Inhibition of Corneal Inflammation by an Acylated Superoxide Dismutase Derivative

Eiko Ando,* Yukio Ando,t Masayasu lnoue,t Yoshimasa Morino,t Ryuji Kamafa,* ond Ryoichi Okomuro"

Superoxide radicals and their metabolite(s) have been postulated to play an important role in the pathogenesis of inflammation. Hence, superoxide dismutase (SOD) has been used to reduce tissue injury caused by reactive oxygens. However, protection of the cornea and other ocular tissues from oxygen toxicity could not be achieved by administering SOD presumably due to its unfavorable in vivo behavior. To scavenge superoxide radicals on the outer surface of corneal epithelial cells, the authors synthesized an acylated SOD derivative (AC-SOD) by linking capric acid. When instilled into rabbit eyes, a significant amount of AC-SOD remained bound to the corneal surface for a fairly long time. Intracorneal injection of lipopolysaccharide (LPS) triggered infiltration of polymorphonuclear leukocytes (PMNs) to the cornea and induced severe keratitis. Topical administration of AC-SOD to the LPS-treated cornea markedly inhibited the infiltration of PMNs and suppressed the occurrence of keratitis. Under identical conditions, topically administered SOD was rapidly removed by tears and, hence, did not inhibit LPS-induced keratitis. When the number of PMNs in the systemic circulation was reduced by intravenous administration of hydroxyurea, LPS-induced keratitis was inhibited markedly. These results indicate that superoxide radicals and circulating PMNs might play a critical role in LPS-induced keratitis. Invest Ophthalmol Vis Sci 31:1963-1967, 1990

Reactive oxygen species play an important role in Materials and Methods the pathogenesis of various ocular diseases, such as cataract,1 uveitis,2"5 and retinopathy of prematurity.6 Materials Keratitis caused by bacterial infection is often asso- Human erythrocyte-type SOD was purified by the 7 8 ciated with corneal abscess, ulcer, and erosion. ' method of Gartner et al." Capric acid was obtained Since a large number of polymorphonuclear leuko- from Wako Pure Chemical, Osaka, Japan. The LPS cytes (PMNs) migrate to the inflammatory cornea, ofSerratia marcescens came from Biological Labora- reactive oxygen species derived from PMNs may in- tory (Campbell). Hydroxyurea was from Sigma, St. crease tissue injury. To test a possible role of reactive Louis, MO. The l25I-labeled Bolton-Hunter reagent oxygen species in corneal inflammation, we tested the (2200 Ci/mmol) was from New England Nuclear, effect of superoxide dismutase (SOD) on endotoxin- Boston, MA. The AC-SOD was synthesized by link- induced keratitis, an animal model for aseptic cor- ing 5 mol of capric acid to the lysyl e-amino groups of 9 neal inflammation. Preliminary experiments re- SOD as described previously.10 Both SOD and AC- vealed that topically instilled SOD was rapidly re- SOD samples were labeled with the 125I-labeled Bol- moved by tears from the corneal surface and, hence, ton-Hunter reagent as described.12 The specific radio- did not have a protective action in situ. activity of the two enzyme samples was 200,000 To stabilize the enzyme on the corneal surface, we cpm/mg protein. synthesized an acylated SOD derivative (AC-SOD) that binds to cell surface membrane via its acyl Animals moiety.10 Topically instilled AC-SOD efficiently bound to the corneal surface and significantly sup- Male Japanese albino rabbits (2-2.5 kg) were fed pressed the lipopolysaccharide (LPS)-induced kera- laboratory chow and water ad libitum. Use of the titis in the rabbit, particularly in its early stages. animals conformed to the ARVO Resolution on the Use of Animals in Research. From the Departments of ""Ophthalmology and tBiochemistry, Kumamoto University Medical School, Kumamoto, Japan. LPS-induced Keratitis of the Rabbit Reprint requests: Dr. M. Inoue, The Department of Biochemis- try, Kumamoto University Medical School, 2-2-1 Honjo, Kuma- Under pentobarbital anesthesia (50 mg/kg), 10 ng moto 860, Japan. of LPS dissolved in 10 jul of saline solution was in-

