Synergistic Uveitic Effects of Tumor Necrosis Facfor-a and Interleukin-

Investigative Ophthalmology & Visual Science, Vol. 33, No. 7, June 1992 Copyright © Association for Research in Vision and Ophthalmology

Synergistic Uveitic Effects of Tumor Necrosis Facfor-a and InterleukinLloyd N. Fleisher, Jenny D. Ferrell, and M. Christine McGahan Tumor necrosis factor (TNF) and interleukin-1 (IL-1), cytokines with multiple, overlapping biologic activities, have been shown to interact synergistically in nonocular tissues. To test the hypothesis that coinjection of TNF and IL-1 interact synergistically in the eye, low, marginally inflammatory doses of human recombinant TNF-a (4000 U), IL-1/? (40 U), and TNF-a + IL-10 (TNF-o/IL-1/9) were injected into the vitreal chamber of the rabbit eye, and inflammation was assessed at 6, 24, 48, and 168 hr post-cytokine injection. TNF-a/IL-1/? induced an anterior uveitis that was barely detectable at 6 hr, increased at 24 hr, peaked at 48 hr, and largely resolved by 168 hr. Synergy was observed for infiltration of inflammatory leukocytes into aqueous humor at 24 and 48 hr and for protein and prostaglandin E levels in aqueous humor at 48 hr. Based upon protein levels in vitreous humor, TNF-a/IL-1/f? also induced a posterior uveitis. This posterior uveitis was not apparent until 48 hr and then increased significantly at 168 hr. At 48 and 168 hr, the effects of TNF-a/IL-1/8 on protein levels in vitreous humor were consistent with a synergistic interaction. Results of separate experiments using higher dose combinations of TNF-a/IL-1/S and a longer time course suggested that the effects of TNF-a/IL1/8 on the blood vitreous barrier persisted beyond 168 hr. The results of this study support the hypothesis that TNF-a and IL-1/? interact synergistically when injected into the rabbit eye. Release of these cytokines by inflammatory leukocytes and ocular tissues is likely to be an important component of naturally occurring uveitis. Invest Ophthalmol Vis Sci 33:2120-2127, 1992

Cytokines have been described as "the language cells used to communicate during the evolution of an inflammatory response."1 Among the numerous cytokines released during an inflammatory event, interleukin-1 (IL-1) and tumor necrosis factor-a (TNF-a) are considered primary cytokines because they initiate a cascade of events integral to the inflammatory process. For example, TNF-a induces release of IL-1. 23 IL-1 and TNF-a induce release of secondary cytokines such as IL-6,4'5 IL-8,6 and a monocyte chemotactic and activating factor.7 They also induce release of bioactive lipids such as eicosanoids289 and platelet-activating factor,10'1112 and increase expression of adhesion molecules on vascular endothelial cells.13-14'15 Recent evidence supports important roles for IL-1 and TNF-a in the ocular inflammatory response. From the Department of Anatomy, Physiological Sciences, and Radiology, North Carolina State University, College of Veterinary Medicine, Raleigh, North Carolina. Supported by National Institutes of Health Grant EY-08688 and a grant from the State of North Carolina. Presented in part at the meeting of the Association for Research in Vision and Ophthalmology, Sarasota, Florida, May 1991. Submitted for publication: September 23, 1991; accepted January 16, 1992. Reprint requests: Lloyd N. Fleisher, Department of Anatomy, Physiological Sciences, and Radiology, North Carolina State University, College of Veterinary Medicine, Raleigh, NC 27606.

When injected intravitreally, IL-1 and TNF-a induce inflammatory responses that differ in terms of time course and types of inflammatory cells infiltrating the anterior uvea and aqueous humor.161718 Furthermore, IL-1 activity has been detected in aqueous humor of rabbits with antigen-induced uveitis19 and in subretinal fluid and vitreous humor of humans with retinal detachment.20 Intravitreal injection of IL-621 or IL-822 also induces uveitis, and IL-6 activity has been detected in rat serum and aqueous humor after footpad injections of endotoxin21 and in aqueous humor of patients with Fuch's heterochromic cyclitis and toxoplasmic uveitis.23 IL-1 and TNFa can exert synergistic effects when administered together.24 This is consistent with the similarity of their biologic effects and the likelihood that they are released in close spatial and temporal proximity. Based upon these considerations, we examined the hypothesis that cointravitreal injection of low doses of TNF-a and IL-1 {would exert synergistic inflammatory effects in the rabbit eye. Materials and Methods Induction and Assessment of Inflammation; Removal of Intraocular Fluids and Tissues All experiments were conducted in accordance with the ARVO Resolution on the Use of Animals in

