Characterization of Oxidative Stress in Various Tissues of Diabetic and Galactose-fed Rats

Rats fed a galactose-rich diet have been used for several years as a model for diabetes to study, particularly in the eye, the effects of excess blood hexoses. This study sought to determine the utility of galactosemia as a model for oxidative stress in extraocular tissues by examining biomarkers of oxidative stress in galactose-fed rats and experimentally-induced diabetic rats. Sprague-Dawley rats were divided into four groups: experimental control; streptozotocin-induced diabetic; insulin-treated diabetic; and galactose-fed. The rats were maintained on these regimens for 30 days, at which point the activities of catalase, glutathione peroxidase, glutathione reductase, and superoxide dismutase, as well as levels of lipid peroxidation and reduced and oxidized glutathione were determined in heart, liver, and kidney. This study indicates that while there are some similarities between galactosemic and diabetic rats in these measured indices of oxidative stress (hepatic catalase activity levels and hepatic and renal levels of oxidized glutathione in both diabetic and galactosemic rats were significantly decreased when compared to normal), overall the galactosemic rat model is not closely parallel to the diabetic rat model in extra-ocular tissues. In addition, several effects of diabetes (increased hepatic glutathione peroxidase activity, increased superoxide dismutase activity in kidney and heart, decreased renal and increased cardiac catalase activity) were not mimicked in galactosemic rats, and glutathione concentration in both liver and heart was affected in opposite ways in diabetic rats and galactose- fed rats. Insulin treatment reversed/prevented the activity changes in renal and cardiac superoxide dismutase, renal and cardiac catalase, and hepatic glutathione peroxidase as well as the hepatic changes in lipid peroxidation and reduced and oxidized glutathione, and the increase in cardiac glutathione. Thus, prudence should be exercised in the use of experimentally galactosemic rats as a model for diabetes until the correspondence of the models has been more fully characterized.

Rats fed a galactose-rich diet have been used for several years as a model for diabetes to study, particularly in the eye, the effects of excess blood hexoses. This study sought to determine the utility of galactosemia as a model for oxidative stress in extraocular tissues by examining biomarkers of oxidative stress in galactose-fed rats and experimentallyinduced diabetic rats. Sprague-Dawley rats were divided into four groups: experimental control; streptozotocin-induced diabetic; insulin-treated diabetic; and galactose-fed. The rats were maintained on these regimens for 30 days, at which point the activities of catalase, glutathione peroxidase, glutathione reductase, and superoxide dismutase, as well as levels of lipid peroxidation and reduced and oxidized glutathione were determined in heart, liver, and kidney. This study indicates that while there are some similarities between galactosemic and diabetic rats in these measured indices of oxidative stress (hepatic catalase activity levels and hepatic and renal levels of oxidized glutathione in both diabetic and galactosemic rats were significantly decreased when compared to normal), overall the galactosemic rat model is not closely parallel to the diabetic rat model in extra-ocular tissues. In addition, several effects of diabetes (increased hepatic glutathione peroxidase activity, increased superoxide dismutase activity in kidney and heart, decreased renal and increased cardiac catalase activity) were not mimicked in galactosemic rats, and glutathione concentration in both liver and heart was affected in opposite ways in diabetic rats and galactose-fed rats. Insulin treatment reversed/prevented INTRODUCTION The defining property of diabetes mellitus (DM) is the inability of the body to regulate glucose metabolism. This inability arises as a result of either insufficient or nonexistent insulin production, as in Type I DM, or tissue insensitivity to insulin, as in Type II DM. This disruption of normal metabolic regulatory processes causes both short-term complications, such as hyperglycemia, and long-term complications, such as arteriosclerosis