Decreased ascorbic acid entry into cornea of streptozotocin-diabetic rats and guinea-pigs

ExpTp. Eye RFS. (1992) 55, 337-344

Decreased Ascorbic Streptozotocin-diabetic

Acid

JOSEPH

Entry into Cornea of Rats and Guinea-pigs

DIMATTIO

Department of Physiology and Biophysics, New York University School of Medicine, 550 First Ave., New York, NY 70076, and New York Veterans Administration Hospital, 423 East 23rd St., New York, NY 70070, U.S.A. (Received

Chicago

3 September

7997 and accepted

in revised form 72 November

7997)

High L-ascorbic acid (AA) levels in aqueous humor and intraocular tissues including lens and cornea are thought to protect against the harmful effects of the photochemical and ambient oxidation reactions involving oxygen and its radicals. Our pulse-chase studies follow a bolus of radiolabeled test molecules including [i4C]L-ascorbic acid and [3H]L-glucose (L&I) introduced into the blood at time t = 0, and determinethe time-dependentconcentrationsof these labeledmoleculesas they move into aqueous humor, cornea1endothelium and stroma tissues.Calculatedentry and exit rate constantsprovide a representativemeasureof the functional stateof passiveand carrier mediatedtransport mechanisms in situ in normal and diabetic animals. Diabeticrats werecategorizedin termsof lengthof time exposedto a uniform, monitoredstreptozotocin (stz) diabetesas: short term (lo-20 days); mid-term (40-60 days): and long term (100 +days). In the rat, we observedlittle change in entry rate of L-glu (a passivemarker) into aqueoushumor [control Ki = 0.0216 If: 0.0021 (n = 14)/mid-term stz-diabetesKi = 0.0202 f 0.0027 (n = lo)] and a modest decreasein the entry rate of AA into aqueoushumor [control KAi= 0.02 31+ 0.0022 (n = 14)/mid-term s&diabetesK,, = 0.0201$_00034 (n = lo)]. At cornea1endothelium,we noted a significantdecrease in the active movement of AA [control K, = 0.614+0.053 (n = 14)/mid-term stz-diabetesK, = 0.220+0.026 (n = 9)] while the passiveL-glu entry rate remainedessentiallyunchanged.Thus, our data suggestthat the stz-diabeticrat endotheliumdemonstratesimpairedability to bring AA into the corneawhile passivemovement,asmonitoredwith L-glucose,remainedunchanged.AA movementinto stromawas apparently passiveand decreasedin the cornea1stromaof diabetic rats. Experimentswith diabeticguinea-pigsdemonstrateda diminishedability to bring AA into aqueoushumor and a further decreasein concentrating AA in cornea1endothelium.Thus, in both speciesour data indicate that the ability to bring AA actively into cornea1endotheliumand stroma is compromisedby diabeteswhile L-glucosepassivemovementinto the cornea remainedunchanged. Key words: diabetes; ascorbic acid: aqueoushumor : cornea; endothelium; stroma: vitamin C; transport: guinea-pig; streptozotocin.

1. Introduction The concentration of ascorbic acid (AA) is high in various ocular tissues including the cornea (Maurice and Riley, 1970). Aside from its significant presence, little is known concerning the biological function of ascorbic acid in the cornea. AA has been reported to stimulate the active transport of chloride in the cornea (Buck and Zadunaisky. 19 75) and may be involved in the mediation of neutral amino acid transport into the cornea (Scott and Friedenthal, 1973). Recent studies suggest that some species including humans and guinea-pig have high AA levels in aqueous humor as a result of active transport presumably by the epithelium of the iris-ciliary body (Becker, 1967; DiMattio and Streitman, 1987 ; Socci and Delamere, 1988 ; Chu and Candia, 1988). In vivo studies in our laboratory with guinea-pigs have indicated an active transport carrier mechanism for ascorbic acid entry into aqueous, which is specific for reduced ascorbate (not dehydroascorbate) and probably independent of * Please usethe VA Hospitaladdress for correspondence. 00144835/92/080337+08

sos.oo/o

the facilitated glucose transport mechanism known to be present between blood and aqueous (DiMattio, 1984, 1989a, b). Although no clear picture of AA function in the eye has evolved, many studies have noted its potential importance. Possibleroles attributed to AA include being a direct chemical co-factor in collagen hydroxylation reactions (Stetton, 1949 ; Axelrod, Udinfriend and Brodie, 195 1; Zannoni and La Du, 1960) ; an antioxidant in concert with glutathione and superoxide dismutase (Bhuyan and Bhuyan, 1977; Varma, Ets and Richards, 1977; Delamere, Paterson and Cotton, 1983; Giblin et al., 1984; Riley, Schwartz and Peters, 1986 ; Delamere and Williams, 1987) ; a radiation absorber (Ringvold, 19 75) ; and a redox coupler involved in mediating the hexose monophosphate shunt (Mapson and Moustafa, 1956; Varma, Bauer and Richards, 1987). Recent work from our laboratory indicates that in the rat, where AA is produced as required, little or no carrier-mediation of AA transport into aqueoushumor is present. We noted, however, active carrier-mediated uptake of AA at the level of the lens epithelium, with only a slow passageof AA on into the lens interior 0 1992 AcademicPressLimited

