To our knowledge, this is the first study to report that prolonged fructose consumption increases 24-h circulating uric acid concentrations, GGT activity and RBP-4 levels in humans. Importantly, these findings also complement our previously reported results (in these same subjects) which demonstrated that hepatic fractional de novo lipogenesis (DNL) and plasma lipids and lipoprotein concentrations increased in subjects consuming fructose-sweetened beverages, but remained unchanged in subjects consuming glucose-sweetened beverages for 10 wks . Despite comparable weight gain (~1-2% of initial body weight), subjects consuming fructose primarily exhibited increases of visceral adipose tissue (VAT), whereas only subcutaneous adipose tissue (SAT) was increased in subjects consuming glucose . In addition, indices of insulin sensitivity and glucose tolerance were decreased during 10 wks of fructose consumption, but were unaffected by isocaloric consumption of glucose .
Prolonged fructose consumption significantly increased both fasting uric acid concentrations and 24-h uric acid exposure. The increase of fasting uric acid concentrations exhibited after 10 wks of fructose consumption is consistent with previous reports of the effects of up to 5 wks of fructose consumption at 20-30% of total calories in humans [24–26]. Reiser et al. demonstrated that consumption of 20% of energy from fructose for 5 wks as part of a diet designed to maintain body weight increased fasting uric acid levels in men  and our results support these findings.
The mechanism by which fructose consumption leads to increased uric acid concentrations is thought to be initiated by the depletion of adenosine triphosphate (ATP) and inorganic phosphate (Pi) resulting from unregulated production of fructose-1-phosphate and glyceraldehyde-3-phosphate from fructose, which bypasses the rate-limiting step of glycolysis, phosphofructokinase (PFK) [3, 27]. Fructose-induced depletion of ATP and Pi leads to a concomitant increase of purine nucleotide degradation and, subsequently, uric acid production . It has also been reported that, in addition to increased nucleotide degradation, prolonged fructose consumption promotes increases in the incorporation of glycine into urate suggesting that upregulation of de novo purine nucleotide synthesis may also contribute to fructose-induced increases of uric acid . While uric acid is a potent and physiologically relevant antioxidant, recent evidence suggests that elevated levels may be associated with increased oxidative stress, and there is still considerable debate as to which of these roles is more important in the context of metabolic disease .
Elevations of fasting uric acid levels have been shown to be associated with hypertriglyceridemia and insulin resistance [30, 31]. Our results are not consistent with these findings as we did not detect any relationships between changes in fasting uric acid levels or 24-h uric acid exposure and the previously reported marked increases of TG exposure or reductions of insulin sensitivity in subjects consuming fructose . Also in contrast to the findings of previous investigations [12, 15], increases of fasting uric acid concentrations and 24-h uric acid exposure were not associated with elevations of RBP-4 levels or GGT activity. Although we did not detect a correlation between elevations of uric acid and changes of TG exposure, insulin sensitivity, RBP-4 levels or GGT activity, it is important to point out that our sample size was fairly modest and thus these findings do not rule a contribution of uric acid to the reported changes in these metabolic parameters. Additional intervention studies are needed to investigate the possibility that these changes are related. It is interesting that fasting uric acid concentrations (but not 24-h uric acid exposure) also increased significantly in subjects consuming glucose-sweetened beverages for 10 wks despite the fact that these subjects did not exhibit any of the adverse changes measured in subjects consuming fructose . This finding suggests that changes of 24-h uric acid exposure may be a more sensitive indicator of increased metabolic dysfunction than changes of fasting levels.
No previous studies have investigated the effects of glucose or fructose consumption on circulating levels of RBP-4. There is increasing evidence suggesting that RBP-4 may be an important link between increases of visceral adiposity and insulin resistance . In animal studies RBP-4 has been clearly shown to reduce glucose uptake and impair insulin signaling in muscle, as well as to increase hepatic glucose production via induction of the gluconeogenic enzyme phosphoenolpyruvate carboxykinase (PEPCK) . In humans, increases of circulating RBP-4 are strongly associated with insulin resistance in adipose tissue , and elevations of circulating RBP-4 are predictive of a diagnosis of metabolic syndrome . While hepatocytes are the primary source of RBP-4, it has been demonstrated that adipocytes can also contribute significantly to circulating concentrations of RBP-4 . Reduced expression of the insulin-stimulated glucose transporter, GLUT4, in adipocytes has been shown to lead directly to increased adipocyte secretion of RBP-4 . Moreover, expression of RBP-4 in humans is significantly greater in visceral adipose tissues (VAT) compared with subcutaneous adipose tissue (SAT) and is associated with an increase of adipocyte size .
Circulating RBP-4 concentrations are significantly elevated following 10 wks of fructose consumption (Table 2) and these changes are correlated with increases of postprandial TG (Table 3, Figure 3). We have hypothesized that fructose consumption can promote reductions in insulin sensitivity by providing substrate for hepatic DNL leading to hepatic triglyceride accumulation, PKC activation, and increased hepatic insulin resistance  and have suggested that this mechanism is responsible for the reductions in insulin sensitivity previously reported in these same subjects . Although the reported increases of RBP-4 were relatively modest, our results suggest the possibility that fructose-induced increases of RBP-4 levels may also have contributed to the previously reported reductions of insulin sensitivity in these subjects .
