Pathogenesis of diabetic retinopathy.

Diabetic retinopathy involves anatomic changes in retinal vessels and neuroglia. The pathogenetic mechanism responsible for retinopathy is imperfectly understood, but much of the mechanism is apparently reproduced by experimental diabetes in animals and by chronic elevation of blood galactose in nondiabetic animals. The evidence that retinopathy is a consequence of excessive blood sugars and their sequelae is consistent with a demonstrated inhibition of retinopathy by strict glycemic control in diabetic dogs. However, retinopathy in the dog model has shown a tendency to resist intervention by strict control. Biochemical and pathophysiological sequelae of hyperglycemia possibly critical to the development of retinopathy in humans and animal models are being studied in many laboratories. Retinopathy occurs in experimental galactosemia in the absence of the renal hypertrophy, mesangial expansion, and glomerular obliteration typical of diabetes in humans and dogs, implying that retinopathy and nephropathy differ appreciably in pathogenesis.

Aldose reductase inhibition fails to prevent retinopathy in diabetic and galactosemic dogs.

To investigate a possible role of excessive polyol production in the pathogenesis of diabetic retinopathy, 16 ALX-induced diabetic dogs and 20 experimentally galactosemic dogs were randomly assigned to 5 yr of treatment with either sorbinil, an aldose reductase inhibitor, or a placebo. The severity of hyperglycemia in sorbinil-treated and placebo groups was monitored throughout the 5-yr study by assay of glycosuria and nonenzymatically glycated plasma protein and HbA1 needed in an effort to avoid confounding possible group differences in hyperglycemia severity with possible drug effects. Inhibition of polyol production by sorbinil was monitored in erythrocytes throughout the study and also in retina and other tissue obtained at autopsy. Trypsin digests of retinal vessels were compared after 60 mo of diabetes and after 42 and 60 mo of galactosemia. In diabetic dogs, development of retinopathy was not significantly influenced by a sorbinil dose (20 mg.kg-1 x day-1) sufficient to prevent elevation of sorbitol levels in retina and other tissue. Likewise, in dogs made experimentally galactosemic for 42-60 mo, administration of sorbinil (60-80 mg.kg-1 x day-1) had no significant effect on the development of retinopathy notwithstanding prevention of 93-96% of the polyol elevation in retina and other tissue. Retinal capillary basement membrane was significantly thicker than normal in diabetic and in galactosemic dogs and was not significantly influenced by administration of sorbinil in either dog model. Thus, no evidence was found that the development of retinopathy is critically dependent on excessive polyol production or accumulation.

Experimental galactosemia produces diabetic-like retinopathy.

Six normal dogs were made galactosemic by feeding a 30% D-galactose diet, and were followed up to 5 yr. For comparison, 10 normal dogs and 10 alloxan-diabetic dogs were concurrently fed the diet less the galactose supplement. Retinopathy occurred in each of four dogs glactosemic 3 or more yr, and was absent at lesser durations of galactosemia, and from normal dogs not given the galactose supplement. The retinopathy was marked by saccular capillary aneurysms, hemorrhages, nonperfused or acellular vessels, tortuous hypertrophic capillaries, loss of capillary pericytes, and other lesions typical of diabetic patients and alloxan-diabetic dogs. In galactose-fed dogs, blood galactose varied between 0 (fasted) and 250 mg/dl (postprandial), and glycosylated hemoglobin levels became supranormal. In contrast to diabetic dogs, blood levels of glucose, free fatty acids, and branched-chain amino acids were not elevated in the galactosemic dogs, and their serum insulin seemed normal. The results suggest that the level of blood hexose is itself an important determinant of retinopathy.

Platelet aggregation in experimental diabetes and experimental galactosemia.