2120

No. 7

CYTOKINE SYNERGY IN THE EYE / Fleisher er ol

Research. Male New Zealand white rabbits, weighing approximately 2-3 kg, were restrained in a paper diaper and the eyes were anesthetized with proparacaine HC1 (Allergan Pharmaceuticals, Inc., Irvine, CA). One eye of each rabbit was injected intravitreally with 50 fi\ of human recombinant TNF-a (specific activity: 107 U/mg protein; Amgen Biologicals, Thousand Oaks, CA), human recombinant IL-1/3 (specific activity: 5 X 108 U/mg protein; Amgen Biologicals, Thousand Oaks, CA), or TNF-a + IL-1/3. The contralateral eye received an equal volume of sterile vehicle consisting of phosphate buffered saline containing 0.025% low-endotoxin bovine serum albumin (Fraction V; Sigma Chemical Co., St. Louis, MO). Injections were made through a 30 G needle attached to a syringe, approximately 3 mm posterior to the limbus. Inflammation was assessed over time by biomicroscopic examination of the anterior segment with a slitlamp. Iridal hyperemia and aqueous humorflarewere rated on a scale of 0-3. 25 At the termination of each experiment, animals were killed by exposure to an atmosphere of 100% CO2. The eyes were removed, aqueous and vitreous humors were aspirated into heparin-treated syringes, and iris-ciliary bodies were removed, gently blotted, weighed, and frozen at -80°C. Aliquots of aqueous humor (5-10 /A) were removed for determination of protein and total leukocyte cell number. Aqueous humor (100 n\) then was centrifuged in a micro-centrifuge (Model 235c; Allied Fisher Scientific, Pittsburgh, PA) for 60 seconds and the supernatant was aspirated and transferred to polypropylene centrifuge tubes for prostaglandin assay. The pellet, containing formed blood elements, was used for differential analysis of leukocytes. Aqueous and Vitreous Humor Protein Concentration, Aqueous Humor Total and Differential Leukocyte Number, and Iris-Ciliary Body Myeloperoxidase Activity Protein concentration in aqueous and vitreous humor was determined by the Lowry method26 modified for reduced volumes. Total leukocyte number in aqueous humor was determined using a hemacytometer. Differential analysis of leukocytes was determined by resuspending the aqueous humor pellet in Hank's Balanced Salt solution (JRH Biosciences, Lenexa, KS) and using a Cytospin 2 (Shandon Inc, Pittsburgh, PA) to deposit the cells on a slide. Slides were allowed to dry and then stained with Diff-Quik (Baxter, McGaw Park, IL), which gives results similar to a Wright-Giemsa stain. Differential counts then were determined by light microscopy. Myeloperoxidase (MPO) activity in iris-ciliary body, a marker for

2121

neutrophils,27 was determined as described by Bradley et al.27 Extraction of Aqueous Humor and Radioimmunoassay of Prostaglandin E Prostaglandin E (PGE) was extracted from aqueous humor supernatants into ethyl acetate and measured by radioimmunoassay as previously described.28 The antiserum used in this assay does not distinguish between PGE, and PGE2. Data Analysis Effects of cytokine treatments at a single time point and the effects of a particular cytokine treatment over the 168 hr time course were determined using oneway analysis of variance. The significance of mean differences then was determined using Tukey's omega test. Significance of mean differences between paired eyes (ie, cytokine-injected vs. cytokine vehicle-injected eyes) was determined using a paired Student's t-test. To determine whether cointravitreal injection of TNF-a + IL-1/3 caused synergistic inflammatory effects, an unpaired Student's t-test, using pooled variances, was used to test whether the algebraic mean of individual cytokine treatments (ie, meanTNFa + meanIL_1/S) was different from the mean when both cytokines were coinjected into the eye (ie, meanTNFa ) Results Determination of Appropriate Doses ofTNF-aandIL-1/ff To determine doses of TNF-a and IL-1 /? that produced marginally observable ocular inflammation, dose-response experiments were performed and inflammation was assessed biomicroscopically over 7-10 d. Results of these experiments indicated that 5000 U of TNF (Figs. 1A and IB) and 50 U of IL-10 (Figs. 2A and 2B) were appropriate doses. With 5000 U of TNF-a, disruption of the blood-aqueous humor barrier and dilation of iridal blood vessels, as indicated by aqueous flare and iridal hyperemia, respectively, never exceeded a rating of 0.5 and essentially were resolved by 168 hr post-cytokine injection. Similar results were observed with 50 U of IL-1/3, except that the inflammatory response essentially was resolved by 72 hours post-cytokine injection. With these doses of TNF-a and IL-10 as guidelines, a dose-response experiment was conducted using three combinations of TNF-a/IL-1/J: 2500 U/25 U, 4000 U/40 U, and 5000 U/50 U (Figs. 3A and 3B). At 24 hr post-intravitreal injection of these cytokine combinations, significant uveitis was observed at all doses.

INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / June 1992

2122

6

24 48 72 TIME (HOURS)

168 240

6

24 48 72 168 240 TIME (HOURS)

Vol. 33

and the salient aspects of the time courses for cytokine-injected eyes are illustrated in Figure. 4 (A-D). Small numbers of leukocytes were observed in aqueous humor at 6 hr post-TNF-a/IL-1/3 injection. Leukocyte infiltration increased dramatically at 24 hr. At 48 hr, leukocyte number was significantly greater than at 6 or 168 hr (Fig. 4A). Furthermore, the effect of TNF-a/IL-1/3 on infiltration of leukocytes into aqueous humor was synergistic at 24 and 48 hr, because the means at these time points were significantly greater than the sum of the means of each cytokine treatment alone (Fig. 4A). Differential analysis of leukocytes in aqueous humor after TNF-a/IL-1/3 injection revealed approximately equal numbers of neutrophils (43%) and monocytes (38%) and smaller numbers of lymphocytes (19%) at 24 hr (Fig. 5). By 48 hr, the cellular response was largely monocytic (62%) with smaller, but approximately equal numbers of neutrophils (20%) and lymphocytes (18%). Analysis of MPO activity in iris-ciliary body revealed no differences between cytokine-injected and contralateral, cytokine vehicle-injected eyes (Table 1), no differences between different cytokine treatments at a sin-

Fig. 1. Effects of different doses of intravitreally injected TNFa over a 240-hr time course (n = 3-5). (A) Iridal hyperemia; (B) aqueous flare.

