Experimental staphylococcic keratitis

Michael J. Hogan, Victor Diaz-Bonnet, Masao Okumoto, and Samuel J. Kimura

The human clinical manifestations of staphylococcic infections of the eyelids, conjunctiva, and cornea are discussed. The corneal lesions occurring during the course of staphylococcic in- fections of the lid develop after the disease has been present in the eyelid for several years, but are not always correlated with the degree of activity. The peripheral infiltrations and ulcerations either are due to hypersensitivity or are a result of some product produced by Staphylococcus aureus. Repeated instillations of staphylococcic antigen and toxin onto the corneas of nonimmunized rabbits failed to produce the superficial epithelial keratitis which is so common in human beings with chronic staphylococcal infections of the lid. Injection of a small challenging dose of staphylococcic antigen into the center of the corneas of rabbits, which had been immunized over a period of time with subcutaneous staphylococcic antigen or toxoid, produced corneal rings resembling those in antigen-antibody reactions of the cornea. The incidence of rings was twice as common in those animals immunized with killed and dis- rupted staphylococci as in those immunized with toxoid. The ring infiltrations which were produced appeared within 24 to 48 hours, were complete or incomplete, lasted 3 to 5 days, and gradually disappeared without scarring. A second ring developed in a few animals. Intra- corneal injection of similar doses of toxoid into another series of similarly immunized animals failed to produce ring infiltrations.

T ment of the lids in other cutaneous dis- _Lhe staphylococcus is one of the most eases, such as acne rosacea, is frequently important eye pathogens. Lesions produced associated with secondary staphylococcic by this organism on the lid, conjunctiva, infections. Most often, however, the lid and cornea have been studied for many margins are affected primarily by a staphy- years. lococcic infection which produces bleph- Cutaneous staphylococcic infections af- aritis. Seborrheic blepharitis is another fecting the lids, such as the pyodermas common disease of the eyelid which often which produce folliculitis, furuncles, and becomes secondarily infected with patho- cellulifis often are associated with second- genic staphylococci. The infection in bleph- ary conjunctivitis and keratitis. Involve- aritis may extend to the conjunctival sac to produce a conjunctivitis which is either acute or chronic. Alterations of the con- From the Department of Ophthalmology and the Francis I. Proctor Foundation for Research in junctival inflammation may or may not oc- Ophthalmology, University of California School cur with exacerbations and remissions of of Medicine, San Francisco, Calif. the inflammation of the lid. Since staphy- This investigation was supported in part by the lococci are not epithelial parasites, it has National Council to Combat Blindness Grant- been postulated that they produce kerato- in-Aid No. G-238, and by a Special Fellowship, conjunctivitis through the action of potent BT-666, from The National Institute of Neuro- exotoxins.1 Allen2 showed experimentally logical Diseases and Blindness, United States Public Health Service, Bethesda, Md. that Staphylococcus aureus toxin, when 267

Downloaded from jov.arvojournals.org on 09/28/2021 268 Hogan et al. Investigative Ophthalmology April 1962