J. DIMATTIO

338

(DiMattio, 1989a, b). In the guinea-pig, which is perhaps a better model for human AA study since it also relies on diet for AA, only reduced ascorbate (not dehydroascorbic acid) was found to be concentrated in aqueous and then further concentrated in lens epithelium by carrier-mediated active transport mechanisms (DiMattio. 1989 a, b). Cornea1 involvement in diabetes has been well documented (Friend and Thoft, 1984 ; Tuft and Coster, 1990) with confirmed reports of decreased epithelium regeneration rates in diabetics (Foulks et al., 1979; Fukishi, Merola and Tanaka, 1980; Friend, Ishii and Thoft, 1982). Aldose reductase inhibitors were found to enhance regeneration in galactosemic rats (Kinoshita et al., 1979). These and other studies (Lass, Spurney and Dutt, 198 5 ; Shetlar, Bourne and Campbell, 1989) suggest potential parallels for diabetic effects between lens and cornea. For this reason we undertook to explore the potential changes in uptake of ascorbate that occur as a consequence of the diabetic process in streptozotocin-diabetic rats. Previous in vivo studies suggest that the cornea1 endothelium of normal rat and guinea-pig demonstrates a significant active uptake of ascorbate (DiMattio and Streitman, 1988: DiMattio, 1991). In this paper, evidence is presented indicating that ascorbate uptake decreases in both cornea1 endothelium and stroma of streptozotocin diabetic rats and guinea-pigs while passive movement (as measured by L-glucose) into the cornea remains unchanged or increases modestly. This decreased delivery of ascorbate to the cornea could signal or result in a state of metabolic or oxidative stress in the diabetic cornea.

2. Materials

and Methods

Transport methods essentially follow those previously published (DiMattio and Zadunaisky, 198 1; DiMattio, 1984, 1989a, b, 1992). All experiments were performed using male, albino Sprague-Dawley rats (2 7 5-3 2 5 g) or English short-haired guinea-pigs (250490 g) under sodium pentobarbital (Nembutal) anesthesia (40 mg kg-‘, i.p.) using methods conforming to the ARVO Resolution on the Use of Animals in Research. Methods for rendering rats diabetic with streptozotocin (stz) administered via the tail vein have been published previously (DiMattio et al., 1981). Guineapigs were made diabetic using modified methods of Schlosser, Kapeghian and Verlangieri (1984). Male English short-haired guinea-pigs weighing between 250-325 g were fasted for 24 hr prior to treatment. They were injected. subcutaneously, with 2 II of Rinsulin using a solution of 10 IJ ml-‘. After about 2 hr, when signs of hypoglycemia (muscle twitching) became apparent, they were anesthetized by i.p. injection of 10 mg ml-l pentobarbital at a dose rate of 40 mg kg-‘. Streptozotocin (Sigma Chemical Co., St

Louis, MO.) 150 mg kg-’ was then administered via a lateral saphenous vein using 1.0 ml citric acid/normal saline (pH 4,O) as the vehicle. This procedure ensures that the competition by n-glucose with streptozotocin at target pancreas cells favored drug access to receptors. n-glucose levels in ocular humors and plasma were determined using a n-glucose specific diagnostic kit supplied by Sigma, which is reliable and utilizes the coupled reactions catalyzed by hexokinase and glucose-h-phosphate dehydrogenase and read at 340 nm on a Beckman DU spectrophotometer (Bonder and Mead, 1974). Transport Experiment

