Retinal Polyol and Myo-Lnosifol in Galactosemic Dogs Given an Aldose-Reducto.Se Inhibitor

Retinal Polyol and Myo-lnosifol in Galactosemic Dogs Given an Aldose-Reducto.se Inhibitor

Timothy 5. Kern and Ronald L. Engerman

Galactitol and myo-\nosito\ concentrations were measured in retinas, erythrocytes, and skeletal muscle of experimentally galactosemic dogs receiving a placebo or the aldose reductase inhibitor, sorbinil, for 5 yr. The concentration of galactitol was increased more than 30-fold in the retina and other tissues by galactosemia, and the increase was inhibited 90-96% in all tissues by sorbinil. The concentration of free myo-inos\to\ was greater than normal in retinas of galactosemic dogs, and its concentration was not altered by the aldose-reductase inhibitor. The mjw-inositol concentration likewise was greater than normal in the retinas of dogs that were diabetic for 2-4 months. The marked inhibition of polyol production and accumulation in the retina of sorbinil-treated galactosemic dogs was not associated with a comparable inhibition of retinopathy. Invest Ophthalmol Vis Sci 32:3175-3177,1991

Experimental elevation of blood galactose concen- to receive either the aldose reductase inhibitor, sor- tration in normal dogs leads to a retinopathy that is binil, or to remain untreated. The dogs were fed twice morphologically similar to that in diabetic dogs and daily (8 AM and 6 PM) to maintain blood galactose humans.12 One mechanism by which hyperglycemia levels elevated as high as possible throughout the day. in diabetes or experimental galactosemia might cause Sorbinil was given orally twice a day, usually at a dose retinopathy is through excessive polyol production of 60-80 mg/kg/day, 30 min before each feeding. The and accumulation. This hypothesis is being investi- dose of the drug was selected for maximal inhibition gated with aldose reductase inhibitors in numerous of erythrocyte galactitol accumulation without caus- studies involving diabetic and galactosemic dogs3"5 ing significant systemic effects (data not shown). The 6 7 and rats and diabetic humans. The conclusions dose was adjusted monthly for changes in body reached by these studies have been contradictory. In weight. Erythrocyte hexitol was measured six times a none of these studies, however, was the extent to year throughout the 5 yr of the study.8 All blood and which retinal polyol production was inhibited quanti- tissues samples were collected after an overnight fast fied. We report the effect of long-term experimental and 12-14 hr after the preceding dose of sorbinil. galactosemia on retinal polyol and myo-inositol con- A special effort was made to ensure that the two centrations in dogs and the effect of an aldose-reduc- experimental groups were comparably galactosemic tase inhibitor on such metabolic abnormalities. because the severity of the hyperglycemia is a major determinant of the development of the retinopathy.2 Materials and Methods Concentrations of hemoglobin A, (Isolab, Akron, This experiment conformed to the ARVO Resolu- OH) and nonenzymatically glycated plasma protein tion on Use of Animals in Research. Normal dogs, 1 xh (fructosamine; Roche, Nutley, NJ) were measured six to 2'/2 yr of age without ocular disease, were assigned times a year, the 24-hr excretion of reducing sugar in randomly to the experimental galactosemic group urine was quantified twice a week, and the consump- (with a 30% galactose diet) or to the control group. tion of the galactose diet was measured twice daily. Galactosemic animals were subdivided prospectively One eye and samples from peroneal nerve and biceps femoris were obtained surgically at 42 months of galactosemia, and the other eye was removed at autopsy From the Department of Ophthalmology, University of Wiscon- (at 60 months of galactosemia). Most of each retina sin-Madison, Madison, Wisconsin. Supported in part by Public Health Service research grant was used for histologic analysis, and the small samples EY00300 from the National Eye Institute and by an unrestricted of retina available for biochemical assay were pooled grant from Research to Prevent Blindness, Inc. by animal (samples collected at 42 and 60 months) Disclosure of Proprietary Interest: None. and frozen (-70°) until assay. Retinal galactitol and Submitted for publication: March 13, 1991; accepted July 12, myoinositol were quantified by gas chromatography 1991. 9 Reprint requests: Timothy S. Kern, PhD, Department of Oph- and mass spectroscopy. Our conclusions were not af- thalmology, University of Wisconsin-Madison, 1300 University Av- fected whether the results were expressed relative to enue, Madison, WI 53706-1532. DNA or protein. Galactitol in other tissues was quan-

3175

Downloaded from iovs.arvojournals.org on 09/29/2021 3176 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / December 1991 Vol. 32