We have suggested that the differential effects of fructose and glucose on regional adipose deposition may be explained in part by the increased sensitivity of SAT relative to VAT to insulin-stimulated lipoprotein lipase (LPL) activation , and by our reported observations that insulin responses were decreased in subjects consuming fructose and increased in subjects consuming glucose . The differential changes in RBP-4 levels in subjects consuming fructose or glucose are consistent with this mechanism considering that reductions of post-meal insulin exposure in subjects consuming fructose would lead to decreased expression of GLUT4 in adipocytes, which would be expected to be associated with an increase in the production and secretion of RBP-4 from adipose tissue (VAT in particular). Thus, it is possible that reduced insulin exposure in subjects consuming fructose led to increased plasma RBP-4 levels directly by decreasing expression of GLUT4 in adipose tissue, and indirectly by increasing deposition of TG into VAT, leading to increased visceral adipocyte size, and increased secretion of RBP-4. The significant reduction of circulating RBP-4 concentrations observed in subjects consuming glucose-sweetened beverages is consistent with the observations that glucose consumption did not result in increased DNL, postprandial hypertriglyceridemia, accumulation of VAT or reduced postprandial insulin exposure .
The finding that increases of RBP-4 during fructose consumption were larger in men than in women was not unexpected considering that TG responses and increases of VAT deposition were also considerably greater in men  (Figure 2). We have also reported that women exhibited greater decreases in insulin sensitivity than men in response to fructose consumption, and have hypothesized that this may be due to decreased rates of VLDL production and secretion, leading to larger increases of hepatic lipid . Our findings suggest that the increased rate of VLDL production/secretion following fructose consumption in men is accompanied by an increase of RBP-4 levels. These elevations of RBP-4 levels are likely the result of increased accumulation of VAT and decreased insulin-stimulated GLUT4 expression in adipocytes. These data suggest that in men fructose consumption likely contributes to increases of hepatic insulin resistance both directly by providing substrate for hepatic DNL leading to increased hepatic TG accumulation and decreased hepatic insulin sensitivity , and indirectly by increasing visceral adiposity , while in women it is primarily a direct effect mediated by increased hepatic DNL and lipid content.
There have been very few prospective investigations of the effects of fructose consumption on the activity of the liver enzymes GGT, AST and ALT in humans. However, the results from one recent study suggest that short-term fructose overfeeding does not alter AST or ALT activities , and our results following 10 wks of fructose consumption are in agreement with these findings. Reports on the effects of fructose consumption on GGT activity in humans are not available, and the results of animal studies are not likely to be physiologically relevant since these investigations have primarily been conducted rats, which have been reported to have hepatic and plasma GGT activities over 20-fold lower than humans . The results of the present study demonstrate that prolonged fructose consumption leads to marked increases of GGT activity in older, overweight/obese adults (Table 2).
We also report that fructose-induced increases of plasma GGT activity are positively associated with increases of 24-h TG exposure, peak TG exposure and the postprandial TG peak reported previously in these same subjects  (Table 3, Figure 3). Martin et al. first demonstrated that GGT activity is strongly associated with postprandial plasma TG levels , which has since been confirmed by other investigators . The authors hypothesized that the upregulation of hepatic microsomal enzymes that results in increased TG synthesis and DNL is accompanied by a concomitant increase in GGT activity and speculated that this process could be initiated by excessive carbohydrate intake . The increases of GGT activity in subjects consuming fructose support the mechanism proposed by Martin et al. However, since GGT activity significantly decreased during 10 wks of glucose consumption, and these changes were not related to measures of TG exposure, we suggest that, under energy balanced conditions, it is not carbohydrate in general (as speculated by Martin et al.) but rather intake of fructose specifically that mediates increases of hepatic DNL, TG synthesis, and GGT activity.
GGT has been well established as a reliable marker of increased hepatic lipid content and hepatic insulin resistance . In addition, elevations of plasma GGT activity are positively associated with increases of visceral adiposity, and this relationship is independent of hepatic fat content [43, 44]. It has been suggested that fructose consumption may contribute to increases of hepatic insulin resistance both directly, by providing substrate for hepatic DNL leading to increased triglyceride accumulation and novel PKC activation  and indirectly by increasing visceral adiposity . Based on the established associations between GGT activity, increased hepatic lipid content and visceral adiposity it is possible that either or both of these mechanisms may contribute to increases of GGT activity during fructose consumption. However, the association of fructose-induced increases of GGT activity with measures of TG exposure, but not with increases of visceral adiposity, suggests that an increase of intrahepatic lipid may be a primary mechanism. It should also be noted that while increased GGT activity is considered a sensitive marker of increased intrahepatic lipid, it lacks the specificity of other liver enzymes such as ALT. Elevations of GGT activity have also been shown to be associated with increased oxidative stress, all-cause mortality, and mortality from cancer and diabetes as opposed to liver disease alone . It is interesting but not clear why the liver enzymes GGT, ALT and AST decreased significantly from values measured during the baseline complex carbohydrate diet in subjects consuming glucose-sweetened beverages.
Elevations of plasma GGT activity are predictive of the development of metabolic syndrome  and it has been suggested that elevated GGT activity should be included as an additional diagnostic risk factor for metabolic syndrome . Our data suggest that under energy balanced conditions, consumption of fructose at 25% of energy requirements for 10 wks leads to greater increases of plasma GGT activity in subjects with less than 3 MSRFs compared to those with metabolic syndrome (≥3 MSRFs), despite comparable baseline values in both groups (see Results). The reasons for this differential response are unclear.