A C T A O P H T H A L M O L O G I C A V O L . 5 9 1 9 8 1

Department of Ophthalmology' (Chairman: M . N . Luxenberg M.D.), and the Depament of Physiology2 (Chairman: Robert Little, M . D.)

Medical College of Georgia Augusta, G e o r p 30912 USA

HYDROGEN PEROXIDE AND CORNEAL ENDOTHELIUM

BY

DAVID S. HULL1, STEVE CSUKAS', KEITH GREEN'

and VELDA LIVINGSTON'

Because of recent evidence of low levels of hydrogen peroxide in the aqueous humor, studies were performed to determine levels of corneal endothelial toxicity as well as factors modifying toxicity. Perfusion of cornea endothelial cells for 3 h with varying concentrations of hydrogen peroxide demonstrated a threshold of toxicity at a nominal concentration between 0.3 and 0.5 mM HzOz. The toxic effect resulted in rapid corneal swelling as well as disruption of endothelial cell cytoplasm and organelles. Both the physiologic and anatomic toxic effects of 0.5 mM HzOz could be blocked with 5400 U/ml catalase. Exposure of corneas to 20 mM HzOz for 10 min in the presence of EDTA - Fe+3 resulted in an enhancement of corneal swelling rate more rapid than that which resulted from a 10 min exposure to 20 mM H202 alone. Neither the presence of ascorbic acid nor the absence of glutathione and adenosine had an effect on the cornea swelling rate which occurred during a 3 h perfusion of endothelium with 0.3 mM HzOz. Chelated iron had no effect on the corneal swelling rate induced by phototactivation of rose bengal presensitized cornea endothelial cells.

Kq words: hydrogen peroxide - cornea - endothelium - free radical - peroxidation

Recent work has demonstrated low levels of hydrogen peroxide in the aqueous humor of humans and rabbits and it has been attempted to associate this finding with cataract formation (Spector et al. 1980; Bhuyan & Bhuyan 1978; Bhuyan & Bhuyan 1977; Bhuyan & Bhuyan 1979). It has also been shown that perfusion of corneal endothelial cells in the specular microscope with rose bengal followed by

Received on January 6th. 198 1.

409

exposure to light results in corneal swelling, whereas corneas perfused in the dark with rose bengal do not swell (Hull, Strickland & Green 1979). The photodynami- cally induced endothelial change has been found to be oxygen dependent and at least in part secondary to hydrogen peroxide produced by the dismutation reaction of superoxide free radical which is catalyzed by superoxide dismutase (Hull, Strickland & Green 1979; Hull et al. 1981).

Because of the known presence of hydrogen peroxide in the aqueous humor and its potential for inducing endothelial damage as demonstrated in the photoactivated model (Hull, Strickland & Green 1979; Hull et al. 1981) it was the purpose of this series of experiments to determine the concentration of hydrogen peroxide that causes corneal endothelial cell damage. In addition, because of the known effect of iron on enhancement of peroxidation in some models, and the fact that the aqueous humor of hyphema patients may contain iron, it was determined what effect chelated iron had on the threshold of endothelial hydrogen peroxide toxicity and photodynamically induced change, and determined if ascorbic acid and glutathione had a modifying influence.

