cytokine expression in a rat model of staphylococcus aureus...

IOVS, December 1998, Vol. 39, No. 13 Reports 2785

2. Eskridge JB, Perrigin DM, Leach NE. The Hirschberg test: correla-tion with corneal radius and axial length. Optom Vis Sci. 1990;67:243-247.

3. Miller JM, Mellinger M, Greivenkemp J, Simons K. VideographicHirschberg measurement of simulated strabismus. Invest Ophthal-mol Vis Sci. 1993:34:3220-3229.

4. Riddell PM, Hainline L, Abramov I. Calibration of the Hirschberg testin human infants. Invest Ophtbalmol Vis Sci. 1994;35:538-543.

5. Hasebe S, Ohtsuki H, Tadokoro Y, Okano M, Furuse T. The reli-ability of a video-enhanced Hirschberg test under clinical condi-tions. Invest Ophthalmol Vis Sci. 1995;36:2678-2685.

6. Zadnik K, Mutti DO, Adams AJ. The repeatability of measurementof the ocular components. Invest Ophthalmol Vis Sci. 1992;33:2325-2333.

7. Gordon RA, Donzis PB. Refractive development of the human eye.Arch Ophthalmol. 1985; 103:785-789-

8. Fontana ST, Brubaker RF. Volume and depth of the anterior cham-ber in the normal aging human eye. Arch Ophthalmol. 1980;98:1803-1808.

9. Inagaki Y. The rapid change of corneal curvature in the neonatalperiod and infancy. Arch Ophthalmol. 1986;104:1026-1027.

10. LarsenJS. The sagittal growth of the eye: ultrasonic measurementof the axial diameter of the lens and the anterior segment frombirth to puberty. Ada Ophthalmol (Copenh). 1971 ;49:427-440.

11. Saunders RA, Andrews CJ. Refractive changes in children undergeneral anesthesia. / Pediatr Ophthalmol Strabismus. 1981; 18:38-41.

12. Shun PJT, Ko L, Ng C, Lin S. A biometric study of ocular changesduring accommodation. AmJ Ophthalmol. 1993;115:76-81.

13. Kornbleuth W, Aladjemoff L, Magora F. Influence of general anes-thesia on intraocular pressure in man. Arch Ophthalmol. 1959;6l:84-87.

Cytokine Expression in a RatModel of Staphylococcus aureusEndophthalmitisMichael J. Giese,1'2 Humphrey L Sumner,1

Judith A. Berliner,2 and Bartly J. Mondino1

PURPOSE. TO examine the ability of viable Staphylococcusaureus to induce the production of tumor necrosis factor(TNF>a, interleukin (IL)-l/3, cytokine-induced neutrophilchemoattractant (CINC), and interferon (IFN)-y after intra-vitreal injection.

METHODS. Experimental rat eyes were injected with a 25-jalvolume of approximately 80 colony-forming units of via-ble S. aureus; control eyes received sterile saline. Eyeswere graded daily for signs of clinical inflammation andwere removed 6, 24, 48, and 72 hours after injection. Onegroup was prepared for histologic analysis, and vitreouswas removed from the other group for cytokine analysis,using standard enzyme-linked immunosorbent assay pro-cedures.

RESULTS. TNF-a, IL-1/3, CINC, and IFN-y were detected inexperimental vitreous samples at increased levels thatpeaked at 24 hours. TNF-a, IL-1/3, and CINC declined at 48hours, but IFN-y remained elevated. At 72 hours, levelsreturned to baseline. Statistically significant elevations ofTNF-a, IL-1/3, and CINC were detected in experimentalsamples at 24, but not at 6 and 48 hours compared with

From the 'Jules Stein Eye Institute and Department of Ophthal-mology, and the 2Department of Pathology and Laboratory Medicine,University of California Los Angeles School of Medicine.

Supported in part by Grant EYO4606 from the National EyeInstitute (BJM), Bethesda, Maryland; a Senior Scientific InvestigatorAward (BJM) from Research to Prevent Blindness, New York, NewYork; and the Wasserman Fund, UCLA, Los Angeles, California. MJG isa 1996-1997 American Optometric Foundation Ezell Fellow.