1963

Downloaded from iovs.arvojournals.org on 09/29/2021 1964 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Ocrober 1990 Vol. 31

Table 1. LPS-induced keratitis Hinson et al.14 On day 5, the number of circulating PMNs decreased to less than 36% of the control Pathologic events Score* group. On day 6, LPS was injected into the cornea as Corneal edema described, and time-dependent changes in the inflam- Slight + 1 mation score (Table 1) were determined. Administra- Mild +2 Severe +3 tion of hydroxyurea was continued for another week Corneal abscess after LPS injection; the number of circulating PMNs Slender ring ( 1/2 circle) +2 Wide ring (Corneal ulcer Results 1 /2 area of corneal abscess +2 Fate of Topically Instilled SOD and AC-SOD Disc +3 To assess the binding of enzyme samples to the * Inflammation score was expressed as the sum of these values. corneal surface, 125I-labeled SOD or AC-SOD was topically instilled onto the intact cornea. As shown in jected into both sides of the central stroma of the Figure 1A, the amount of AC-SOD remained asso- rabbit cornea by using a microsyringe (30-G needle). ciated with the cornea was significantly larger than After injection of LPS, the right side of the cornea was that of SOD. Binding of the two SOD samples was topically instilled with 640 /xg of either SOD or AC- also determined in inflamed corneas pretreated with SOD dissolved in 80 JX\ of saline six times a day (every LPS (Fig. IB). Interestingly, binding of SOD and 4 hr). The left side of the eye received the same vol- ume of saline. At the indicated times after LPS administration, the grade of corneal inflammation was determined according to the inflammation score (Table 1). Microinjection of LPS was done at approx- imately 1:00 PM.

Binding of SOD and AC-SOD to the Cornea 3.0 D) Six days after injection of LPS, 320 ng of 125I-labeled SOD or AC-SOD was topically instilled onto the intact and inflamed corneas. Under pentobarbital anesthesia, the central part of the cornea (10 mm in V) diameter) was excised with a trephine after 30 min of CD instillation. After washing the excised cornea with 3 ml of ice-cold saline, tissue-associated radioactivity was determined in a Packard autogamma-scintilla- tion spectrophotometer model-5130 (IL). Q O CO Thiobarbituric Acid-Reactive Metabolites • in LPS-Injected Cornea I • The LPS-induced keratitis consists of four stages: 0 early, active (ring abscess), late (disciform ulcer and B neovascularization), and scarring stages. At each Fig. 1. Fate of topically instilled enzymes on the corneal surface. stage, the cornea was excised with a trephine (7 mm (A) Under pentobarbital anesthesia (50 mg/kg), l25I-labeled SOD or in diameter) without being contaminated by the area AC-SOD was topically instilled to the cornea of the intact rabbit with neovascularization. Thiobarbituric acid-reactive (160 /xg/cornea). Thirty minutes after instillation, the cornea was excised, washed with 3 ml of ice-cold saline, and studied for tissue- metabolite (TBAR) levels in the inflamed corneas associated radioactivity. (B) Six days after intracorneal injection of 13 were determined by the method of Uchiyama et al. LPS (jug/cornea), radioactive SOD samples were instilled to the inflammatory cornea. The cornea was excised, washed in 3 ml of Depletion of the Circulating PMNs ice-cold saline, and tested for tissue-associated radioactivity. De- tailed conditions are described in the text. The data show the mean Rabbits were intravenously injected with hydrox- ± SEM derived from three eyes. Open columns = SOD-treated yurea (100 mg/kg/day) for 5 days by the method of groups; spotted columns = AC-SOD-treated groups. *P < 0.05.