However, at the 5000 U/50 U dose, infiltration of leukocytes into aqueous humor was significantly greater than at either of the lower doses, and protein concentration in aqueous humor was significantly higher than at the 2500 U/25 U dose. None of the dose combinations tested induced significant changes in vitreous humor protein concentration at the 24 hr time point.

6


96

168


96

168

Time Course for TNF-a/IL-10-Induced Uveitis Based upon the results of the dose-response studies, 4000 U of TNF-a and 40 U of IL-1/3 were chosen to study the time course of TNF-a/IL-ljS-induced inflammation over a 168 hr (7 d) period and to determine whether this cytokine combination induced synergistic inflammatory effects compared to separate groups of animals injected with TNF-a or IL-1/3. The individual means ± standard error of the mean for all cytokine-injected eyes (right eyes) and cytokine vehicle-injected eyes (left eyes) are listed in Table 1,

Fig. 2. Effects of different doses of intravitreally injected IL-l/3 over a 168-hr time course (n = 2-6). (A) Iridal hyperemia; (B) aqueous flare.


No. 7

** I

s

u CC Ul

Fig. 4C). Furthermore, at 48 and 168 hr, the effect of TNF-a/IL-1/3 on vitreous protein concentration was synergistic (Fig. 4C). It should be noted that TNF-a/ IL-l/?-induced iridal hyperemia also was most pronounced at 24 and 48 hr, although somewhat more vasodilation was observed at 24 hr (Table 1). Intravitreal injection of TNF-a/IL-1/3 caused PGE levels in aqueous humor to significantly increase by 6 hr, reach a maximum at 48 hr, and return to baseline levels at 168 hr. Aqueous humor PGE levels in TNFa/IL-ljfl-injected eyes were greater than corresponding cytokine vehicle-injected eyes at all time points except 168 hr (Table 1). Furthermore, at 48 hr, the effect of TNF-a/IL-1/3 on aqueous humor PGE concentration was synergistic (Fig. 4D).

3500 3000

I

oC

2500 2000

GL

I

1500 1000 500 0

2500/25 4000/40 5000/50 DOSE OF TNFa / IL-1 p (UNITS)

20

it

18

Inflammatory Effects of IL-l/ft and TNFa Alone

16 14 12 10 8 X.

6

1

4 2 0 AQUEOUS B

2123

L

2500/25 4000/40 5000/50 DOSE OF TNFa / IL-1P (UNITS)

CZD

Fig. 3. Inflammatory effects of three dose combinations of IL-1/3/ TNFa at 24 hr postintravitreal injection (n = 7-16). (A) Iridal hyperemia and cells in aqueous humor; (B) protein in aqueous and vitreous humors. *Significantly different from the low combination dose. **Significantly different from the other two combination doses.

gle time point (Table 1), and no changes over time for a particular cytokine treatment (Table 1). Treatment with TNF-a/IL-1/3 induced disruption of blood-ocular barriers, resulting in elevated protein concentration in the anterior and posterior segments of the eye. However, the time courses for these changes were distinctly different. In aqueous humor, protein concentration was significantly greater in TNF-a/IL-l/3-injected eyes compared to contralateral, cytokine vehicle-injected eyes at all time points (Table 1). Protein concentration in TNF-a/IL-1/3injected eyes increased significantly at 24 hr and was maximally elevated at 48 hr. At 48 hr, aqueous protein concentration was significantly greater than at any other time point, and the effect of the combined cytokine treatment was synergistic (Fig. 4B). In vitreous humor, protein concentration did not increase until 48 hr post-TNF-a/IL-1/3 injection. It continued to increase up to 168 hr, at which time it was significantly greater than at any other time point (Table 1,

The inflammatory effects of intravitreally injected IL-1/3 (40 U) or TNF-a (4000 U) were mild compared to those produced by TNF-a/IL-1/3 (Table 1). For the sake of simplicity, those inflammatory effects significantly altered by TNF-a or IL-1/? over time are described in this section but are not annotated in Table 1. For TNF-a, small, but significant increases in aqueous humor leukocyte number occurred at 24 and 48 hr. At both of these time points, cells consisted primarily of monocytes, 69% and 78%, a smaller number of lymphocytes, 28% and 22%, and few if any neutrophils, 3% and 0%. Vitreous humor protein also was elevated at 168 hr. No other inflammatory parameter was significantly altered by TNF-a treatment. IL-10 had somewhat more inflammatory effects than TNFa. Iridal hyperemia and aqueous humor PGE levels were significantly increased at 24 hr, aqueous humor leukocyte number was significantly increased at 24 and 48 hr, and vitreous humor protein concentration was significantly elevated at 168 hr. Regarding leukocytes in aqueous humor, the cells consisted of neutrophils (57%), monocytes (32%), and small numbers of lymphocytes (11%) at 24 hr. By 48 hr, cells were primarily monocytes (70%) and lymphocytes (28%). Only 2% were neutrophils. Discussion The results of the present investigation indicate that coinjection of human recombinant TNF-a and IL-1/3 into the vitreal chamber of the rabbit eye induced inflammatory effects consistent with a synergistic interaction between these cytokines. Using doses of TNFa (4000 U) and IL-1/3 (40 U), which produced only marginal inflammatory effects when injected alone, significant inflammation was induced when the two cytokines were injected together. In nonocular systems, synergy between TNF-a and IL-1 has been dem-