Submitted for publication February 17, 1998; revised April 21 andJuly 8, 1998; accepted August 11, 1998.

Proprietary interest category: N.Reprint requests: Michael J. Giese, Jules Stein Eye Institute, 100

Stein Plaza, DSERC 3-143, UCLA School of Medicine, Los Angeles, CA90095-7000.

levels in saline control samples (P < 0.03). A statisticallysignificant increase in IFN-y was detected at 24 and 48hours compared with control levels (P < 0.03). In exper-imental animals, clinical inflammation and inflammatorycells peaked at 24 hours, persisted at 48 hours, and beganto decline thereafter. Neutrophils were the predominantinflammatory cell detected at 24 (72.3% of cells) and 48(60.1%) hours. By 72 hours, the total number of inflam-matory cells had decreased by 75.0%, and the cellularinfiltrate had changed so that neutrophils equaled mono-cytes-macrophages.

CONCLUSIONS. S. aureus induced the expression of TNF-a,IL-1/3, CINC, and IFN-y. The time course of these cytokinelevels could account for the clinical inflammatory re-sponses and the entry and decline of vitreous cells in thismodel of bacterial endophthalmitis. (Invest OphthalmolVis Sci. 1998;39:2785-2790)

S taphylococcus aureus is a common cause of postoperativeendophthalmitis1 and comprises 99% of the bacterial iso-

lates reported in the Endophthalmitis Vitrectomy Study.2 En-dophthalmitis produced by S. aureus, a Gram-positive bacte-rium, is associated with early onset and a poor visualoutcome.1'3 Direct tissue damage by the bacteria and the hostinflammatory response are responsible for the tissue destruc-tion associated with endophthalmitis. Cytokines are importanthost mediators of the inflammatory-immune response, andactivation of these responses in the immediate postinfectionperiod is critical in eliminating infectious agents and in decreas-ing tissue damage. Whole fixed and heat-killed staphylococciand bacterial components (e.g., protein A, lipoteichoic acid,peptidoglycan, and a-toxin) from a variety of Gram-positivebacteria, including S. aureus, have been shown to induce thein vitro production of interleukin (IL)-1> -4, -6, and -8; tumornecrosis factor (TNF)-a, and interferon GFN)-y from mononu-clear cells, lymphocytes, endothelial cells, and neutrophils.45

Therefore, these mediators may be involved in the early phasesof endophthalmitis. A principal goal of these studies was togain knowledge about the local, early host immune response toS. aureus infections of the eye. In particular, we examined theability of viable S. aureus to induce the production of vitreouscytokines in an established rat model of endophthalmitis.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933204/ on 06/12/2018

2786 Reports IOVS, December 1998, Vol. 39, No. 13

Cytokines can be generally grouped into four categories orphases of inflammatory activity: recognition, recruitment, re-moval, and repair.6 The principal cytokines produced early andrapidly in response to infection and involved in recognition areTNF-a and IL-1. Tumor necrosis factor has been implicated inthe pathogenesis of experimental and clinical uveitis, cornealneovascularization, proliferative vitreoretinopathy, and centralnervous system demyelination of the type found in optic neu-ritis.7 Interleukin-1 seems to be involved in retinal inflamma-tory diseases and has been implicated in the recruitment ofleukocytes into the retina and in the disruption of the ocularvascular barriers.8 In addition, IL-1 activity has been detected inthe vitreous of patients with diabetic retinopathy.9 Cytokinesassociated with the recruitment phase of the inflammatoryresponse (called chemokines) are responsible for mobilizationof leukocytes into sites of inflammation. Cytokine-induced neu-trophil chemoattractant (CINC) is a rat chemokine with func-tions equivalent to those of human IL-8 that has been identifiedin the aqueous humor of rats with endotoxin-induced uveitis.10

CINC production can be induced by TNF and IL-1, which alsoinduce IL-8 production. Cytokines associated with the removalphase comprise a group of proteins that activate and inducethe differentiation of other cells that enhance the clearance ofantigens or infectious agents. These proteins include IFN-y andIL-2. Interferon-y may also function as a proinfiammatory cyto-kine and has been detected in ocular tissues from patients withsympathetic ophthalmia or uveitis and in the vitreous of pa-tients with acquired immune deficiency syndrome who havevitritis and retinitis.1'