Downloaded from iovs.arvojournals.org on 09/29/2021 No. 10 INHIDITION OF CORNEAL INJURY DY ACYLATED SOD / Ando er ol 1965

Fig. 2. Effect of SOD and AC-SOD on corneal inflammation. Under pentobarbital anesthesia, 10 ng of LPS was injected into the corneal stroma of the rabbit. After 30 min, 80 ^1 of saline or 640 jig of AC-SOD or SOD was topically instilled to the cornea six times a day. Photographs were taken 6 days after LPS injection. (A, left) saline- treated group; (B, right) AC-SOD-treated group.

AC-SOD to the LPS-injected cornea increased mark- markedly inhibited the increase in TBAR levels of edly compared with that in the intact cornea. Again, LPS-treated cornea (Fig. 5). significantly larger amounts of AC-SOD bound to the injured cornea than did SOD. The amounts of AC- Effect of Depleting PMNs in the Circulation SOD associated with the conjunctiva and sclera were also higher than those of SOD (data not shown). To understand the role of PMNs in the propagation of the inflammatory process occurring in the Effect of AC-SOD on Corneal Inflammation avascular cornea, circulating PMNs were depleted by hydroxyurea. Interestingly, the inflammatory score To examine the effect of SOD samples on corneal markedly decreased in PMN-depleted animals espe- inflammation, the enzyme samples were topically in- cially at an early stage of keratitis (Fig. 6). stilled onto the LPS-treated cornea. As shown in Fig- ure 2, AC-SOD significantly suppressed corneal in- Discussion flammation while SOD had no appreciable effect. The protective effect of AC-SOD was remarkable par- Our study shows that AC-SOD binds to the mem- ticularly at early stages of inflammation (Fig. 3). No brane surface of the cornea and effectively suppresses such protective effect was observed with SOD at any LPS-induced keratitis, particularly at its early stages. time tested. Thus, superoxide radical and/or its metabolite(s) might play an important role in the pathogenesis of TBAR Levels in LPS-injected Cornea keratitis. Among various ocular diseases, uveitis has been reported to be suppressed by intravenous ad- The process of LPS-induced keratitis can be di- ministration of SOD.1"4 Since the cornea is an avas- vided into four stages as described. To understand the cular tissue, negligible amounts would be delivered to biochemical events occurring during inflammation, the cornea when the therapeutic agent is adminis- TBAR levels in the cornea were determined at differ- tered intravenously. Hence, both SOD and AC-SOD ent stages of inflammation. As shown in Figure 4, did not have an antiinflammatory effect on LPS-in- TBAR levels in the cornea showed a marked change duced keratitis when they were given intravenously during the inflammatory process, increasing signifi- (data not shown). cantly at the active stage (B and C) and returning to It has been well documented that negatively normal ranges at the late (D) and scarring stages (E). charged glycocalyx is anchored onto the epithelial cell Thus, lipid peroxidation seemed to occur particularly surface of the cornea.15 Thus, the glycocalyx may at the initial stages of LPS-induced keratitis. To test minimize the binding of negatively charged mole- whether superoxide radicals play a critical role in cules, such as SOD, to the intact corneal surface due LPS-induced lipid peroxidation, TBAR levels in the to electrostatic repulsion. However, the barrier func- cornea were also determined during the active stage tion of cell-surface glycocarix would be decreased in animals treated with saline or AC-SOD. The latter when the epithelial cells were challenged with hazard-

Downloaded from iovs.arvojournals.org on 09/29/2021 1966 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Ocrober 1990 Vol. 31

