Endotoxin-Induced Uveitis is Partially Inhibited by Anti–IL-8 Antibody Treatment

Abstract

purpose. To examine the potential therapeutic effect of a neutralizing anti–IL-8 monoclonal antibody in endotoxin-induced uveitis (EIU) in the rabbit.

methods. An anti–IL-8 antibody (WS-4) was injected intravitreal 2 hours before, simultaneously with, or 6 hours after endotoxin challenge in rabbits. Eyes were examined for clinical signs of inflammation, and aqueous humor (AH) was sampled to study cellular infiltration and protein content. Leukocyte subset analysis was performed on Giemsa-stained AH cytospins. Histologic grading of inflammation was performed on hematoxylin-eosin–stained sagittal sections of enucleated eyes. In separate experiments, animals received the anti–IL-8 antibody simultaneously with the endotoxin challenge, before repeated anterior chamber paracentesis was performed (at 6, 12, 24, 48, and 72 hours after injection) to estimate the kinetics and durability of changes in total cell count and protein concentration in AH.

results. Anti–IL-8 therapy caused a decrease in the clinical and histologic grade of inflammation in EIU. The mean cell count in the AH at the peak of inflammation (24 hours) in eyes receiving endotoxin only was 6419 ± 1165/μl (mean ± SE) compared to 2546 ± 573/μl in rabbits treated simultaneously with 250 μg of anti–IL-8 antibody (P < 0.05). The protein concentration in the AH was not significantly altered by anti–IL-8 treatment. Kinetic analysis of the leukocyte count in the AH demonstrated persistent inhibition of leukocyte accumulation (range, 60%–91% compared to control eyes) by the anti–IL-8 antibody administered simultaneously with endotoxin. This inhibition was sustained for up to 72 hours after injection.

conclusions. Anti–IL-8 antibody treatment partially blocks EIU in rabbits. A consistent decrease in the recruitment of polymorphonuclear leukocytes into the anterior chamber was obtained when neutralizing antibody was injected simultaneously with endotoxin. These findings suggest that IL-8 contributes to the chemotactic signal for the recruitment of leukocytes in EIU.

Interleukin (IL)-8 is a prototypic α (C-X-C subfamily) chemokine and a potent stimulus for polymorphonuclear (PMN) leukocyte recruitment and activation. It is produced in settings of inflammation by a wide range of cells, including monocyte–macrophages, lymphocytes, vascular endothelium, dermal fibroblasts, keratinocytes, synovial cells, hepatocytes, ¹ and the ocular epithelium of cornea and retina. ^{2 3 4} IL-8 production is typically stimulated by endotoxin or proinflammatory cytokines, such as IL-1 or tumor necrosis factor (TNF)-α, and is associated with a rapid influx of PMN over 4 to 6 hours. IL-8 promotes PMN leukocyte recruitment by regulating the expression of the integrin family of leukocyte adhesion receptors, ⁵ resulting in neutrophil adherence to vascular endothelium, transendothelial migration, and haptotactic migration through extracellular matrix to sites of inflammation. ^{6 7}  

Previous studies have reported the anti-inflammatory effects of neutralizing anti–IL-8 antibody treatment in animal models of neutrophil-dominant inflammation, including a rat model of IgG immune complex–induced lung injury ⁷ and a rabbit model of ischemia–reperfusion lung injury. ⁸ We recently documented significantly increased levels of several chemokines, including IL-8, in the aqueous humor (AH) of patients with acute anterior uveitis (AU). ⁹ In an attempt to evaluate further the role of this chemokine in the regulation of inflammation in acute AU, we have studied the effects of neutralizing anti–IL-8 antibody on the induction of endotoxin-induced uveitis (EIU) in rabbits. 

Materials and Methods

Rabbit Model of EIU

Female New Zealand white rabbits aged 8 to 12 weeks, weighing 2.5 to 3 kg, were obtained from Combined Universities Laboratory Animal Services (CULAS), Sydney, Australia. Experiments conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Preliminary dose–response studies determined that a dose of 100 ng of endotoxin (lipopolysaccharide [LPS] from Salmonella typhimurium; Difco Laboratories, Detroit, MI) was optimal for the induction of uveitis. The endotoxin was administered by intravitreal injection. Each experimental group consisted of 3 to 4 rabbits (6–8 eyes). To control for animal-to-animal variability in this model, the right eye was always considered as the experimental eye (i.e., endotoxin plus neutralizing antibody), whereas the left eye acted as control (i.e., endotoxin plus control antibody). Preliminary experiments revealed that a contralateral untreated eye remains reliably unaffected after induction of EIU in the other eye. 