Few investigators have evaluated the intraocular cytokineresponse to Gram-positive bacteria. Successful treatment ofbacterial endophthalmitis requires not only the elimination ofthe infecting organism and its components, but also control ofthe host immune response. In the present study, we haveexpanded our investigations in experimental endophthalmitisby showing the presence of TNF-a, IL-1/3, CINC, and IFN-y indie vitreous of eyes injected with viable 5. aureus.

described protocol.12 Right eyes of experimental and controlrats received 25-juJ intravitreal injections containing 80 colony-forming units per milliliter and NS, respectively. Twenty-fivemicroliters of the same S. aureus suspension was plated ontofive blood agar plates to confirm the approximate number ofcolony-forming units of S. aureus injected.

Clinical Inflammatory Scores

Eyes of all rats were examined daily with a slit lamp biomicro-scope and direct ophthalmoscope and before death at 6, 24,48, and 72 hours after treatment. Clinical inflammation wasgraded according to a scoring system developed by Peyman etal.13 and Chan et al.14

Cytokine Detection

Vitreous was removed from eyes of 134 rats at 6, 24, 48, and 72hours after injection of NS or viable S. aureus. Sera wereobtained according to a previously described protocol.12 Righteyes were removed and immediately frozen on dry ice. Frozenglobes were bisected posterior to the iris, the lens was liftedout, and the vitreous was removed and placed in conical tubeson ice. Vitreous samples from three eyes were pooled forexperimental and control groups. Pooled samples were spun atl6,000g for 30 minutes at 4°C. The supernatant was removedand stored at — 70°C until assayed for cytokines.

Samples were evaluated using commercially purchasedstandard sandwich enzyme-linked immunosorbent assays(ELISAs; Biosource, Camarillo, CA) for the detection of TNF-a(sensitivity 4 pg/ml), IL-1/3 (sensitivity 3 pg/ml), and IFN-y(sensitivity 13 pg/ml). A vitreous ELISA was developed todetect CINC and modified from a previously described proto-col (sensitivity 50 pg/ml).10

HistopathologyThirty-nine rats were killed at 6, 24, 48, and 72 hours and theireyes prepared according to standard methods for histologic

MATERIALS AND METHODS

Bacterial Strains and Components

In this study, a prototypic strain of S. aureus was used toinduce endophthalmitis; it was a gift from Ambrose Cheung,Rockefeller University, New York, New York. This strain main-tains its hemolytic pattern when propagated on sheep eryth-rocyte agar, and produces a-, /3- and 5-hemoIysins; lipases; andfibronectin-binding protein. The isolate used in these studieswas maintained in sheep red blood at —70°C. When ready foruse, bacteria were plated onto rabbit or sheep blood agarplates. An isolated colony was subcultured into sterile trypticsoy broth. Bacteria were centrifuged and washed with Hpopo-lysaccharide-free 0.85% sterile normal saline (NS). Dilutions ofbacteria were made with NS, according to a previously de-scribed method.12

Intravitreal Injections

Inbred female Lewis rats (Harlan Sprague-Dawley, San Diego,CA) raised in a pathogen-free environment were used in thisstudy. Studies were conducted in accordance with the ARVOStatement for the Use of Animals in Ophthalmic and VisionResearch. Injections were performed according to a previously

FIGURE 1. Diagram defining the 4 fields in the vitreous cham-ber used to quantitate inflammatory cells. Anterior chamber: 1,2, 3, 4.



analysis. To quantify infiltrating cells, we counted cells in thevitreous in three random 5-ju<m sections per eye. One sectionwas located superior to the optic nerve, the second containedthe optic nerve, and the third was inferior to the optic nerve.Thus, we focused on the areas where cells enter into the eye.Neutrophils, lymphocytes, and monocytes-macrophages werecounted at 400 X in four denned high-powered fields in thevitreous chamber, using an eyepiece measuring grid. Cellsfrom 12 fields (4 per section) per eye were totaled and used tocalculate the percentage of each cell type per eye (Fig. 1).None of the eyes used in the analysis contained large vitreousor subretinal hemorrhages or showed crystalline lens damage.