At time zero (t = 0), a double-labeled bolus of radiolabeled test substances was introduced into the circulation via a cannulated femoral vein using 0.5 ml of saline as the vehicle. The specific activity of test molecules did not vary significantly during the short experimental period of 13 min since the concentrations of unlabeled test molecules remained stable and it is presumed that transport and metabolic processes do not favor either labeled or unlabeled test molecules. From t = 0 to t = T, blood samples were removed from the circulation at 1, 2, 3, 5, 7, 9, 11 and 13 min in the 13-min transport experiments. In longer-term experiments samples were taken at 3-5 min intervals. At time t = T. the animal was killed with an iv. overdose of pentobarbital. The eyes were quickly enucleated and samples of aqueous humor (lo-20 /tl) obtained through a cut into the corneoscleral limbus within 2-5 min post-mortem. The whole cornea was then removed and the endothelium dissected under a light microscope to the level of Descemet’s membrane by gentle scraping but leaving the bulk of Descemet’s membrane remaining with stroma. Combined endothelium samples from two eyes ranged from 0.40 to 2.00 mg wet weight, whereas stroma samples ranged from 3.00 to 7.00 mg and were treated individually for greatest accuracy. The endothelium and stroma samples were quickly weighed and set to dry (6O”C, 24 hr). The following day the dry samples were weighed with a Mettler electrobalance and returned to the oven until no further change in dry weight was observed (usually 3 days) and a stable weight was used to estimate tissue concentration. Tissue samples were dissolved in solubilizer (0.5 N quaternary ammonium hydroxide in toluene ; Beckman BTS 4 50) and isotopic concentrations (cpm mg-’ dry weight) of labeled test species determined using Dimiscent scintillation cocktail (National Diagnostics). Average water content was estimated using 12 cornea1 endothelium and stromal samples and used to convert cpm mg-’ dry weight to cpm ml-l tissue water in rate calculations. The blood samples were centrifuged ( 1000 9 for 10 min) and 1O-/J samples of clear plasma of clear plasma were counted using Liquescent

ASCORBIC

ACID

ENTRY

INTO

CORNEA

OF

DIABETIC

ANIMALS

scintillation cocktail (National Diagnostics). Doublechannel counting for 14C and 3H was used and the results were computer processed to correct for quenching and channel spillover and uniformly reported as dpm ml-‘. The plasma data were fitted to a double exponential decay function with time of the form: CI’ = A + &-*lt + Ce-*G

be lo-20 days and a high exposure to be 100 days or more. The level of the diabetes, as estimated by plasma and urine glucose levels, was generally comparable from one animal to another beyond the second day post streptozotocin (stz) administration (normal, 156+_9 mg glucose per 100 ml: 30-day diabetic, 426 _+66 mg glucose per 100 ml : 60-day diabetic, 456+82 mg glucose per 100 ml; lOO+day diabetic, 461+_ 78 mg glucose per 100 ml). Plasma glucose levels in stz-diabetic guinea-pigs were not as elevated as with rats (normal guinea-pig, 178 _+ 18 mg glucose per 100 ml; 30 +day diabetic, 346 + 37 mg glucose per 100 ml). Therefore, we use the number of days of relatively uniform hyperglycemia as an index of level of exposure. We found that cornea1 endothelium % water content did not vary significantly with animal weight and was taken as an average 79.4_+ 1.9 (n = 22) to calculate concentrations of labeled molecules (dpm ml-’ water) from dried endothelium tissue samples. Water content of endothelium from diabetic animals was not found to be significantly different from control values. Diabetic cornea1 stroma demonstrated a decreased water content with exposure to periods of stz-diabetes as compared to control animals. Figure 1 illustrates our results with control and diabetic animals. This decrease in cornea1 water content in diabetic rats was not significant to a level of P < 0.02 but persisted with increasing periods of exposure to stz-diabetes. Appropriate values were used to calculate lens tissue concentrations in dpm ml-’ tissue water. Radio TLC analysis was performed with control and diabetic animals to determine the state of labeled AA introduced via the blood circulation. As reported previously in control rats (DiMattio, 1989). only the reduced form of AA could be detected in aqueous humor. As part of this study, we found that only reduced AA was present in aqueous humor of control

(1)

where A, N. C, b,. b, are determined constants and A has been shown to be the plasma concentration at t = ~8, C,,( cx; ). Graphically determined constants were used as first-guess approximations for a ‘best fit’ determination of the same constants via curve fitting the above data using Asystant software (Macmillan Software Co.. NY). The calculated plasma constants along with aqueous, cornea1 endothelium and stroma concentration data and estimates of steady-state (SS) levels determined in longer-term experiments of 4 hr duration were used to calculate transport rate constants using linear equations in a manner paralleling previous modeling of ocular humors and lens and cornea (DiMattio and Zadunaisky, 198 1; DiMattio, 1984, 1989, 1992). The relevant SS concentration ratio estimates used to calculate relevant rate constants are also reported. An SS ratio of 1.0 is expected for passive or facilitated diffusion, whereas an SS ratio significantly higher than 1.0 would suggest an active, energy-requiring, transport process. Radio thin-layer chromatographic (TLC) techniques were used to monitor the radiolabeled molecules in aqueous humor. Appropriate standards and labels were run on TLC plates and scanned with a Radiomatic Radioscanner RS with RTLC analysis software. Labeled L-glucose was run on G-plates (Analtech, Newark. DE J using benzene : methanol : acetone : acetic acid (70: 20 : 5 : 5) and ran with an Rf of 0.178 with methods outlined previously (DiMattio, 1989). Since ascorbic acid is unstable in air and/or water, all samples were spotted and analysed under 100% nitrogen and confirmed using glass TLC silica gel H plates custom prepared by Analtech with 6% metaphosphoric acid to stabilize the ascorbic acid and prevent oxidation during analysis. The metaphosrun with phoric acid H plates were acetonitrile : butyronitrile : water (66 : 3 3 : 2) and gave an Rf of 0 16 for ascorbic acid.