Materials and Methods

Group I : Hydrogen peroxide dose response

Adult albino rabbits weighing three kg were killed with an overdose of intravenous sodium pentobarbital. The eyes were enucleated together with the conjunctival sac and eyelids, after which the corneas were mounted in a specular microscope (Maurice 1968; Dikstein & Maurice 1972; Maurice 1972; McCarey, Edelhauser & Van Horn 1973). This instrument allows constant perfusion of the endothelial surface at 37°C and + 15 mM Hg pressure, observation of the corneal endothelium, and sequential measurements of corneal thickness. Silicone oil (Dow Chemical Company, Midland, Michigan) was placed on the epithelial surface to prevent evaporation. Corneal swelling rates were determined by linear regression analysis and a comparison of the swelling rates of experimental and control corneas was made by an analysis of co-variance (Snedecor & Cochran 1967; Steel & Torrie 1960). Control corneas were perfused constantly with Krebs-Ringer bicarbonate with added 0.299 mM reduced glutathione (Sigma Chemical Company) and 0.501 mM adenosine (Sigma Chemical Company) (Dikstein & Maurice 1972 ; McCarey, Edelhauser & Van Horn 1973; Green & Green 1969). All solutions were made up in deionized water, but no special precautions were taken to remove trace metals from other reagents. The experimental corneas were perfused for 1 h with the control solution and then, when corneal thickness had stabilized, they were perfused with nominal concentrations of hydrogen peroxide (30% stock, certified A.C.S., Fisher Scientific Corp.) ranging from 0.1 mM to 1.0 mM dissolved in

410

Krebs-Ringer bicarbonate solution with adenosine and glutathione and pH adjus- ted to 7.3. In one set of experiments with 0.5 mM HzOz, 5400 U/ml (200 pg/ml) catalase (Sigma Chemical Co., 27,000 units/mg) was added to the perfusing solution. At the end of each experiment, the corneas were fixed in two per cent glutaraldehyde in phosphate buffer and submitted for scanning and transmission electron microscopy.

Group 1 1 : Chelated ferric chloride, glutathione and ascorbic acid

Corneas were mounted in the specular microscope as in Group I. All corneas were perfused for 3 h with 0.3 mM hydrogen peroxide in Krebs-Ringer bicarbonate solution with adenosine and glutathione and other additives as noted below.

Subset A corneas were perfused with either 5 mM ascorbic acid, or 5 mM ascorbic acid in combination with 0.1 1 mM EDTA - 0.1 mM FeCb (Svingen, O’Neal & Aust 1978) in order to determine whether ascorbic acid could enhance the effect of what was known to be a subtoxic level of hydrogen peroxide (as determined in Group I).

Subset B corneas were perfused with 0.11 mM EDTA - 0.1 mM FeCb. Subset C control corneas were perfused in Krebs-Ringer bicarbonate without

adenosine and glutathione. Experimental corneas were perfused with 0.1 1 mM EDTA and 0.1 mM FeCh in Krebs-Ringer bicarbonate without adenosine and glutathione in order to eliminate any possible glutathione contribution to the response seen in Subset B.

Group 111: Ferric and ferrous chelates

Corneas were mounted on specular microscope mounting rings. They were then held endothelial side down in a beaker containing one of the following solutions dissolved in Krebs-Ringer bicarbonate with added glutathione and adenosine: (A) 0.1 1 mM EDTA and 0.1 mM FeCb, (B) 0.11 mM EDTA and 0.1 mM FeCL, (C) 1.7 mM ADP and 0.1 mM FeCb, (D) 1.7 mM ADP and 0.1 mM FeC12. In all instances the solutions were at pH 7.3 and ferrous chelates were made up in Ringer previously bubbled 45 rnin with nitrogen (Svingen, ONeal& Aust 1978).

Hydrogen peroxide was then added to the beakers in 5 sequentially divided aliquots over a period of one rnin so as to make a final concentration of 20 mM. Corneas were allowed to soak 10 min, were removed and rinsed for 3 rnin in Krebs-Ringer bicarbonate with glutathione and adenosine, mounted in the specular microscope and perfused for 3 h with Krebs-Ringer bicarbonate with glutathione and adenosine.

Controls were immersed for 10 min in 20 mM HzOz in the same manner as experimentals except that the soaking solutions did not contain iron chelates. These experiments were performed to determine if iron chelates would accelerate hydrogen peroxide toxicity.