Statistics

We compared mean clinical scores by mixed-model repeatedmeasures analysis of variance (ANOVA). This model correctlytook into consideration the within and between animal vari-ability. The Fisher-Tukey least significant difference criterionwas used to compute post hoc t tests in this model. We used anonparametric Kruskal-Wallis one-way analysis of variancewith post hoc P computed by Fisher-Tukey's method formultiple comparisons to compare mean total cell number,mean percentage of inflammatory cells, and mean cytokineconcentration. Tabulated data are presented as mean ± SEM.Significance was determined at P < 0.05.

RESULTS

Clinical Inflammation

Six hours after injection, 14 (93.3%) of 15 rats in the experi-mental group and 15 (83-3%) of 18 rats in the control group didnot show signs of inflammation (Fig. 2). Mild iris hyperemia inthe remaining rats of both groups accounted for the lowinflammatory scores.

After 24 hours, 41 (91.1%) of 45 animals in the experi-mental group showed signs of endophthalmitis. The conjunc-tiva and cornea remained normal in this group; the mostmarked inflammatory changes were noted in the iris and theanterior and vitreous chambers. Mildly to markedly dilated irisvessels (mean, 1.40 ± 0.11) were seen on ophthalmoscopicexamination. Experimental animals also showed a grade 1 to 2anterior chamber response (mean, 1.10 ± 0.12) manifested byanterior chamber cells and fibrin along with posterior syn-echia. Thirty-four (75.6%) of 45 eyes contained a vitreousexudate. A statistically significant increase in the inflammatoryresponse was present in the experimental rats in comparisonwith that in the control rats (P < 0.0001).

At 48 hours, the clinical inflammatory scores in experi-mental eyes had slightly declined. There was a decrease in theanterior chamber response (mean, 0.76 ± 0.12) but the irisfindings (mean, 1.40 ± 0.13), and vitreous exudates remainedapproximately the same. A statistically significant difference ininflammation between experimental and control animals wasstill present (P < 0.0001).

At 72 hours, a marked decline in clinical inflammatoryfindings in experimental (P < 0.0001) and control groups fromthe value at 48 hours was noted. In the experimental group,most of the decrease in inflammatory scores was accounted forby a decline in the anterior chamber response. Only 1 (6.7%) of15 animals showed anterior chamber inflammation. Vitreousexudates were still present in 11 (73.3%) of 15 rats, but a

4.5 -i

0.0

48 72

Hours after Injection

FIGURE 2. Mean clinical inflammatory scores of rats injectedwith Staphylococcus aureus or saline.

reduction in size was noted. The inflammatory response in theexperimental group was still significantly greater than that seenin the control group (P < 0.0001).

Cytokines

At 6 hours, TNF-a was detected in experimental and saline con-trol vitreous samples (Fig. 3A), but not in vitreous samples fromuninjected eyes. At 24 hours, a greater than 10-fold increase inTNF-a was detected in the experimental group compared withthat in the control group (P < 0.03). This represented a greaterthan fourfold increase compared widi the 6-hour level and agreater than 600-fold increase compared with die uninjectedlevel. At 48 hours, die TNF-a level declined from die 24-hour peakby 81.8% in the experimental group but still was significantlyelevated when compared widi that in the control group (P <0.01). When experimental and control samples were compared at72 hours, no significant difference in the TNF-a levels was ob-served. No TNF-a was detected in die sera of experimental andcontrol rats at any of the time points.

Interleukin-1/3 was found in all experimental vitreous sam-ples at 6, 24, 48, and 72 hours (Fig. 3B). At 24 hours, experi-mental levels of IL-1/3 were more than 4000 times higher thancontrol levels (P < 0.03). The level of IL-1/3 in the experimentalgroup peaked at 24 hours, decreased by approximately 78.5%at 48 hours, and continued to decrease at 72 hours. Interleu-kin-1/3 was below detectable limits in saline-injected eyes andwas not detected in sera of experimental and control animals atany time point.