339

“d 3

80

3. Results Male Sprague-Dawley rats were made diabetic with streptozotocin at SO-64 days of age. Diabetic animals evidenced clear symptoms including weight loss. hyperglycemia (as monitored by elevated glucose levels in blood and urine ; ketone levels also monitored) and low insulin levels. We note that some 20% of our animals died before reaching 100 days of diabetes. In order to estimate a fair or comparabre level of exposure to the diabetes, we considered a low exposure level to

Age (days)

FIG. 1. Cornea1 stroma % water reported as functions of age (days) for control (0) and stz-diabetic (0) rats. Values are listed as means +_s.o. with each point representing a minimum of six animals. Animals were rendered diabetic with streptozotocin at about 60 days (52-63 days). 27-2

J. DIMATTIO

340 TABLE

Thirteen minute entry rate constants

I

(1 per min) from plasmato aqueoushumor in

CA/c,*

control

K,, (1 per min)

and

diabetic ruts

Transport SS

[%]L-ascorbic acid Control

Short-term diabetes(lo-20 days) Mid-term Sk-diabetes(40-60 days) Long-term &-diabetes ( 100 + days) [3H]L-glucose Control Short-term diabetes(lo-20 days) Mid-term &-diabetes (40-60 days)

0.572+0.037 (14) 0.539rfIO.046 (8) 0.524t0.048 (10) 0.49 5 + 0.044 (9)

0.0231$_0.021 (14) 0.0218+0.032 (8) 0.0201 f0.034 (IO) 0.0198 kO.026

0.35820.27 (14) 0.341 kO.38

@0216_$0.021 (14) 0.0207iO.030

0.32710.36

CA/C,

data and calculated

rate constant

Entry rate

constants

1.0

(9) 1.0 1.0 1.0

(10)

0,330+0.42

1.0

0.0222 kO.034

(9)

are listed as means+

TABLE

I.0

0.0202 + 0.02 7

(9) * Reported

1.0

(8)

(81 (10)

Long-termstz-diabetes(100 + days)

I.0

S.D.M.

from experiments

of 13-14

II

(1 per min) from aqueoushumor to

corned

endotheliumin

control

K, { 1 per min) G.ndolC** _~--.~~---__-_-__~_____ [‘%$-ascorbic acid Control

3.28 kO.22 (14) 3.39kO.26

Short-term diabetes(lo-20 days)

0.614iO.053 (14) 0.598+0.062

(8)

Mid-term stz-diabetes(40-60 days) Long-termstz-diabetes(100 + days)

min duration.

and diabetic rats Transport SS ~ ~~~~ _ ~ _~~~_. ~ 5.0 5.0

(8)

2.20+0.18 (10)

0.420f0.035 (10)

P < 0.001 1.48f0.15

0,220?0.026

5.0

P < 0.001 5.0

19)

(9) P < 0.001

P < 0.00

1

[3H]L-glucose

Control

1.32kO.16 (14)

Short-term diabetes(lo-20 days) Mid-term s&diabetes(40-60 days)

1.30+0.1?3 (8) 1.27fO.14

Long-termstz-diabetes(100 + days)

110) 1~17$012

C,,,,/C,

data and calculated

rate constant

are listed as means*

guinea-pigs. Moreover, diabetic animals of both species demonstrated the presence of only radiolabeled reduced AA in aqueous humor. No radiolabeled dehydroascorbic acid could be detected in aqueous humor of control and diabetic rats or guinea-pigs. Table I reports average aqueous to plasma concentration ratios found at 13-14 min post-bolus input, t = 0, and calculated ‘Plasma to aqueous ’ entry rate constants, K, for L-glucose and ascorbic acid (AA) for both control and s&z-diabeticrats with varying periods

1.5

0~511+0~059 (8)

1.5

0.506 iy 0.050

1.5

(10)

1.5

0.4 76 + 0.040

(9) P < 0.02 * Reported

0~525+0~053 (14)

(9)