2124

Vol. 33

Table 1. Effects of intravitreal injections of TNFa (4000 U), IL-10 (40 U), TNFa + IL-1/3 and cytokine vehicle Time after intravitreal injection 6hr

Intravitreal injection: OS eye ODeye TNFa IL-l/S

OD OS* OD OS

TNFa/IL-1/3 OD 24 hr

TNFa

OS OD

IL-10

OD

OS OS TNFa/IL-1/3 48 hr

168 hr

IL-1/3

OD OS OD OS OD

TNFa/IL-1/3

OS OD

TNFo

OS OD

IL-1/3

OD

TNFa/IL-1/3

OD OS

TNFo

OS OS

Iridal hyperemia 0.0 (6) 0.0 (6) 0.0 (6) 0.0 (6) 0.2 (6)f 0.0 (6) 0.3 (6) 0.0 (6) 0.9 (5)t 0.0(5)

1.8(9)tt 0.1(9) 0.3 (7) 0.1(7) 0.1(7) 0.1(7) 1.3(12)t* 0.1 (12) 0.0 (4) 0.0 (4) 0.0 (4) 0.0 (4) 0.0 (4) 0.0 (4)

Iris-ciliary body MPO (/mg tissue)

Aqueous cells

Ovl) 33 ± 33 (6) 28 ± 29 (6) 3 ± 3 (6) 0(6) 67 ± 38 (6) 3 ± 2 (6) 367 ± 141(6)$ 1 ± 1 (6) 224 ± 50 (5)t 0(5) 2185 ± 540 (8)f*

0(8) 491 ±316(7)$ 1 ± 1 (7) 183 ±48(7)$ 13 ± 6 (7) 3350 ± 378 (12)t* 26 ±20 (12) 5 ± 5 (4) 2 ± 2 (4) 0(4) 0(4) 180 ± 24 (4)ft 0(4)

0.33 0.29 0.32 0.29 0.27 0.22 0.27 0.39 0.28 0.25 0.47 0.35 0.43 0.33 0.28 0.34 0.45 0.41 0.34 0.35 0.37 0.30 0.25 0.30

± 0.03 (6) ± 0.04 (6) ±0.01 (6) ± 0.02 (6) ± 0.06 (6) ± 0.06 (6) ± 0.02 (5) ± 0.05 (5) ± 0.04 (5) ± 0.05 (5) ±0.13 (9) ± 0.04 (9) ±0.18 (7) ± 0.07 (7) ± 0.05 (7) ± 0.04 (7) ±0.10(12) ±0.09(12) ± 0.02 (4) ± 0.03 (4) ± 0.04 (4) ± 0.02 (4) ± 0.05 (4) ± 0.05 (4)

Aqueous protein (fg/trf)

Vitreous protein (ng/nO

Aqueous PGE (ng/ml)

2.2 ± 0.6 (6) 1.6 ±0.4 (6) 2.1 ±0.4(6) 2.3 ± 0.8 (6) 3.2 ± 1.4(6)$ 1.1 ±0.1 (6) 4.0 ± 0.9 (6) 2.8 ± 1.1 (6) 10.2 ± 4.3 (5)t 1.6 ±0.3 (5) 13.7 ± 2.7 (9)$§ 1.0 ±0.1 (9) 2.5 ± 1.2(7) 1.6 ±0.6 (7) 3.0 ± 0.7 (7)$ 1.2 ±0.2 (7) 18.6 ± 1.8(12)f$ 1.3 ± 0.1 (12) 1.3 ±0.2 (4) 1.0 ±0.2 (4) 1.3 ±0.2 (4) 1.1 ±0.1 (4) 2.6 ± 0.5 (4)t$ 1.0 ±0.1 (4)

0.4 ± 0.05 (6) 0.4 ± 0.02 (6) 0.5 ± 0.02 (6) 0.5 ±0.10 (6) 0.5 ± 0.03 (6) 0.5 ± 0.04 (6) 0.4 ± 0.05 (6) 0.4 ± 0.05 (6) 0.4 ± 0.08 (4) 0.4 ± 0.06 (4) 0.6 ± 0.07 (8) 0.4 ± 0.03 (8) 0.3 ± 0.03 (7) 0.3 ± 0.04 (7) 0.6 ±0.10 (7) 0.4 ± 0.07 (7) 1.3 ± 0.15 (12)t$ 0.4 ±0.05 (12) 1.3 ±0.27 (4)$ 0.5 ± 0.04 (4) 1.1 ±0.09(4)$ 0.5 ± 0.03 (4) 4.0 ± 0.24 (4)f$ 0.5 ± 0.03 (4)