CINC was detected in experimental vitreous samples at 6,24, 48, and 72 hours, whereas lower levels were detected insaline-injected control samples (<912.0 pg/ml). No CINC was



C = Saline ControlE = Experimental

6E 24C 24E 48C 48 E 72C 72E 6E 24C 24E 48C 48E 72C 72E

Hours after InjectionFIGURE 3. Concentration of TNF-a (A), IL-1/3 (B), cytokine-induced neutrophil chemoattractant (CINC) (C), and IFN-y (D) in thevitreous of rats injected with viable Staphylococcus aureus or saline at 6, 24, 48, and 72 hours. For S. aureus-in]ecte<\ experimentaleyes (E), n = 5; for saline-injected control eyes (C) and for cytokines TNF-a, IL-1/3, and IFN-y, n = 3, except for CINC, for whichn = 4. Each n is composed of three pooled vitreous samples.

detected in uninjected vitreous samples (Fig. 3C). At 6 hours,a greater than 2500-fold increase was detected in comparisonwith levels in uninjected samples, whereas a 2.75-fold increasewas recorded compared with the level in control samples. At24 hours, a statistically significant increase was seen in exper-imental samples compared with the level in control samples(P < 0.02). The level of CINC peaked at 24 hours, decreased by80.0% at 48 hours, and approached control levels by 72 hours.No concentrations of CINC were detected in the sera of ex-perimental and control rats at any of the time points.

Interferon-y was found in experimental vitreous samplesat 24, 48, and 72 hours (Fig. 3D). Experimental vitreous IFN-ylevels were significantly greater at 24 (P < 0.03) and 48 (P <0.01) hours when compared with those in control vitreoussamples. The level of IFN-y peaked at 24 hours, decreasedslightly at 48 hours, and returned to baseline at 72 hours. Astatistically significant decrease in level was noted between the48-hour and 72-hour time points (P < 0.01). No IFN-y wasdetected in the vitreous of saline-injected eyes or in the sera ofsaline- or 5. aureus-injected animals at any of the time points.

Histopathology

Essentially, no inflammatory cells were observed in the vitre-ous of experimental eyes at 6 hours and in control eyes at all

time points. At 24 hours, the mean total cellular response(31 ± 5) in the vitreous chamber of experimental eyes signif-icantly increased from the 6-hour time point (P < 0.005; Fig.4). The primary infiltrating cells were neutrophils (72.3%)followed by monocytes-macrophages (12.7%) and lymphocytes(2.3%). The clinically visible vitreous exudate in experimentaleyes was primarily composed of neutrophils. At 48 hours, adecrease in total cell number (25 ± 2) was noted comparedwith that at the 24-hour time point in the experimental group.At this time, neutrophils comprised the primary cell population(60.1%) in experimental vitreous. The monocyte and lympho-cyte populations increased as a percentage of total cells fromthe 24-hour time point. At 72 hours, the total number ofinflammatory cells decreased (19 ± 3) from that recorded atthe 48-hour time point. At this time, a 1.5-fold decrease in theneutrophil percentage was noted, whereas the monocyte-mac-rophage and lymphocyte populations increased as a percent-age of total cells. No plasma cells were detected at any of thetime points.

Inflammatory cells were seen around the ciliary body andtracking posteriorly into the vitreous chamber and around theoptic nerve head. Most experimental eyes and a few controleyes showed breakdown of the blood-retinal barrier mani-fested as a proteinaceous exudate seen in the anterior or



(IA13S

ls(±

itory

Cel

it In

flam

ma

Mea

n P

erce

t

1 UU /0 —

90% -

80% -

70% -

60% -

50% -

40% -

30% -

20% -

10% -

0% -

—•— Neutrophils—•— Monocytes/Macrophages••••*-• Lymphocytes

I

/

/

/

i

Aj -

• i

5

1 • ' i

Mean TotalCell Number

0.00

24 48

Hours After Injection

31±5 25±2

72

19±3

FIGURE 4. Mean percentage of inflammatory cells per eye inthe vitreous of experimental eyes at 6 (n = 6), 24 (n = 7), 48(n = 6), and 72 (n = 5) hours after injection of Staphylococcusaureus and mean total cells per eye listed below graph. Per-centage of inflammatory cells do not add up to 100.0% becausesome cells were unidentifiable.

vitreous chambers or in both. Clinical inflammatory findings inanimals used for histologic analysis at 6, 24, 48, and 72 hourswere similar to the inflammatory scores in the animals used forcytokine ELISAs (data not presented).