P < 0.02 S.D.M.

from experiments

of 13-14

min duration,

of exposure to Sk-diabetes. L-glu (MW 180) is used as a measure of passive movement. It is not appreciably metabolized in the eye and is comparable in size to ascorbic acid (MW 176). Our results indicate a possible small decreasein AA entry rate with diabetes with a small subsequent increase in AA entry rate ( - 14.3 % in the 100 +day group of diabetic animals). We note that the changes in AA movement into aqueous seen with diabetic animals were modest at best and not very significant. The blood-aqueous humor entry

ASCORBIC

ACID

ENTRY

INTO

CORNEA

OF

DIABETIC

341

ANIMALS

TABLE III Mean entry rate constants (1 per min) from cornea endothelium to stroma in control and diabetic rats KS (1 per min)

Transport SS

0~550~0~122 (14) 0.539+0.092 (8) 0,440 f 0.078 (10) 0.318+0.071 (9)

0.241 kO.053 (14) 0.235+0.046 (8) 0.191&-DO35 (10) 0.132+0.026 (9) P < 0.001

0.7

0.781+0.116 (14) 0.753+0.098 (8) 0.572+0082 (10) 0.482kO.071 (9)

0.371 kO.053 (14) 0.352*0.049 (8) 0.2 52 + 0.044 (10) 0.206+0.034 (9) P < O*OOl

0.9

CsICEndO* [‘4C]L-ascorbic acid Control Short-term diabetes (lo-20

days)

Mid-term stz-diabetes (40-60 days) Long-term Sk-diabetes (100 + days) [3H]L-glucose Control Short-term diabetes (lo-20

days)

Mid-term stz-diabetes (40-60 days) Long-term stz-diabetes (100 + days)

* Reported C,jC,,,,,

data and calculated

rate constant

are listed as means

constant, K,. for L-glu did not vary appreciably with exposure to diabetes although animals exposed to 100 + days of diabetes did evidence a 3 % increase in K,. The similarity in size and entry rate of L-glu and AA supports previous data suggesting that AA entry into aqueous humor of the rat is via passive movement (DiMattio, 1989). Table 11 reports analogous data and ‘aqueous humor to cornea1 endothelium ’ entry constants for Lglu and AA for normal and stz-diabetic rats exposed to varying periods of exposure to stz-diabetes. Although both AA and L-glu enter cornea1 endothelium quickly, the level and rate of entry of AA in normal cornea1 endothelium is significantly higher than that of L-glu. Stz-diabetes (up to 60 days) significantly decreased the rate of AA entry into endothelium at the same time that passive movement, monitored with L-glu, remained essentially unchanged. The transport steadystate values listed were derived from experiments of 30, 60 and 120 min duration (DiMattio, 1992) and used in calculating the entry constant reported. The experimental C,,,,,,/C, ratio (13 min post-bolus), indicative of movement into endothelium, is also significantly decreased (P < 0.00 1) in animals exposed to more than 40-60 days of diabetes. Table III reports stroma data and ‘cornea1 endothelium to interior stroma ’ entry constants for AA and L-glu of normal and stz-diabetic rats. In control rats, AA moves into the cornea1 stroma more slowly than L-glu suggesting non-specific bulk or passive movement. In addition steady-state estimates from 30, 60, 120 min and 4 hr experiments reveals that no accumulation of labeled molecules occurs and that some regions of the stroma are inaccessible to labeled

~s.D.M.

from

experiments

of 13-14 TABLE

0.7 0.7 0.7

0.9

0.9 15

min duration

IV

Guinea-pig plasma to aqueous humor entry rate

constants, K,,

[‘%&ascorbic Control Diabetict [3H]L-glucose Control Diabetic

CA/c,*

KAi (1 per min)

2.98kO.36

0.0696 f 0.007

SS

acid 10.0

(10)

(10)

1.10_+0.076 (8) P < 0.001

0.0352_fO.O038 (81 P < 0.001

10.0

0~0116f0~0009 (12) 0.0129+0.0013 (8)

I.0

0.260+0.024 (12) 0.276+0.029 (8)

1.0

* Concentration ratios and rate constants are reported as mean f SD. with the number of experiments below in parentheses. Results are based on experiments of 20 _+ 1 min duration. t Animals were monitored diabetic for 30+ days (40-74 days diabetic).

molecules, keeping the steady state from achieving a simple diffusion ratio of 1-O which would indicate uniform equilibrium with endothelium water and aqueous humor. Results with diabetic animals indicate that both test molecules move into the stroma more slowly than in control animals. Tables IV to VI report analogous data with diabetic guinea-pigs. In Table !LVwe note a significant decrease in the rate of ascorbic acid entry into aqueous humor, probably reflecting a decrease in the rate of active movement of AA. Table V indicates that the entry rate of AA into the endothelium of the diabetic guinea-pig is significantly

decreased

as compared

to control

342

J. DIMATTIO TABLE

V

Guinea-pig aqueous humor to cornea1 endothelium entry rate constants, K,,

[“C]L-ascorbic acid Control 2.25 kO.29 (10) 1.03+0.21 Diabetic? (7) P < 0.001 [3H]L-glucose 1.27+0,21 Control (12) 1.19io.23 Diabetic (8) N.S.