41 1

Table I. Effect of HzOz concentration on corneal swelling rate during 3 h perfusion of endothelium

HzOz mM Corneal swelling

confidence limits Additives N rate p/h ? 95%

A. 0.1 none 4 5 + 2 B. 0.9 none 5 3 f 3 C. 0.5 none 6 26 f 7'

0.5 5400 U/ml catalase 7 8 f 21 D. 1.0 none 6 61 ? 13

1. P < 0.0 1 within group C Table I

Group IV: Ferric chelates and photodynamic change

Corneas were mounted in the specular microscope and following a 1 h stabilization period were perfused with 5 x 10% rose bengal dissolved in Krebs-Ringer bicar- bonate with glutathione, adenosine and either 1.7 mM ADP - 0.1 mM FeCb or 0.11 mM EDTA - 0.1 mM FeCb. Corneas were exposed to a 25 watt incandescent light at

Fig. 1. Scanning electron micrograph of rabbit cornea endothelial cells following 3 h perfusion with 0.3 mM HzOz. Posterior surface of cells is flat and mosaic pattern is maintained. (Original

magnification x 1000).

412

5 cm for 1 min as has been previously described (Hull, Strickland & Green 1979; Hull et al. 1981). A water filled petri dish was placed between the light source and the cornea to absorb heat generated by the bulb (Hull. Strickland & Green 1979; Hull et al. 198 1). Following light exposure the chambers were flushed and perfused for 3 h with Krebs-Ringer bicarbonate with glutathione, adenosine, and either 1.7 mM ADP - 0.1 mM FeCb or 0.11 mM EDTA - 0.1 mM FeCb.

Controls were treated as experimentals, however perfusing solutions did not contain chelated iron. These experiments were performed to determine the effects of these experimental manipulations on an endothelium which was prephotosensi- tized with rose bengal.

Results

Group I

Control corneas generally swelled less than 5 p/h. Corneas did not swell more rapidly than controls when perfused with either 0.1 mM or 0.3 mM hydrogen peroxide (Table 1). Perfusion with 0.5 mM hydrogen peroxide resulted in a corneal swelling rate of 26 f 7 p/h which was statistically higher than the paired control value of 0.2 f 1 p/h P < 0.01. Perfusion with 0.5 mM hydrogen peroxide and 5400 U/ml catalase resulted in a swelling rate of 8 f 2 p/h which was less than the swelling rate of 26 f 7 p/h with 0.5 mM HzOz alone (P < 0.01).

Electron microscopy of corneas following 3 h perfusion with 0.3 mM HzOz showed maintenance of normal endothelial cell architecture (Fig. 1). Perfusion for 3 h with 0.5 mM HzOz resulted in swelling of cellular cytoplasm and disruption of organelles (Figs. 2 and 3). Endothelial cells perfused for 3 h with 0.5 mM HzOz and 5400 U/ml catalase demonstrated an undamaged cellular architecture (Fig 4).

Group II

In the presence of 0.3 mM hydrogen peroxide (which was known from Group I to be below the toxic threshold), neither 5 mM ascorbic acid, nor 0.1 1 mM EDTA - 0.1 mM FeCb, nor a combination of 5 mM ascorbic acid and 0.11 mM EDTA - 0.1 mM FeCb accelerated corneal swelling rates (Table 2). The absence of glutathione and adenosine (with or without 0.11 mM EDTA - 0.1 mM FeCh) also had no accelerating effect on corneal swelling rates.

Group 111

Corneas exposed to 20 mM HzOz for 10 min in the presence of 0.1 1 mM EDTA - 0. I mM FeCb swelled more rapidly than paired controls P < 0.01 (Table 3, Subset A). EDTA - FeClz, ADP - FeCb and ADP - FeCL did not cause acceleration of corneal swelling rates following 10 min exposure to 20 mM HzOz.

413

Fig. 2. Scanning electron micrograph of rabbit cornea endothelial cells following 3 h perfusion with 0.5 mM HzOz. Cells appear ruptured and mosaic pattern of cell borders is lost. (Original

magnification x 1000)

Fig. 3 . Transmission electron micrograph of cornea endothelial cells following 3 h perfusion with 0.5 mM HzOz. There is swelling of the endoplasmic reticulum, mitochondria, and nuclei, The

posterior cell border is irregular. (Original magnification x 7380).