DISCUSSION

5. aureus is a virulent pathogen that can produce endoph-thalmitis with poor visual outcome.13 Few investigators haveevaluated the host intraocular immune response to 5. aureusand its components. Previous studies have shown that S. au-reus and its components can induce in vitro production ofvarious cytokines.4'5 Injection of purified TNF and IL-1 intorat and rabbit vitreous can elicit ocular inflammatory re-sponses.15"17 However, none of these studies reflect the in-traocular response to viable bacteria in endophthalmitis. Toclarify the early inflammatory response during S. aureus en-dophthalmitis, our study was focused on cytokine involvementelicited by intraocular injection of live bacteria.

Our results show that intravitreal injection of viable S.aureus induced the production of vitreous TNF-a, IL-1/3, CINC,and IFN-y and clinical signs of endophthalmitis. In the exper-imental group, cytokine levels peaked at 24 hours as did valuesin clinical inflammatory findings and total vitreous cell num-bers. The elevated levels of vitreous TNF-a, IL-1/3, and CINC at6 hours and the lack of systemic levels in experimental samplessuggest that their production is local in response to bacteriaand/or components. Potential sources include vitreous hyalo-cytes18; endothelial cells of retinal, iris, and ciliary body ves-

sels; and resident macrophages. At 6 hours, a small but insig-nificant increase in TNF-a and CINC was noted in NS controlvitreous samples. This may have been produced as a result ofinjection trauma. The increase in TNF-a, IL-1/3, CINC, andIFN-y at 24 hours may be related to the constant inflammatorystimulus present in the vitreous as a result of bacterial growthand release of cell wall and secreted components.,At this time,inflammatory cells that have entered the eye may augment thecytokine response. 5. aureus has been shown to be a morepotent inducer of TNF-a production from granulocytes19 andmononuclear leukocytes than has lipopolysaccharide.5 In ad-dition, macrophages have been shown to produce TNF-a afterS. aureus exposure.5 The initial source of IFN-y in the vitreousat 24 hours may be from natural killer cells that have beenshown to produce IFN-y in a T-cell-independent manner afterstimulation with 5. aureus}0 Breakdown of the blood-retinalbarrier seen at 24 hours could be partially accounted for byTNF-a's ability to alter tight junctions in endothelial cells in-ducing leakage of plasma proteins and water.21

Eyes infected with S. aureus showed primarily a neutro-phil response, which was expected. Studies have shown thatmice injected with TNF-a produce neutrophil- and monocyte-specific chemokines,22 and in vitro cell culture studies23 haveshown that heat-killed S. aureus induces the expression of IL-6and IL-8 but not the expression of chemokines specific forother inflammatory cells. In addition, injection of purifiedCINC into the anterior chamber of Lewis rats results in theemigration of neutrophils into the aqueous humor.10 Our his-tologic data are similar to those in other in vivo studies inwhich intravitreal injections of TNF and IL-1 induced infiltra-tion of neutrophils.1517 The predominant neutrophil responsecould also be related to an early differential expression ofadhesion molecules that favor neutrophil emigration. At 48hours, TNF-a, IL-1/3, and CINC levels declined rapidly from the24-hour peak, whereas clinical inflammation and IFN-y levelsremained elevated. This prolonged clinical response could becaused by bacterial products (e.g., toxins) and other hostinflammatory mediators (e.g., leukotriene B4) that may bepresent in the vitreous and may be involved in later responses.The cellular response at 48 hours was characterized by adecrease in neutrophils and an increase in lymphocytes andmonocytes-macrophages. Their subsequent activation by 5. au-reus products may account for the 48-hour IFN-y level. Thisresponse was followed at 72 hours by a shift in cell type to apopulation of approximately equal numbers of neutrophils andmonocytes-macrophages. This shift may be partially related toneutrophil apoptosis, decreased production of neutrophil che-motactic agents such as CINC, production of other cytokinesthat change adhesion molecule expression, and production ofother chemotactic agents that favor lymphocyte and mono-cyte-macrophage infiltration. The continued presence of neu-trophils may be related to the ability of TNF-a, IL-1/3, and IFN-yto prolong neutrophil survival.24 The presence of IFN-y at 48hours may enhance antibody formation by B lymphocytes andhelp downregulate aspects of the inflammatory response,whereas the rapid decline in TNF-a and IL-1/3 levels, as notedin our model, would also be beneficial to the host. The cyto-toxic effects of TNF are enhanced in the presence of othercytokines, particularly IFN-y.23 When present together, TNF-aand IL-1/3 induce a more pronounced inflammatory responsethan when present separately.17 Thus, the observed decreasein TNF-a and IL-1/3 at the time of IFN-y increase would enable



