The first four articles in this series concerned intraperitoneal glucose tolerance tests (1), the time course of changes in food intake, body weight, plasma glucose and insulin concentrations and insulin resistance (2), the secretory behaviour of isolated pancreatic islets (3) and the metabolism of D-glucose in isolated pancreatic islets (4) in fructose-fed rats. Comparisons were made between the results for control female rats exposed from the 8th week after birth and for the ensuing 8 weeks to a diet containing 64% (w/w) starch and 5% (w/w) sunflower oil (Ssun rats) and female rats exposed over the same period to a diet containing 64% D-fructose and 5% sunflower oil (Fsun rats) or 3.4% sunflower oil mixed with 1.6% salmon oil (Fsal rats) or safflower oil (Fsaf rats). The selection of these experimental conditions was based on an extensive examination of the relevant literature, as documented in a review study dealing mainly with the influence of nutritional factors on the fructose-induced metabolic syndrome (5). The present report extends these observations to a first set of post-mortem investigations dealing inter alia with: the percentage of glycated hemoglobin in blood and liver glucokinase activity relating to carbohydrate metabolism; plasma concentration and liver content of cholesterol, triglycerides and phospholipids in respect to lipid metabolism; and protein, albumin, urea and creatinine concentrations in plasma in respect to nitrogen-rich compounds.

The four groups of female Wistar rats used in the present study were the same as those defined in the first report in this series. At sacrifice, blood was sampled by cardiac puncture following overnight fasting.

The methods used for the measurement of the plasma concentration of albumin (2), blood concentration of glycated hemoglobin (2), liver protein content (6) and hepatic glucokinase activity (7) have been described previously. The kits used for the measurement of the plasma concentration of urea and creatinine (Biocon, Vöhl-Marienhagen, Germany), calcium (Spinreact S.A., Girona, Spain), iron (Randox Laboratories Ltd., Crumlin, UK), plasma and liver triglycerides and phospholipids (Spinreact S.A.) plasma and liver cholesterol (Biolabo, Maizy, France), plasma HDL-cholesterol (Biocon) were obtained from the sources cited in parentheses.

The plasma albumin concentration (2) and percentage of glycated hemoglobin (2) are listed in the present report for purpose of comparison. All results are presented as the mean value ± SEM (standard error of the mean) together with the number of individual determinations (n). The statistical significance of a difference between mean values was assessed using the Student’s t-test.

The results recorded in the present study are shown in Fig. 1. It documents that, for the majority of variables, specifically the percentage of glycated hemoglobin (Fig. 1A), liver cholesterol, triglyceride and phospholipid content (Fig. 1B) and plasma concentration of albumin, urea, creatinine, phospholipids, triglycerides and total cholesterol (Fig. 1C), the measurements recorded in the Fsun rats were invariably higher than those determined in the Ssun rats. In the fructose-fed rats, the substitution of part of the sunflower oil by an equal amount of either safflower oil or salmon oil minimized the relative magnitude of such increases, with salmon oil having the greatest minimizing effect. Conversely, the substitution of starch by D-fructose in the diet containing sunflower oil decreased the plasma concentration of calcium, iron and HDL-cholesterol (Fig. 1D). The effect of D-fructose was again opposed by the substitution of some of the sunflower oil by either safflower oil or salmon oil, the latter having the stronger effect. Two variables, that is, the blood hemoglobin concentration and liver protein content, were not significantly affected by the dietary manipulations (Fig. 1E). Finally, the activity of liver glucokinase was markedly higher in the Fsal rats than in the other three groups of rats (Fig. 1F). Thus, in several respects, the partial substitution of sunflower oil by either safflower oil or salmon oil corrected, at least to some extent, the biochemical perturbations found in the metabolic syndrome of the fructose-fed rats (Fig. 1A-C) and, on occasion, even increased the parameter under consideration above the level found in the control Ssun rats (Fig. 1D and F).