Fig. 4. Transmission electron micrograph of cornea endothelial cells following 3 h perfusion with 0.5 mM HzOz and 5400 U/ml catalase. Cellular organelles are well preserved, the posterior

cell border and intercellular spaces are intact. (Original magnification x 7380).

Group IV

Ferric chelates did not accelerate the rate of corneal swelling associated with photoactivation of rose bengal presensitized cornea endothelial cells. (Table 4). It is to be remembered that corneas perfused with rose bengal in the dark do not undergo significant swelling (Hull, Strickland & Green 1979).

Discussion

It is known that normal cell metabolic processes involve the production of superoxide anion and hydrogen peroxide (Aust, Roerig & Pederson 1972; Loschen et al. 1974). These metabolic products are potentially damaging to the cell, are present in the aqueous humor, and conceivably may be related to aging and degenerative changes, as well as cataract formation (Pryor 1978; Bhuyan & Bhuyan 1978; Bhuyan & Bhuyan 1977; Bhuyan & Bhuyan 1979). NormaIly protective mechanisms in the cell include superoxide dismutase which results in the produc- tion of hydrogen peroxide from superoxide anion.

415

Table 2. Relationship of ascorbic acid, EDTA - Fe+s, and adenosine - gluta- thione to corneal swelling rate during 3 h perfusion with 0.3 mM HzOz. All solutions contained 0.3 m M HZOZ in Krebs-Ringer bicarbonate with 0.501 mM adenosine and 0.299 mM glutathione unless otherwise

indicated.

Corneal swelling rate p/h 2 95% confidence limit I N l Compound and

concentration

A.

B.

C.

5 mM ascorbic acid 5 5 + 1 0.3 mM HZOZ 5 mM ascorbic acid 5 7 2 3 0.1 1 mM EDTA 0.1 mM FeCb 0.3 mM HzOz

0.11 mM EDTA 0.1 m M FeCb 0.3 mM HzOz

6

Ringer (with 0.3 mM HzOz) 6

Ringer without adenosine 5 and glutathione but with 0.3 m M HzOz

0.1 1 mM EDTA 0.1 mM FeCh

No adenosine and glutathione

0.3 mM HzOz

5

10 2 2

10 f 3

12 f 2

8 + 3

P > 0.01 within groups A, B, C. Table 2.

HOi +O; + H+ + HzOz + 02 Hydrogen peroxide is subsequently removed from the system by catalases and peroxidases. Superoxide dismutase, catalase, and glutathione peroxidase have been shown to be present in eye tissues including the cornea endothelium (Bhuyan & Bhuyan 1978; Bhuyan & Bhuyan 1977). It was the purpose of this series of experiments to relate hydrogen peroxide concentration to endothelial cell damage.

This study has demonstrated that the threshold of toxicity resulting from short-term exposure to hydrogen peroxide on cornea endothelium is between 0.3 and 0.5 mM and that the effect is blocked by 5400 U/ml catalase indicating the protective effect of catalase on a presumed direct hydrogen peroxide effect. Recent

416

Tahle 3 . Effect of 10 min soak of corneas in 20 mM HzOz in the presence of chelates of Fe+2 and Fe+3. Swelling rates determined during subse-

quent 3 h perfusion in Krebs-Ringer bicarbonate.

Iron Chelate Cornea swelling rate

p/h f 95% confidence limits

A. 0.11 mM EDTA 0.1 mM FeCh Control

B. 0.11 mM EDTA 0.1 mM FeClz Control

C. 1.7mMADP 0.1 mM FeC13 Control

D. 1.7 m u ADP 0.1 mM FeC1z Control

25 f 3'

17 f 2'

I6 f 3

15 f 2

17 f 3

15 f 2

17 f 3

1 8 f 1

1 , different from paired control P < 0.01.

work has demonstrated rabbit and human aqueous humor levels of hydrogen peroxide ranging from 25-60 p ~ , which is about one-tenth the concentration found in this study necessary to induce acute endothelial toxicity (Spector et al. 1980; Bhuyan & Bhuyan 1978). Studies have indicated a role of hydrogen peroxide