and microangiopathies. Ill While both behavioral modification and medication have had success in controlling aspects of the disease, many of the chronic complications *Corresponding author. Tel.: 812-855-3201, Fax: 812-855-4436, e-mail: watkins@indiana.edu 211 continue to elude treatment. Notably, DM patients are at significantly higher risk for cardiovascular disease, retinopathy, peripheral neuropathy, and nephropathy. Ill Research has long implicated free radical oxidative stress resulting from hyperglycemia as one source of these long-term complications. I2-41 Circulating glucose, unable to enter tissues due to the deficit of or insensitivity to insulin, collects in excess in both the blood and non-insulin dependent tissues. Oxidative stress may be created as the overabundant hexoses engage in metal catalyzed oxidation, Isl participate in glycation reactions with proteins and lipoproteins, I61 or enter the polyol pathway resulting in the conversion of glucose into sorbitol and the release of free radicals. I71 For nearly two decades, experimentally induced hypergalactosemia has been used as a model to study the morphological and biochemical changes that retinal tissue undergoes as a result of high.blood sugars. I8,91 The retinopathy that develops under these conditions is morphologically similar to that which occurs in diabetes, I1,11 and the stability of the hypergalactosemic state facilitates long-term studies, which are difficult in experimentally diabetic animals. The elegance of this model lies in looking only at the effects of high blood sugar, excluding the complex array of metabolic and hormonal imbalances that arise in DM. Il In spite of the great utility of this model in mimicking diabetic retinopathy, the whole-animal effects of hypergalactosemia have been less thoroughly examined. Although galactosemia has been used for years in biochemical and morphological studies in many tissues, investigations into the interactions of galactosemia and oxidative stress in extra-ocular tissues are minimal in the literature, and it is unclear whether the galactose-fed rat can serve as a model for the oxidative stresses in DM. This study hypothesizes that focusing solely on excess hexose, as in experimental galactosemia, does not provide a reasonable model of the oxidative stresses experienced in various tissues by diabetic rats.

Materials and Animals
All chemicals were purchased from Sigma (St. Louis, MO). Male Sprague-Dawley rats (175-200g) were procured from Harlan Sprague Dawley (Indianapolis, IN). After 7 days acclimatization, the rats were randomly divided into four treatment groups of 15 rats. One group was left untreated as normal controls. In two groups, diabetes was induced by a bolus intravenous STZ injection into a saphenous vein (100mg/kg in freshly prepared 10mM sodium citrate, pH 4.5). One group of diabetic rats was treated with insulin (2-4 units PZI insulin/day sc) from Eli Lilly (Indianapolis, IN). The fourth group of 15 was fed a diet of 50% galactose (Dyets, Bethleham, PA). All other rats were given Purina Rat Chow (No. 5012, St. Louis, MO) and water ad libitum. Diabetic rats having serum glucose levels < 350 mg/dl, as determined by Sigma glucose (HK) kit, were excluded from the study. Animal experimentation and husbandry adhered to the U.S. Public Health Service guidelines. I21

Tissue Collection and Assay
Thirty days after the beginning of treatments, animals were anesthetized with halothane (2-4%, ih). Blood was drawn by cardiac puncture for terminal blood glucose concentrations and for quantitation of hemoglobin Alc (HblAc) levels with Sigma Kit 441-B. Liver, kidney, and heart were excised and immediately frozen to -70C. Frozen tissue from each rat was homogenized (5%w/v) in ice-cold 0.1 M Tris-HC1 buffer (pH 7.4) and assayed for degree of lipid peroxidation by measuring the concentration of thiobarbituric acid reactive substances (TBARS). I131 The remaining homogenates were centrifuged at 100,000 g for 1 hour, and supernatants from each liver, kidney, and heart were assayed for activity levels of catalase, I141 superoxide dismutase, I51 glutathione peroxidase, I161 and glutathione reductase. [71 Another 0.25 g aliquot of each tissue was homogenized in 3.75 ml ice-cold 0.1 M sodium phosphate-5 mM EDTA (pH 8.0), then l ml 