1 II

8 •

6 / • ? A • y \ \\ '',

c | | || o X 4 CO * || ||| || E E * ; ; 1 Fig. 5. Effect of AC-SOD on lipid peroxidation. Rabbits were 2 '0. !i % % injected with 10 ng of LPS to the corneal stroma. Then, 80 n\ of I || || saline or 640 ng of SOD or AC-SOD was topically instilled six times {':• 1 a day. After 7 days of LPS injection, TBAR levels in the cornea | 5 •;• / ll ' ••: ' were determined. The results are the means ± SEM derived from k '* •:• ! three eyes. Other conditions were the same as in Figure 3. (A) 1 i •;* > 'v K ;t; ^ •.* / Control group; (B) AC-SOD-treated group. P < 0.01. 1V/i*1 ' •:• / 0 i'c. i 0-/ / '345678 ^ 1 9 Time after LPS injection (day) ous membrane perturbers, such as LPS. Such patho- Fig. 3. Evaluation of corneal inflammation. After 10 ;ug of LPS logic changes on the corneal surface might account was injected into the corneal stroma of the rabbit, 80 n\ of saline or for the increased binding of both SOD and AC-SOD 640 Mg of AC-SOD or SOD was topically instilled six times a day. to the injured cornea. Although binding of both SOD At the indicated times after injection, the grade of corneal inflammation was evaluated according to the inflammation score as de- and AC-SOD markedly increased in the inflamed scribed in the text. The data show the mean ± SEM derived from cornea, AC-SOD, but not SOD, was effective in sup- four to seven eyes. Open columns = control group; shaded columns pressing keratitis. This suggests that the amount of = SOD-treated group; spotted columns = AC-SOD-treated group. superoxide radicals generated by LPS-injection might *P < 0.05. be larger than that dismutated by the bound SOD. Alternatively, amphipathic AC-SOD might be local- ized closer to the membrane lipid bilayer of corneal * 20 epithelial cells, and hence, it might inhibit oxida- tive injury of the cells more effectively than hydro- philic SOD. The reason why topically administered AC-SOD was not antiinflammatory at later stages of keratitis is 10 not clear at present. The AC-SOD metabolizes superoxide radicals to hydrogen peroxide. Although hy- I drogen peroxide in plasma could be metabolized to I I water by cell-associated catalase and glutathione per- 16 III s c oxidase, no such metabolism would occur on the outer surface of the avascular cornea. Superoxide rad- 0 4 7 14 25 icals and hydrogen peroxide could be converted to Time after LPS injection (day) hazardous metabolites with a fairly long half-life.17 Fig. 4. Thiobarbituric acid-reactive metabolites. After injecting Thus, other metabolites with a longer half-life than 10 jig of LPS into the corneal stroma of the rabbits, the cornea was the oxygen species which have a short half-life, such excised at four different stages of inflammation: early (B), active (C), late (D), and scarring (E) stages. TBAR levels in the intact (A) as superoxide radicals and hydroxyl radicals, may and inflammatory corneas were measured as described in the text. play a critical role in the LPS-induced keratitis in its The data show the mean ± SEM derived from three eyes. *P < 0.01. later stages.

Downloaded from iovs.arvojournals.org on 09/29/2021 No. 10 INHIBITION OF CORNEAL INJURY DY ACYLATED SOD / Ando er ol 1967

olite(s) generated by oxygen radicals in the later stages of keratitis should be studied further. Key words: superoxide, acylated-SOD, lipopolysaccharide, corneal inflammation, polymorphonuclear leukocytes