Tahk 4 . Effect of chelated FeCb on corneal swelling induced by photoactivation

of endothelial cells presensitized with 5~ 1 0 - 6 ~ rose bengal. ~~

Iron Chelate Cornea swelling rate 1 p/h ? 95% confidence limits 1

A. 1.7 mM ADP 0.1 mM FeCb Control

B. 0.1 1 mM EDTA 0.1 mM FeCh Control

4 28 k 3

4 28 f 5

4 27 f 5

4 26 f 6

417 Acta ophthal. 59, 3

in alteration of the ocular lens (Fukui et al. 1980; Varma, Kurnar & Richards 1979; Bhuyan & Bhuyan 1978), but the effect of long-term lower concentrations of hydrogen peroxide on cornea endothelial aging and degenerative changes is not yet known. It is of interest that the threshold of toxicity of an organic peroxide, t-butyl hydroperoxide, on human erythrocytes is 0.75 mM which is the same order of magnitude found in this experiment for hydrogen peroxide (Corry, Meiselman & Hochstein 1980). Exposure of erythrocytes to a lower concentration of t-butyl hydroperoxide for an extended period of time, however, also resulted in an alteration of red cell properties (Corry, Meiselman & Hochstein 1980).

Work originally done by Haber & Weiss in 1934 and recently modified has shown that the extremely reactive hydroxyl radical can be formed by the interaction of superoxide anion with hydrogen peroxide in the presence of iron chelates and may explain OH. production in vivo (Borg et al. 1978; Pryor 1978; Haber & Weiss 1934; Halliwell 1976; McClune & Fee 1976; McCord 8c Day 1978).

a) Fe+? (chelate) + HzOz + H+ + Fe+S (chelate) + OH. + HzO b) Fe+-" (chelate) + 0 2 -+ Fe+z (chelate) + 0 2

C) HzOz + Oz+ Hf OH. + 0 2 + HzO

The hydroxyl radical formed from this reaction is an extremely reactive oxidant and could explain hydrogen peroxide toxicity, not by direct attack, but rather by formation of the very reactive hydroxyl radical species. This Fenton-type reaction is of potential clinical interest because of the possibility of bound iron in haemoglobin reacting with aqueous humor hydrogen peroxide in hyphema patients resulting in hydroxyl radical production and anterior segment cellular alteration or destruction.

Group 111 corneas were exposed to toxic levels of HzOz (20 mM) for Id min in the presence of both ferrous and ferric chelates and only exposure to EDTA - Fe+3 resulted in an acceleration of corneal swelling (Table 3 compare Subgroup A with B, C, D). Previous work has demonstrated that EDTA - iron is a better lipid peroxide propagating agent than ADP - iron, and that in the peroxidation of microsomal lipid, maximum rates were found with EDTA - Fe+3 (Svingen, ONeal & Aust 1978; Pederson & Aust 1973).

Chelated iron also enhances the peroxidation of polyunsaturated fatty acids with the ferric ion (Fef3) being required for the first step in the oxidative cleavage of polyunsaturated fatty acids and the formation of phospholipid peroxides (Poyer & McCay 1971). The catalytic activity of Fef3 can be strongly stimulated in some model systems by addition of a reducing agent such as ascorbic acid (which is present in aqueous humor) which regenerates Fe+2 (Svingen, ONeal & Aust 1978; Schneider, Smith & Hunter 1964; Barber 1966; Orrenius, Dallner & Ernster 1964). In this model system however, Group I1 corneas did not swell at rates more rapid than controls when perfused for 3 h with slightly below toxic threshold concentra- tions of hydrogen peroxide (0.3 mM) in combination with chelated iron either with

418

or without 5 mM ascorbic acid (Table 2, Subgroups A and B). In addition removal of adenosine and glutathione from the perfusion media had no effect on corneal swelling rate (Table 2, Subgroup C). I t was also of interest that photoactivation of endothelium in the presence of iron chelates did not result in corneal swelling rates more rapid than that caused by photooxidation alone (Group IV, Table 4). Thus in the current endothelial experimental model system, if lipid peroxidation is a primary event, propagation by Fe+3 and resultant physiologic and anatomic alteration of endothelial cells has not yet been demonstrated.