Platelet aggregation and related plasma factors have been studied in experimentally diabetic dogs, experimentally galactosemic dogs, and in normal dogs. Platelet aggregation, when induced in vitro by ADP (up to 22 microM) or collagen (up to 4 micrograms/ml), was not significantly different from normal in the diabetic or galactosemic dogs. Plasma (platelet-deficient) from the diabetic or galactosemic dogs did not enhance ADP-induced aggregation of normal canine platelets. Fibrinogen concentration in blood tended to be elevated in hyperglycemic diabetic dogs, and was significantly correlated with the levels of HbA1 and blood glucose, whereas galactosemic animals had normal fibrinogen concentrations. The quantity of 6-keto prostaglandin F1 alpha released from aortas of diabetic dogs was less than that released from aortas of normal dogs. Since alloxan-diabetic dogs and galactosemic dogs develop microaneurysms, occluded capillaries, and other retinal lesions characteristic of diabetic retinopathy in human patients, these studies suggest that retinopathy can develop in the absence of extraordinary in vitro platelet aggregation.

Progression of incipient diabetic retinopathy during good glycemic control.

To assess the extent to which the progression of diabetic retinopathy can be arrested by improved glycemic control, 35 normal dogs were randomly divided into a nondiabetic and three alloxan-induced diabetic groups prospectively identified according to glycemic control: poor control for 5 yr (PC), good control for 5 yr (GC), and poor control for 2.5 yr followed by good control for 2.5 yr (PGC). To achieve good control, insulin was given twice daily together with a measured diet so that hyperglycemia and glucosuria were mild and infrequent, and HbA1 was comparable to normal. Retinal capillary aneurysms and other lesions developed during 60 mo of poor control (group PC) and were inhibited if good control was begun promptly within 2 mo (group GC). In group PGC, retinopathy was absent or equivocal at 2.5 yr of poor control and, surprisingly, was found to develop subsequently despite good glycemic control. Retinopathy in group PGC was greater at autopsy than at 2.5 yr and was greater than in group GC. The results indicate that retinopathy may be preventable but tends to resist arrest even in its incipient stages, before more than the first few aneurysms have appeared.

Dogs with induced or spontaneous diabetes as models for the study of human diabetes mellitus.

Diabetes mellitus can be produced readily in dogs by any of several chemical or surgical methods, and in recent years genetically diabetic dogs have also become available for study. These models are suitable for investigating a wide variety of questions relevant to human diabetes mellitus. Especially noteworthy is the occurrence in diabetic dogs of a number of ocular and other complications typical of human diabetes mellitus. Procedures by which canine models of diabetes may be established or maintained have been outlined in this report.

Pharmacological inhibition of diabetic retinopathy: aminoguanidine and aspirin.

Effects of aminoguanidine and aspirin on the development of retinopathy have been examined in 5-year studies of diabetic dogs. Either agent was administered daily in doses of 20-25 mg. kg(-1). day(-1). Because severity of hyperglycemia greatly influences development of the retinopathy, special effort was devoted to maintaining comparable glycemia in experimental and control groups. The retinal vasculature was isolated by the trypsin digest method, and retinopathy was assessed by light microscopy. Diabetes for 5 years resulted, as expected, in saccular capillary aneurysms, pericyte ghosts, acellular capillaries, retinal hemorrhages, and other lesions. Administration of aminoguanidine essentially prevented the retinopathy, significantly inhibiting the development of retinal microaneurysms, acellular capillaries, and pericyte ghosts compared with diabetic controls. Aspirin significantly inhibited the development of retinal hemorrhages and acellular capillaries over the 5 years of study, but had less effect on other lesions. Although diabetes resulted in significantly increased levels of advanced glycation end products (AGEs) (namely, pentosidine in tail collagen and aorta, and Hb-AGE), aminoguanidine had no significant influence on these parameters of glycation. Nitration of a retinal protein was significantly increased in diabetes and inhibited by aminoguanidine. The biochemical mechanism by which aminoguanidine has inhibited retinopathy thus is not clear. Aminoguanidine (but not aspirin) inhibited a diabetes-induced defect in ulnar nerve conduction velocity, but neither agent was found to influence kidney structure or albumen excretion.

Kidney morphology in experimental hyperglycemia.

