJ neutrophil and alveolar elastolytic activ- Clin Ophthalmol. 1971; 25:607-619. ity in vitro. Relative resistance of macrophage elastoly- 42. Bora NS, Bora PS, Gobleman CL, Fleischman JB, Ka- tic activity to serum and alveolar proteinase inhibitors. plan HJ. Identification and molecular characteriza- J Clin Invest. 1984;74:1693-l700. tion of a novel circulating protein in pars planitis. Ab- 27. Lungarella G, Menegazzi R, Gardi C, et al. Identifica- stract. Exp Eye Res. 1992;55(suppl. 1):111. tion of elastase in human eosinophils: immunolocali- 43. Mahida YR, Ceska M, Effenberger F, Kurlak L, Lind- zation, isolation and partial characterization. Arch Bio- ley I, Hawkey CJ. Enhanced synthesis of neutrophil- chemBiophys. 1992;292:128-135. activating peptide-1/interleukin-8 in active colitis. 28. Davis JL, Solome D, Nussenblatt RB, Palestine AG, Chan CC. Immunocytochemical staining of vitreous Clin Sci. 1992;82:273-275. cells. Ophthalmology. 1991;99:250-256. 44. Gross V, Andreesen R, Leser HG, et al. Interleukin-8 29. Hogan MJ, Wood IS, Godfrey WA. Aqueous humour and neutrophil activation in acute pancreatitis. EurJ cytology in uveitis. Arch Ophthalmol. 1973; 89:217- Clin Invest. 1992;22:200-203. 220. 45. Smolin G, O'Connor GR. Immunologic diseases af- 30. Elner VM, Strieter RM, Elner SG, Baggiolini M, Lind- fecting the uveal tract and retina: Behcets disease. Oc- ley I, Kunkel SL. Neutrophil chemotactic Factor (IL- ular immunology. Philadelphia: Lea & Febiger; 8) by cytokine-treated retinal pig- 1981:290-292. ment epithelial cells. AmJPathol. 1990; 136:745-750. 46. Feldmann M, Brennan FM, Chantry D, et al. Cytokine 31. de Vos AF, Hoekzema R, Kijlstra A. Cytokines and assay: role in the pathogenesis of autoimmunity. Im- uveitis, a review. Curr Eye Res. 1902;ll:581-597. munol Rev. 1991; 119:105-123.

Downloaded from iovs.arvojournals.org on 09/30/2021