Susceptability of a specific membrane to peroxidative attack depends on the composition and amount of phospholipid in the membrane. It is known that lysosomes contain less lipid and peroxidize slower than mitochondria and microsomes which have a higher lipid content (Desai, Sawant & Tappel 1964). The precise lipid content of cornea endothelial cell plasma membrane is at this time poorly understood, however recent advances in membrane isolation may be of help in the future identification of specific lipid content (Zam, Cerda & Polack 1980). It is knoen that hydrogen peroxide can also cause oxidation of sulfhydryl compounds in addition to causing peroxidation of polyunsaturated fatty acids (Fridovich 1976). In addition it is also possible that a direct effect of radical species may occur on the membrane proteins which could affect membrane transport and change the endothelial permeability characteristics. Such knowledge might provide some answers to the data revealed in the present investigation.

Acknowledgments

Supported in part by Research Grants EY02386 (Dr. Hull) and EY01413 (Dr. Green) from the National Eye Institute, in part by a Research Grant from the Georgia Lions Lighthouse Foundation, Inc., and in part by the Lions Clubs of Augusta, Georgia. A Wang 2200 Computer, used for statistical data evaluation was provided through a Research to Prevent Blindness, Inc., grant.

References

Aust S D, Roerig D L, Pederson T C (1972): Evidence for superoxide generation by NADPH-Cytochrome C reductase of rat liver microsomes. Biochem Biophys Res Comm 47: 1133- 1137.

Barber A A (1966): Lipid peroxidation in rat tissue homogenates: interaction of iron and ascorbic acid as the normal catalytic mechanism. Lipids 1 : 146- 15 1.

Bhuyan K C & Bhuyan D K (1977): Regulation of hydrogen peroxide in eye humors. Effect of 3-amino- 1 H- 1,2,4-triazole on catalase and glutathione peroxidase of rabbit eye. Biochemica et Biophysica Acta 497: 64 1-65 1 .

Bhuyan K C & Bhuyan D K (1978): Superoxide dismutase of the eye. Relative functions of superoxide dismutase and catalase in protecting the ocular lens from oxidative damage. Biochimica et Biophysica Acta 542: 28-38.

419 27'

Bhuyan K C & Bhuyan D K (1979): Mechanism of cataractogenesis induced by 3-amino-1H- 1,2,4-triazole I : Morphology and histopathology of cataract and the role of catalase in the regulation of H202 in the eye. In: Caughey W S (ed). Biochemical and Clinical Aspects of Oxygen, Symposium, Colorado State University, September 24-29, 1978. New York, Academic Press, pp 785-796.

Borg D C, Schaich K M, Elmore J J & Bell J A (1978): Cytotoxic reactions of free radical species of oxygen. Photochem and Photobiol28: 887-907.

Corry W D, Meiselman H J & Hochstein P (1980): t-butyl hydroperoxide induced changes in the physicochemical properties of human erythrocytes. Biochim Biophys Acta 597: 224 - 234.

Desai I D, Sawant P L & Tappel A L (1964): Peroxidative and radiation damage to isolated lysosomes. Biochim Biophys Acta 86: 277-285.

Dikstein S & Maurice D M (1972): The metabolic basis to the fluid pump in the cornea, J Physiol (Lond), 221: 29-41.

A, editor, Free Radicals in Biology, New York, Academic Press, Inc., Vol I, p 239.

choline uptake by the lens, Invest Ophth and Vis Sci 19, Suppl. p 13.

Physiol217: 635-641.

Haber F & Weiss J (1934): The catalytic decomposition of hydrogen peroxide by iron salts. ProcRSocLondSerA 147: 332-351.