Animals were anesthetized using 0.4 to 0.8 ml of ketamine (20–40 mg/kg) and 0.1 to 0.2 ml of xylazine (1–2 mg/kg) injected intramuscularly in the hind limb. Amethocaine (0.5%) administered topically to the eyes supplemented the general anesthesia. Before injection, the eye was exposed by retracting the upper lid. A 30-gauge needle was inserted transconjunctivally at the 12 o’clock position, 3 to 4 mm posterior to the limbus, and 20 μl of the endotoxin solution was injected using a disposable syringe. The animals were observed closely until recovery from anesthesia. The eyes subsequently were monitored for redness, photophobia, and lacrimation. Each eye was examined using diffuse and oblique illumination and indirect ophthalmoscopy for signs of uveitis. Clinical scoring of AU was done using Hogan’s classification. ¹⁰ Animals were killed by intracardiac injection of 2 ml of euthanasia solution (phenobarbital) at a concentration of 350 mg/ml. The eyes were enucleated immediately after death and stored in 10% buffered formalin for 24 to 48 hours before processing. AH was aspirated before enucleation (see AH Sampling). 

AH Sampling

Immediately after death, 200 to 250 μl of AH was aspirated from the anterior chamber using a 30-gauge disposable insulin syringe, avoiding injury to the iris or lens. To inhibit fibrin clot formation, freshly prepared plasmin at a final concentration of 50 nM was added to the AH and incubated at 37°C for 90 to 120 minutes. The plasmin reaction was stopped using 100 μl of 2 mM aprotinin (Trasylol; Bayer Australia Ltd., Sydney). The AH cell count was performed in duplicate using a hemocytometer under 100× magnification. The AH was then centrifuged, and the supernatant was stored at −70°C, and the cell pellet was reconstituted in buffer and spread on a glass slide for Giemsa staining and differential cell count. 

Histopathologic Examination

Hematoxylin and eosin–stained sections of 3 to 5 μm thickness were prepared from paraffin-embedded blocks of the enucleated eyes to characterize the histopathologic features of inflammation. Sections were examined for the presence of keratic precipitates, inflammatory cells in anterior chamber, and ciliary body and retina and for altered vascularity of the iris and ciliary body. Infiltrating cells were counted under 200× magnification in contiguous fields across the whole section and were assigned to a histologic grade on a semi-logarithmic scale using the following criteria: grade 0, no cells/field; grade 1, 1 to 10 cells/field; grade 2, 11 to 30 cells/field; grade 3, 31 to 100 cells/field; and grade 4, 101 to 300 cells/field. The saggital section included two fields each of the ciliary body, anterior chamber angle, and iris and ora serrata and one field each of anterior chamber, pupillary area, vitreous, and the retina. A differential count of the infiltrating cells in the experimental eyes was performed by counting the mononuclear and PMN cell proportions under 400× magnification in five randomly selected fields. The mean value for each subset was then calculated and presented as a percentage of the total. 

Protein Assays

Protein levels in the AH were measured using a standard bicinchonic acid microtiter assay protocol (Bio-Rad; Pierce, Rockford, IL). 

IL-8–Induced Ocular Inflammation

Sixty microliters of recombinant human IL-8 (gifted from Dennis Taub’s laboratory at NIH) in a dose of 1 μg was administered into the anterior chamber or in four divided doses into four quadrants of the ciliary body of the rabbit eyes (n = 6 to 8 eyes). The severity of ocular inflammation was assessed using clinical and histopathologic grading. AH was aspirated at different time points to estimate the cell accumulation and the protein content. 

Neutralizing Antibody to IL-8

A murine monoclonal antibody against human IL-8 (WS-4) and a control antibody were prepared in the Department of Pharmacology, Cancer Research Institute, Kanazawa University, Kanazawa, Japan. This anti–IL-8 antibody has previously been shown to neutralize the biological activity of rabbit IL-8 and to abrogate early neutrophil accumulation in a rabbit model of arthritis. ¹¹ These antibodies were purified and tested for endotoxin contamination by the Limulus amebocyte lysate test (Associates of Cape Cod, Falmouth, MA). Endotoxin contamination of less than 10 pg/ml was considered acceptable, as our previous dose–response studies of EIU in the rabbit have shown no significant inflammatory response to intravitreal injection of endotoxin doses at or below this range. 