The determinants responsible for these diet-induced changes in metabolic variables remain, to some extent, a matter of speculation, as recently reviewed elsewhere (5). The following considerations should not be ignored, however. In terms of carbohydrate metabolism and as judged from the percentage of glycated hemoglobin, the present findings are consistent with the well-known perturbation of glucose homeostasis in fructose-fed rats, currently attributed to an increase in hepatic glucose output and peripheral insulin resistance (5). This supports the view that the dietary supply of ω3 fatty acids to the Fsal rats and, to a lesser extent, that of ω6 fatty acids, in this case mainly C18:2ω6 to the Fsal rats (1) may oppose such an unfavorable effect of the fructose-rich diet (5). Incidentally, in the Fsal rats, the glycogen content of the liver did not differ (98.5±7.5%; n=6) from that in the Ssun rats (100.0±8.3%; n=6), whilst being decreased (P<0.007) to 71.8±4.9% (n=11) of the same reference value in the other fructose-fed rats (data not shown). The major novel information provided by the present study in respect to carbohydrate metabolism relates to the measurements of liver glucokinase activity. The results of such measurements, as given in Table III, document a higher activity of glucokinase in the liver of the Fsal rats than in the other 3 groups of rats (P<0.05 or less). The higher liver glucokinase activity in the Fsal rats, as compared with either the Fsun or Fsaf rats, is reminiscent of the higher glucokinase activity found in the pancreatic islet homogenates of control rats, as compared with second generation rats depleted in long-chain polyunsaturated ω3 fatty acids, a difference tentatively ascribed to inhibition of the enzyme by endogenous long-chain fatty acyl-coenzyme A in the ω3-depleted rats (7). Indeed, in the present study, the Fsal rats were the sole fructose-fed animals exposed to a diet containing sizeable amounts of long-chain polyunsaturated ω3 fatty acids (1), this coinciding with a lower liver triglyceride content (P<0.001) in the Fsal rats than in the Fsun or Fsaf rats. Furthermore, in the fructose-fed rats, a significant negative correlation prevailed between the individual values for liver glucokinase activity and liver triglyceride content (r=−0.580; n=17; P<0.02).

In terms of lipid metabolism, the salient findings comprised: the increased plasma concentration and liver content of cholesterol, triglycerides and phospholipids, but decreased plasma concentration of HDL-cholesterol in the Fsun rats, as compared with the Ssun rats; and the partial or complete correction of these perturbations in the Fsaf and Fsal rats, particularly in the latter. The internal consistency of these findings is further documented by the negative correlation found at the individual level between the total cholesterol concentration and that of HDL-cholesterol in plasma (r=−0.537; n=23; P<0.01). Likewise, the likely relationship between liver and plasma lipids is supported by the finding of close correlations between the individual values for the liver content of cholesterol, triglycerides and phospholipids and the plasma concentration of the same lipids (Fig. 2). The correlation coefficient in those three cases averaged 0.846±0.055 (n=3) with, in all cases, a probability well below 0.001 (n=23 in all three cases). Notably, the individual negative xy products represented no more than 1.6±0.8% (n=3) of the corresponding positive xy products, no negative individual xy product being observed between the liver content and plasma concentration of triglycerides.

No clear difference in the protein content of the liver, heart, kidney or adipose tissue was observed among the four groups of rats. Nevertheless, the plasma albumin concentration was somewhat higher (P<0.02 or less) in each group of fructose-fed rats than in the control starch-fed rats (Table I). It was significantly lower, however, in the Fsal rats than in the Fsun or Fsaf rats (P<0.001). Likewise, the plasma protein concentration was slightly higher (P<0.05) in the fructose-fed rats than in the control starch-fed rats (data not shown). The correlation between the individual values for plasma albumin and protein concentrations was close to achieving statistical significance (r=+0.389; n=23; P<0.07).

The urea and creatinine plasma concentrations were increased in the Fsun rats, with a trend towards normalization in the Fsaf and Fsal rats. Relative to the corresponding mean values found in Ssun rats (100.0±5.5%; n=12), the measurements for the Fsun rats averaged 141.6±10.1% (n=12; P<0.002) and decreased to 132.0±8.9% (n=12) in the Fsaf rats and 120.6±7.0% (n=10) in the Fsal rats. The latter two mean values were significantly higher (P<0.04 or less), however, than that determined in the Ssun rats. These findings may indicate incipient nephropathy in the fructose-fed rats.

Finally, the plasma concentration of calcium and iron also underwent dual changes. Relative to the mean value in Ssun rats (100.0±5.2%; n=12), that in the Fsun rats was decreased (P<0.001) to 75.5±3.5% (n=12) and was distinctly different (P<0.01 or less) from those in Fsaf (95.1±5.8%, n=12) and Fsal rats (115.5±5.5%; n=10). The latter value was not significantly higher (P<0.06) than that recorded in the control Ssun rats. Changes in the plasma contribution of albumin are unlikely to act as determinants of changes in the plasma concentration of calcium, the correlation coefficient between these two variables (r=−0.390; n=23; P<0.07) being negative. As an alternative hypothesis, it may be speculated that the decrease in plasma calcium and iron concentration indicates an impaired intestinal absorption of these metals. To our knowledge, however, an enteropathy in animal models of diabetes mellitus has only been well-documented in type 1 diabetic animals such as BB rats (8–10), in which case it was proposed that the enteropathy plays a role in the development of type 1 diabetes, rather than resulting from it (11).

In summary, the present set of post-mortem data collected in rats exposed for 8 weeks to a fructose-rich diet and distinct oils confirms certain classical findings and reveals some unexpected features of the metabolic syndrome prevailing in fructose-fed rats, whilst duly documenting the favorable effect of diets enriched with selected long-chain polyunsaturated ω3 and ω6 fatty acids on minimizing the fructose-induced perturbations in several metabolic variables.