dropped into the eye, can result in super- time to time, especially in the chronic vari- ficial punctate erosions of the corneal epi- eties of conjunctivitis. It tends to be more thelium. pronounced on arising in the morning, Except for fine diffuse punctate super- when the symptoms also are worse. During ficial epithelial erosions which stain with the day the erosions heal and symptoms fluorescein, the human cornea rarely is af- diminish. Scrapings of the lid margin fected during all these staphylococcic in- usually show an abundance of gram-posi- fections. Occasionally, during the course of tive cocci, and cultures verify the presence some infections of the lid however, mar- of pathogenic Staphylococcus aureus. Con- ginal corneal infiltrations and ulcerations junctival scrapings show epithelial cells, occur. Most often cultures from the ulcers lymphocytes, polymorphonuclear leuko- show no growth of organisms. It has been cytes, and, at times, staphylococci. postulated that these peripheral corneal Peripheral marginal infiltrations and lesions are due to the effects of staphylo- ulcerations also may appear during the coccic toxin, either as a direct toxic effect, quiet or active phase of the blepharitis. 3 J or to hypersensitivity. Woods and Burky Thygeson7 found Staphylococcus aureus on demonstrated that hypersensitivity to the lids in 133 of 156 catarrhal ulcers as- staphylococcic toxin may produce conjunc- sociated with chronic catarrhal con- tivitis. Very little, however, has been done junctivitis, and indicated the importance to explain the pathogenesis of the corneal of conjunctivitis in the production of the lesions in human beings. Germuth and co- corneal lesions. Single or multiple gray, 5 workers showed that an opaque ring could rounded, or crescentic lesions may develop be produced in the peripheral cornea of the in the peripheral corneal stroma, separated rabbit with bovine serum globulin and from the limbus by a narrow rim of normal albumin. This demonstrates that immuno- corneal tissue (Fig. 1). Any portion of the logic phenomena can be produced in the peripheral cornea may be affected, and at corneal stroma of experimental animals. times the multiple lesions extend to form 0 The Wessely phenomenon in experi- partial ring infiltrates. Only occasionally a mental animals also provides evidence that complete ring infiltration occurs around the hypersensitivity plays a part in the de- corneal periphery.7 The infiltration is 1 to velopment of experimental infiltrations of 2 mm. in breadth, and rarely extends the peripheral cornea. deeper than the midstroma. Ulceration is This study is one of a series designed to common (catarrhal ulcer). Recurrences are develop an experimental model in rabbits common over a period of years. Vasculari- to gain further information on the patho- zation occurs from the limbus. Cultures genesis of staphylococcic keratitis. from the ulcers are negative. Cytologic studies show acute inflammatory cells and Clinical manifestations necrotic corneal debris. There is a rapid It is known that corneal lesions are un- response of the lesions to corticosteroid common in children. It also is known that therapy. This suggests either that the in cases of staphylococcic blepharitis the lid lesions are due to some type of antigen- inflammation may be mild at the time the antibody reaction or that they are toxic and keratitis develops. are benefited by the anti-inflammatory The epithelial keratitis most often pro- corticosteroid effect. duces mild symptoms, such as smarting or Similar corneal ulcerations have been re- burning. It is characterized by punctate ported rarely with Moraxella lacunata and staining, which involves the lower half of by Hemophilus aegyptius, but in these con- the cornea, but occasionally it is more ditions the organism can be recovered from diffuse, even involving the upper limbal the conjunctiva. area. The staining varies in intensity from Thygeson7 reported 14 cases of ring in-

Downloaded from jov.arvojournals.org on 09/28/2021 Volume 1 Experimental staphylococcic keratitis 269 Number 2

Fig, 1. Marginal infiltrate secondary to staphylococcus.

filtration and ulceration of the cornea in sterile. It was placed in a sterile container and his series. The flora of the conjunctiva and refrigerated at 4° C. Toxoid. A commercial preparation of toxoid * lid margin was normal in 12 cases, and in 2 containing 10,000 units per milliliter was used. a coagulase-positive staphylococcus was Toxin. Crude staphylococcus toxinf was used found. It is known that similar lesions can (Lot No. 42925-199), and was preserved with also occur during the course of bacillary thimerosal 0.01 per cent. dysentery, periarteritis, and ulcerative Rabbits. Black Dutch rabbits of medium size colitis. were purchased from various animal suppliers. The eyes were examined with a slit lamp prior to carrying out the experimental work. All had Materials normal corneas. Antigen. A strain of mannitol- and coagulase- positive, hemolytic Staphylococcus aureus, isolated Methods from the lid of a patient with an external hordeolom was used. The organism was inocu- Immunization. Immunization with Staphylococ- lated on nutrient agar, incubated for 24 hours at cus aureus was carried out as follows: A partially 37° C., and harvested by scraping the colonies heat-killed suspension was mixed with an equal and suspending the organisms in saline solution. volume of Freund's complete adjuvant, f One milli- The resulting suspension showed a plate count of liter of this mixture was given subcutaneously 1010 live organisms per milliliter. A portion of each week for a total of 6 weeks. Agglutinin titers this antigen was heated for 2 hours at 60° C. for were determined at the end of 6 weeks, and were immunization of some rabbits. It was found that positive in all animals at a titer of 1:32 or more. heating the suspension for 2 hours at 60° C. failed Immunization of another series of rabbits was to kill all the organisms. Therefore, the freshly carried out by giving 1.0 ml. of toxoid and prepared suspension was placed in a sonic oscil- Freund's adjuvant subcutaneously each week for lator* for 2 hours at 0° to -3° C. The frequency 6 weeks. The presence of precipitin antibody to was 10 kilocycles per second at 250 watts. This toxin was demonstrated by use of the agar dif- failed to kill all organisms. Therefore, the sus- fusion technique. The results showed strongly pension was centrifuged at 1,000 r.p.m. for 10 positive bands with all sera. minutes. The supernatant fluid was then passed Challenge. The immunized and normal control through a millipore filter (H.A. 0.47 micron pore rabbits were challenged either by local instillation size). The filtrate was cultured and found to be *Lederle Laboratories, Penr] River, N. Y, "Raytheon Model DF 101. fDifco Laboratories, Detroit, Mich.