0.63 ± 0.22 (6) 0.39 ±0.18 (6) 0.24 ± 0.05 (6) 0.21 ±0.05(6) 1.60 ±0.67 (6)$ 0.29 ±0.15 (6) 0.49 ±0.14 (6) 0.84 ± 0.37 (6) 2.63 ± 1.11(5) 0.61 ±0.06(5) 2.07 ± 0.51 (9)$§ 0.19 ±0.05 (9) 0.62 ± 0.29 (6) 0.25 ± 0.07 (6) 0.36 ±0.11(7) 0.43 ± 0.07 (7) 2.26 ± 0.39 (12)f$ 0.36 ±0.09 (12) 0.15 ±0.03 (4) 0.14 ±0.01 (4) 0.40 ±0.14 (4) 0.38 ±0.14 (4) 0.29 ± 0.09 (4) 0.19 ±0.04 (4)

Data expressed as mean ± SEM for the number of animals in parenthesis. * All OS eyes received injections of cytokine vehicle, t Significantly different from TNFa- and IL-1/S-injected groups at that time point (P < 0.05).

$ Significantly different from the contralateral (OS), cytokine vehicle-injected eye (P < 0.05). § Significantly different from the TNFa-injected group at that time point (P < 0.05).

onstrated for such diverse actions as induction of shock29 and a Shwartzman-like reaction,30 cytotoxicity,3132'33 increased vascular permeability,34 infiltration of polymorphonuclear leukocytes into the peritoneal cavity,35 production of other cytokines such as granulocyte/macrophage- and granulocyte-colony stimulating factors,36 production of collagenase by synovialfibroblasts,9proliferation of thymocytes,37 suppression of lipoprotein lipase activity in adipocytes,38 antibody-mediated glomerular injury,39 and enhancement of prostaglandin release by endothelial cells,840 fibroblasts,9 and macrophages.2 In the present investigation, TNF-a/IL-1/3-induced uveitis was first evident in the anterior segment and, based upon vitreous humor protein concentration, it later progressed to the posterior segment of the eye. TNF-a/IL-1/3-induced anterior uveitis was mild at 6 hr, intensified significantly between 24 and 48 hr, and was largely resolved by 168 hr, the endpoint of this study. Synergy between TNF-a and IL-1/3 was observed for infiltration of inflammatory leukocytes into aqueous humor at 24 and 48 hr and for protein and PGE levels in aqueous humor at 48 hr (Figs. 4A,B,D). Failure of TNF-a/IL-1 /3 treatment to significantly increase MPO activity in iris-ciliary body at those time points when total leukocyte number in

aqueous humor was increased probably was a reflection of the small numbers of neutrophils in aqueous humor—940 (43% of 2185) at 24 hr and 670 (20% of 3350) at 48 hr (Table 1, Figs. 4A and 5). Apparently, in a highly vascularized tissue such as iris-ciliary body, larger numbers of neutrophils are required to significantly increase MPO activity. This is consistent with the recent report by our group that MPO activity in iris-ciliary body only increased when cell number in aqueous humor was considerably higher than observed in the present study and was made up primarily of neutrophils.41 TNF-a/IL-1 /3induced posterior uveitis was first apparent at 48 hr when vitreous humor protein levels (1.3 ± 0.15 mg/ ml) more than doubled compared to those observed at 6 and 24 hr. By 168 hr, vitreous humor protein concentration increased an additional three-fold (compared to 48 hr) to 4.0 ± 0.24 mg/ml (Fig. 4C). At 48 and 168 hr, vitreous humor protein levels were significantly higher than the sum of the individual cytokine treatments, suggesting a synergistic effect of TNF-a/IL-1/3 on disruption of the blood-vitreous barrier. In a separate series of experiments, vitreous humor protein levels measured 240 hr after intravitreal injection of TNF-a/IL-1/3 (5000 U/50 U and 10,000 U/100 U) were 5.6 ± 0.8 (n = 3) and 10.0


No. 7

- A -

TNFa

-O-

IL-1 P

- • -

-A-

TNFa/IL-ip

TNFa

2125

-o-

IL-ip

-°-

TNFa/IL1-p

3000 •

2000 -

o CC a. 0

A

24

6

24

48

- A -

TNFa

-O-

IL-1P

48

168

TIME (HOURS)

TIME (HOURS) -D-

- A -

TNFo/IL-1p

TNFo

-D-

TNFo/IL-ip

UJ

a a

O UJ

24

48

168

24

48

168

TIME (HOURS)

TIME (HOURS)

Fig. 4. Comparison of the inflammatory effects of intravitreally injected TNFa (4000 U), IL-1/3 (40 U), and TNFa/IL-10 (4000 U/40 U) over a 168-hr time course (n = 4-12). *Significantly different from the 6- and 168-hr time points for that treatment. **Significantly different from all other time points for that treatment. §Significantly different from the algebraic mean of TNFa + IL- ljS treatments at that time point (synergy). (A) Cells in aqueous humor; (B) protein concentration in aqueous humor; (C) protein concentration in vitreous humor; (D) PGE concentration in aqueous humor.