25% phosphoric acid was added. After vortexing for 10sec, the samples were centrifuged at 100,000 g for 30min. The resulting supernatants were assayed for concentrations of both reduced (GSH) and oxidized (GSSG) glutathione. I81 Protein concentration in all samples was determined by the method of Lowry. [ 19] Statistical Analysis For each tissue, means and standard errors of the means were calculated for all groups. Data were analyzed by one-way analysis of variance followed by Duncan's test. Significance was set at p 0.05. In the table and figures, all experimental groups (diabetic, insulin-treated diabetic, and galactose-fed) were compared to the untreated control group and significant differences are indicated by * Table I shows general characteristics of the experimental animals. Overall body weight of diabetic rats and galactosemic rats was decreased in comparison to that of normal rats. However, because the liver weight of the galactosemic rats was also significantly lower than normal, the liver-to-body-weight ratio, while elevated in diabetic rats, was normal in galactose-fed rats. As expected, the blood glucose concentrations in diabetic rats were significantly higher than in normal controls, and there was a parallel increase in glycosylated hemoglobin, as measured by HbAlc. The galactosemic rats showed no significant increase or decrease in blood glucose, but the apparent increase in HbAlc was not significant at p < 0.05. In retrospect this is plausible because, while glycosylation secondary to excess glucose is best measured with HbAlc, the glycosylation secondary to galactosemia is more detectable on the HbAla and HbAlb fractions. [21 Insulin treatment reversed/prevented the diabetes-induced changes in body weight, liver/body weight ratio, serum glucose concentrations and HbAlc levels.

RESULTS
Comparison of biomarkers of oxidative stress within the rats revealed some trends. Hepatic catalase activity levels in both diabetic and galactosemic rats were significantly decreased when compared to normal (Fig. 1). Hepatic and renal levels of GSSG were significantly diminished compared to normal in both diabetic and galactosemic rats (Fig. 2).
However, several effects of diabetes (increased glutathione peroxidase activity in liver, increased superoxide dismutase activity in kidney and heart, decreased renal and increased cardiac catalase activity, and decreased renal glutathione reductase activity, Fig. 1; increased lipid peroxidation in liver and heart, Fig. 2) were not mimicked in galactosemic rats (although these values were higher in treated  Effects of 30 days of streptozotocin-induced diabetes, insulin-treated diabetes, and 50% galactose diet on enzyme activities in liver, kidney, and heart. One unit of superoxide dismutase is defined as the quantity of superoxide dismutase required to produce 50% inhibition of the rate of reduction of cytochrome C under assay conditions. One unit of catalase is defined as the amount of enzyme that liberates half the peroxide oxygen from a hydrogen peroxide solution in 100 seconds at 25C. One unit of glutathione reductase and glutathione peroxidase is equivalent to one nmol NADPH oxidized per minute. Values represent means s.e.m, of 6 to 15 animals per group. *Significantly different from normal control, p < 0.05. animals than in control rats). Curiously, GSH concentration in both liver and heart was affected in opposite ways in diabetic rats and galactose-fed rats (Fig. 2). Insulin treatment reversed/prevented the activity changes in renal and cardiac superoxide dismutase, renal and cardiac catalase, and hepatic glutathione peroxidase (Fig. 1). Interestingly, renal glutathione reductase after insulin administration was decreased below the levels in normal and untreated diabetic rats. As Figure 2 shows, insulin treatment of diabetic rats reversed/ prevented the hepatic changes in TBARS, GSH and GSSG as well as the increase in cardiac GSH. Effects of 30 days of streptozotocin-induced diabetes, insulin-treated diabetes, and 50% galactose diet on concentrations of thiobarbituric acid reactive substances (TBARS), reduced glutathione (GSH) and oxidized glutathione (GSSG) in liver, kidney, and heart. Values represent means s.e.m. of 6 to 15 animals per group. *Significantly different from normal control, p < 0.05.