repair without toxicity. The decrease in TNF-a levels may berelated to production of soluble TNF receptors, decreasedbiosynthesis, a short half-life of TNF mRNA, TNF protein deg-radation, and the production of antiinflammatory cytokines(e.g., IL-10, IL-12), which decrease TNF-a production in mac-rophages.25 Constant stimulation of monocytes by bacterial orhost products may induce a refractory state inhibiting theproduction of TNF.25 Other cytokines, for example IL-4 andinactivating substances such as those detected in patients in-jected with lipopolysaccharide, may produce inhibitory effectson IL-1/3 production, resulting in a reduced level.25

In summary, we have identified four inflammatory-im-mune cytokines in the vitreous of rat eyes infected with 5.aureus. The ocular inflammatory-immune response to intraoc-ular bacteria is multifactorial. This study provides a more com-plete understanding of cytokine mediators produced in the eyein the period of early response to S. aureus infection. How-ever, important questions remain to be studied. Do additionalproinflammatory cytokines other than the ones detected in thisstudy play a role in 5. aureus endophthalmitis? Are S. aureusfactors responsible for induction of these cytokines? What isthe ocular source of these cytokines? What is the association ofthese cytokines with pathologic changes in the eye? Our re-sults should be helpful in subsequent studies of the role ofinflammatory cytokines in the pathogenesis of endophthalmitisand evaluation of new treatment strategies.

Acknowledgments

The authors thank John Zagorski for kindly providing the anti-CINCantibody; Ben Glasgow and Alissa Riesner for histologic support; Man-uel Munguia, Edison Chiu, and David Shum for assistance with animals;Jeff Gornbein for statistical assistance; and Yoon Cho for technicalassistance.

References

1. Mao LK, Flynn HW Jr, Miller D, Pflugfelder SC. Endophthalmitiscaused by Staphylococcus aureus. Am J Ophthalmol. 1993;ll6:584-589.

2. Han DP, Wisniewski SR, Wilson LA, et al. Spectrum and suscepti-bilities of microbiologic isolates in the endophthalmitis vitrectomystudy. AmJ Ophthalmol. 1996; 122:1-17.

3. Kattan HM, Flynn HW Jr, Pflugfelder SC, Robertson C, ForsterRK. Nosocomial endophthalmitis survey: current incidence ofinfection after intraocular surgery. Ophthalmology. 1991 ;98:227-238.

4. Henderson B, Wilson M. Cytokine induction by bacteria: beyondlipopolysaccharide. Cytokine. 1996;8:269-282.

5. Ferrante A, Staugas REM, Rowan-Kelly B, et al. Production oftumor necrosis factors alpha and beta by human mononuclearleukocytes stimulated with mitogens, bacteria, and malarial para-sites. Infect Immun. 1990;58:3996-4003.

6. Kunkel SL, Lukacs N, Strieter RM. Cytokines and inflammatorydisease. In: Sirica AE, ed. Cellular and Molecular Pathogenesis.Philadelphia: Lippincott-Raven Publishers; 1996:23-38.

7. Cunningham ET, Stalder A, Paolo Sanna P, et al. Localization oftumor necrosis factor receptor messenger RNA in normal and

Herpes simplex virus-infected mouse eyes. Invest Ophthalmol VisSci. 1997;38:9-15.