To evaluate the role of hyperglycemia in the pathogenesis of diabetic nephropathy, the kidneys from dogs experimentally galactosemic for 5 yr have been compared with the kidneys from age-matched normal dogs and dogs with alloxan-induced diabetes for 5 yr. The width of glomerular capillary basement membrane and the quantity of plasma protein immunohistochemically demonstrable in the basement membrane were supranormal in the galactosemics, as they were in the diabetics. In contrast, kidney weight, mesangial volume, and the prevalence of obliterated glomeruli, glomerular exudates, and mesangial nodules in the galactosemic animals were comparable to those of normal animals and clearly were less than observed in the insulin-deficient diabetic animals. These galactosemic dogs are known to have developed a retinopathy morphologically indistinguishable from that of diabetic patients and dogs. Thus, galactosemia sufficient to produce diabetic-like lesions in the glomerular basement membrane and retina was found to be nevertheless insufficient to elicit several renal abnormalities that are typical of diabetes. The polyol concentration in erythrocytes was greater than normal in the galactosemics and the diabetics and was greatest in the galactosemics. The absence of mesangial expansion, glomerular obliteration, and nephromegaly in galactose-fed dogs raises the possibility that these abnormalities in diabetes are not a result of excessive polyol pathway activity.

Relationship of microvascular disease in diabetes to metabolic control.

Dogs were made alloxan-diabetic and randomly distributed into either of two prospective treatment groups. In one group it was intended that the metabolic signs of diabetes be controlled poorly, and commercial insulin was administered in doses inadequate to prevent chronic, severe hyperglycemia and glucosuria. In the other group it was intended that the metabolic disorder be well controlled, and the animals received food and commercial insulin twice daily such that the hyperglycemia and glucosuria became mild or infrequent. Experimental improvement of the carbohydrate disorder was accompanied by amelioration of hyperlipemia and other clinical signs of deficient insulin activity. By 60 months of diabetes, retinal capillary aneurysms, pericyte ghosts, obliterated vessels, and other microvascular abnormalities typical of diabetes were apparent in each animal of the poor-control group. Better control was found to reduce significantly the incidence and severity of microvascular lesions. The data suggest that the mechanism responsible for diabetic retinopathy is initiated as a result of deficient insulin activity and that the development of the microvascular complications of diabetes are preventable and may be inhibited by careful control of the metabolic disorder.

Abnormalities of retinal metabolism in diabetes or experimental galactosemia. III. Effects of antioxidants.

Effects of antioxidants on hyperglycemia-induced alterations of retinal metabolism were evaluated in rats diabetic or experimentally galactosemic for 2 months. Oxidative stress was estimated by measuring lipid peroxides (measured as thiobarbituric acid reactive substances [TBARS]) in retina and plasma. Erythrocyte osmotic fragility, another measure of oxidative stress, also was determined in the same groups of rats. In diabetic rats, TBARS were elevated by 74% in retina and 87% in plasma. In galactose-fed rats, TBARS were significantly elevated in retina (P < 0.05), but were normal in plasma. The administration of supplemental dietary ascorbic acid and alpha-tocopherol acetate for 2 months prevented the elevation of retinal TBARS and the decrease of Na(+)-K(+)-ATPase and calcium ATPase activities in retinas of diabetic animals without having any beneficial effect on plasma TBARS. In galactosemic rats, these antioxidants had a partial beneficial effect on the activity of retinal Na(+)-K(+)-ATPase, but failed to have any effect on calcium ATPase. The beneficial effects of antioxidants in diabetes and experimental galactosemia were not caused by the amelioration of hyperglycemia or retinal polyol accumulation. Erythrocyte osmotic fragility was increased by more than twofold in diabetes, but was normal in experimental galactosemia, and antioxidants prevented diabetes-induced increases in erythrocyte osmotic fragility-Diabetes-induced increased oxidative stress and subnormal ATPase activities in the retina can be inhibited by dietary supplementation with antioxidants.

Characterization of the mechanism for the chronic activation of diacylglycerol-protein kinase C pathway in diabetes and hypergalactosemia.