Halliwell B (1976): An attempt to demonstrate a reaction between superoxide and hydrogen peroxide. FEBS Lett 72: 8- 10.

Hull D S, Strickland E C & Green K (1979): Photodynamically induced alteration of cornea endothelial cell function. Invest Ophthalmol and Vis Sci 18: 1226- 1231.

Hull D S, Green K, Csukas S & Livingston V (1981): Photodynamic alteration of cornea endothelium : relation to bicarbonate fluxes and oxygen concentration. Biochim Biophys Acta 640: 231-239.

Loschen G, Azzi A, Richter C & Flohe L (1974): Superoxide radicals as precursors of mitochondrial hydrogen peroxide. FEBS Lett 42: 68-72.

Maurice D M (1968): Cellular membrane activity in the corneal endothelium of the intact eye. Experientia 24: 1094- 1095.

Maurice D M (1972): The location of the fluid pump in the cornea. J Physiol (Lond), 221: 43-54.

McCarey B E, Edelhauser H F & Van Horn D L (1973): Functional and structural changes in the corneal endothelium during in-vitro perfusion. Invest Ophthalmol 12: 410-417.

McClune G J & Fee J A (1976): Stopped flow spectrophotomereic observation of superoxide dismutation in aqueous solution. FEBS Lett 67: 294-298.

McCord J M & Day E D (1978): Superoxide-dependent production of hydroxyl radical catalyzed by iron-EDTA complex. FEBS Lett 86: 139- 142.

Orrenius S, Dallner G & Ernster L (1964): Inhibition of the TPNH-linked lipid peroxidation of h e r microsomes by drugs undergoing oxidative demethylation. Biochem Biophys Res Commun 14: 329-334.

Pederson T C & Aust S D (1973): The role of superoxide and singlet oxygen in lipid peroxidation promoted by xanthine oxidase. Biochem Biophys Res Commun 52 :

Poyer J L & McCay P B (1971F): Reduced trephosphopyridine nucleotide oxidase-catalyzed

Fridovich I (1976): Oxygen radicals, hydrogen peroxide, and oxygen toxicity. In: Pryor, W

Fukui H N, Jernigan H M, Goosey J D & Kinoshita J H (1980): Peroxidative effects on the

Green K & Green M A (1969): Permeability to water of rabbit corneal membranes. Am J

1071-1078.

alterations of membrane phospholipids, J Biol Chem 246: 263-269.

420

Pryor W A (1978): The formation of free radicals and the consequences of their reactions in-vitro. Photochem and Photobiol28: 787-801.

Schneider A K, Smith E E & Hunter F E (1964): Correlation of oxygen consumption with swelling and lipid peroxide formation when mitochomdria are treated with the swelling- inducing agents Fez+, glutathione, ascorbate, or phosphate. Biochem 3: 1470- 1477.

Snedecor G W & Cochran W G (1967): Statistical Methods, 6th edition. The Iowa State University Press, Ames, Iowa, p 419.

Spector A. Anderson E I, Garner W H & Kaplen J (1980): HzOz and human cataract. Invest Ophth and Vis Sci 19, Suppl. p 14.

Steel R G D & Torrie H J (1960): Principles and Procedures of Statistics. New York, McCraw-Hill Book Co., p 171.

Svingen B A, ONeal F 0 & Aust S D (1978): The role of superoxide and singlet oxygen in lipid peroxidation. Photochem and Photobiol28: 803-809.

Varma S D, Kumar S & Richards R D (1979): Light-induced damage to ocular lens cation pump: prevention by Vitamin C Proc Natl Acad Sci USA 76: 3504-3506.

Zam 2 S, Cerda J & Polack F M (1980): Isolation of the plasma membrane from corneal endothelial cells. Invest Ophth and Vis Sci 19: 648-652.

Authors address: David S. Hull, M.D., Department of Ophthalmology, Medical College of Georgia, Augusta, Georgia 30912 USA.

42 1