Neutralizing Experiments

Experiments were conducted to study the effects of anti–IL-8 antibody on the evolution of EIU in the rabbit. Each set of experiments was performed on a minimum of 6 to 8 eyes (3 to 4 rabbits) and a maximum of 24 eyes. Varying doses of anti–IL-8 antibody (50, 100, 250, 500, and 1000 μg) and 100 ng of LPS in the volume of 20 μl were injected intravitreal into the right eye, whereas the left eye was injected with LPS and control antibody at the same concentrations. All animals were killed at 24 hours after injection. In other experiments, the kinetics of EIU over 48 hours was studied after injecting the chosen dose of anti–IL-8 antibody (250 μg) in groups of animals that were killed at 6, 12, 24, and 48 hours after LPS administration. In the subsequent experiments an anti–IL-8 antibody was administered at three time points: 2 hours before, simultaneously with, or 6 hours after endotoxin challenge. All three groups of animals were killed at 24 hours after endotoxin challenge, irrespective of the time of anti–IL-8 antibody injection. Inflam-matory activity in the eyes was documented as described above. 

To examine the evolution of EIU using 250-μg dose of anti–IL-8 antibody, the eyes were treated simultaneously with endotoxin injection in a group of three to four animals. Repeated anterior chamber paracentesis was performed at 6, 12, 24, 48, and 72 hours after injection. Animals were killed at the end of the observation period, 72 hours after the endotoxin challenge. In addition to estimating leukocyte exudation and protein content in AH, clinical evaluations were performed to assess the severity of ocular inflammation. 

Statistical Analysis

Experimental and control eye pairs formed the basis of the statistical analyses, using two-tailed paired Student’s t-tests, χ² distribution, and the nonparametric test of Wilcoxon matched pairs. 

Results

IL-8–Induced Inflammation

Recombinant human IL-8–induced acute inflammation, when injected into the eyes of experimental rabbits. The human chemokine induced significant rabbit leukocyte accumulation in the AH within 12 hours of intraocular injection, with a cell count of 1900 ± 240/μl (mean ± SE) after ciliary body injection and 1145 ± 124/μl after AH injection (P < 0.05). By comparison, saline injected eyes remained inflammation free. The clinical and histologic scores of the injected eyes ranged from 3 to 4 in all cases. Protein levels in the AH were raised (20±3.5 mg/ml in the ciliary body–injected group and 17.7 ± 3.2 mg/ml in the AH–injected group) in comparison to saline-injected eyes (4.1 ± 1.2 mg/ml). 

Neutralizing Experiments

Initial experiments were designed to optimize the dosage and kinetics of the effect of anti–IL-8 antibody on the inflammatory response measured at 6 to 48 hours after endotoxin injection in groups of animals. Subsequent experiments evaluated the effect of anti–IL-8 antibody administered at three time points: 2 hours before, simultaneously with, or 6 hours after endotoxin challenge. Another kinetic study was performed to examine the evolution of EIU in a group of animals (using 250-μg dose of anti–IL-8 antibody), wherein the inflammatory response was observed at 6, 12, 24, 48, and 72 hours post-LPS injection to measure leukocyte counts and protein concentration. Preliminary experiments revealed that repeated aqueous sampling during the evolution of EIU did not change significantly the total or differential count of inflammatory cells or protein concentration in the AH over a time period of 6 to 72 hours. 

Dose Response and Time Course of Anti–IL-8 Antibody Inhibition of EIU

Anti–IL-8 antibody treatment produced a significant decrease in the AH leukocyte count in EIU when endotoxin and anti–IL-8 antibody were administered simultaneously. The significant inhibition (63.2% ± 35.6%, mean ± SE) of the cell count in the treated eyes measured at 24 hours after injection was obtained with the 250-μg dosage of antibody (P < 0.05) (Fig. 1) . The mean protein level in the AH of control antibody–treated eyes was 31.4 ± 2.1 mg/ml (mean ± SE), which was not significantly different from the mean of 38 ± 4.3 mg/ml for the treated eyes at the peak of inflammatory activity. The histologic grading of tissue sections obtained at 24 hours after injection showed a median grade of 3 (31–100 cells/field) in the treated eyes, whereas for the control eyes, the median grade of inflammation was 4 (101–300 cells/field). Compared to controls, four of six anti–IL-8 antibody–treated eyes showed a decrease in the histologically evaluated inflammation by at least one grade. 