Similar vascular pathological conditions are observed in diabetic animals and those with diet-induced hypergalactosemia. Both diabetes and hypergalactosemia are believed to cause vascular dysfunction via a common biochemical mechanism. In this study, we have found that both diabetes and hypergalactosemia in the short term (2-4 months) can increase total diacylglycerol (DAG) levels by 52 +/- 9 and 74 +/- 13% in the retina and aorta, respectively, of diabetic dogs, and by 94 +/- 9 and 78 +/- 11% in the retina and aorta, respectively, in dogs with hypergalactosemia as compared with normal control animals (P < 0.01). The elevation of DAG levels was maintained for 5 years in the aortas of diabetic and hypergalactosemic dogs. To characterize the mechanism of the DAG increases, we have determined that total DAG levels were significantly increased in cultured macro- and microvascular cells exposed to elevated glucose (22 mM) and galactose (16.5 mM) levels. These increased levels were not prevented by sorbinil, an aldose reductase inhibitor. One of the sources of the increased DAG levels was probably derived from de novo synthesis from both hexoses as determined by radiolabeling studies. Intracellularly, the DAG elevation activated protein kinase C (PKC) activity with increases of 58 +/- 12% (P < 0.05) and 66 +/- 8% (P < 0.01) in the membrane fraction of cultured aortic smooth muscle cells exposed to elevated glucose and galactose levels, respectively. These findings have clearly demonstrated a possible common biochemical mechanism by which hyperglycemia and hypergalactosemia can chronically activate the DAG-PKC pathway in the vasculature and could be a possible explanation for the development of diabetic vascular complications.

Evidence of a glycemic threshold for the formation of pentosidine in diabetic dog lens but not in collagen.

The relationship between long-term glycemic control and the advanced Maillard reaction was investigated in dura mater collagen and lens proteins from dogs that were diabetic for 5 years. Diabetic dogs were assigned prospectively to good, moderate, and poor glycemic control and maintained by insulin. Biochemical changes were determined at study exit. Mean levels of collagen digestibility by pepsin decreased (NS) whereas collagen glycation (P < 0.001), pentosidine cross-links (P < 0.001), and collagen fluorescence (P = 0.02) increased with increasing mean HbA1 values. Similarly, mean levels of lens crystallin glycation (P < 0.001), fluorescence (P < 0.001), and the specific advanced lens Maillard product 1 (LM-1) (P < 0.001) and pentosidine (P < 0.005) increased significantly with poorer glycemic control. Statistical analysis revealed very high Spearman correlation coefficients between collagen and lens changes. Whereas pentosidine cross-links were significantly elevated in collagen from diabetic dogs with moderate levels of HbA1 (i.e., 8.0 +/- 0.4%), lens pentosidine levels were normal in this group and were elevated (P < 0.001) only in the animals with poor glycemic control (HbA1 = 9.7 +/- 0.6%). Thus, whereas protein glycation and advanced glycation in the extracellular matrix and in the lens are generally related to the level of glycemic control, there is evidence for a tissue-specific glycemic threshold for pentosidine formation, i.e., glycoxidation, in the lens. This threshold may be in part linked to a dramatic acceleration in crystallin glycation with HbA1 values of > 8.0% and/or a loss of lens membrane permeability. This study provides support at the molecular level for the growing concept that glycemic thresholds may be involved in the development of some of the complications in diabetes.

Abnormalities of retinal metabolism in diabetes or experimental galactosemia: V. Relationship between protein kinase C and ATPases.