Table 1. Comparison of severity ofgalactosemia in placebo- and sorbinil-treated galactosemic dogs Duration HbA, Glycated plasma Group n (mos) protein (nmol/g) Galactose + placebo 9 60 6.4:t0.4 66 ±6 Galactose + sorbinil 10 60 6.7:1:0.4 64 ±4 Normal 6 60 5.7:h0.2 55 ±4

tified by gas chromatography of trimethylsilyl deriva- The concentration of galactitol in erythrocytes tives on capillary columns.8 tended to correlate with retinal polyol concentration Other dogs were made diabetic with alloxan and in the same animals. In sorbinil-treated galactosemic kept in poor glycemic control for 2-4 months. Insulin dogs only, or in all galactose-fed dogs (placebo- and was given daily in doses insufficient to prevent sorbinil-treated dogs pooled together), the retinal ga- chronic hyperglycemia and glucosuria (0.3-1.3 units/ lactitol concentration was correlated significantly kg/day). Retinal polyol (sorbitol) and myo-inositol in with the erythrocyte galactitol concentration just be- the diabetic and nondiabetic dogs (both, n = 5) were fore death (P < 0.05 and P <0.01, respectively) and quantified as described. the average concentration during the 5 yr of study (P Comparisons of the experimental groups by stu- < 0.05 and P <0.01, respectively). The correlation dent t-test and the nonparametric Mann-Whitney U between retinal and erythrocytic galactitol in the pla- test gave similar conclusions. The results were ex- cebo-treated galactosemic group was not statistically pressed as the mean ± the standard deviation. significant. The concentration of free myo-inositol in the retina Results was higher than normal in the galactosemic dogs (Ta- ble 2; P < 0.05). Sorbinil administration throughout The two galactosemic groups were comparably ga- the 5 yr ofgalactosemia did not alter the elevated reti- lactosemic, as indicated by their similarity with re- nal myo-inositol concentration, despite its effect on spect to hemoglobin A, and nonenzymatically gly- the galactitol concentration. cated plasma protein (Table 1). Likewise, the amount of reducing sugar excreted daily and the amount of Dogs diabetic for 2-4 months likewise had a higher galactose consumed daily were similar between the than normal retinal myo-inositol concentration (non- two galactosemic groups. diabetic, 98 nmol/mg DNA ± 14; diabetic, 132 ± 14; P < 0.01). The concentration of polyol in the retinas The galactitol concentration in the placebo-treated of the diabetic dogs (4.3 nmol/mg DNA ± 2.0) was galactosemic group was more than 30-fold greater much less than in the galactosemic group. than normal in each of the tissues examined, and sorbinil inhibited hexitol accumulation in each tissue by Discussion 90-96% over the 5 yr of the study (Table 2; both, P < 0.01). The hexitol elevation in the retinas of the A 30% galactose diet in these dogs increased the galactosemic dogs was inhibited 96% by the aldose hexitol concentration in the retina and other tissues reductase inhibitor. The actual amount of inhibition by more than 30-fold. This accumulation of polyol each day probably was greater, because the tissues was much greater than that observed in diabetes, a were collected long after (12-14 hr) the preceding result perhaps of the poor metabolism of galactitol dose of sorbinil. No galactitol was detected in the ret- compared with sorbitol. This excess production and inas from three of seven sorbinil-treated galactosemic accumulation of polyol probably is concentrated in dogs. relatively few retinal cell types, particularly ganglion

Table 2. Effect of sorbinil on hexitol (galactitol + sorbitol), galactitol and myo-inositol concentrations in tissues of representative galactose-fed dogs Retina% Erythrocyte* Skeletal muscle^ (nmol hexitol (nmol hexitol (nmol galactitol (nmol myoinositol pergHb) per g protein) per mg DNA) per mg DNA) Galactose + placebo 1475 ±517 (9) 222 ± 148 (5) 23.3 ± 7.2 (6) 101 ± 17(6) Galactose + sorbinil 134 ± 34(10) 25 ± 9 (5) 1.4 ± 1.7(7) 112 ±24 (7) Normal 43 ± 11 (6) Not detected (2) 0.5 ± 0.9 (5) 75 ± 14(5)

n in parenthesis. t At 42 months. * Six assays/year for 5 years. i Forty-two- and 60-month samples pooled by animal.

Downloaded from iovs.arvojournals.org on 09/29/2021 No. 13 RETINAL SUGARS IN GALACTOSEMIA / Kern and Engerman 3177