8. Claudio L, Martiney JA, Brosnan CF. Ultrastructural studies of theblood-retina barrier after exposure to interleukin-1 or tumor ne-crosis factor-a. Lab Invest. 1994;70:850-86l.

9. Abu El Asrar AM, Maimone D, Morse PH, Gregory S, Reder AT.Cytokines in the vitreous of patients with proliferative diabeticretinopathy. Am J Ophthalmol. 1992;114:731-736.

10. Guex-Crosier Y, Wittwer AJ, Roberge FG. Intraocular produc-tion of a cytokine (CINC) responsible for neutrophil infiltrationin endotoxin induced uveitis. Br J Ophthalmol. 1996;80:649-653.

11. De Vos AF, Hoekzema R, Kijlstra A. Cytokines and uveitis, areview. Curr Eye Res. 1992;ll:581-597.

12. Ravindranath RMH, Mondino BJ, Adamu SA, Pitchekian-Halabi H,Hasan SA, Glasgow BJ. Immunopathologic features of Staphylococ-cus aureus endophthalmitis in the rat. Invest Ophthalmol Vis Sci.1995;36:2482-2491.

13- Peyman GA, Paque JT, Meisels HI, Bennett TO. Postoperativeendophthalmitis: a comparison of methods for treatment and pro-phylaxis with gentamicin. Ophthalmic Surg. 1975;6:45-55.

14. Chan CC, Hikita N, Dastgheib K, Whitcup SM, Gery I, NussenblattRB. Experimental melanin-protein-induced uveitis in the Lewisrat. Immunopathologic processes. Ophthalmology. 1994; 101:1275-1280.

15. De Vos AF, van Haren MAC, Verhagen C, Hoekzema R, Kijlstra A.Tumour necrosis factor-induced uveitis in the Lewis rat is associ-ated with intraocular interleukin 6 production. Exp Eye Res. 1995;60:199-207.

16. Bamforth SD, Lightman SL, Greenwood J. Ultrastructural analysisof interleukin-1 j3 induced leukocyte recruitment to the rat retina.Invest Ophthalmol Vis Sci. 1996;38:25-35.

17. Fleisher LN, Ferrell JB, McGahan C. Synergistic uveitic effects oftumor necrosis factor-a and interleukin-1 j3. Invest Ophthalmol VisSci. 1992;33:2120-2127.

18. Fischer BM, Fleisher LN, Tompkins W, Ma J, McGahan MC. Endo-toxin induced tumor necrosis factor like activity from bovinehyalocytes. [ARVO Abstract]. Invest Ophthalmol Vis Sci. 1992;33(4):S1114. Abstract nr 2106.

19. Mandi Y, Endresz V, Krenacs L, Regely K, Degre M, Beladi I. Tumornecrosis factor production by human granulocytes. Int Arch Al-lergy Appl Immunol. 1991;96:102-106.

20. Billiau A. Interferon-y. Biology and role in pathogenesis. Adv Im-munol. 1996;62:6l-130.

21. Bamforth SD, Lightman S, Greenwood J. The effect of TNF-alphaand IL-6 on the permeability of the rat blood-retinal barrier in vivo.Ada Neuropathol. 1996;91:624-632.

22. Tessier PA, Naccache PH, Clark-Lewis I, Gladue RP, Neote KS,McColl SR. Chemokine networks in vivo. Involvement of C-X-Cand C-C chemokines in neutrophil extravasation in vivo in re-sponse to TNF-a. / Immunol. 1997;159:3595-3602.

23. Kinnaert P, De Wilde JP, Bournonville B, Husson C, Salmon I.Direct activation of human peritoneal mesothelial cells by heat-killed microorganisms. Ann Surg. 1996;224:749-755.

24. Collatta F, Re F, Polentarutti N, Sozzani S, Mantovani A. Modulationof granulocyte survival and programmed cell death by cytokinesand bacterial products. Blood. 1992;80:2012-2020.

25. Tracey KJ. Tumor necrosis factor. In: Remick DG, Friedland JS,eds. Cytokines in Health and Disease. 2nd ed. New York: MarcelDekker; 1997:223-238.


cytokine expression in a rat model of staphylococcus aureus...

Documents