±1.1 mg/ml (n = 6; mean ± SEM), respectively. Although higher TNF-a/IL-1/3 doses were used in these experiments, the results suggest that TNF-a/IL-1/3—A— NEUTROPHILS

- O LYMPHOCYTES

- D MONOCYTES

UJ

O

800

24 48 TIME (HOURS)

168

Fig. 5. Differential analysis of leukocytes in aqueous humor over a 168-hr time course after intravitreal injection of TNFa/IL-1/3 (4000 U/40 U)(n = 4-12).

induced disruption of the blood-vitreous barrier persists beyond 168 hr and may even intensify during this time. A progressive posterior uveitis has been reported after intravitreal injection of high doses of TNF-a (20,000 U;18100,000-500,000 U17), but not of IL-1.16 These high doses of TNF-a also induced an anterior uveitis. However, it differed from that observed in the present study because it persisted well beyond 168 hr and was accompanied by extensive uveal cellular infiltration and cellular accumulation in aqueous humor. It also should be noted that in an earlier investigation in which human recombinant IL-1 a was used, 12.5 U produced extensive anterior uveitis.16 In the present investigation, 40 U of IL-1 /3 produced a weak uveitic response, suggesting that IL-1 a is more potent than IL-1/3 in the rabbit eye. This may be related to IL-1 a being a membrane-bound form, while IL-1/? is released from cells.42 Although the present study did not address how TNF-a/IL-lj8 causes synergistic ocular inflammatory

2126


effects, several mechanisms are likely to be involved. These include effects on vascular endothelium, leading to increased permeability and expression of adhesion molecules for leukocytes, generation of chemotactic substances for leukocytes, enhanced production of prostaglandins and platelet-activating factor, and generation of other cytokines. TNF-a/IL-l/3-induced stimulation of prostaglandin release could contribute to the vasodilation observed in this study. IL-1 has been shown to stimulate release of PGE2 from corneal fibroblasts,43 and TNF-a and IL-1 stimulate release of PGE2 from conjunctival fibroblasts.44 Furthermore, TNF-a and IL-1/3 promote leukocyte chemotaxis by inducing expression of adhesion molecules such as intercellular adhesion molecule-1 (ICAM-1),13 endothelial cell adhesion molecule-1,14 and vascular cell adhesion molecule-115 on the surface of endothelial cells. TNF-a and IL-1 recently have been reported to increase expression on ICAM-1 on cultured human corneal endothelium and stimulate neutrophil adherence to these cells.45 Endothelial cell adhesion molecules bind to receptors on leukocytes, trapping them in the circulation. The trapped leukocytes are attracted to the site of inflammation, possibly as a result of TNFa- and IL-1-induced release of leukocyte chemotactic factors such as IL-8,6 IL-6,4>46>47 and a monocyte chemotactic and activating factor.7 IL-8 is a neutrophil activator and chemotactic factor for neutrophils and T lymphocytes,4849 whereas IL-6 is a macrophage activator and chemotactic factor for T lymphocytes.5051 It should be noted that IL-6 and IL-8 induce inflammation when injected intravitreally,2122 and that IL-6 appears to be synthesized intraocularly during uveitis induced by injection of endotoxin into the footpad of the rat.21 In summary, low, marginally inflammatory doses of TNF-a plus IL-lj8 have been shown to exert synergistic inflammatory effects when injected into the vitreal chamber of the rabbit eye. The inflammatory response begins in the anterior segment and progresses to the posterior segment as the anterior uveitis resolves. Release of TNF-a and IL-1 and their potential synergistic interactions could be important components of the etiology of naturally occurring uveitis. Key words: tumor necrosis factor, interleukin-1, synergy, uveitis, cytokines Acknowledgements The authors would like to thank Angela Grimes and Maribeth Smith for excellent technical assistance.

References 1. Taylor R: Cytokine networks: Immunobiology surfaces. Journal of the National Institutes of Health 3:71, 1991.