K,, (1 per min)

SS

0.221 kO.31 (10) 0.0 74* (1.018 (7) P < 0.001

5.0

0.344+0.048 (12) 0.330f0.056 (84

1.5

5.0

I.5

N.S.

* Concentration ratios and rate constants are reported as means _+ S.D. with the number of experiments below in parentheses. Results are based on experiments of 20 + 1 min duration. t Animals were monitored diabetic for 30 + days ( 30-74 days).

VI Guinea-pig cornea1 endothelium to stroma entry rate constants, K,? TABLE

hi

(1 per min) [W]L-Ascorbic acid Control 0.380+0.049 (10) Diabetict 0.182+0.023 (8) P < 0.001 [SH]L-glucose Control 0.498$_0.044 (12) Diabetic? 0,562&0054 (8)

SS

0.108_+0.010 0.6 (10) 0.051 kO.021 0.6 (8) P < 0.001 0.137F0.012 (12) 0~143~0.016 (8)

1.0 I.0

* Concentration ratios and rate constants are reported as mean k S.D. with the number of experiments below in parentheses. Results are based on experiments of 20 + 1 min duration. t Animals were monitored diabetic for 30 + days (30-74 days).

animals. Thus, while a 20 min C,,,,/C, ratio for AA of 2.2 5 indicates active uptake, this uptake is depressed in diabetic endothelium. L-glu entry into endothelium is fast and passive but may involve some molecular binding leading to accumulation and result in a steady-state ratio somewhat greater than the 1.0 expected from simple diffusion. Whatever sum processes occur in normal guinea-pig affecting L-glucose movement, these processes appear intact in that no change in L-glu entry rate was observed in the diabetic guinea-pig. Similarly, Table VI reports a decrease in the rate of movement of AA from endothelium into the cornea1 stroma of diabetic guinea-pigs. This process is not active since no accumulation occurs and probably reflects the decrease in active uptake and release by endothelium allowing for a lower rate of net move-

ment into the diabetic stroma. The movement from endothelium to stroma of passive marker L-glucose remains essentially unchanged in diabetic animals. 4. Discussion Previous studies with guinea-pigs have indicated that active transport mechanisms bring L-ascorbic acid into the aqueous humor (Becker, 1967 : DiMattio and Streitman, 198 7 ; Chu and Candia. 1988 ; Socci and Delamere, 1988). Recent work from our laboratory indicated further that only L-ascorbic acid entered aqueous regardless of whether a pulse of labeled AA or dehydroascorbic acid had been introduced into the blood (DiMattio, 1989). We suggested that the reduced form of AA could be the required or biologically active form of the vitamin. No active AA transport into aqueous humor was found with normal rats. The observation of active uptake of AA by lens epithelium of both rats and guinea-pig (DiMattio. 1989) and the present data suggesting active uptake by the cornea1 endothelium appear to support the idea put forth by Varma that AA in aqueous humor could provide a supply of this molecule for use by intraocular tissue such as lens and cornea IVarma, 1987). The present study indicates that the rate of active movement of reduced AA into aqueous humor of guineapigs is decreased significantly in diabetic animals. In the rat where AA movement into the aqueous is primarily passive little change in AA entry rate in diabetic rats was observed. This study reports that in both rats and guinea-pigs the active transport mechanism bringing AA into the cornea1 endothelium is compromised in stz-diabetes. Furthermore, the rate of appearance of radiolabeled molecules in the cornea1 stroma is decreased in diabetic animals. The movement of AA into the stroma does not appear to be an active process since entry rates in control animals are less than those of passive marker L-glucose and no accumulation was noted. The movement of AA molecules was slower in both diabetic rats and guinea-pigs. In these same diabetic animals, the movement of L-glucose into the cornea1 endothelium, which is passive and indicative of endothelial permeability, does not change appreciably. This suggests that cornea1 endothelium permeability remains stable in experimental diabetic animals. These results support recent in vitro work indicating that neither Type I nor Type II diabetes had any effect on the endothelium permeability of human donor corneas (Watsky, McDermott and Edelhauser. 1989). Thus, our data indicate that diabetes results in a decrease in carrier-mediated active movement of AA by ciliary epithelium in guinea-pigs and by the cornea1 endothelium of both rats and guinea-pigs. This further suggests that possible antioxidant functions of AA in the cornea are compromised in the diabetic. In addition. other requirements for AA such as in the formation of Descemet’s membrane could also be