DISCUSSION
Experimentally galactosemic animals have proven to be a productive model for the study of diabetic retinopathy, extensive studies having shown that the effects of high blood hexoses accurately mimic the morphological lesions of DM in the retina. [11'211 Galactose, flooding into the retinal tissue in excess, initiates non-enzymatic glycation and enters into the polyol pathway to generate free radicals and induce oxidative damage, as well as activating protein kinase C and other diabetes-like abnormalities. [7, Further, therapeutics research has found this model to be useful in determining the utility of experimental drugs in controlling this exaggerated polyol pathway. I81 While antioxidants do not promise a cure for chronic complications of DM, many studies [26][27][28][29][30][31][32] have shown they have some success in ameliorating these complications in vitro. Some researchers have sought to utilize the experimentally galactosemic model for research into antioxidant therapies in the eye, [9'33] a viable strategy since previous research has so clearly shown the utility of this model in this tissue. However, galactosemic retinopathy does not always respond to treatment in the same way that diabetic rats do. [25] The elegance of this model, its ability to isolate high blood hexoses and the sequelae of this hyperhexosemia from the other metabolic imbalances found in DM, makes it appealing to research outside of the realm of the eye. Some studies have begun to do just that, and have found that experimental galactosemia is indeed useful in antioxidant therapeutics research in other tissues, where the administration of a mixture of antioxidants did effectively control cardiac abnormalities in both diabetic and galactosemic animals. I311 However, in spite of these successes, this study urges caution in the expanded use of this model. Galactose feeding does render the whole animal hypergalactosemic, not just the eye or the heart. [22] Yet, the responses of other tissues to excessive blood hexose have not been fully characterized. This study sought to evaluate oxidative stress as seen in this model, and did find some similarities between galactosemia and STZ-induced diabetes, particularly in the liver. However, interactions of glutathione in the two models are disparate. The concentration of hepatic GSSG and activity of glutathione peroxidase are both affected by diabetes but not by galactose feeding, and the effects of diabetes and galactose on the concentrations of GSH are opposed.
The overall level of concordance between the two groups is not convincing. Diabetes exerts an extensive systemic effect, acting at multiple sites with equal intensity. On the other hand, the galactose model relies on intestinal absorption of galactose, which then is passed via the portal circulation into the systemic circulation. The end result is that in the galactose-fed model, the liver is exposed to the highest concentration of the oxidative stressor and thus may bear the brunt of the damage. This could explain the significant similarities between the models in the liver, and the lack of consonance in the other organs.
The apparent lack of statistical significance in the elevated level of HbAlc in galactose-fed rats (Tab. I) is an artifact of the statistical analysis. If all four groups are analyzed, only the diabetic group is significantly different from normal (p < 0.001); the galactosemic group is not significantly different (p >0.05). However, if only the normal and galactosemic groups are compared, then the galactosemic HbAlc rises to the level of significance (p<0.001). In other studies where only these two groups are compared and where hyperglycemia is maintained for [12][13][14][15][16][17][18][19][20][21][22] months, [34,351 total glycated hemoglobin levels in the galactose-fed group appear to be elevated from normal levels.
Thus, in spite of some agreement, the composite picture presented in this study shows that galactose feeding does not render a model that closely mimics the effects of diabetes in extraocular tissues. It seems likely that hyperglycemia alone, as isolated in the galactosemia model, is not sufficient to explain the oxidative stress in liver, kidney and heart of diabetic rats. Further comparisons between the experimentally hypergalactosemic model and the diabetic model could focus on isolating other potential mechanisms of diabetic pathology, such as nonenzymatic protein glycation and exaggeration of the polyol pathway, [8,36,37] as well as examining specific cellular and molecular indices.
In conclusion, in light of the overarching divergence of biomarkers of oxidative stress between the models of the diabetic animal and the galactosemic animal, we agree with previous studies I381 that urge caution in the use of this model in studies of hyperglycemia as seen in DM.