Vol. 33

2. Bachwich PR, Chensue SW, Larrick JW, and Kunkel SL: Tumor necrosis factor stimulates interleukin-1 and prostaglandin E2 production in resting macrophages. Biochem Biophys Res Commun 136:94, 1986. 3. Nawroth PP, Bank I, Handley D, Cassimeris J, Chess L, and Stern D: Tumor necrosis factor/cachectin interacts with endothelial cell receptors to induce release of interleukin 1. J Exp Med 164:1363, 1986. 4. Sironi M, Breviario F, Proserpio P, Biondi A, Vecchi A, Van Damme J, Dejana E, and Montovani A: IL-1 stimulates IL-6 production in endothelial cells. J Immunol 142:549, 1989. 5. Sanceau J, FalcofFR, Beranger F, Carter DB, and Wietzerbin J: Secretion of interleukin-6 (IL-6) by human monocytes stimulated by muramyl dipeptide and tumour necrosis factor alpha. Immunology 69:52, 1990. 6. Larsen CG, Anderson AO, Oppenheim JJ, and Matsushima K: Production of interleukin-8 by human dermalfibroblastsand keratinocytes in response to interleukin-1 or tumour necrosis factor. Immunology 68:31, 1989. 7. Larsen CG, Zachariae CO, Oppenheim JJ, and Matsushima K: Production of monocyte chemotactic and activating factor (MCAF) by human dermalfibroblastsin response to interleukin 1 or tumor necrosis factor. Biochem Biophys Res Commun 160:1403, 1989. 8. Zavoico GB, Ewenstein BM, Schafer AI, and Pober JS: IL-1 and related cytokines enhance thrombin-stimulated PGI2 production in cultured endothelial cells without affecting thrombin-stimulated von Willebrand factor secretion or platelet-activating factor biosynthesis. J Immunol 142:3993, 1989. 9. Meyer FA, Yaron I, and Yaron M: Synergistic, additive, and antagonistic effects of interleukin-1 beta, tumor necrosis factor alpha, and gamma-interferon on prostaglandin E, hyaluronic acid, and collagenase production by cultured synovial fibroblasts. Arthritis Rheum 33:1518, 1990. 10. Bussolino F, Breviario F, Tetta C, Aglietta M, Mantovani A, and Dejana E: Interleukin 1 stimulates platelet-activating factor production in cultured human endothelial cells. J Clin Invest 77:2027, 1986. 11. Camussi G, Bussolino F, Salvidio G, and Baglioni C: Tumor necrosis factor/cachectin stimulates peritoneal macrophages, polymorphonuclear neutrophils, and vascular endothelial cells to synthesize and release platelet-activating factor. J Exp Med 166:1390, 1987. 12. Valone FH and Epstein LB: Biphasic platelet-activating factor synthesis by human monocytes stimulated with IL-1/3, tumor necrosis factor, or IFN-7. J Immunol 141:3945, 1988. 13. Pober JS, Gimbrone MA Jr, Lapierre LA, Mendrick DL, Fiers W, Rothlein R, and Springer TA: Overlapping patterns of activation of human endothelial cells by interleukin 1, tumor necrosis factor, and immune interferon. J Immunol 137:1893, 1986. 14. Bevilacqua MP, Pober JS, Mendrick DL, Cotran RS, and Gimbrone MA Jr: Identification of an inducible endothelial-leukocyte adhesion molecule. Proc Natl Acad Sci USA 84:9238, 1987. 15. Carlos TM, Schwartz BR, Kovach NL, Yee E, Rosso M, Osborn L, Chi RG, Newman B, Lobb R, and Harlan JM: Vascular cell adhesion molecule-1 mediates lymphocyte adherence to cytokine-activated cultured human endothelial cells. Blood 76:965, 1990. 16. Rosenbaum JT, Samples JR, Hefeneider SH, and Howes EL Jr: Ocular inflammatory effects of intravitreal interleukin-1. Arch Ophthalmol 105:1117, 1987. 17. Rosenbaum JT, Howes EL Jr, Rubin RM, and Samples JR: Ocular inflammatory effect of intravitreally-injected tumor necrosis factor. Am J Pathol 133:47, 1988.

No. 7


18. Fleisher LN, Ferrell JB, and McGahan MC: Ocular inflammatory effects of intravitreally injected tumor necrosis factor-alpha and endotoxin. Inflammation 14:325, 1990. 19. Prendergast RA, Libre PE, Orencole SF, and Dinarello CA: Ocular IL-1 generation in immunogenic uveitis. Invest Ophthalmol Vis Sci 29(suppl):65, 1988. 20. Davis JL, Jalkh AE, Roberge F, Caspi R, Flynn HW Jr, Schepens CL, and Nussenblatt RB: Subretinal fluid from human retinal detachment contains interleukin 1. Invest Ophthalmol Vis Sci 29(suppl):396, 1988. 21. Hoekzema R, Murray PI, Van Haren MAC, Helle M, and Kijlstra A: Analysis of interleukin-6 in endotoxin-induced uveitis. Invest Ophthalmol Vis Sci 32:88, 1991. 22. Ferrick MR, Thurau SR, Oppenheim MH, Herbort CP, Ni M, Zachariae COC, Matsushima K, and Chan CC: Ocular inflammation stimulated by intravitreal interleukin-8 and interleukin-1. Invest Ophthalmol Vis Sci 32:1534, 1991. 23. Murray PI, Hoekzema R, Van Haren MAC, De Han FD, and Kijlstra A: Aqueous humor interleukin-6 levels in uveitis. Invest Ophthalmol Vis Sci 31:917, 1990. 24. Le J and Vilcek J: Tumor necrosis factor and interleukin 1: Cytokines with multiple overlapping biological activities. Lab Invest 56:234, 1987. 25. Hogan MJ, Kymura SJ, and Thygeson P: Signs and symptoms of uveitis. I. Anterior uveitis. Am J Ophthalmol 47:155, 1959. 26. Lowry OH, Rosebrough NJ, Fair AL, and Randall RJ: Protein measurement with the Folin phenol reagent. J Biol Chem 193:265, 1951. 27. Bradley PP, Priebat DA, Christensen RD, and Rothstein G: Measurement of cutaneous inflammation: Estimation of neutrophil content with an enzyme marker. J Invest Dermatol 78:206, 1982. 28. Fleisher LN, Ferrell JB, Smith MG, and McGahan MC: Lipid mediators of tumor necrosis factor-a-induced uveitis. Invest Ophthalmol Vis Sci 32:2293, 1991. 29. Dinarello CA, Okusawa S, and Gelfand JA: Interleukin-1 induces a shock-like state in rabbits: Synergism with tumor necrosis factor and the effect of cyclooxygenase inhibition. Prog Clin Biol Res 299:203, 1989. 30. Movat HZ, Burrowes CE, Cybulsky MI, and Dinarello CA: Acute inflammation and a Shwartzman-like reaction induced by interleukin-1 and tumor necrosis factor. Synergistic action of the cytokines in the induction of inflammation and microvascular injury. Am J Pathol 129:463, 1987. 31. Mandrup PT, Bendtzen K, Dinarello CA, and Nerup J: Human tumor necrosis factor potentiates human interleukin 1mediated rat pancreatic beta-cell cytotoxicity. J Immunol 139:4077, 1987. 32. Last BK, Homon CA, Faanes RB, and Merluzzi VJ: Synergistic and overlapping activities of tumor necrosis factor-alpha and IL-1. J Immunol 141:527, 1988. 33. Pukel C, Baquerizo H, and Rabinovitch A: Destruction of rat islet cell monolayers by cytokines. Synergistic interactions of interferon-gamma, tumor necrosis factor, lymphotoxin, and interleukin 1. Diabetes 37:133, 1988. 34. Abe Y, Sekiya S, Yamasita T, and Sendo F: Vascular hyperpermeability induced by tumor necrosis factor and its augmentation by IL-1 and IFN-gamma is inhibited by selective depletion of neutrophils with a monoclonal antibody. J Immunol 145:2902, 1990. 35. Sayers TJ, Wiltrout TA, Bull CA, Denn A3, Pilaro AM, and Lokesh B: Effect of cytokines on polymorphonuclear neutrophil infiltration in the mouse. Prostaglandin- and leukotriene-