Downloaded from jov.arvojournals.org on 09/28/2021 270 Hogan et al. Investigative Ophthalmology April 1962

or intracorneal injection of the antigen and its 6 corneas with 0.01 per cent thimerosal, and 6 corresponding control fluid. corneas with sterile normal saline. Local instillation. One drop of sonically dis- (3) Twenty normal nonimmunized rabbit eyes rupted and filtered staphylococcus antigen was were also injected with each of the challenge applied to the right cornea of 3 of the staphylo- fluids; 4 eyes were used for each fluid. coccus-immunized rabbits from a 1 ml. Luer All animals were, subjected to careful pre- tuberculin syringe every 5 minutes for 10 times. challenge examination by ordinary illumination The left eyes of these rabbits were treated simi- and with a slit lamp. Subsequent examinations larly with sterile normal saline. Three nonimmu- were carried out immediately after injection and nized rabbits were given the same antigen in the daily thereafter. same way to the right eye and sterile normal saline solution was instilled into the left eye. Results Three of the toxoid-immunized rabbits were chal- lenged in the right eye by instillation of the Local instillation. The conjunctival in- crude staphylococcal toxin. As a control, a solu- stillation of staphylococcic antigen in both tion of 0.01 per cent thimerosal (the preservative immunized and nonimmunized rabbits pro- for the toxin) was instilled in the left eyes. At the duced a very mild conjunctivitis for several same time 3 nonimmunized rabbits received drops hours after the drops. Corneal punctate of toxin in the right eye and thimerosal in the left eye in the same manner. staining failed to occur in any of the Intracorneal injection. (1) Thirty-one eyes suit- animals. The instillation of crude staphy- able for challenge were selected from the group lococcic toxin in toxoid-immune and normal of rabbits immunized with the staphylococcic rabbits failed to produce any corneal antigen. A 1 ml. tuberculin syringe with a 30- changes. gauge needle was used to make an injection of 0.01 to 0.02 ml. in the center of the corneal Intracorneal instillation (Table I). stroma. Twenty-four corneas were challenged with Staphylococcus-immunized rabbits. Thir- the staphylococcal antigen, 4 with undiluted crude teen of the 31 eyes challenged by intra- toxin, and 3 with normal saline. corneal injection developed a ringlike (2) Similarly, 68 rabbit eyes immunized with corneal opacity which was thought to be toxoid were challenged by intracorneal injection an antigen-antibody reaction (Fig. 2). The as follows: 23 corneas with staphylococcic antigen, 12 corneas with commercial staphylococcic toxoid, 13 eyes which developed a ring were all in 21 corneas with varying dilutions of crude toxin, the group challenged with staphylococcic

Table I Challenged with Staph. Control Toxin Toxoid filtrate Saline fluid Ring Uveitis Staphylococcus- immunized 0 4 (withkeratitis) (31 eyes) 24 13 23 0 0

oxoi d-immu ni zed 21' 21 (68 eyes) 12 11 23 21 0 5

Normal 0 4(3 with keratitis) (20 eyes) 0 0 0 1 0 0 0 0 °Of the 21 eyes 3 received undiluted toxin; 3, a 1:1 solution; 8, a 1:4 solution; and 7 a 1:8 solution.

Downloaded from jov.arvojournals.org on 09/28/2021 Volume 1 Experimental staphylococcic keratitis 271 Number 2

Fig. 2. Ring infiltrate in toxoid-immunized rabbit challenged with staphylococcal antigen.