Effect of Anti–IL-8 Antibody Administered at Different Time Points during the Induction of EIU

The anti–IL-8 antibody (250 μg) administered before, simultaneously with, or after endotoxin caused a reduction in the mean cell count in AH. The posttreatment group showed a 76% ± 40.8% reduction, whereas simultaneous and pretreatment were less effective (60.3% ± 33% and 37% ± 6.8%, respectively) (Fig. 2) . In the former group, the AH cell count decreased from 3925 ± 1167/μl in the control antibody–treated eyes to 940 ± 128/μl in the anti–IL-8 antibody–treated eyes. Of these, only the simultaneous treatment produced a statistically significant effect (P < 0.05, Wilcoxon matched-pairs test). Including the simultaneously treated group, there was no significant change in the AH protein concentration in the pre- and posttreated rabbits (P > 0.05). 

Histologic changes in both treated and control groups were consistent with the cell count results. The histologic grade in the eyes treated with anti–IL-8 antibody 2 hours before endotoxin injection had a median grade of 2 (10–30 infiltrating cells/field), whereas the control eyes had a median grade of 3 (31–100 cells/field). In the animals treated with anti–IL-8 antibody simultaneously or after endotoxin, the treated eyes had a median grade of 3 (31–100 cells/field) compared to a median grade of 4 (101–300 cells/field) for the control antibody–treated eyes. 

Evolution of Ocular Inflammation after Anti–IL-8 Treatment

The course of ocular inflammation after anti–IL-8 antibody treatment was studied in group of animals by examining AH for cellular accumulation and protein level at time points from 6 to 72 hours, after simultaneous administration of anti–IL-8 antibody and endotoxin. There was a reduction in the AH leukocyte count at all time points, with significant changes at the peak of inflammation and at 48 and 72 hours after injection. A substantial and consistent inhibition of cell infiltration (range, 60.3% ± 33% to 91% ± 12.2%, mean ± SE) was seen from as early as 6 hours to 72 hours after injection (Fig. 3) . During the evolution of EIU treated with anti–IL-8 antibody, there was no significant change in the level of protein in AH. 

At the peak of inflammation (24 hours), the mean ratio of PMN leukocytes to monocytes was 68:32 for the eyes treated with anti–IL-8 antibody compared to 76:24 for the untreated eyes. The inhibition in the cellular infiltration by PMN cells was statistically significant at 24 hours (P < 0.02, χ ² distribution) and 48 hours (P = 0.01) after treatment with anti–IL-8 antibody in comparison to the group injected with control antibody. However, at a later stage of EIU (72 hours), the ratio of PMN leukocytes and mononuclear cells was reversed (11:89), maintaining a similar proportions in both the treated (anti–IL-8 antibody) and untreated (control antibody) groups, unlike at 24 to 48 hours. The absolute number of cells (monocytes and polymorphs) was not performed. 

Effect of Anti–IL-8 Antibody Treatment on Clinical Uveitis Grade

When compared to the control eyes, the anti–IL-8 antibody–treated eyes showed reduced signs of inflammation in the form of congestion, lacrimation, keratitis, flare, and aqueous cells in the anterior chamber. The control eyes had a mean clinical uveitis score of 4 with flare, cellular accumulation, and membrane formation in the AC, whereas the treated eyes had a mean clinical score of 2 at the peak of AU. 

Discussion

Administration of IL-8 in vivo has been reported to produce inflammation in several biological systems. ¹² This study documents the induction of leukocyte recruitment, with clinical and histologic signs of ocular inflammation, after recombinant human IL-8 injection into the rabbit eye. This finding supports previous observations, that IL-8, whether injected exogenously or released from monocytes and other cells (uveal, corneal, or retinal pigment epithelial cells) during inflammation, is a key contributor to leukocyte recruitment. de Boer et al. ¹³ reported significant levels of IL-8 in the vitreous fluid of patients with uveitis. Similarly, we have recently reported elevated levels of IL-8 in AH of patients with active AU. ⁹  

Anti–IL-8 antibody partially blocks the development of EIU by neutralizing locally produced IL-8 in the eye. As the animal-to-animal variability in this model is significant, this study adopted a matched (treated/control) eye pair strategy to evaluate the anti–IL-8 antibody treatment. The clinical features and cellular infiltration in the anterior uvea during EIU in the rabbits were significantly reduced after treatment with anti–IL-8 antibody compared to administration of control antibody. 