36.

37. 38.

39.

40.

41. 42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

2127

independent induction of infiltration by IL-1 and tumor necrosis factor. J Immunol 141:1670, 1988. Leizer T, Cebon J, Layton JE, and Hamilton JA: Cytokine regulation of colony-stimulating factor production in cultured human synovial fibroblasts: I. Induction of GM-CSF and GCSF production by interleukin-1 and tumor necrosis factor. Blood 76:1989, 1990. Hurme M: Both interleukin 1 and tumor necrosis factor enhance thymocyte proliferation. EurJ Immunol 18:1303, 1988. Ogawa H, Nielsen S, and Kawakami M: Cachectin/tumor necrosis factor and interleukin-1 show different modes of combined effect on lipoprotein lipase activity and intracellular lipolysis in 3T3-L1 cells. Biochim Biophys Acta 1003:131, 1989. Tomosugi NI, Cashman SJ, Hay H, Pusey CD, Evans DJ, Shaw A, and Rees AJ: Modulation of antibody-mediated glomerular injury in vivo by bacterial lipopolysaccharide, tumor necrosis factor, and IL-1. J Immunol 142:3083, 1989. Burch RM and Tiffany CW: Tumor necrosis factor causes amplification of arachidonic acid metabolism in response to interleukin 1, bradykinin, and other agonists. J Cell Physiol 141:85, 1989. McGahan MC and Fleisher LN: Cellular response to intravitreal injection of endotoxin and xanthine oxidase in rabbits. Graefes Arch Clin Exp Ophthalmol 1991 (in press). Inamura N, Sone S, Okubo A, Kunishiga E, Nakanishi M, and Ogura T: Tumor cytotoxicity of human monocyte membrane bound interleukin 1 alpha induced by synergistic actions of interferon gamma and synthetic acyltripeptide, FK 565. Cancer Immunol Immunother 28:164, 1989. Shams NBK, Sigel MM, Davis JF, and Ferguson JG: Corneal epithelial cells produce thromboxane in response to interleukin 1 (IL-1). Invest Ophthalmol Vis Sci 27:1543, 1986. Mitchell MD, Rodriquez P, and Feldman ST: The effect of cytokines on prostaglandin production by conjunctival fibroblasts. Invest Ophthalmol Vis Sci 31(suppl):513, 1990. Elner VM, Elner SG, Pavilack MA, Todd R3, Yue BY, and Huber AR: Intercellular adhesion molecule-1 in human corneal endothelium. Modulation and function. Am J Pathol 138:525, 1991. Grossman RM, Kreuger J, Yourish D, Granelli-Piperno A, Murphy DP, May LT, Kupper TS, Sehgal PB, and Gottlieb AB: Interleukin 6 is expressed in high levels in psoriatic skin and stimulates proliferation of cultured human keratinocytes. Proc Natl Acad Sci USA 86:6367, 1989. Jirik FR, Podor TJ, Hirano T, Kishimoto T, Loskutoff DJ, Carson DA, and Lotz M: Bacterial lipopolysaccharide and inflammatory mediators augment IL-6 secretion by human endothelial cells. J Immunol 142:144, 1989. Baggiolini M, Walz A, and Kunkel SL: Neutrophil-activating peptide-1/interleukin 8, a novel cytokine that activates neutrophils. J Clin Invest 84:1045, 1989. Larsen CG, Anderson AO, Appella E, Oppenheim JJ, and Matsushima K: The neutrophil-activating protein (NAP-1) is also • chemotactic for T lymphocytes. Science 243:1464, 1989. Lotz M, Jirik F, Kabouridis P, Tsoukas C, Hirano T, Kishimoto T, and Carson DA: B cell stimulating factor 2/interleukin 6 is a costimulant for thymocytes and T lymphocytes. J Exp Med 167:1253, 1988. Jablons DM, Mule JJ, Mclntosh JK, Sehgal PB, May LT, Huang CM, Rosenberg SA, and Lotze MT: IL-6/IFN-0-2 as a circulating hormone. Induction by cytokine administration in humans. J Immunol 142:1542, 1989.