In the retinas of diabetic animals, protein kinase C (PKC) activity is elevated, and Na+-K+-ATPase and calcium ATPase activities are subnormal. These abnormalities are also present in another model of diabetic retinopathy, experimental galactosemia. We have investigated the relationship between hyperglycemia-induced abnormalities of PKC and ATPases using a selective inhibitor of beta isoform of PKC (LY333531). Diabetes or experimental galactosemia of 2 months' duration resulted in > 50% elevation of PKC activity in the retina, and administration of LY333531 prevented the elevation. In retinas of the same rats, the LY333531 prevented hyperglycemia-induced decreases of both Na+-K+-ATPase and calcium ATPase activities. Retinal microvessels, the main site of lesions in diabetic retinopathy, likewise showed elevated activity of PKC and inhibition of ATPases in diabetes and in experimental galactosemia, and administration of LY333531 to diabetic animals prevented these abnormalities. PKC activity in sciatic nerves, in contrast, became subnormal in diabetes and experimental galactosemia, and LY333531 had no effect on PKC activity in the sciatic nerve. PKC activity in the cerebral cortex was not affected by diabetes or experimental galactosemia. The results suggest that diabetes-induced reductions in Na+-K+-ATPase and calcium ATPase in the retina are mediated in large part by PKC-beta. The availability of an agent that can normalize the hyperglycemia-induced increase in PKC activity in the retina should facilitate investigation of the role of PKC in the development of diabetic retinopathy.

Ocular complications.

Ocular complications of diabetes in humans are reviewed briefly, and experimental models available for study of the complications are described. Potentially suitable models include not only diabetic animals, but also nondiabetic animals in which analogous lesions have been demonstrated. Many abnormalities of the lens, cornea, iris, and retina comparable to those of diabetes in humans may be observed in diabetic animals, although all abnormalities are not necessarily observed in every species. Retinal changes, in particular, may occur in diabetic animals of several species, but only in large animals (dogs, primates) have saccular capillary aneurysms been reproduced consistently, together with other retinal changes typical of diabetes in humans. A few examples of the uses of animal models are offered, and attention is called to a lack of animal models of proliferative diabetic retinopathy and of rubeosis iridis.

Abnormalities of retinal metabolism in diabetes or experimental galactosemia. IV. Antioxidant defense system.

Activities of enzymes that protect the retina from reactive oxygen species were investigated in experimentally diabetic rats and experimentally galactosemic rats, two animal models known to develop vascular lesions consistent with diabetic retinopathy. Diabetes or experimental galactosemia of 2 months duration significantly decreased the activities of glutathione reductase and glutathione peroxidase in the retina while having no effect on the glutathione synthesizing enzymes glutathione synthetase and gamma-glutamyl cysteine synthetase. Activities of two other important antioxidant defense enzymes-superoxide dismutase (SOD) and catalase-also were decreased (by more than 25%) in retinas of diabetic rats and galactosemic rats. Administration of supplemental antioxidants, vitamins C and E, for the 2 months prevented the diabetes-induced impairment of antioxidant defense system in the retina. In experimentally galactosemic rats, the supplemental antioxidants were not as effective: SOD activity was normalized, but the enzymes of the glutathione redox cycle were only partly restored, and the subnormal catalase activity was unaffected. Diabetes or experimental galactosemia results in significant impairment of the antioxidant defense system in the retina, and exogenous antioxidant supplementation can help alleviate the subnormal activities of antioxidant defense enzymes.

Abnormalities of retinal metabolism in diabetes or experimental galactosemia. VI. Comparison of retinal and cerebral cortex metabolism, and effects of antioxidant therapy.

Metabolic abnormalities observed in retina and in cerebral cortex were compared in diabetic rats and experimentally galactosemic rats. Diabetes or experimental galactosemia of 2 months duration significantly increased oxidative stress in retina, as shown by elevation of retinal thiobarbituric acid reactive substances (TBARS) and subnormal activities of antioxidant defense enzymes, but had no such effect in the cerebral cortex. Activities of sodium potassium adenosine triphosphatase [(Na,K)-ATPase] and calcium ATPase became subnormal in retina as well as in cerebral cortex. In contrast, protein kinase C (PKC) activity was elevated in retina but not in cerebral cortex in the same hyperglycemic rats. Dietary supplementation with an antioxidant mixture (containing ascorbic acid, Trolox, alpha-tocopherol acetate, N-acetyl cysteine, beta-carotene, and selenium) prevented the diabetes-induced and galactosemia-induced elevation of retinal oxidative stress, the elevation of retinal PKC activity and the decrease of retinal ATPases. In cerebral cortex, administration of the antioxidant diet also prevented the diabetes-induced decreases in (Na,K)-ATPase and calcium ATPases, but had no effect on TBARS and activities of PKC and antioxidant-defense enzymes. The results indicate that retina and cerebral cortex differ distinctly in their response to elevation of tissue hexose, and that cerebral cortex is more resistant than retina to diabetes-induced oxidative stress. The greater resistance to oxidative stress in cerebral cortex, as compared to retina, is consistent with the resistance of cerebral cortex to microvascular disease in diabetes, and with a hypothesis that oxidative stress contributes to microvascular disease in diabetes. Dietary supplementation with these antioxidants offers a means to inhibit multiple hyperglycemia-induced retinal metabolic abnormalities.