A dose of 250 μg of anti–IL-8 antibody was adequate in inhibiting inflammation throughout the time course of induction of EIU, including at the peak of inflammation (24 hours after injection). There was a significant reduction in total leukocyte accumulation and change in the ratio of PMN leukocytes to monocytes during the evolution of EIU. In experimental arthritis animal model, a higher dose of 500 μg of IL-8 monoclonal antibodies produced a significant inhibition of PMN infiltration into rabbit synovial fluid during the early phase of joint inflammation, rather than at the peak of the disease process. ¹¹  

Both in vitro and in vivo studies, in endotoxin-induced pleurisy in rabbits, addition of anti–IL-8 antibody or desensitization of the neutrophils to IL-8, significantly decreased PMN migration in response to pleural fluid as the chemoattractant source. ^{14 15} This inhibitory effect has been further corroborated in similar in vitro studies in ocular inflammation and experimental arthritis. ¹¹ A study ¹³ reported inhibition of PMN migration by 41% to 79% when the vitreous samples obtained from uveitis- and nonuveitis-affected eyes were treated with a monoclonal antibody to IL-8. It is interesting to note that excessive accumulation of IL-8 itself at the interface of the endothelium and the circulating neutrophil may inhibit accumulation of leukocytes, perhaps by decreasing the chemotactic gradient or desensitization of the neutrophils. ⁷  

Disruption of the blood aqueous barrier (BAB) and changes in vascular permeability lead to the accumulation of cells and protein in the AH. Recruitment of PMN leukocytes into the extravascular space is regulated by chemoattractants, including α-chemokines, particularly IL-8. Reduction in the uveal cellular infiltrate after anti–IL-8 antibody therapy highlights the chemotactic role of IL-8 in leukocyte trafficking into the anterior chamber. Usually BAB disruption starts early, 1 to 6 hours after endotoxin injection, ¹⁶ resulting in protein accumulation earlier than leukocyte infiltration. This sequence of events was not affected by treatment with anti–IL-8 antibody, which did not lead to a significant change in protein exudation. Eyes with IL-8–induced inflammation showed raised AH proteins in contrast to the previous studies, in which the anterior chamber injection of CINC (cytokine-induced neutrophil chemoattractant), which shares similarity with IL-8, did not alter the protein exudation in AH. ¹⁷  

This divergence in the effect of anti–IL-8 antibody treatment suggests that other factors, including cytokines released in response to LPS, mediate the changes in vascular permeability responsible for protein exudation. These factors may remain unaffected by a reduction in the IL-8 concentration after treatment with neutralizing antibody to IL-8. 

Anti–IL-8 antibody administered at different time points during the induction of EIU consistently produced a reduction in inflammation, which was most evident when the anti–IL-8 antibody treatment was given concurrently with endotoxin. Similarly, the cellular infiltration observed from 6 to 72 hours post-LPS injection was consistently inhibited by anti–IL-8 antibody administered simultaneously with the endotoxin. Preliminary studies in our laboratory have demonstrated that in rabbit EIU, the leukocyte infiltration persists until 96 hours post-LPS injection. In this study, the persistent inhibition of leukocyte accumulation indicates that a single administration of anti–IL-8 antibody at the onset of inflammation may be sufficient to neutralize locally produced IL-8 throughout the evolution of the AU. 

 Submitted for publication September 28, 1998; revised April 5, 1999; accepted May 5, 1999.

 Commercial relationships policy: N.

¹  Present address: Department of Pharmacology, Cancer Research Institute, Kanazawa University, Kanazawa, Japan.

 

Figure 1.

 

Different doses of anti–IL-8 antibody mixed with LPS (100 ng) were administered intravitreally in experimental eyes. Control eyes received LPS plus isotype matched control antibody. AH leukocyte count was measured 24 hours after the injections and the mean reduction in leukocyte count (expressed as percentage of the control) compared. The average AH cell count in control eyes using different doses varied as follows: 50 μg, 2325 ± 1525 (mean ± SE); 100 μg, 2645 ± 1976; 250 μg, 6419 ± 1165; 500 μg, 5280 ± 1320; and 1000 μg, 6259 ± 1570.

Figure 2.

 

Anti–IL-8 antibody (250 μg) was injected at three different time points in relation to LPS (100 ng) administration. Animals were studied at 24 hours post-endotoxin treatment. The mean (±SE) cell count of these three groups was compared. There was a statistically significant difference between the groups in which anti–IL-8 antibody was administered simultaneously with endotoxin.

Figure 3.

 

Changes in AH leukocyte count evaluated by repeated anterior chamber paracentesis from 6 to 72 hours after simultaneous administration of LPS (100 ng) and anti–IL-8 antibody (250 μg). Error bars, SEM.