ASCORBIC

ACID

ENTRY

INTO

CORNEA

OF

DIABETIC

compromised. Humans, like guinea-pigs, cannot produce AA but demonstrate significant levels of AA in aqueous humor (Cole and Riley, 1970). Therefore, we suggest that our results are likely to apply to humans as regards endothelial uptake of AA and depressed delivery rates of AA to diabetic endothelium and stroma of cornea. This could have harmful consequences as regards maintaining cornea1 health in diabetics as may be reflected in decreased sensitivity, stromal edema and defective re-epithelialization observed in diabetic corneas (Friend and Thoft, 1984; Tuft and Caster. 1990).

Acknowledgements The author is indebted to Mr Jack Streitman for expert technical assistance. The computer assistance of Mr Al Drayton is gratefully acknowledged. In addition, the author wishes to thank Dr V. Fisher for V.A. Hospital facilities without which this work would not have been possible. This work was funded by Research Grant ROl EY04418, from the National Institutes of Health (NEI), Bethesda, MD, U.S.A.

References Axelrod. J.. Udinfriend. S. and Brodie, B. B. (1951). Ascorbic acid in aromatic hydroxylation. III. Effect of ascorbic acid on the hydroxylation of acetanilide, aniline and antipyrine in vivo. 1. Pharmacol. Exp. Ther. 111, 176-81. Becker,B. (1967). Ascorbatetransport in guinea-pigeyes. Invest. Ophthalmol.

6, 410-l

5.

Bhuyan, K. C. and Bhuyan, D. K. (1977). Regulation of hydrogen peroxide in eye humors. Effect of 3-amino1H-l .2,4-triazole on catalase and glutathione peroxidase of rabbit eye. Biochim. Biophys. Acta 497. 641-51.

Bonder, K. J. L. and Mead. D. C. (1974). Evaluation of glucose-6-phosphate dehydrogenasefrom Leuconostoc mesenteroidesin hexokinasemethod for determining glucosein serum.Clin. Chem. 20, 58-62. Buck, M. G. and Zadunaisky,J. A. (1975). Stimulationof ion transport by ascorbicacid through inhibition of 3:5Cyclic-AMP phosphodiesterase in cornea1epithelium and other tissues.Biochim. Biophys. Acta 389. 251-60. Chu, T. C. and Candia. 0. (1988). Active transport of ascorbateacrossthe isolatedrabbit ciliary epithelium. Invest. Ophthalmol.

Iris. Sci. 29, 594-9.

Cole, D. F. and Riley. M. V. (1970). Aqueous and ciliary body. In Biochemistry oj’ the Eye (Ed. Graymore, C.N.). Pp. 105-72. AcademicPress:London. Delamere.N. A., Paterson.C. A. and Cotton. T. R. (1983). Lens cation transport and permeability changesfollowing exposureto hydrogen peroxide. Exp. Eye Res. 37, 45-53. Delamere.N. A. and Wilhams. R. N. (1987). Comparative aspectsof ascorbatein the aqueoushumor. Invest. Ophthalmol. Vis. Sci. 28 (Suppl.), 74. DiMattio, J. (1984). In vivo entry of glucoseanalogsinto lens and cornea of the rat. bwest. Ophthalmol. Vis. Sci. 2 5, 160-5. DiMattio. J. (1989a). A comparativestudy of ascorbicacid entry into aqueousand vitreous humorsof the rat and guineapig. Invest. Ophthalmol. Vis. Sci. 30. 3647. DiMattio. J. (1989b). Active transport of ascorbicacid into lensepitheliumof the rat. Exp. EyeRes. 49, 873-85. DiMattio. J. (29921. Ascorbic acid entry into cornea of rat and guinea-pig.Cornea 11. 53-67.

ANIMALS

343

DiMattio, J.. Altszuler, N., Ellis, S. and Zadunaisky, J. A. (198 1). Glucosetransport acrossocular barriersof the streptozotocin-diabeticrat. Diabetes 30, 903-7. DiMattio. J. and Streitman, J. (1987). A comparisonof vitamin C and glucosetransport acrossocular barriers of rats and guinea-pigs.Invest. Ophthalmol. Vis. Sci. 28, 74. DiMattio, J. and Streitman,J. (1988). A comparativestudy of in vivo ascorbicacid and glucosemovementinto cornea of rats and guinea-pigs.Invest. OphthalmoI. KS. Sci. 29, 48. DiMattio. J. and Zadunaisky,J. A. (1981). Glucosetransport into the ocular compartmentsof the rat. Exp. Eye Res. 32,

517-32.