and microvascular cells, such as endothelia and peri- failure of an aldose reductase inhibitor to inhibit (or cytes, known in dogs to possess aldose reductase.8'10"12 even tend to inhibit) the severity of retinopathy de- The accumulation of polyol in diabetes or galacto- spite 96% inhibition of polyol accumulation in the semia is associated with a marked loss of mj/0-inositol retinas of these galactosemic dogs raises important in several tissues including the lens and peripheral questions about the role of the polyol pathway in the nerve.13 Polyol accumulation is associated with the pathogenesis of diabetic-like retinopathy. loss of myo-inositol also from the retina of diabetic 91415 Key words: galactosemia, aldose reductase inhibitor, retina, rats and rabbits. According to our observations, polyol, rayoinositol, dog however, rayoinositol was not lost from the retinas of galactosemic or diabetic dogs. To our knowledge, Acknowledgments there is no evidence that the retinopathy that occurs The authors thank Pfizer, Inc., Groton, Connecticut, for in galactosemic or diabetic dogs is related to a loss of providing sorbinil; Drs. J. Williamson and R. Tilton, Wash- myoinositol. It is possible that a small subset of cells ington University, St. Louis, Missouri, for assistance in anal- or an intracellular pool of myo-inositol in the retinas ysis of retinal sugars; and M. Larson and M. Garment for of galactosemic or diabetic dogs might become myo- skillful technical assistance. inositol deficient. References The present study shows that sorbinil effectively crosses the blood-retinal barrier and inhibits polyol 1. Engerman RL and Kern TS: Experimental galactosemia pro- duces diabetic-like retinopathy. Diabetes 31:26A, 1982. production in the retina. The drug inhibited polyol 2. Engerman RL and Kern TS: Experimental galactosemia pro- accumulation by 90-96% in the retinas and other tis- duces diabetic-like retinopathy. Diabetes 33:99, 1984. sues of galactosemic dogs. A similar degree of inhibi- 3. Engerman RL and Kern TS: Determinants of experimental tion of polyol accumulation in the erythrocytes and diabetic retinopathy. In Diabetes 1988, Larkins R, Zimmet D, other tissues of these animals suggests that the percent and Chisholm D, editors. New York, Elsevier, 1989, pp. 221 — 224. inhibition of erythrocyte polyol levels by sorbinil may 4. Engerman RL and Kern TS: Effect of sorbinil on retinopathy offer a convenient estimate of the drug's effect on al- in diabetic dogs and galactosemic dogs. ARVO Abstracts. In- dose reductase activity in the retina, at least in galac- vest Ophthalmol Vis Sci 31(Suppl):124, 1990. tosemia. 5. Kador PF, Akagi Y, Takahashi Y, Ikebe H, Wyman M, and Despite the marked inhibition of polyol production Kinoshita JH: Prevention of retinal vessel changes associated with diabetic retinopathy in galactose-fed dogs by aldose reduc- and accumulation in the retina of sorbinil-treated ga- tase inhibitors. Arch Ophthalmol 108:1301, 1990. lactosemic dogs, we observed no inhibition of retinop- 6. Robison WG, Jr, Nagata M, Laver N, Hohman TC, and Kino- athy in these animals.3'4 None of the characteristic shita JH: Diabetic-like retinopathy in rats prevented with an lesions of diabetic or galactose-induced retinopathy, aldose reductase inhibitor. Invest Ophthalmol Vis Sci 30:2285, including microaneurysms, pericyte ghosts, or retinal 1989. 7. Sorbinil Retinopathy Trial Research Group: A randomized capillary basement membrane thickening, were inhib- trial of sorbinil, an aldose reductase inhibitor, in diabetic reti- ited quantitatively in our sorbinil-treated dogs at ei- nopathy. Arch Ophthalmol 108:1234, 1990. ther 42 or 60 months of galactosemia. Others, how- 8. Kern TS and Engerman RL: Hexitol production by canine ever, concluded that the development of retinopathy retinal microvessels. Invest Ophthalmol Vis Sci 26:382, 1985. in galactose-fed dogs was delayed by aldose reductase 9. Eades DM, Williamson JR, and Sherman WR: Rapid analysis 5 of sorbitol, galactitol and mw-inositol from biological sources. inhibitors. The difference in the conclusions between JChromatogr 490:1, 1989. these two studies seems not related to differing polyol 10. Kern TS and Engerman RL: Distribution of aldose reductase levels; blood polyols were inhibited at least as much in in ocular tissues. Exp Eye Res 33:175, 1981. our study as in that by Kador et al.5 It was not until 11. Kennedy A, Frank RN, and Varma SD: Aldose reductase activ- several hours after administration of the drug that ity in retinal and cerebral microvessels and cultured vascular cells. Invest Ophthalmol Vis Sci 24:1250, 1983. they inhibited blood polyols to a degree similar to the 12. Akagi Y, Terubayashi H, Millen J, Kador PF, and Kinoshita 93% inhibition observed in our dogs. JH: Aldose reductase localization in dog retinal mural cells. Because galactitol is poorly metabolized, its con- Curr Eye Res 5:883, 1986. centration in a tissue might provide an estimate of the 13. Stewart MA, Kurien MW, Sherman WR, and Cotlier EV: Inosi- amount of flux through aldose reductase in that tis- tol changes in nerve and lens of galactose fed rats. J Neurochem 15:941, 1968. sue. The higher-than-normal concentration of hexitol 14. Tilton RG, Chang K, Weigel C, Eades D, Sherman WR, Kilo in the retinas of our sorbinil-treated galactosemic C, and Williamson JR: Increased ocular blood flow and l25I-al- dogs thus suggests that the flux through aldose reduc- bumin permeation in galactose-fed rats: Inhibition by sorbinil. tase in these animals was not normalized entirely and Invest Ophthalmol Vis Sci 29:861, 1988. may be up to threefold higher than normal for part of 15. MacGregor LC, Rosecan LR, Laties AM, and Matchinsky FM: Altered retinal metabolism in diabetes: I. Microanalysis of the day. The possibility that failure to attain 100% lipid, glucose, sorbitol, and myo-inositol in the choroid and in inhibition of polyol production might be sufficient to the individual layers of the rabbit retina. J Biol Chem cause retinopathy cannot be excluded. However, the 261:4046, 1986.

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