antigen. All 24 which received this antigen peated instillations of staphylococcic anti- had a coincident marked uveitis with fibrin gen as well as toxin is puzzling. Allen2 was deposits, aqueous flare, and cells. The able to produce conjunctivitis and an epi- saline-injected corneas cleared within a thelial keratitis by the repeated local in- few hours after the injection and remained stillation of toxin into the conjunctival sac clear. The 4 eyes challenged with crude of various animals and human beings. We toxin developed marked uveitis, keratitis, are repeating this experiment in animals and corneal necrosis at the site of injection. with toxins from known strains of patho- Toxoid-immunized rabbits. Six of the 68 genic staphylococci. It is possible our re- eyes challenged intracorneally developed a sults differ from those of Allen2 because ring" peripheral to the injection site. The 6 of the method of preparation of the antigen eyes were in the group of rabbits chal- and the toxin. lenged with staphylococcic antigen. All At the present stage it is impossible to rabbits, however, developed a marked correlate the observations in this series of uveitis, except for those challenged with experimental animals with the type of normal saline and 0.01 per cent thimerosal. lesion seen in human eyes. A number of Normal rabbits. None of the 20 eyes differences are noted, principally in the used developed a ringlike reaction. Five development of the rings, their location, out of 20 eyes developed a uveitis. Four and method of fading. In the human eye eyes injected with toxin had corneal the peripheral cornea is believed to be the necrosis and a marked uveitis. One eye site of the antigen-antibody reaction. Ex- challenged with staphylococcal antigen de- cept for circumferential spread, lesions in veloped a very mild uveitis. human beings do not migrate once they are fully developed, whereas in the rabbit Discussion the corneal ring varies in position. Most The failure of the nonimmunized rabbits often a single ring forms, and, as successive to develop epithelial keratitis after re- rings are formed, they seem to migrate

Downloaded from jov.arvojournals.org on 09/28/2021 272 Hogan et al. Investigative Ophthalmology April 1962

toward the central cornea before fading immunized animals it produced uveal in- completely. flammation. The rings observed in the corneas were Microscopic examination of the anterior either complete or incomplete, appeared portion of the eyes of the affected rabbits within 24 to 48 hours, were visible for 3 to showed the cornea to be markedly edema- 8 days, and varied in breadth and density. tous and thickened. There was a diffuse Corneal scarring did not result at their sites infiltration of polymorphonuclear leuko- of appearance. Some scarring nearly always cytes. These cells formed heavy aggrega- occurred at the site of injection. tions at the site of ring formation. The It is apparent that only the intracorneally anterior chamber contained a considerable injected staphylococcic antigen produced number of acute and chronic inflammatory corneal rings in both the staphylococcus as cells and the iris also showed considerable well as the toxoid-immunized animals. acute inflammation. Toxoid is a crude antigen containing modi- fied toxin as well as other products of REFERENCES staphylococcic cell bodies. It is demon- 1. Morax, V., and Elmassian, M.: The role of strated by this study that immunization toxins in the production of inflammation of the conjunctiva, Ann. d'ocul. 122: 81, 1899. can be carried out with toxoid. The failure 2. Allen, J. H.: Staphylococcic conjunctivitis. Ex- of intracorneally injected toxoid to produce perimental reproduction with staphylococcus rings in immunized animals may be due toxin, Am. J. Ophth. 20: 1025, 1937. to the small amount of antigen injected. As 3. Woods, A. C: Clinical problem of allergy in evidence of the greater effectiveness of the relation to conjunctivitis and iritis, Arch. Ophth. staphylococcic bacterial antigen, corneal 17: 1, 1937. 4. Burky, E. L.: Staphylococcus toxin and anti- rings developed about twice as frequently toxin, Internat. Clin. 3: 258, 1936. in the staphylococcus-immunized group as 5. Germuth, F. G., Maumenee, A. E., Pratt-John- in the toxoid-immunized group. son, J., Senterfit, L. B., Van Arnam, C. E., and The keratitis in the affected animals was Pollack, A. C: Observations on the site and mechanisms of antigen-antibody interaction in accompanied by edema, haziness, and anaphylactic hypersensitivity, Am. J. Ophth. 46: thickening of the stroma. It lasted for (Pt. 2) 282, 1958. several weeks, then healed. The uveitis 6. Wessely, K.: Uber anaphylaktische Erschein- produced by toxin injected intracorneally ungen an der Hornhaut (Experimentelle was interesting in that it ran a violent Erzeugung einer parenchymatosen Keratitis durch artfremdes Serum), Miinchen. med. course, at times with an associated Wchnschr. 58: 1713, 1911. hypopyon. Toxoid given to nonimmunized 7. Thygeson, P.: Marginal corneal infiltrates and animals produced no uveitis, whereas in ulcers, Tr. Am. Acad. Ophth. 51: 198, 1946.

Downloaded from jov.arvojournals.org on 09/28/2021