References

1. Scharf J and Dovrat A: Superoxide dismutase molecules in human cataractous lenses. Ophthalmic Res 18:332, 1986. 2. Rao NA, Fernandez MAS, Cid LL, Romero JL, and Sevanian A: Retinal lipid peroxidation in experimental uveitis. Arch Ophthalmol 105:1812, 1987. 3. Rao NA, Romero JL, Fernandez MAS, Sevanian A, and Marak GE: Role of free radical in uveitis. Surv Ophthalmol 32:209, 1987. 4. Rao NA, Sevanian A, Fernandez MAS, Romero JL, Faure JP, Kozak Y, Till GO, and Marak GE: Role of oxygen radicals in experimental allergic uveitis. Invest Ophthalmol Vis Sci 28:886, 1987. 5. Rao NA, Calandra AJ, and Sevanian A: Modulation of lens- induced uveitis by superoxide dismutase. Ophthalmic Res 18:41, 1987. 6. Yagi K: Lipoperoxide level of the retina of chick embryo ex- posed to high concentration of oxygen. Clin Chim Acta 80:355, 1977. 7. Kamata R and Nakashima H: Serratial keratitis. Ganka (Oph- thalmology) 20:863, 1978. 3 4 5 6 7 8 9 8. Kamata R, Matumoto K, Okamura R, Yamamoto T, and Time after LPS injection (day) Maeda H: The serratial 56K. protease as a major pathogenic Fig. 6. Effect of PMN depletion on keratitis. Rabbits were intra- factor in serratial keratitis: Clinical and experimental study. venously injected with hydroxyurea (100 mg/kg) for 5 days. On day Ophthalmology 92:1452, 1985. 6, LPS was injected into the cornea as described in the text. At the 9. Onodera T, Kumagai S, Watanabe T, and Tazawa Y: Experi- indicated times after LPS administration, the grade of corneal in- mental rabbit corneal ring formation caused by bacterial en- flammation was determined according to the inflammation score dotoxin. Folia Ophthalmol Jpn 36:1013, 1985. as described in the text. The data show the mean ± SEM derived 10. Ando Y, Inoue M, Utsumi T, Morino Y, and Araki S: Synthe- from five to six eyes. Open columns = control group; spotted col- sis of acylated SOD derivatives which bind to biomembrane umns = PMN-depleted group. *P < 0.01; **P < 0.05. lipid surface and dismutate extra-cellular superoxide radicals. FEBS Lett 240:216, 1988. 11. Gartner A, Hartmann HJ, and Wester U: A simple, rapid and Since endotoxin enhances the production of che- effective isolation of erythrocyte Cu2+Zn2+-superoxide dismu- motactic factors,15 a large number of PMNs migrated tase. Biochem J 221:549, 1984. 12. Langone JJ: Radioiodination by use of the Bolton-Hunter and to the injured site of the cornea at the early stage of related reagents. Methods Enzymol 70:221, 1979. inflammation and induced corneal abscess in its later 13. Uchiyama M and Mihara M: Determination of malondialde- 9 stage. Thus, although keratitis represents an inflam- hyde precursor in tissues by thiobarbituric acid test. Anal Bio- mation occurring in avascular cornea, it is not sur- chem 86:271, 1979. prising that depletion of the circulating PMNs mark- 14. Hinson JM, Hutchison AA, Ogletree ML, Brigham KL, and Snapper JR: Effect of granulocyte depletion on altered lung edly inhibited the infiltration of PMNs and LPS-in- mechanics after endotoxemia in sheep. Respirat Environ Ex- duced tissue injury (Fig. 6). It has been well ercise Physiol 55:92, 1983. documented that the activated PMNs enhance the 15. Mishima S: Clinical pharmacokinetics of the eye. Invest inflammatory reaction by releasing both reactive ox- Ophthalmol Vis Sci 21:504, 1981. ygen species18 and various lysosomal enzymes.19 Pre- 16. Agar NS, Sadrzadeh MN, Hallaway PE, and Eaton JW: Eryth- rocyte catalase: A somatic oxidant defence? J Clin Invest liminary experiments in our laboratory revealed that 77:319, 1986. topical administration of various inhibitors for lyso- 17. Halliwell B: Superoxide-dependent formation of hydroxyl rad- somal proteases significantly decreased LPS-induced icals in the presence of iron chelates. FEBS Lett 92:321, 1978. keratitis, particularly in its later stages. Thus, other 18. Sato E, Takahara Y, Sasaki J, Matsuno T, and Utsumi K: factors, such as lysosomal enzymes derived from Selective inhibition of stimulation responses of neutrophils by membrane modulators. Cell Struct Funct 11:125, 1981. PMNs, might be responsible for the propagation of 19. Mondino BJ, Brown SI, Rabin BS, and Bruno J: Alternate the inflammatory reaction in its later stages. Possible pathway activation of complement in a Proteus mirabilis ulcer- involvement of such enzymes and hazardous metab- ation of the cornea. Arch Ophthalmol 96:1659, 1978.

Downloaded from iovs.arvojournals.org on 09/29/2021