Foulks,G. N.. Theft, R. A., Perry. H. D. and Tolentino. F. I. (1979). Factors related to cornea1epithelium complication after closed vitrectomy in diabetics. Arch. Ophthalmol.

97. 1076-82.

Friend, J.. Ishii, Y. and Theft, R. A. ( 1982). Cornea1 epitheliumchangesin diabeticrats. Ophthalmic. Res. 14, 269-76.

Friend, J. and Theft, R. A. (1984). The diabetic cornea. In Ocular Manifistation of Diabetes. (Ed. Pavan, P.R.). Pp. 11l-23. Little. Brown and Co.: Boston. Fukushi, S., Merola, L. 0. and Tanaka, M. (1980). Reepithelialization of denudedcorneas in diabetic rats. Exp. Eye Res. 31, 611-16. Giblin, F. J.. McCready,J. P., Kodama,T. and Reddy, V. N. 11984). A direct correlation between the levels of ascorbicacidand H,O, in aqueoushumor. Exp. Eye Res. 38.

87-93.

Kinoshita,J. H., Fukushi, S., Kador. P. F. and Merola, L. 0. (1979). Aldose reductasein diabetic complicationsof the eye. Metabolism 28, 462-8. Lass, J. I-i., Spurney, R. V. and Dutt. R. M. (1985). A morphologic and fluorophotometric analysis of the cornea1endothelium in type I diabetesmelitus and cystic fibrosis.Am. 1. Ophthalmol 100, 783-8. Mapson. L. W. and Moustafa. E. M. (1956). Ascorbic acid and glutathione asrespiratory carrier in the respiration of pea seedlings.Biochem. 1. 62, 248-55. Maurice. D. M. and Riley, M. V. (19 70). The cornea. In Biochemistry ofthe Eye. (Ed. Graymore.C.N.).Pp. l-95. AcademicPress: London. Riley, M. V., Schwartz, C. A. and Peters, M. I. (1986). Interactions of ascorbateand H,O,: implicationsfor in vitro studiesof lens and cornea. Curr. Eye Res. 5. 207-16.

Ringvold, A. (1975). Distribution of ascorbicacid in the ciliary body of the albinorabbit, guinea-pigand the rat. Acta Ophthaimol.

53 75 l-9.

Schlosser,M. J., Kapeghian. J. C. and Verlangieri, A. J. (1984). l!%ectsof streptozotocinin the maleguinea-pig: a potential model for studying diabetes.Life Sci. 35, 649-55.

Scott, W. N. and Friedenthal.D. F. (1973). A proposedrole for ascorbatein the transport of amino acidsand ions into cornea. Exp. Eye Res. 15, 683-92. Shetlar, D. J.. Bourne, W. M. and Campbell,J. (1989). Morphologic evaluation of Descemet’s membraneand cornea1endotheliumin diabetesmellitus.Ophthalmology 96,

247-55.

Socci,R. R. and Delamere,N. A. (1988). Characteristicsof ascorbatetransport in the rabbit iris-ciliary body. Exp. Ege Res. 46, 853-61.

Stetton, M. R. (1949). Someaspectsof the metabolismof hydroxyproline studiedwith the aid of isotopicnitrogen. 1. Biol. Chem. 181. 31-7. Tuft. S.J. andCaster,D. J. (1990). The cornea1endothelium. Eye 4, 389-424.

Varma, S.D. (1987). Ascorbicacid and the eye with special

344 reference to the lens. (Third Conference on Vitamin C.) Ann. N.Y. Acad. Sci. 498, 280-307. Varma, S. D., Bauer, S. A. and Richards, R. D. (1987). Hexose monophosphate shunt in rat lens: stimulation by Vitamin C. Invest. Ophthalmol. Vis. Sci. 28, 1164-9. Varma. S. D., Ets, T. K. and Richards, R. D. (1977). Protection against superoxide radicals in rat lens. Ophthalmic Res. 9, 421-31.

J. DIMATTIO

Watsky, M. A., McDermott, M. L. and Edelhauser, H. F. (1989). In vitro cornea1 endothelial permeability in rabbit and human: the effect of age, cataract surgery and diabetes. Exp. Eye Res. 49, 751-67. Zannoni, V. G. and La Du, B. N. (1960). Tyrosylurea resulting from inhibition of p-hydroxyphenylpyruvic acid oxidase in Vitamin C-deficient guinea-pigs. 1. Bid. Chem. 235. 2667-71.