Hexitol production by canine retinal microvessels.

Hexitol-producing activity has been quantitated in microvessels isolated from retina and cerebral cortex of dogs, a species known to develop a retinopathy similar to that seen in diabetic patients. Both the retinal and cerebral microvessels produce galactitol from galactose, but at a rate several-fold less than that seen in canine lens epithelium. This hexitol-producing activity of the microvessels is not an artifact of erythrocyte contamination and can be inhibited by the aldose reductase inhibitor, Sorbinil.

Permeability and patency of retinal blood vessels in experimental diabetes.

Increased permeability of retinal blood vessels in human diabetic retinopathy is well known clinically. Its morphologic equivalent is unknown. In dogs with 5 years of poorly controlled alloxan diabetes and nonproliferative diabetic retinopathy comparable to that of man, permeability and patency of retinal blood vessels were tested with the protein tracer horseradish peroxidase and evaluated by electron microscopy. A breakdown of the blood-retinal barrier was found associated with extensive tracer leakage around retinal blood vessels. Tracer had seemingly permeated endothial junctions, and was not transported through the endothelial cytoplasm. Blood vessels which had lost their endothelial cells and were partially occluded by glial cells retained some patency to tracer. These findings suggest the following. (1) Endothelial tight junctions are not a static cell specialization but one that can open due to chronic metabolic or osmotic factors prevailing in diabetes. Opened tight junctions may account for plasma leakage seen clinically in human diabetic retinopathy. (2) In the absence of endothelial cells perfusion does not necessarily end abruptly. The tracer method and electron microscopy may show details of vascular obstruction that are not readily demonstrated clinically.

Retinal polyol and myo-inositol in galactosemic dogs given an aldose-reductase inhibitor.

Galactitol and myo-inositol 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-inositol was greater than normal in retinas of galactosemic dogs, and its concentration was not altered by the aldose-reductase inhibitor. The myo-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.

Galactose-induced retinal microangiopathy in rats.

PURPOSE: The suitability of the galactose-fed rat as a model of diabetic retinopathy was examined in nondiabetic rats fed diets enriched with either 30% or 50% galactose for up to 2 years. METHODS: Retinal capillaries were examined by light and electron microscopy, and the prevalence or severity of diabetic-like lesions was quantitated. RESULTS: Histologic evaluation of trypsin digests of retina revealed significantly greater than normal frequencies of pericyte ghosts and acellular capillaries at both 15 and 23 months receiving a 50% galactose diet. Similar lesions were observed in rats receiving a 30% galactose diet for 23 months. Capillary basement membrane thickening, dilated hypercellular capillaries (or intra-retinal microvascular abnormalities), and foci of vascular cells appeared in rats fed 50% galactose, but saccular microaneurysms characteristic of retinopathy in diabetic patients, diabetic dogs, and experimentally galactosemic dogs were not observed. Administration of the aldose reductase inhibitor, Sorbinil, to rats fed 50% galactose resulted in a significant inhibition of cataract and of galactitol accumulation in nerve and blood (by more that 90%) and retina (by 62%), but did not inhibit development of the retinal microvascular lesions. CONCLUSIONS: Two years of galactosemia in rats seems to reproduce only a portion of the lesions characteristic of diabetic retinopathy in patients or dogs. Nevertheless, lesions characteristic of at least the early stages of retinopathy clearly do develop in this galactosemic rat model, and are not restrained by inhibition of retinal polyol accumulation by 62%.